Comprehensive Cytopathology E-Book
2098 pages
English

Vous pourrez modifier la taille du texte de cet ouvrage

Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus

Comprehensive Cytopathology E-Book

-

Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus
2098 pages
English

Vous pourrez modifier la taille du texte de cet ouvrage

Description

This best-selling book provides you with a comprehensive guide to the diagnostic applications of exfoliative and aspiration cytology. The book takes a systemic approach and covers the recognized normal and abnormal cytological findings encountered in a particular organ. Appropriate histopathological correlations and a consideration of the possible differential diagnosis accompany the cytological findings. The book is lavishly illustrated, making it the perfect practical resource for daily reference in the laboratory.
  • Provides an accessible guide to diagnostic investigation and screening.
  • Includes a summary of major diagnostic criteria and discusses the pitfalls and limitations of cytology.
  • Utilizes a consistent chapter structure to make finding the answers you need quick and easy.
  • Provides updates to crucial chapters to keep you on top of the latest diagnosis and techniques.
  • Incorporates differential diagnosis tables for easy comparison/contrast of diagnoses.
  • Offers more than 1800 full-color images depicting a full range of normal and abnormal findings.
  • Discusses new concepts on molecular basis of neoplasia.
  • Explores the role of cytogenetics in cancer development.

Sujets

Ebooks
Savoirs
Medecine
Médecine
United States of America
Oncogén
Intraepithelial neoplasia
Hodgkin's lymphoma
Squamous intraepithelial lesion
Oncology
Cirrhosis
Meningitis
Bethesda System
Robotics
Kaposi's sarcoma
Photocopier
Liver
Viral disease
Surgical suture
Thyroid nodule
Endometrial biopsy
Vaginal intraepithelial neoplasia
Phase contrast microscopy
Endophthalmitis
Chromosome abnormality
Non-small cell lung carcinoma
Viral
Immunocytochemistry
HPV vaccine
Pericardial effusion
Cervical intraepithelial neoplasia
Dysplasia
Opportunistic infection
Fluorescent in situ hybridization
Mediastinum
Neoplasm
Carcinoma in situ
Uveitis
Cholangiocarcinoma
Retinal detachment
Thymoma
Hashimoto's thyroiditis
Urinalysis
Chromosomal translocation
Review
Retroperitoneal space
Flow cytometry
Physician assistant
Retinoblastoma
B-cell chronic lymphocytic leukemia
Terminology
Renal cell carcinoma
Pancreatic cancer
Pleural effusion
Cytogenetics
Lumbar puncture
Ovarian cyst
Biopsy
Vitrectomy
Lesion
Pleurisy
Carcinoma
Soft tissue
Adenocarcinoma
Medical imaging
Mentorship
Salivary gland
Internal medicine
Ascites
Endoscopy
Bladder cancer
Genital wart
Cytopathology
Human papillomavirus
Cyst
Esophagus
Fixative
Karyotype
Artifact
Methylation
Lymph node
Lymphoma
Non-Hodgkin lymphoma
Respiratory system
Intestine
Differentiation
Ophthalmology
Urinary system
Obstetrics and gynaecology
Camera
Pneumonia
X-ray computed tomography
Multiple sclerosis
Philadelphia
Peritoneum
Stomach
Confocal
Pancreas
Infection
Thyroid
Ovary
Microscope
Microscopy
Mechanics
Chemotherapy
Chromatin
Breast
Adrenal gland
Japan
Moving
Anus
Antibodies
Pathology
Gene
Effusion
Neuraxis
Clustering
Small
Illumination
India
Assay
City
Electronic
Automation
Clientélisme (Rome)
Inflammation
Fluorescence
Baltimore
DNA
Réaction en chaîne par polymérase
Gênes
Copyright
Éthanol
Colon

Informations

Publié par
Date de parution 18 septembre 2008
Nombre de lectures 0
EAN13 9781437719628
Langue English
Poids de l'ouvrage 8 Mo

Informations légales : prix de location à la page 0,1495€. Cette information est donnée uniquement à titre indicatif conformément à la législation en vigueur.

Exrait

COMPREHENSIVE CYTOPATHOLOGY
Third Edition

Marluce Bibbo, MD ScD FIAC FASCP
Professor of Pathology and Cell Biology, Jefferson Medical College, Director, Cytopathologym, Thomas Jefferson University Hospital, Philadelphia, PA, USA

David Wilbur, MD
Associate Professor of Pathology, Harvard Medical School, Director, Cytopathology, Massachusetts General Hospital, Boston, MA, USA
SAUNDERS
Copyright
First edition 1991
Second edition 1997
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the Publishers. Permissions may be sought directly from Elsevier's Health Sciences Rights Department, 1600 John F. Kennedy Boulevard, Suite 1800, Philadelphia, PA 19103-2899, USA: phone: (+1) 215 239 3804; fax: (+1) 215 239 3805; or, e-mail: healthpermissions@elsevier.com. You may also complete your request on-line via the Elsevier homepage ( http://www.elsevier.com ), by selecting ‘Support and contact’ and then ‘Copyright and Permission’.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
A catalog record for this book is available from the Library of Congress

Notice
Medical knowledge is constantly changing. Standard safety precautions must be followed, but as new research and clinical experience broaden our knowledge, changes in treatment and drug therapy may become necessary or appropriate. Readers are advised to check the most current product information provided by the manufacturer of each drug to be administered to verify the recommended dose, the method and duration of administration, and contraindications. It is the responsibility of the practitioner, relying on experience and knowledge of the patient, to determine dosages and the best treatment for each individual patient. Neither the Publisher nor the author assume any liability for any injury and/or damage to persons or property arising from this publication.
The Publisher
Printed in China
Last digit is the print number: 9 8 7 6 5 4 3 2 1
Commissioning Editor: William Schmitt
Development Editor: Nani Clansey
Project Manager: Rory MacDonald
Designer: Stewart Larking
Illustration Manager: Gillian Richards
Illustrator: Marion Tasker
Marketing Manager(s) (UK/USA): Kathy Neely/Lisa Damico
Contributors

Fadi W. Abdul-Karim, MD, Associate Professor of Pathology and Orthopedics, Department of Pathology, University Hospitals of Cleveland, Cleveland, OH, USA

Tahseen Al-Saleem, MD, Director of Hematopathology and Flow Cytometry, Senior Member, Division of Medical Science, Fox Chase Cancer Center, Philadelphia, PA, USA

Anniek J.M. van Aspert – van Erp, PhD CFIAC, Research Associate, Department of Pathology, University of Nijmegen, Ravenstein, The Netherlands

Peter H. Bartels, PhD, Professor Emeritus of Optical Sciences, Department of Optical Sciences, University of Arizona, Tucson, AZ, USA

Simon Bergman, MD, Associate Professor of Pathology, Department of Pathology, Wake Forest University Health Sciences, Winston-Salem, NC, USA

Marluce Bibbo, MD ScD FIAC FASCP, Professor of Pathology and Cell Biology, Department of Pathology, Anatomy and Cell Biology, Jefferson Medical College, TJU, Director of Cytopathology, Thomas Jefferson University Hospital, Philadelphia, PA, USA

Sandra H. Bigner, MD, Medical Director, Laboratory Corporation of America, Burlington, NC, USA

Lukas Bubendorf, MD, Professor of Pathology, Institute for Pathology, Division of Cytology, University Hospital Basel, Basel, Switzerland

Johan Bulten, MD PhD MIAC, Assistant Professor of Pathology, Department of Pathology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands

Alicia L. Carter, MD, Fellow Cytopathology, Duke University Medical Center, Burlington, NC, USA

Mamatha Chivukula, MD, Assistant Professor of Pathology, Magee-Womens Hospital of UPMC, Pittsburg, PA, USA

Luiz M. Collaço, MD, Professor of Pathology, Federal University of Parana, Evangelica Faculty of Parana, Bigorrilho, Curitiba, Brazil

Terence J. Colgan, MD, Professor of Laboratory Medicine and Pathobiology, Mount Sinai Hospital, Toronto, ON, Canada

David J. Dabbs, MD, Professor and Chief of Pathology, Department of Pathology, Magee-Womens Hospital of UPMC, Pittsburgh, PA, USA

Magnus von Knebel Doeberitz, MD, Professor of Pathology, Department of Applied Tumor Biology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany

Craig E. Elson, MD FCAP, Director of Cytopathology, Department of Pathology, Research Medical Center, HCA Midwest Division, Kansas City, MO, USA

Michael S. Facik, MPA CT(ASCP), Cytopathology Laboratory Supervisor, Department of Cytopathology, University of Rochester Medical Center, Rochester, NY, USA

Brendan T. Fitzpatrick, MD, Attending Pathologist, Our Lady of Lourdes Medical Center, Camden, NJ, USA

William J. Frable, MD, Professor of Pathology, Virginia Commonwealth University, Richmond, VA, USA

Hugo Galera-Davidson, MD FIAC, Professor of Pathology, Faculty of Medicine, University of Seville, Hospital Universitario Virgen Macarena, Seville, Spain

Kim R. Geisinger, MD FCAP, Professor of Pathology, Department of Pathology, Wake Forest University School of Medicine, Winston-Salem, NC, USA

Ben J. Glasgow, MD, Associate Professor of Ophthalmology and Pathology, Department of Pathology, UCLA Center for Health Sciences, Los Angeles, CA, USA

Katharina Glatz-Krieger, MD, Professor of Pathology, University Hospital of Basel, Basel, Switzerland

Ricardo González-Cámpora, MD FIAC, Professor of Pathology, Faculty of Medicine, University of Seville, Hospital Universitario Virgen Macarena, Seville, Spain

Hans Jurgen Grote, MD, Assistant Professor, Institute of Cytopathology, Heinrich-Heine University, Dusseldorf, Germany

Prabodh K. Gupta, MD FIAC, Professor, Pathology and Laboratory Medicine, Cytopathology and Cytometry Section, Department of Pathology and Laboratory Medicine, Hospital of the University of Pennsylvania, Philadelphia, PA, USA

Pierre Heimann, MD PhD, Pathologist, Service d'Anatomie Pathologique, Cytologie, Cytogénétique, Institut Jules Bordet, Brussels, Belgium

William W. Johnston, MD FIAC, Professor Emeritus of Pathology, Duke University, Durham, NC, USA

Ruth L. Katz, MD, Professor of Pathology, Director, Image Analysis and Research Cytopathology, Department of Pathology, MD Anderson Hospital, Houston, TX, USA

Catherine M. Keebler, ScD(hon) CFIAC, Registrar, International Academy of Cytology, Chicago, IL, USA

William H. Kern, MD FIAC, Clinical Professor Emeritus of Pathology, Department of Pathology, LAC/USC Medical Center, Pasadena, CA, USA

Larry F. Kluskens, MD PhD, Assistant Professor, Director of Cytology, Department of Pathology, Rush University Medical Center, Chicago, IL, USA

Savitri Krishnamurthy, MD, Associate Professor, Department of Cytopathology, MD Anderson Cancer Center, University of Texas, Houston, TX, USA

Oscar Lin, MD PhD, Associate Attending, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Diane B. Mandell, CT (ASCP) CFIAC, Supervisor, Cytology Service, Center for the Health Sciences, University of California at Los Angeles, Los Angeles, CA, USA

Cindy McGrath, MD, Assistant Professor of Pathology and Laboratory Medicine, Cytopathology Section, University of Pennsylvania Medical Center, Philadelphia, PA, USA

C. Meg McLachlin, MD, Associate Professor, Pathology and OBGYN, University of Western Ontario, Medical Leader, Surgical Pathology, London Health Sciences Centre, London, ON, Canada,

Christopher R.B. Merritt, MD, Professor of Radiology, Thomas Jefferson University Hospital, Philadelphia, PA, USA

Kiran F. Narsinh, MD, Fellow Eye Pathology, Department of Pathology, JSEI,UCLA Medical Center, Los Angeles, CA, USA

Joseph F. Nasuti, MD, Attending Pathologist, Dianon Systems, Division of LabCorp, Bridgeport, CT, USA

Ritu Nayar, MD, Associate Professor of Pathology, Northwestern University, Feinberg School of Medicine, Northwestern Memorial Hospital, Chicago, IL, USA

Bernard Naylor, MB ChB FIAC, Professor Emeritus of Pathology, The University of Michigan, Ann Arbor, MI, USA

Wai-Kuen Ng, MBBS FRCPA FHKCPath FHKAM(Pathology) FIAC, Honorary Clinical Associate Professor, Department of Pathology, Queen Mary Hospital, The University of Hong Kong, Hong Kong SAR, China

José Schalper Perez, MD MIAC, Professor of Pathology, San Sebastian University School of Medicine, Chief, Division of Anatomical Pathology and Cytopathology, Concepcion, Chile

Reda S. Saad, MD PhD, Associate Professor, Department of Pathology and Laboratory Medicine, Allegheny General Hospital, Pittsburgh, PA, USA

Jan F. Silverman, MD FCAP, Professor and Chairman, Department of Pathology and Laboratory Medicine, Allegheny General Hospital, Pittsburgh, PA, USA

Lambert Skoog, MD PhD, Professor of Clinical Cytology, Department of Pathology and Clinical Cytology, Karolinska University Hospital Solna, Stockholm, Sweden

Diane Solomon, MD, Division of Cancer Prevention, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA

Theresa M. Somrak, JD CT(ASCP) CFIAC, Director, Cytopathology Education Consortium Activities, Chicago, IL, USA

Kari Syrjanen, MD PhD FIAC, Professor, Department of Oncology & Radiotherapy, Turku University Hospital, Turku, Finland

Edneia M. Tani, MD PhD, Associate Professor Clinical Cytology, Department of Pathology and Cytology, Karolinska University Hospital Solna, Stockholm, Sweden

Liang-Che Tao, MD FRCPC, Professor Emeritus of Pathology and Laboratory Medicine, Department of Pathology, Indiana University, Camano Island, WA, USA

Alain Verhest, MD PhD FIAC, Professor of Pathology (retired), Brussels, Belgium

G. Peter Vooijs, MD PhD FIAC, Professor of Pathology, Scientific Director, Technological Medicine, University of Twente, Enschede, The Netherlands

Nicolas Wentzensen, MD, Department of Applied Tumor Biology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany

David C. Wilbur, MD, Associate Professor of Pathology, Harvard Medical School, Director, Cytopathology, Massachusetts General Hospital, Boston, MA, USA

Moira D. Wood, MD, Assistant Professor of Pathology, Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA, USA

Bin Yang, MD PhD, Assistant Professor, Department of Pathology, The Cleveland Clinic Foundation, Cleveland, OH, USA

Grace C.H. Yang, MD FIAC, Professor of Clinical Pathology and Laboratory Medicine, Papanicolaou Cytology Laboratory, Department of Pathology and Laboratory Medicine, New York Presbyterian Hospital-Weill Cornell Medical Center, New York, NY, USA

Nancy A. Young, MD, Director, Outpatient Laboratory, Senior Member, Division of Medical Science, Fox Chase Cancer Center, Philadelphia, PA, USA

Maureen F. Zakowski, MD FCAP, Professor of Pathology, Memorial Sloan-Kettering Cancer Center, Department of Pathology and Cytology, New York, NY, USA

Lucilia Zardo, MD, Director, Section for Technology in Cytopathology, Division of Pathology, National Cancer Institute, Rio de Janeiro, Brazil
Preface
Since the publication of the second edition of Comprehensive Cytopathology in 1997 new trends in the practice of gynecologic and non gynecologic cytology have emerged. Liquid based preparations and HPV testing have been introduced in gynecologic cytology, immunocytochemistry, clinical cytogenetics and molecular techniques have seen expanded applications to non gynecologic cytology, and fine needle aspiration under sonographic, computer tomographic or fluoroscopic guidance for cytologic diagnosis has become an indispensable component of the workup of body lesions, with more FNA's being performed by radiologists, and endoscopists in close cooperation with cytologists.
The table of contents reflects additions and changes in topics to provide detailed discussions of the significant advances in the field. Examples include New concepts on molecular basis of neoplasia ( chapter 1 ), The role of genetics in cancer development ( chapter 2 ), Evaluation of the sample in smears and liquid based preparations ( chapter 5 ), FNA of Mediastinum ( chapter 26 ), FNA of Pediatric Tumors ( chapter 29 ), Virtual Microscopy ( chapter 33 ), Automation of Cervical Cytology ( chapter 34 ), Molecular Techniques ( chapter 36 ). Among the unique features of this edition is the large number of more than 1700 color illustrations depicting the main diagnostic entities with applications of adjunct techniques to cytopathology. New contributors from institutions around the world ( chapters 1 , 2 , 3 , 6 , 9 , 10 , 18 , 21 , 24 , 25 , 26 , 28 , 29 , 33 , 34 , 35 & 36 ) have joined the distinguished roster of authors not only to write new chapters but also to rewrite them.
I would like to thank Dr David Wilbur for editing the chapters in the FGT section of CC3E and for his valuable input. Thanks to all contributors for their efforts and outstanding contributions. Special thanks to Nani Clansey, Development Editor, for her enormous help in the development of this work and William R Schmitt, Editor, for his support and also to Rory MacDonald, Project Manager, Stewart Larking, Designer and to Gillian Richards, Illustration Manager in Elsevier Ltd.
In this third edition of Comprehensive Cytopathology we have attempted, through an international group of experts to present a state of the art work. Again we placed in a single comprehensive book discussions on general cytology, diagnostic exfoliative cytology of all body sites, entities and diagnostic challenges in fine needle aspiration of various organs, effects of therapy on cytologic specimens and special methodologies and technologies in cytology. We hope that our attempt to incorporate advances in the field while keeping emphasis on cytomorphology will be valuable to all professionals interested in cytopathology.

Marluce Bibbo, MD
Dedication
To George L Wied and Stanley F. Patten, our great mentors and friends, and our families for encouragement and support.

Marluce Bibbo, David Wilbur
Acknowledgments
Some of the material in the third edition is derived from chapters in the second edition. The editors acknowledge the contributions of the following authors to the previous edition.

George L. Wied

Karen H. van Hoeven

Dorothy Rosenthal

Torsten Lowhagen

Hi Young Hong

Lisa M. Bibb
Specific acknowledgment is made for permission to use the following figures:
Ch 2 , Figs 5 & 6 : Courtesy of Applied Imaging Corporation
Ch 5 , Fig 12 : Courtesy of Dr Zubair W Balloch, Philadelphia, PA, USA
Ch 5 , Fig 14 : Courtesy of Dr Yener Erozan, Baltimore, MD, USA
Ch 5 , Figs 18 & 19 : Courtesy of Dr Corrado Minimo, Philadelphia, PA, USA
Ch 7 , Fig 11 : Courtesy of Dr Belur Bhagavan, Baltimore, MD, USA
Ch 7 , Figs 23 & 75 : Courtesy of Dr S Bhambhani, New Delhi, India
Ch 7 , Fig 25 : Courtesy of Dr K Kapila and K Verma, New Delhi, India
Ch 9 , Fig 16 a: Courtesy of Robert H Young, Boston, MA, USA
Ch 10 , Figs 18 , 19 , 37 & 38 : Courtesy of K. Shimizu, Japan
Ch 10 , Figs 20 , 21 , 39 & 40 : Courtesy of Y Norimatsu, Japan
Ch 14 , Fig 13 : Courtesy of Dr Jan Silverman, Pittsburg, PA, USA
Ch 14 , Fig 16 : Courtesy of Dr Robert Petras, Cleveland, OH, USA
Ch 14 , Fig 22 : Courtesy of Josh Sickel MD, Mountain View, CA, USA
Ch 15 , Figs 3 , 9 & 14 : Courtesy of Dr PT Chandrasoma, Los Angeles, CA, USA
Ch 15 , Fig 16 : Courtesy of Dr Yener Erozan, Baltimore, MD, USA
Ch 15 , Fig 44 : Courtesy of Dr Demetrius Bagley, Philadelphia, PA, USA
Ch 16 , Figs 2 , 4 , 7 , 10 & 16 Courtesy of Dr Keith Volmar, Chapel Hill, NC, USA
Ch 18 , Fig 49 : Courtesy of Dr Syed Z Ali, Baltimore, MD, USA
Ch 19 , Fig 31 : Courtesy of Dr Sudha R Kini, Detroit, MI, USA
Ch 19 , Fig 36 : Courtesy of Dr Suzanne M Selvaggi, Madison, WI, USA
Ch 19 , Fig 64 : Courtesy of Dr Praboth K Gupta, Philadelphia, PA, USA
Ch 19 , Fig 66 : Courtesy of Ms Jamie L Covell, Charlottesville, VA, USA
Ch 19 , Fig 67 : Courtesy of Dr Cesar V Reyes, Mines, IL, USA
Ch 19 , Fig 69 : Courtesy of Ms Sharon Hicks, Martinez, CA, USA
Ch 19 , Fig 72 : Courtesy of the late Dr James N Landers, Detroit, MI, USA
Ch 19 , Fig 94 : Courtesy of Dr Andrew A Renshaw, Miami, FL, USA
Ch 19 , Fig 126 : Courtesy of Mr. Tarring A Seidel, Danville, PA, USA
Ch 19 , Fig 137 : Courtesy of Dr Dorothy L Rosenthal, Baltimore, MD, USA
Ch 19 , Fig 139 : Courtesy of Dr Bella Maly, Jerusalem, Israel
Ch 19 , Fig 140 : Courtesy of Dr Maria C Gamarra, Buffalo, NY, USA
Ch 24 , Figs 9 & 10 : Courtesy of Dr Janet F Stastny, Outpatient Cytopathology Center, Johnson City, TN, USA
Ch 24 , Figs 14 & 16 : Courtesy of Dr Nivaldo Medeiros, Emeritus Attending, Clinical Hospital of the University of Sao Paulo, Brasil
Ch 27 , Fig 9 : Courtesy of Dr Ritu Nayar, Northwestern University, Chicago, IL. USA
Ch 27 , Fig 15 A: Courtesy of Gary Rust, MD, Humble, TX, USA
Ch 27 , Fig 36 B: Courtesy of Dr Stephen Somerville, Longview, TX, USA
Ch 27 , Fig 45 : Courtesy of Dr Alan Heimann, Stony Brook University, NY, USA
Ch 27 , Fig 56 : Courtesy of Dr Bruce MacKay, MD Anderson Cancer Center, Houston, TX, USA
Ch 28 , Fig 100 : Courtesy of Dr Syed Z Ali of Johns Hopkins Hospital, Baltimore, MD, USA
Ch 31 , Fig 10 (right): Courtesy of Dr J Reagan, Case Western Reserve Cytopathology Laboratory, Cleveland, OH, USA
Ch 34 , Figs 4 AB, 5 & 17 : Courtesy of James Linder MD, Cytyc Corporation
Ch 34 , Figs 10 AB: Courtesy of Dr Mathilde Boon, Leiden, Netherlands
Ch 35 , The authors would like to acknowledge Jeff Richmond, MD, Cytology Fellow, University of Pittsburgh Medical Center (UPMC), Marluce Bibbo, MD, and Mary Blumberg, MD, Surgical Pathology Fellow, UPMC, for their kind contribution of some of the figures; and Sean Toddy, CT, American Society for Clinical Pathology (ASCP), and Peter Atanasoff, CT, ASCP, Product Manager, Cytyc Corp, for providing the Cellient Automated cell block system pictures.
Table of Contents
Copyright
Contributors
Preface
Dedication
Acknowledgments
Part I: GENERAL CYTOLOGY
Chapter 1: The Cell: Basic Structure and Function
Chapter 2: Basic Cytogenetics and the Role of Genetics in Cancer Development
Chapter 3: Cytologic Screening Programs
Chapter 4: Diagnostic Quality Assurance in Cytopathology
Chapter 5: Evaluation of the Sample in Smears and Liquid-Based Preparations
Part II: DIAGNOSTIC CYTOLOGY
Chapter 6: The Bethesda System for Reporting Cervical Cytology
Chapter 7: Microbiology, Inflammation and Viral Infections
Chapter 8: Benign Proliferative Reactions, Intraepithelial Neoplasia, and Invasive Cancer of the Uterine Cervix
Chapter 9: Glandular Lesions of the Uterine Cervix
Chapter 10: Endometrial Lesions, Unusual Tumors and Extrauterine Cancer
Chapter 11: Vulva, Vagina, and Anus
Chapter 12: Peritoneal Washings and Ovary
Chapter 13: Respiratory Tract
Chapter 14: Alimentary Tract (Esophagus, Stomach, Small Intestine, Colon, Rectum, Anus, Biliary Tract)
Chapter 15: Urinary Tract
Chapter 16: Central Nervous System
Chapter 17: Eye
Chapter 18: Cytology of Soft Tissue, Bone, and Skin
Chapter 19: Pleural, Peritoneal, and Pericardial Effusions
Chapter 20: Fine-Needle Aspiration Biopsy Techniques
Chapter 21: Imaging Techniques
Chapter 22: Salivary Glands and Rare Head and Neck Lesions
Chapter 23: Thyroid
Chapter 24: Lymph Nodes: Cytomorphology and Flow Cytometry
Chapter 25: Breast
Chapter 26: Mediastinum
Chapter 27: Kidneys, Adrenals, and Retroperitoneum
Chapter 28: Liver and Pancreas
Chapter 29: Pediatric Tumors
Chapter 30: Effects of Therapy on Cytologic Specimens
Part III: SPECIAL TECHNIQUES IN CYTOLOGY
Chapter 31: Cytopreparatory Techniques
Chapter 32: Light Optical Microscopy
Chapter 33: Virtual Microscopy
Chapter 34: Automation in Cervical Cytology
Chapter 35: Immunocytochemistry
Chapter 36: Molecular Techniques
Index
Part I
GENERAL CYTOLOGY
CHAPTER 1 The Cell: Basic Structure and Function

Magnus von Knebel Doeberitz, Nicolas Wentzensen

Contents
PART I: BASIC STRUCTURE AND FUNCTION OF MAMMALIAN CELLS
Overview
Nucleus
Contents of the Nucleus
Nuclear Morphology
Hematoxylin
Nucleoli
Nuclear Envelope and Nuclear Shape
Cytoplasm and Plasmalemma
Cytoplasmic Stain
Endoplasmic Reticulum
Golgi Apparatus
Mitochondria
Lysosomes
Cytoskeleton, Centrosome
Cell Membrane, Receptors, and Signal Transduction
Cell Junctions
Cell Growth and Division
PART II: THE MOLECULAR BASIS OF NEOPLASIA
Overview
Principles of Malignant Transformation
Cancer-related Genes
The Major Pathways of Carcinogenesis
Carcinogenesis induced by Papillomavirus Infections
Basic Structure of the Virus and Its Genome
Epidemiology of HPV Infections
The Role of the HR-HPV E6 and E7 Genes
Progression of HPV-Infected Epithelial Cells to Invasive Cancer Cells
Concluding Remarks

PART I: BASIC STRUCTURE AND FUNCTION OF MAMMALIAN CELLS

Overview
Cells are the basic structural and functional units of all living organisms. The estimations about the total cell count of a human body vary widely; a number as large as 10 14 seems conceivable. Although the principal components of all cells are very similar, the differentiation of cells results in a wide variation of cellular morphology and function.
The smallest human cells by diameter are spermatozoas with a size of ~3 μm, followed by the anucleate erythrocytes (8 μm). The largest cells are female oocytes that can be as large as 35–40 μm and are visible to the naked eye. Motor neurons are extremely long cells, with their axons reaching from the spine to the distal extremities (up to 1.4 m length).
Most cells can only be functional in large united structures, such as organs or suborganic structures. Other cells, mainly of hemato- or lymphopoietic origin, are mobile and active as single cells, although in many cases, their functionality is dependent on interaction with other cells.
Cytopathology studies diseases on the cellular level. While in histopathology, cells are assessed in the spatial context, in virtually all cytological applications, they are removed from their spatial context and must be assessed isolated or as cell sheets.
Apart from the loss of the spatial information, cells can have considerably altered morphology when taken out of the united structures. Cell–cell contacts are important features that build the shape of a cell. Many structural elements within a cell are connected to proteins that are attached to other cells or the basal membrane. This must be taken into account when cells are compared in histological and cytological assessments.
Another important difference between histology and cytology is based on the fact that in histology, tissue sections that are plain two-dimensional are assessed, while in most cytology applications, complete cells that still have some three-dimensional features, although they might appear flat in the microscope, are analyzed. Based on these facts, the transfer of histological morphology to the picture seen in cytology can only be limited.
In this section, an overview of the most important cellular structures and functions relevant for cytopathology are presented ( Fig. 1.1 ). We have assembled the most important information on cellular structures by describing the regular function in brief, the relevance for cytopathology, and the morphology in normal and abnormal cells. For more detailed information about cellular structures and functions, a cell biology textbook is recommended.

Fig. 1.1 Schematic presentation of an epithelial cell displaying the most important structures discussed in this chapter.

Nucleus
The nucleus contains the genomic DNA, histones, and several proteins that are responsible for DNA replication, repair, and transcription of genetic information ( Fig. 1.2 A).

Fig. 1.2 Contents of the nucleus, DNA. (A) A nucleus displaying nucleoli, euchromatin, and heterochromatin. (B) Two nucleosomes consisting of DNA coiled around histone proteins. (C) The structure of double-stranded DNA. Organic bases are connected to a sugar–phosphate backbone. Complementary bases (A-T, C–G) are held together by hydrogen bonds.
The assessment of a cell's nucleus is one of the most important tasks in cytopathology. The size of the normal nucleus is highly variable, depending on the underlying cell type. In many malignant cells, nuclei are considerably enlarged. Apart from nuclear size, the chromatin density, the nuclear membrane, and the presence of nucleoli are important features of nuclear morphology and will be described in detail.
The nucleus contains about 25% dry substance, of which 18% is DNA plus a similar amount of histone proteins. The rest of the dry substance contains the non-histone proteins, nucleoli, and the nuclear membrane.

Contents of the Nucleus

DNA
The genetic information of organisms is coded in deoxyribonucleic acid (DNA). DNA is a long stretch of single nucleotides connected by a sugar–phosphate backbone ( Fig. 1.2 C). The genetic information is stored in specific sequences consisting of four different bases: adenine, guanine, thymine, and cytosine (A, G, T, C). A triplet of bases is coding for an amino acid that constitutes the basic component of proteins. Although in principle the triplet code allows for 64 different variations, only 20 protein-building amino acids exist. Many amino acids are coded by multiple base triplets. The genetic code is degenerate, thereby tolerating errors in the base sequence to some degree. Two DNA stretches are combined as a double helix; one complete turn is reached after 10 bases. The DNA stretches are not covalently bound, but attached via hydrogen bonds between complementary bases A–T and C–G.
DNA is a very robust and stable molecule, since it must protect the genetic code of an organism. The genetic information is transferred to the ribosomes (the protein production machinery) by ribonucleic acid (RNA) that has 3 important features different from DNA: RNA is based on a ribose backbone, it contains uracil instead of thymidine, and it is usually single-stranded. Compared to DNA, RNA is a rather unstable molecule.
The total DNA of a cell is separated on chromosomes. In total, 22 different chromosomes and two sex chromosomes exist. The chromosomes vary in size and in the content of coding sequences, they are numbered in decreasing order of their size. During the metaphase of mitosis, chromosomes are condensed and can be identified in light microscopy. In transcriptionally active cells, DNA is decondensed and takes up the room of the complete nucleus. When metaphase chromosomes are stained according to Giemsa, a heterogenous pattern of regions with strong staining (G-bands) and regions without staining (R-bands) can be observed. R-bands contain more genes than G-bands and are replicated early during cell division. The banding pattern of chromosomes has been used to determine chromosomal regions by indicating the chromosome number, the position with reference to the centrosome (p = short arm, q = long arm), and the position of the chromosomal banding (e.g. 3q26). In total, the human genome consists of ~3.2 billion bases, coding for approximately 25,000 genes. 1

Nuclear Proteins
Histones are basic proteins that build a structural unit together with the DNA, called the nucleosome ( Fig. 1.2 B). In the nucleosome, 146 base pairs (bp) are coiled around different histone subunits. The main function of the nucleosome is the high-density packing of DNA inside the nucleus, leading to a 50,000-fold increased compactness of DNA as compared to unpacked DNA.
Histone acetylation reduces the affinity between the DNA molecule and histones, leading to increased accessibility of DNA for transcription machinery components such as RNA polymerase and transcription factors. In general, for gene transcription, the DNA needs to be unpacked from the histones.
Besides histones, nuclear non-histone proteins build the nuclear scaffold structure and are involved in DNA transcription and replication.

Nuclear Morphology

Chromatin
Chromatin represents the complex structure of proteins and DNA in the nucleus of non-mitotic cells. There is usually twice as much protein as DNA in a nucleus. Since most cells in the human body are non-mitotic, chromatin is the morphological appearance of most cell nuclei assessed in cytology. The chromatin distribution and organization depends on many different factors, such as cell type, differentiation, metabolism, proliferation status, and, most important in cytopathology, neoplastic transformation.
Two conformations of chromatin are discriminated: euchromatin and heterochromatin. Euchromatin contains transcriptionally active protein-coding DNA regions. Heterochromatin represents the complex of DNA that is densely packed on histones. DNA sections not transcribed are usually stored in heterochromatin. Heterochromatin is further differentiated into constitutional, facultative, and functional heterochromatin. Constitutional heterochromatin consists mainly of highly repetitive DNA stretches in the centromeric region that are supposed to have structural functions but have also been found to express microRNAs that do not code for proteins but are involved in gene regulation. Facultative heterochromatin designates inactivated DNA regions that usually code for proteins but are not necessary in the respective cell, e.g. inactivation of the second X chromosome (Barr body). Functional heterochromatin contains DNA regions not necessary for the respective differentiation of a cell.

Hematoxylin
In many cytological applications, the chromatin is stained with hematoxylin. Hematoxylin is a basic dye extracted from the heartwood of the tree Haematoxylum campechianum . By itself, hematoxylin is a very weak stain. Different mordants, such as potassium alum, are used to generate the typical dark blue or purple staining. Hematoxylin strongly binds to acidic components of a cell, most importantly to the phosphate groups of nuclear DNA; the stained structures are therefore called "basophilic" ( Fig. 1.3 A).

Fig. 1.3 Exemplary pictures of nuclear and cytoplasmic staining. (A) Cervical cells stained with hematoxylin only. (B) Cervical cells stained with Hematoxylin and Eosin. (C) Cervical cells stained according to Papanicolaou.
Based on the nuclear stain, a wide variation of chromatin alterations can be observed, both alterations of structure and staining intensity. Structural aberrations include chromatin margination, i.e. aggregation of chromatin to the nuclear membrane, which is a sign of cell degeneration. Other chromatin alterations are coarsening and clumping that is usually accompanied by chromatin thinning in other regions.
Hyperchromasia, i.e. increased staining intensity, can result from increased chromatin loads or by a decreased nuclear volume, which inversely applies to hypochromasia. In additions, chemical modifications of the chromatin (e.g. during specific stains or cell treatments) can increase the stain uptake, simulating hyperchromasia.

Nucleoli
Nucleoli are small basophilic spherical bodies located in the nucleus. Usually they can be found in the central nuclear region but may also be close to the nuclear membrane. A nucleolus is built by a nucleolus organizing region (NOR) of a specific chromosome. These regions contain the genes for ribosomal RNA subunits that build the protein synthesis machinery. Since in a diploid human cell, in total 10 chromosomes containing NORs exist, in principal 10 nucleoli per nucleus could be present. Usually, only one or two nucleoli are found, since NORs from several chromosomes build a common nucleolus. Nucleoli have two distinctive regions, the pars fibrosa that contains the proteins required for transcription and the pars granulosa that contains the ribosomal precursors. During mitosis, nucleoli disappear and are reconstituted in the daughter cells. Shortly after cell division, a larger number of nucleoli that fuse gradually can be observed.
Depending on the cell type, the presence of nucleoli is physiological or can indicate malignant processes: liver cells that regularly produce a lot of protein can frequently exhibit nucleoli. In reactive or regenerative cells, nucleoli can become more prominent. In hepatocellular carcinoma, usually more than 50% of the cells show prominent, frequently multiple nucleoli. Intestinal epithelial cells also regularly show single nucleoli. In ageing and starving cells, a shrinking of nucleoli can be observed. In cancer cells, nucleoli can vary substantially with regard to size and shape.
In many malignant cells, multiple nucleoli that appear disjoint, odd-shaped, and spiculated can be observed. Proteins associated with nucleolar organizer regions can be visualized by a simple argyrophilic staining method. The structures highlighted by this method are called "argyrophilic nucleolar organizer regions" (AgNORs). Different distributions of AgNORs have been described between normal, dysplastic, and malignant tissues. In several cancer entities, AgNOR aberrations were found to have independent prognostic significance with respect to patient survival. 2 Increased NOR counts have been explained by increased metabolism with a high demand of ribosomes, but also by aneuploidy leading to increasing numbers of NOR regions in cancer cells.

Nuclear Envelope and Nuclear Shape
The nuclear envelope (NE) consists of two lipid membranes. The inner membrane is associated with the telomeres and anchors the chromosomes, while the outer membrane is part of the endoplasmic reticulum. The space between the two lipid layers is called perinuclear cisterna. The nuclear envelope constitutes the nucleus and separates the genomic material from the cytosol. During cell division, the nucleus disappears; the nuclear envelope is broken down to vesicles and is reassembled during telophase. The nuclear envelope builds a strong barrier between nucleus and cytosol; a number of nuclear pore complexes regulate the traffic between both compartments. There can be passive diffusion or active transport; in general, proteins synthesized in the cytoplasm require a specific nuclear signal in order to have access to the nucleus.
Inside the nuclear envelope is a network of chromatin fibrils and a nuclear lamina built from laminins. The nuclear envelope can be visible in light microscopy.
The regular nuclear shape is that of a smooth sphere or spheroid, based on the orderly arrangement of the chromosomes and the nuclear lamina. Many factors can affect the shape of the nucleus: stress, transcriptional, and synthetic activities can disturb the arrangement of interphase chromosomes; DNA amplifications can lead to uneven distribution of the nuclear material and to nuclear enlargements. At the same time, aberrations of the nuclear envelope can lead to alterations of the nuclear skeleton, resulting in altered chromosomal distributions. It has been assumed that changes of the nuclear envelope occur mainly after mitosis, when the nuclear envelope is reassembled. Alterations of the nuclear envelope have been directly linked to oncogene activity: Six hours after transfection with the ret oncogene, increasing cell counts with NE alterations were observed in human thyroid cells, indicating that nuclear alterations may occur even independent of postmitotic re-assembly. 3 NE alterations and the respective nuclear shape are an important diagnostic feature of many malignancies, especially papillary thyroid cancers and different types of leukemias.

Cytoplasm and Plasmalemma
The cytoplasm consists mainly of water (80–85%). The remaining constituents are proteins (10–15%), lipids (2–4%), polysaccharides (1%), and nucleic acids (1%). The cytoplasm is confined to the outside by the plasma membrane, a lipid bilayer, and to the inside by the nuclear membrane. In most cytology applications, normal cells have a homogenous cytoplasm with occasional granules or inclusions.

Cytoplasmic Stain
Eosin is the most common dye to stain the cytoplasm in histology. It is an acidic dye that binds to basic components of a cell, mainly proteins located in the cytoplasm. It gives a bright pink color that contrasts that dark blue nuclear hematoxylin staining ( Fig. 1.3 B). A combination of hematoxylin and eosin is the most frequently used dye in histology. In cytology, frequently, a Pap stain is performed. It consists of a hematoxylin-based nuclear stain and a polychromatic cytoplasmic stain, including Orange G and two polychromic dyes, EA36 and EA50. The cytoplasmic stain results in highly transparent cells, making it possible to assess superimposed cells in a Pap smear. Based on the cell type and the differentiation status, the cytoplasm can be pink–light red (e.g. superficial cervical cells) or light blue–green (e.g. cervical parabasal and intermediate cells) with all variations in between. The nuclei are dark brown or dark blue/violet and the nucleolus appears bright red ( Fig. 1.3 C).

Endoplasmic Reticulum
The endoplasmic reticulum (ER) consists of a single membrane making up for more than half of all internal membranes of the cell ( Fig. 1.4 A). The part of the membrane that faces the cytosol is studded with ribosomes. This part of the ER is called rough ER; the regions without ribosomes are called smooth ER.

Fig. 1.4 Membranous organelles. (A) The rough endoplasmic reticulum. (B) The Golgi apparatus. (C) A mitochondrion.
The main function of the ER is the packaging and delivery of newly synthesized proteins.

Golgi Apparatus
The Golgi apparatus is part of the membrane system that also contains the ER. It consists of stacked membrane-coated cavities, called dictyosomes ( Fig. 1.4 B). The Golgi apparatus is located close to the nucleus and can be very large in secretory cells, where it fills almost the complete cytoplasm. The convex side facing the ER/nucleus is called cis-Golgi; the concave side facing the cytoplasm is called trans-Golgi. From the Golgi apparatus, small vesicles transport products to other cellular sites or the exterior. Inside the structure, complex biochemical operations are being performed most of them resulting in post-translational modifications of synthesized proteins. Several secretory mammalian cell types are characterized by a prominent polarized Golgi apparatus located between the nucleus and the luminal surface: Goblet cells in the respiratory and digestive tract produce large amounts of glycoproteins, pancreatic cells secrete enzymes such as zymogen, and breast cells produce milk droplets.

Mitochondria
Mitochondria produce ATP, the universal fuel of living organisms, by oxidative processing of nutrients. They are located in the cytoplasm and separated from it by a double membrane ( Fig. 1.4 C). On average, an eukaryotic cell contains about 2000 mitochondria. Depending on age and cell type, mitochondrial size can vary between 0.5 and 10 μm. The highest mitochondrial counts can be found in cells with high energy demand, such as muscle cells, nerve cells, or metabolically active cells in the liver. Mitochondria are inherited in non-mendelian fashion via the cytosol of oocytes. During cell division, mitochondria are divided between the two daughter cells. They are genetically semi-autonomous since they possess their own circular genome, but are dependent on a number of proteins encoded by the nuclear DNA.
The Pap stain does not color mitochondria, but iron hematoxylin or acid fuchsin does. A more specific stain for mitochondria is rhodamine 123. Stained mitochondria appear as single spheres or long, branching structures, up to 7 × 0.5 μm in size. Mitochondria can be found in large numbers in hepatocellular carcinoma, resulting in a granular appearance of the cytoplasm. There are many other causes of granular cytoplasm; the underlying cellular components can only be visualized by ultrastructural methods. Since mitochondria represent the energy system of living cells, they are very important in the malignant development. Multiple functional and structural alterations during carcinogenic processes have been described. 4

Lysosomes
Lysosomes are small vesicles derived from the Golgi apparatus; they contain up to 40 acidic enzymes (hydrolases) at a pH 5. The membrane prevents the aggressive enzymes from destroying cellular structures. Although the contents can vary substantially, there are basically no morphological differences between functionally different lysosomes. The main function of lysosomes is the digestion of internal (non-functional cell organelles) and external (food, bacteria, leukocytes, debris) material. The processed material is either released to the cytoplasm, secreted, or stored in lysosomes.
Several storage diseases (e.g. Hunter-Hurler-Syndrome) are characterized by a deficiency of lysosomal enzymes. These disorders lead to accumulation of incompletely digested mucopolysaccharides in the lysosomes.

Cytoskeleton, Centrosome
The cytoskeleton is a complex lattice of various filaments building the cellular structure and shape; it is responsible for dynamic activities such as movement in growth and differentiation ( Fig. 1.5 ). Although is has been thought for a long time that the cytoskeleton is a special feature of eukaryotic cells only, it is becoming more and more clear that also prokaryotes have cytoskeleton-like structures. 5

Fig. 1.5 Display of an epithelial cell with cytoskeleton and cell–cell contacts.
The filaments are required for cell movement (cytokinesis), transport of material across the cell surface, muscle contraction, intracellular transport, and sorting and dividing of replicated chromosomes by the mitotic spindle.
Three main classes of cytoskeletal filaments are distinguished: actin filaments, intermediate filaments, and microtubules ( Fig. 1.5 ).
Actin filaments have a diameter of 7 nm and are built from six different actin types; in muscle cells, actin is functionally linked to myosin. They can be found in all cells, with especially high numbers in fibroblasts and the highest concentrations in muscle cells, since actin is part of the contractile structures. Lamellipodia (bulges of the cell surface for cell motility) and filopodia (enhanced cell surface for absorption) are built from bundled actin. Several glandular tissues such as breast and prostate have contractile myoepithelial cells that can forcibly express the glandular contents.
Intermediate filaments have a diameter of 10 nm, consist of one or more of 19 different cytokeratin molecules, and are the strongest fibers among the cellular filaments. They are mainly responsible for the structural framework of a cell and determine the cell's tensile strength. They build rope-like polymers. Keratins belong to the group of intermediate filaments. In keratinizing epithelial cells, keratin filaments accumulate and are cross-linked by other proteins and disulfide bonds. The keratinizing process starts at the periphery and progresses to the nuclear area. In fully differentiated cells, the nucleus becomes more and more pyknotic and finally dissolves.
Other examples for intermediate filaments are desmin in skeletal muscle, glial filaments in astrocytes, and neurofilaments in axons.
Microtubules are long, hollow tubes with 25 nm diameter assembled from microtubule oligomers originating in a membraneless body, the centrosome. The centrosome is the main microtubule organizing center (MTOC) of a cell and functions as an important regulator of the cell cycle. The centrosome consists of two orthogonally arranged centrioles surrounded by pericentriolar material and is situated between the nucleus and the Golgi apparatus. Upon cell division, each daughter cell receives one centriole. Although in most model organisms, a proper cell division can be achieved without a functional centrosome, an organism requires functional centrosomes to survive in the long term. Aberrant centrosomes are a hallmark of chromosomally instable cancer cells. Because of aberrant formation of the mitotic spindle apparatus, these cells acquire more and more chromosomal aberrations.
Two classes of substances can interfere with the microtubular network: Colchicine prevents the polymerization of microtubules, while paclitaxel interferes with their depolymerization. In a living cell, the microtubular network is continuously polymerized and depolymerized. Therefore, both agents lead to a non-functional spindle apparatus abrogating the cell division and are used as cytotoxins in cancer therapy.

Cell Membrane, Receptors, and Signal Transduction
The basic structure of the cell membrane is a semi-permeable lipid bilayer built from phospholipids, glycolipids, and steroids that contains various proteins floating on one side of or reaching through the complete membrane. The lipid bilayer has a gage of 6 to 10 nm and is barely visible in light microscopy. The cell membrane separates all cellular components from the environment and it assures the spatial and functional entity of a single cell. It allows the cell to persist in environments that would be harmful to the cellular components, such as extreme pH conditions and ion concentrations different from the cytoplasm. The cell membrane controls what is going into and out of a cell; thereby it regulates the import of nutrients and the export of cellular products. The transport is organized by passive (transport via a gradient that does not require energy) and active (transport against a gradient that requires energy) protein channels.
There are different types of membrane proteins. Peripheral membrane proteins only temporarily adhere to the respective cell membrane; they usually interact with integral membrane proteins. Many regulatory subunits of ion channel and transmembrane receptors, as well as enzymes and hormones, are peripheral membrane proteins. In contrast, integral membrane proteins are permanently attached to the membrane. Transmembrane proteins are integral proteins that span both lipid layers; they must contain a hydrophobic part that is placed in the lipid section and hydrophilic intra- and extracellular parts. Typical representatives are ion channels, proton pumps, and G-protein coupled receptors. Lipid-anchored proteins are covalently linked to lipids in the cell membrane; the most common type are G-proteins.
The communication of a cell with the environment (i.e. other cells or the extracellular matrix) is mediated by a wide variation of interacting molecules, usually designated as receptors. The functional principle of receptors is based on the key–lock principle; i.e. a specific receptor requires the binding of a specific ligand (either cell based or freely circulating) to be activated. The transfer of information between cells may be mediated by two different classes of receptors located in the cell membrane: Ionotropic receptors are based on specific ion channels that change the electric potential of cells upon activation ( Fig. 1.6 A). Non-ionotropic receptors have no pores, but are based on transmembrane proteins that stimulate intracellular proteins linked to the receptors and thereby modulate intracellular signal cascades.

Fig. 1.6 Important membrane proteins. (A) Different calcium ion channels controlling the intracellular calcium concentration. (B) G-protein receptors; the example shows adrenergic activation of the adenylate cyclase (AC). (C) Tyrosine kinase receptors; the example shows EGF-mediated phosphorylation of intracellular downstream targets.
The two most important non-ionotropic receptor types are G-protein coupled receptors and tyrosine kinase receptors. G-protein coupled receptors consist of seven transmembrane domains, an extracellular receptor region, and an intracellular part that binds the G-protein ( Fig. 1.6 B). Upon activation by an outside signal, the receptor changes its conformation and releases the G-protein in a not yet fully resolved mechanism. G-proteins, short for guanine nucleotide binding proteins, are the most important proteins involved in second messenger cascades, responsible for many central nervous (vision, olfactory system, neurotransmitters) and immune system functions. Receptor tyrosine kinases activate downstream targets by adding phosphate groups to intracellular proteins ( Fig. 1.6 C). They are frequently promoting cell growth and cell division (e.g. receptors for insulin-like growth factor, epidermal growth factor, EGF). Consequently, many malignant processes are linked to aberrations of G-protein coupled receptors and especially tyrosine kinase receptors. In about 25% of the breast cancers, a subtype of EGF receptors (ErbB2/Her2neu) is overexpressed, which leads to increased signaling of the EGF pathway, resulting in increased cell proliferation. Meanwhile, inactivating antibodies directed against Her2neu are successfully used in breast cancer therapy. Similarly, cetuximab is a monoclonal antibody downregulating ErbB1 signaling that has been approved for special types of colorectal cancer.

Cell Junctions
As complex organisms are built from billions of cells, the cell-to-cell contact is a very important parameter. Depending on the organ or the function of a tissue, cell contacts establish adherence of functionally connected cells, build a barrier against a lumen, and are involved in intercellular communication ( Fig. 1.5 ).
Adherence is based mainly on spot desmosomes (macula adhaerens) consisting of keratin filaments that connect the cytoskeleton of individual cells. A different form of adherence is the adhesion belt (zonula adherens) that connects the apical part of an epithelial cell to another epithelial cell. Hemidesmosomes are located at the basal pole of the cell and attach the cell to the basal membrane.
Tight junctions (zonula occludens) build an impermeable barrier between cells and are typical for structures that have a lumen.
Gap junctions are required for communication; small inorganic molecules and electric signals are exchanged via 1.5-nm pores in the cell membrane. Gap junctions can be quickly established from precursors floating in plasma membrane
The terminal bar describes a light microscopic structure that represents the sum of the adhesion belt and actin filament bundles as well as other protein filaments at the apical end of a cell.
In general, malignancy causes loss of cell-to-cell adherence. It has been shown that the cell adhesion molecule E-cadherin is lost during the formation of some epithelial cancers. The loss of E-cadherin is frequently accompanied by an overexpression of other cadherins, an effect called the “cadherin switch.” Besides the cell adhesion, signal transduction pathways and altered, inducing malignant transformation of the respective cell.

Cell Growth and Division
In mammalian organisms, some cells are created early in embryogenesis and remain unchanged throughout the whole life (e.g. lens of eyes, cells of CNS, heart muscle cells, auditory cells of ear). However, many epithelial tissues, as well as hemato-lymphopoietic cells, depend on a continuous replenishment of their cell pools. Other tissues have retained a regenerative capacity, e.g. liver stem cells, that make it possible to regenerate the organ after tissue damage. The detailed mechanisms of cell division can be recapitulated in a cell biology textbook. In brief, the cell cycle consists of four phases, the G1, S, G2, and M phases. Resting cells are in a constant G0 phase that has a direct transition from the G1 phase. G phases are gap phases, S indicates the synthesis phase in that the genomic material is doubled, and M indicates the phase of mitosis, in that two daughter cells are created ( Fig. 1.7 A). In order to assure a proper cell division with equal distribution of the genomic material in the daughter cells, several checkpoints (after G1, S, and G2) must be passed to allow for a continuation of the cell cycle. If the requirements at a checkpoint are not met, the cell is not allowed to continue; it might even go into apoptosis.

Fig. 1.7 The cell cycle and its morpholgical appearance. (A) Schematic presentation of the cell cycle. (B) Typical anaphase during mitosis of plant cells. (C) Mitosis of transformed epithelial cells in immunohistochemistry.
In light microscopy, mitotic figures represent the M phase, i.e. the distribution of chromosomes into the opposite poles of a cell ( Figs. 1.7 B, C). Mitotic figures are rare in normal tissue; in normal liver, about 1–2 mitoses can be observed per 10,000 cells. In rapidly dividing cancer tissue, the frequency of mitotic figures can be much higher, depending on the underlying cell type.
The regulation of the cell cycle is very complex. In general, uncontrolled activation of the cell cycle is a basic feature of all tumors. The reasons for uncontrolled cell cycle activation can be either the loss of inhibitory gene functions (inactivated tumor suppressor genes) or the activation of cell cycle promoting gene functions (activated oncogenes). The detailed mechanisms of carcinogenesis are described in Part II of this chapter.

PART II: THE MOLECULAR BASIS OF NEOPLASIA

Overview
Single cells in complex organisms work in a hierarchical order determined by their differentiation state. Each single cell fulfills its function through biochemical processes that are highly regulated and act in well-orchestrated pathways. The outgrowth of tumors occurs if the cells have lost the capability to follow these predetermined rules. Growth restriction, a typical feature of normal tissues, is lost; the capability to commit cellular suicide (apoptosis) once certain death signals have been activated is lost. Differentiation pathways that permit the cell to enter irreversible cell cycle arrest become disconnected. Mechanisms that under normal conditions are required to maintain the normal histological architecture of an organ are erroneously activated to feed the outgrowing tumor with essential nutrients.
Tumors may be derived from more or less all human tissues, including epithelium that may be transformed into benign adenomas and malignant carcinomas, mesenchymal tissues that may be transformed into benign fibromas and malignant sarcomas, hematopoietic tissues that may be transformed into lymphoma and leukemia, and finally even germ cells that may be transformed into benign teratomas and malignant teratocarcinomas.
Benign lesions are partially growth restricted: They do not invade, do not metastasize, but do grow locally. In contrast, malignant tumors usually acquire more autonomous growth properties, and develop complex strategies to invade other organs, to disseminate their cells to distant anatomical sites, to evade immunological attack of the host's immune system, and to initiate from disseminated cells again autonomous proliferative metastatic lesions. Following the concept that all these features are mediated by a complex network of genes, the question arises what genes are regulating the process of growth control in cancer cells and why are these genes activated or inactivated in distinct somatic cells that thus acquire the capacity to grow out as cancer cell. Evidently the modifications required to induce benign tumors are by far less complex than those required for an invasive cancer. They usually can be achieved by minor modifications of the genome of the affected cells. In many cases benign tumors are therefore precursors of malignant tumors that require a substantially more complex pattern of genetic modifications ( Fig. 1.8 ).

Fig. 1.8 Transformation of epithelial cells.
The biological phenotype of cancer cells is defined by the expression of certain genes, whereas other genes are not expressed. These phenotype-specific gene expression patterns are referred to as "gene signatures" that differentiate a distinct cell (e.g. cancer cells) from cells with other biological properties (e.g. a normal stem cell that is not transformed). Cancer cells of the same organ origin, e.g. gastric cancers, may display substantially different biological properties that reflect different gene signatures. 6 These changes of the gene expression pattern of cells or tissues can be monitored by gene expression arrays ( Fig. 1.9 ). To convert a normal somatic cell into a full-blown cancer cell many distinct steps are required, during which the gene expression pattern of the respective cell is gradually modified and the phenotype of the cell is transformed into increasingly neoplastic cells that are selected in an ongoing Darwinian selection process. Specific modifications of the gene expression signature trigger the next step of selection.

Fig. 1.9 Analysis of the gene expression signature of a panel of gastric cancer cells with different biological phenotypes. Genes expressed at high levels are indicated in red, those expressed at lower levels are indicated in green. By comparing the expression level of distinct genes (or c-DNAs) among different cell lines, complex differences among the expression of many genes can be monitored in a large number of tissues of cell lines. Using distinct software algorithms tissues or cell lines can be clustered in an hierarchical order that reflects their biological phenotype. The figure shows an example of various gastric cancer cell lines of which some display similar phenotypes: (B) epithelial cell cluster; (C) B lymphocytes cluster; (D) T lymphocytes cluster; (E) fibroblast cell cluster; (F) endothelial cell cluster. The results were visualized and analyzed with TreeView (M Eisen; http://rana.lbl.gov ).
Data and images were taken from Ji et al. 6
Given the complex alterations required to achieve the signature of a full-blown cancer cell and given the many individual selections steps required to transform a normal cells into a cancer cell, transformation cannot be achieve by a linear selection process but depends on higher level mechanisms that allow for major modifications of the genetic code in a relatively brief period of time or restricted number of cell divisions. The integrity of the number and structure of chromosomes, for example, is one particularly important aspect herein. If the mechanisms that maintain the integrity of the chromosomes fail, major genomic modifications may rapidly occur. Since most of these are non-viable, most cells that experience these failures will die in a process called genomic catastrophe. However, some cells may survive the complex modifications induced by malfunction of the mechanisms that preserve chromosomal homeostasis. If the gene signatures of the surviving cells allow for continuous autonomous growth eventually even at distant anatomical sites, the respective cell clones may be selected and their sustained growth may then clinically manifest itself as metastasizing cancer.
To preserve the ordered function of cells in higher order organisms a number of redundant genome protective mechanisms that prevent consequences of genetic catastrophes have been evolved. They primarily constitute organized suicide mechanisms called apoptosis that become activated once major modification of the cellular genome in distinct genetically damaged cells occur. They assure that most cells that undergo genetic catastrophes undergo apoptosis before they can grow out as transformed cancer cell. Thus, outgrowth of a cancer cell still is the rare exception in view of the many billions of proliferating cells that constitute the human body and the many events that trigger genomic catastrophes in damaged cells.

Principles of Malignant Transformation
The development of a cancer cell from a normal cell goes through three basic steps:
(1) Immortalization : In contrast to normal cells, immortalized cells can divide indefinitely, as long as they are supported with nutrients. They still have the same shape as normal cells, and they stop growing when they reach other cells (contact inhibition).
(2) Transformation : Transformed cells are independent of tissue-specific growth factors; they lose their contact inhibition and may grow invasively. Their shape is altered; the specific differentiation is lost.
(3) Metastasis : Metastasizing cells acquire the potential to migrate to distant sites and grow out to tumors.
These paramount changes occur on the level of the individual cancer cell. Subsequently, the establishment of tumors larger than 1–2 mm 3 requires the development of a vascular system that can support the growing tumor with nutrients. In order to achieve this, tumors induce angiogenesis via angiogenic factors such as vascular angiogenic growth factor (VEGF), fibroblast growth factor (FGF), and platelet-derived endothelial growth factor (PDGF).
Tumor growth is based on a complex interplay between the transformed cells and the surrounding tissue: invasive growth is associated with the expression of proteolytic enzymes that degrade the peritumoral stroma, most importantly matrix-metallo-proteinases (MMPs). Furthermore, the immune system is involved in the local control of growing cancers. It is estimated that the majority of malignant cell clones that develop in the human body are eliminated by immune system components directed against the transformed cells, a process called immune-surveillance. Invasive tumor development needs to evade these immune control mechanisms. A number of immune evasion strategies, including loss of antigen presentation machinery components, induction of suppressive T cells, and induction of apoptosis in attacking lymphocytes, have been analyzed.

Cancer-related Genes
Three major groups of genes are involved in carcinogenesis: oncogenes, tumor suppressor genes, and genes that are responsible for DNA repair and stability. 7 Oncogenes are mostly activators of the cell cycle that are strictly controlled in non-malignant cells. Activation can occur by chromosomal translocations that bring an active promoter close to an oncogene that is usually not expressed (e.g. BCR-ABL translocation in leukemia). Gene amplifications frequently lead to overexpression of oncogenes, as it is the case for the MYC gene. In addition, point mutations can lead to continuous activation of oncogenes, e.g. activating BRAF or RET mutations. In general, mono-allelic activation of oncogenes is sufficient for malignant transformation. In contrast, tumor suppressor genes (TSGs) usually require two hits, since the unaffected allele can substitute in part for the mutated allele. Still, in some cases, a partial effect (haploinufficiency) conferred by the loss of one allele has been described. TSGs can be altered by point mutations or by larger chromosomal losses. Typically genes involved in cell cycle control, regulation of preprogrammed cellular suicide (apoptosis), or the maintenance of genomic integrity may serve as TSGs. Repair/stability genes comprise a specific subgroup of TSGs in that they maintain the genetic integrity of the cell. Their loss of function is a prerequisite for the rapid acquisition of the critical mutations required for neoplastic transformation. To the latter group belong among others the mismatch repair genes (MMR), nucleotide excision repair genes, base excision repair genes, and genes involved in chromosomal recombination and segregation, such as BRCA1 and ATM. Germ line mutations of many of the TSGs have been identified as the cause of inherited cancer predisposition genes in familial cancer syndromes (e.g. hereditary nonpolyposis colon cancer (HNPCC), familiar adenomatous polyposis (FAP), multiple endocrine neoplasia II (MENII), BRCA1 and 2 for familial forms of breast cancer, and p53 for the Li-Fraumeni syndrome). In hereditary cancer syndromes, a tumor suppressor gene is inactivated in the germ line and a second hit is necessary to abrogate the function of the respective tumor suppressor gene in individual somatic cells (two-hit-hypothesis). Frequently, the initial gene alteration induces uncontrolled proliferation of the affected cells. In the course of accelerated cell divisions, genetic errors are accumulated and finally lead to malignant transformation of the cells. For many cancer entities, specific pathways of consecutive gene alterations have been described. Two central tumor suppressor genes are affected in many cancers: p53 is a central protein in the control of programmed cell death. Inactivation of p53 seems to be very important for malignant progression to be possible. pRb is a key regulator of cell cycle progression by controlling the E2F protein. The majority of cancers show inactivation of these TGSs themselves. However, specific cellular functions can be abrogated by attacking different components of the respective functional pathway. Thus functional pathways that represent a complex network of different gene functions may be hit on various levels in order to promote tumor development. Many different functional pathways that explain the heterogeneity of the genes affected in sporadic cancers have been described. 7
The accumulation of gene mutations necessary for malignant progression cannot be achieved at the standard mutation rates (1 mutation/million bases) observed in proliferating cells. It seems clear now that this baseline mutation rate is not sufficient for carcinogenesis and that some kind of genetic instability must occur in order to allow for the necessary mutations in cancer cells. This can only been achieved if central pathways that maintain the genetic integrity of a cell are hit. If such cells succeed to survive, they may rapidly accumulate a sufficient number of mutations that permit the neoplastic growth properties. Thus carcinogenesis can formally be subdivided into three major pathways depending on the molecular mechanism that makes it possible for a sufficient number of mutations or other genetic alterations of oncogenes and tumor suppressor genes that initiate and maintain the neoplastic conversion and progression to be accumulated.

The Major Pathways of Carcinogenesis
Alterations of mechanisms that maintain the genetic integrity of the cell's genome thus constitute the least common denominator. Consequently, neoplasia emerges as the failure of genetic functions that control either the composition of whole chromosomes, their number, structure, and distribution during mitosis or, alternatively, the integrity of the genetic information encompassed in the chromosome even if they do not undergo gross numerical or structural alterations. Consequently, cancer is the result of three major mechanisms that destroy the integrity of the genetic information:
1. Chromosomal instability (CIN) —Induced by failure of mechanisms that guarantee the even distribution of chromosomes to the daughter cells that emerge during mitosis;
2. Microsatellite instability (MSI) —Induced by failure of DNA mismatch repair enzymes that proofread and repair errors that occur during the de novo synthesis of DNA in the S-phase of the cell cycle; and finally
3. CpG island methylator phenotype (CIMP) —Induced by failure of the epigenetic control of genes that regulate critical steps in these processes and is often associated with the later development of MSI-induced cancers.
The vast majority of cancers occur via the CIN pathway. The major underlying mechanisms of carcinogenesis mediated by CIN is induced by disturbances of the bipolar character of the mitotic spindle during mitosis and the desegregation of chromosomes during mitosis ( Fig. 1.10 ). 8 During the normal M phase of the cell cycle the chromosomes line up in a plane, the metaphase plate, and associate with spindle fibers of microtubule proteins. The fibers together form a metaphase spindle. They are connected to the kinetochores on the chromosomes, i.e. nucleoprotein bodies associated with the centromeric DNA of the chromosomes, and the centrosomes at the poles of the mitotic cell. In a normal dividing cell the spindle fibers pull each sister chromatid apart toward the centrosomes. This ensures that each emerging daughter cell will get exactly one copy of the sister chromatids to form exactly one new copy of the respective chromosomes in the emerging new daughter cells. This complex mechanism is controlled by various checkpoints that monitor that before the process proceeds to the next step exactly two centrosomes and microtubules spindle apparatus have been formed, that each chromatid in a pair associates with its own distinct half of the spindle. Chromatid separation is not allowed to proceed until all pairs of chromatids are lined up in the metaphase plate. If these checkpoint controls fail, chromosomal segregation may fail and both chromatids may be pulled to one chromosome (non-disjunction). As consequence one of the daughter cells may become haploid for the respective chromosome and the other triploid. Alternatively, one chromosome is completely lost, if the attachment between microtubules and kinetochores fails.

Fig. 1.10 Influence of centrosome aberrations on chromosomal instability. (A) Normal centrosomal distribution with two spindle poles results in equal distribution of the chromosomal material to the daughter cells. (B) Aberrant spindle pole formation leads to unequal distribution of the chromosomal material; as a result, most cells die while some can acquire genetic alterations that lead to a growth advantage and the development of malignant cell clones.
More severe, however, is the impact of failing checkpoints that control the centrosomes themselves. Cancer cells that arise through the CIN pathway are characterized by an uneven number and uneven distribution of centrosomes during mitosis. This results in a total disorder of the normally bipolar spindle apparatus and leads to complex multipolar spindle structures that during mitosis result in disruption of the chromatids and a complex uneven distribution of the chromosomal material to the emerging daughter cells ( Fig. 1.10 ).
The major checkpoints that control these processes appear to act at the transition of the G1 phase of the cell cycle to the S phase and are controlled by cyclins and cyclin-dependent kinase complexes. Interestingly, these processes are targeted by two important viral oncoproteins encoded by high-risk human papillomaviruses (HPV), the HPV E6 and E7 proteins, that induce and maintain transformation of cervical cells as we will learn later in this chapter. This results in failing control of the mitotic processes and in particular distribution of chromosomes during mitosis and severe numerical and structural alterations of the chromosomes of the emerging daughter cells. The affected cells usually commit suicide, induced primarily by p53 gene or related genes. Thus most emerging cancer cells develop molecular mechanisms for evading this cellular suicide mechanism. In the case of papillomavirus-associated cancers it is the HPV E6 protein that binds to and inactivates p53. In other cancers not induced by oncogenic HPVs, p53 functions are usually abrogated by an inactivating mutation or deletion of the p53 gene itself or related genes within the same functional pathway.
Cells that achieve to evade the suicide control may survive the genomic catastrophe and form the initial cells of an outgrowing cancer. The emerging disproportionate distribution of chromosomal material in these transformed cell clones induce various important morphological alterations of the affected cells' nuclei that are the cornerstones of cytopathology ( Fig. 1.11 ). Aneuploid nuclei, coarser texture of the chromatin, changes in the size and shape of the nuclei, hyper- and hypochromasia, and altered shape and number of nucleoli are all immanent consequences of the desegregation of chromosomes during mitosis of cells that have lost control over the strictly bipolar mitotic figures, resulting in chaotic multipolar mitotic spindle complexes ( Fig. 1.10 ).

Fig. 1.11 Comparative genomic hybridization (CGH) and spectral karyotype hybridization (SKY). The average CGH ratio profiles for the diploid cell line DLD-1 and the aneuploid cell line HT29 are presented in (A) and (B). Note the remarkably stable genome of DLD-1 with only three copy number variations (chromosomes 2, 6, and 11). HT29 shows a highly aberrant ratio profile, with copy number alterations occurring on 13 chromosomes. The gains of 7, 8q, 13, and 20q are common aberrations in colorectal carcinomas. SKY of metaphase chromosomes prepared from these cell lines is shown in (C) and (D). No numerical aberrations were identified in the diploid cell line (C), whereas trisomies were common in the aneuploid cell line HT29 (D). All aberrations detected by SKY were also seen by CGH analysis. This indicates that no reciprocal, balanced chromosomal translocations have occurred.
Data and images were taken from Ghadimi et al. 11
Alternatively, cancer cells may arise through the MSI pathway. 9, 10 Cancers that emerges through this pathway are characterized by substantially different biological properties. They usually do not display numerical or structural changes of the chromosomes or the mitotic figures. 11 Cells of these tumors usually divide normally. Consequently, these tumor cells do not display aneuploidy, aberrant mitotic figures or gross alterations of their chromosomes. They usually remain diploid without major morphological alteration of their nuclei. However, errors emerge through a more subtle, superficially less brutal mechanism that in its clinical consequences may end as disastrously as the chromosomal instable cancers. In these cases the cancer cells acquire increasing mutations of the DNA sequence itself. After each round of DNA replication in the S-phase of the cell cycle usually hundreds and thousands of mutations, which need to be checked and repaired before the cell cycle proceeds to avoid very high accumulation, occur in the replicated genetic code due to misannealing and mispairing. Hence, all cells in nature develop a sophisticated proofreading mechanism mediated by the DNA-MMR complex , a multiprotein complex that proofreads and fixes the mutations. If distinct components of this repair complex are lost, the proofreading capacity declines and mutations particularly in thermodynamically sensitive DNA sequences occur. These lead to the rapid accumulation of mutations particularly in repetitive DNA sequences that consist of longer stretches of mononucleotides (mononucleotide repeats). Since these sequences are also commonly referred to as microsatellites, this latter mechanism of carcinogenesis is referred to as microsatellite instability. MSI is observed in up to 15% of colorectal cancers, a subset of endometrial cancers, and a number of urinary tract cancers. But it is also found in a number of endometrial and bladder cancers, leukemias and lymphomas, and skin cancers. It is the hallmark of an inherited cancer predisposition syndrome referred to as hereditary non-polyposis colon cancer syndrome. HNPCC syndrome is characterized by inherited mutations of defect copies of genes that encode components of the DNA-MMR complex. MSI may, however, also occur in sporadic cancers without distinct inherited background. In most of these cases accidental failure of the epigenetic regulation of the genes that encode the components of the MMR complex may fail usually due to altered methylation of the respective promoter sequences.
The third emerging major mechanism of carcinogenesis is referred to as CpG island methylator phenotype . 12 Cancer cells that emerge in frame of this molecular phenotype initially experience alterations of the epigenetic control mechanisms that tell genes of specific cells where and when they should be active or silent. In many instances this is regulated by methylation of specific sites within the DNA by the addition of methyl groups particularly to CpG sites frequently found throughout the whole genome. The addition of these methyl groups either completely blocks the expression of the respective gene by condensing its chromatin structure or may just modify the binding properties of activating or inhibitory factors that activate or repress transcription of the respective gene. The CIMP phenotype is clearly the less well-characterized cancer phenotype so far. Because of the basic mechanism it may end up finally in cancer cells that impress as CIN phenotype due to the transcriptional inactivation of genes involved in chromosomal homeostasis or alternatively as MSI cancers due to the inactivation of genes that maintain the DNA-MMR functions; here the most pronounced example is the hMLH1 gene frequently inactivated via the CIMP phenotype in MSI-positive sporadic colorectal cancers.

Carcinogenesis Induced by Papillomavirus Infections
Human papillomavirus infections play the predominant carcinogenic role for cancers of the lower female genital tract, and in particular cervical cancer. 13 Molecular pathways of how these viruses contribute to neoplastic transformation of epithelial cells have to a large extent been clarified and thus will be considered in more detail in the following paragraphs.
Published reports on the concept that infectious agents may be involved in the pathogenesis of cancers of the female lower genital tract dates back to the middle of the nineteenth century. Domenico Rigoni Stern described his observation that women with frequent sexual contacts with various partners are at substantially higher risk to develop cervical cancer than women who did not have sexual contacts. Since then sexually transmittable agents that may explain this peculiar epidemiological feature of these cancers have been extensively investigated. It lasted until the end of the 1970s when the first truly important clues that made it possible to delineate the formal molecular pathogenesis of cervical cancers were put forward. In 1976 Harald zur Hausen published his hypothesis that cervical cancer and its precursor lesions may be caused by agents similar to those that cause hyperproliferative lesions in the genital tract, the condylomata acuminata or genital warts. 14 This hypothesis initiated an intense chase to track down the putative agents. Genomes and viral particles of a new type of human papillomavirus were identified in biopsies of genital warts and labeled as HPV 6. 15 Shortly later a related HPV type was cloned from DNA samples isolated from laryngeal papillomas and referred to as HPV 11. 16 This HPV type showed substantial homologies with HPV 6 and in subsequent studies both HPV types were found in laryngeal as well as in genital papillomas. 17
Both HPV types were used as probes to look for related DNA sequences in DNA extracted from tumor biopsies and cell lines derived from cervical cancers in further hybridization experiments. These experiments led to the identification of related but clearly distinct HPV sequences in cervical cancer cells. Subsequent cloning and detailed characterization has revealed that these sequences are indeed new types of the HPV group that have since then been referred to as HPV 16 and 18. 18, 19

Basic Structure of the Virus and Its Genome
Human papillomaviruses are small viral particles that constitute viral capsids built up by self-assembling proteins encoded by the L1 and L2 genes of the virus ( Fig. 1.12 ). They lack an envelope and are thus relatively resistant toward environmental hazards. Thus viral capsids measures about 55 nm in diameter and include an about 8,000-bp-long circular episomal genome that is highly twisted (supercoiled circular genomes). This circular genome encompasses three major genetic and functional regions:
• Early region E, which includes about eight different genes (E1–E8);
• Late region L, which includes the two genes that encode the proteins that make up the capsids (L1 and L2); and
• Upstream regulatory regions of the early region (URR), which includes the important regulatory sequences of the early promoter and enhancer that mediates the highly complex regulation of the viral gene expression pattern in their host cells that is so important for all processes related to HPV-related carcinogenesis and that we will discuss later.

Fig. 1.12 Human papillomaviruses. Top: electron microscopy pictures of viral particles and viral DNA. Bottom: Schematic diagram of the viral genome indicating the most viral genes and most important functions.
A second regulatory element that regulates the expression of the late genes is included in sequences that are part of the E7 gene (late promoter).
Papillomavirus types are differentiated in HPV genotypes based on distinct variations of their nucleic acid sequences. A certain stretch within the L1 gene is used as a reference to differentiate different types according to an international agreement on the classification of HPV types. 20
Meanwhile more than 120 different HPV types have been characterized, but is estimated that the true number of papillomaviruses that may infect humans exceed 200. Papillomaviruses are strict epitheliotropic viruses that exclusively infect epithelial cells of the outer surfaces of the human body. Most of the HPV types infect the cutaneous parts of the skin (skin types), whereas about 40 HPV types are preferentially found in lesions in the mucosal surfaces in the lower anogenital tract (mucosa types). 21 Among the mucosa types two different classes of HPV types are distinguished ( Table 1.1 ): (a) the low-risk HPV (LR-HPV) types that cause massive exophytic, hyperplastic wart-like lesions, and (b) the high-risk HPV (HR-HPV) types associated with cancer particularly of the uterine cervix ( Fig. 1.13 ). The latter, however, usually cause only very minor lesions without massive hyperplasia. These lesions rarely exceed the surface of the epithelium in the early stages of the infection. They commonly occur and regress without the infected person realizing the infection. Thus although these infections occur in men and women apparently with the same incidence, clinical consequences that would trigger follow-up usually only occur in women who have developed cervical lesions as part of a carcinogenic process, whereas in men the infections usually occur and regress unnoticed.
Table 1.1 Correlation between phylogenetic and epidemiologic classification of mucosal HPV types Phylogenetic classification Epidemiologic classification   High-risk HPV types Low-risk HPV types High-risk HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 68, 82, 26 ∗ , 53 ∗ , 66 ∗ 70 Low-risk HPV types 73 6, 11, 40, 42, 43, 44, 61, 72, 81, CP6108
The epidemiological classification of these types as probable high risk is based on zero controls and one to three positive cases. 21
∗ Putative high risk types.

Fig. 1.13 Different types of HPV infection. (1) A state of very low viral activity shortly after initial infection through microlesions of the epithelium ("latent phase"). (2) The replicative infection characterized by strong viral gene expression and viral capsid formation in the upper differentiated epithelial layers. (3) The transforming infection characterized by deregulated oncogene expression in the replication competent basal cells. CIN, cervical intraepithelial neoplasia, grade 1–3.
Papillomaviruses infect the basal cells of the epithelium via binding to certain cell-surface glycosaminoglycans expressed on the basal cells of the epithelium ( Fig. 1.14 ). Once they have entered the cell, the viral capsids are broken down and the episomal viral genome is released in the nucleus. Viral early genes are expressed at this stage on a very low and highly controlled level that allows for low copy replication of the genomes on the order of 10 to 50 genomes per infected basal cell; however, massive amplification of the viral genome or even replication of the virus does not occur at this stage of infection in basal cells. Only in certain instances not yet characterized in detail does high-level gene expression of the viral early genes occur in differentiated cells of the intermediate layers of the epithelium. This triggers amplification of the viral genome. Once the cells have reached the superficial cell layers the early genes are shut off and high-level expression of the late genes (L1 and L2) occurs (early–late switch). This results in expression of the respective proteins, packaging, and self-assembly of new mature viral particles that are finally released from the cells once during the normal differentiation process the infected and virus-producing keratinocytes decay into keratin fragments. The replication strategy of the human papillomaviruses is therefore tightly linked to natural differentiation processes of their host cells. For the virus this has two advantages.

Fig. 1.14 Schematic representation of the different phases of an HR-HPV infection: After infection of basal cells of the epithelium the virus may persist in a latent state. Upon differentiation and maturation of the cells, the viral genome may be replicated and novel virus particles are produced and released at the surface of the epithelium. This stage is often characterized by typical koilocytes as a morphological hallmark of virus production. If transcriptional control of the viral genome in the basal and parabasal cells fails, expression of the viral oncogenes in these replicating cells may induce chromosomal instability and thus initiate transformation. Later in the progression to invasive cancers the viral genome often becomes integrated into the host cells' chromosomes as a sign of increasing chromosomal instability.
First, the biology of the primary host cells at the bottom of the epithelium that retain the capacity to multiply and generate new cells is only marginally altered by the infection. HPV infections induce no cytolysis, no inflammation, or other tissue damage in the basal cell compartment of the epithelium. Acute LR-HPVs induce more proliferation of the infected basal cells and cause exophytic lesions that may clinically impress as warts or condylomata seen with comparable incidence in men and women. Proliferation of basal cells is induced by acute HR-HPV types to much less extent since a simple infection almost never causes exophytic lesions. Usually these infections regress spontaneously and clinically unrecognized.
Secondly, due to the fact that the virus is only produced on the very superficial surfaces of the infected epithelium there is very limited contact between viral antigens and the immune system of the infected host. Thus the acute infection induces very little humoral immune responses and serum antibodies are only observed in low titers in some of the infected individuals. They do not induce a protective humoral immunity. Cellular immune responses are only weakly activated, which usually takes a long time during which the virus can multiply and spread until reliable cytotoxic immune functions have been activated to defeat the virus and the lesions it has caused.
The tight association of the replication strategy of the papillomaviruses with the differentiation status of their host cells further allows the virus to multiply and spread with a highly restricted amount of their own genetic information. Complex genetic features that control the restricted expression pattern of the virus during the different differentiation stages are contributed apparently by the host cell and not by the virus; thus the virus has no need to retain genes that may mediate these functions by themselves.

Epidemiology of HPV Infections
HR- and LR-HPV infections are usually acquired in the early years upon uptake of sexual activity. Most studies have been performed in women, but it can be extrapolated that the infection pattern in men is not substantially different from that in young women.The highest infection rates are seen in young women at 18 to 25 years of age. 22 Over time the incidence of HPV infections decreases, but there seems to be a second peak of HPV infections in older women of more than 45 years of age. The infection clearly shows a typical sexually transmitted pattern and depends on personal lifestyle habits, for example, the number of sexual partners, and the frequency of sexual contact with partners who have themselves sexual contact with other partners. Several studies have shown that the cross-sectional incidence of HR-HPV infections ranges between 3 and 25% of the female population. The cumulative infection in average women is calculated to be more than 50% once in their lifetime. Thus a tremendous amount of women (and hence also men) are infected with these agents.
The HPV types 16 and 18 are found in about 70% of all cervical carcinomas, whereas among healthy women the remaining HR-HPV types are more common. This observation suggests that women infected with the two types, HPV 16 and 18, may have a higher risk to developing high-grade precursor lesions or even invasive carcinomas than women who are infected with the other HR-HPV types. 23 This notion is strongly supported by a variety of studies that indicate that women infected with HPV 16 and 18 are more likely to develop high-grade lesions (cervical intraepithelial neoplasia (CIN) 3) in less time than women infected with other HR-HPV types. Moreover, the time of progression of high-grade lesions to invasive carcinomas appears also to depend on the HPV type that causes the high-grade lesions. 23a
HR-HPV infections usually last for several months and in most instances regress spontaneously without causing relevant clinical lesions. 24 The acute HPV infection may impress as CIN 1 lesions in histological sections in that the typical koilocytes indicate the acute replication of the virus in intermediate and superficial cells of the infected epithelium. These acute infections strictly adhere to the regulatory gene expression pattern outlined earlier in that the cellular differentiation stage determines the expression of the viral genes in the epithelium. According to this, limited and highly regulated gene expression is found in basal and parabasal cells of the epithelium. By this replication strategy, HPVs successfully avoid expression of viral genes in proliferation competent epithelial cells that may become transformed stem cells for a cancer. Under "normal" conditions viral genes are only expressed in cells that are irreversibly cell cycle arrested. HPVs thereby can multiply almost unnoticed by the host and spread to many other individuals ( Fig. 1.14 ).

The Role of the HR-HPV E6 and E7 Genes
The causal relationship of HR-HPV types and cervical cancer is based on four major experimental observations:
1. In more than 99 % of cervical cancers genomes of HR-HPV types can be found;
2. Two viral gene, E6 and 7, are in all cervical cancer preserved and expressed;
3. The E6 and E7 genes can transform primary epithelial cells into neoplastic cell clones; and
4. If the expression of these genes is inhibited cervical cancer cells stop to proliferate and lose the neoplastic growth properties.
These facts clearly underline the central role of the viral E6 and E7 genes for the HR-HPV mediated transformation processes.
During the past 20 years basic research on the biochemical activities of both viral proteins have revealed three major aspects that underline their oncogenic activities: the E6 protein of the oncogenic HPV types induces premature degradation of the p53 tumor suppressor gene and thus interfere with its proapoptotic functions 25 ; the E7 protein induces destabilization of the retinoblastoma protein complex and thus allows the cell to evade cell cycle control at the G1/S phase transition thorough the pRB pathway 26 ; and both genes interfere with centrosome synthesis and function that results in desegregation of the chromosomes during mitosis and numerical and structural chromosomal aberrations. 27 As a result of chromosomal instability induced by E6 and E7, HPV genomes may become integrated into the host genome. 28 Both viral oncoproteins interact with many more proteins of epithelial cells that are summarized in Fig. 1.15 . These interactions further support that central role of E6 as anti-apoptotic protein and of E7 as proliferation inducing and cell cycle deregulating factor.

Fig. 1.15 Interactions with cellular proteins and oncogenic functions of the HR-HPV proteins E6 and E7.

Progression of HPV-Infected Epithelial Cells to Invasive Cancer Cells
The progression of pre-neoplastic lesion of the uterine cervix to invasive cancer cells is a well-described process usually subdivided into three main stages referred to as cervical intraepithelial neoplasia 1, 2, and 3 ( Figs. 1.13 , 1.14 ). This classification is based on histopathological criteria that are essentially describing the pathogenic effects of persisting HR-HPV infection in the cervical epithelium. CIN 1 is characterized by minor activation of the proliferation rate of the basal cells, whereas the proliferating cells are still restricted to the lower third of the epithelium. The abundant presence of koilocytes in the more superficial cell layers document the massive production and release of HPV in the more superficial cell layers. In CIN 2 lesions the proliferating cells extend to the two lower thirds of the epithelium. In these lesions the number of koilocytes decreases gradually. CIN 3 lesions are characterized by proliferating cells that extend now into the upper third of the epithelium or in case of the carcinomata in situ lesions (CIS) extend through the full thickness of the epithelium. These diagnostic categories are very useful in clinical practice since they make it possible to subdivide precancerous stages according to their likelihood to progress to invasive cancer. However, since they do not make it possible to visualize the molecular events induced by the deregulated expression of the viral oncogenes in these cells, they do not formally make it possible to subdivide the preneoplastic lesions according to the molecular events involved in the carcinogenic processes.
As discussed above, the expression of HPVs is tightly regulated in basal and parabasal cells of the epithelium. Thus, the initiation of the carcinogenic process is not the infection of basal epithelial cells by HR-HPVs but rather the emergence of epithelial cell clones that fail to downregulate the expression of the viral oncogenes in the basal cell layer. 29 According to this model, preneoplastic HPV-induced lesions can be subdivided into acute virus-producing lesions (CIN 1, acute or replicating HPV infection) and transforming HPV infections that emerge from CIN 1 lesions but may progress via CIN 2 finally to CIN 3/CIS lesions into invasive cervical carcinomas. 30 In terms of the relative risk that each of these different stages carries for progression, this revised molecular model of the pathogenesis of the HPV-induced cancers strongly suggests that as long as no deregulated expression of the viral genes has occurred in the basal cells, the risk for progression to high-grade lesions or even cancer is comparably low. This situation dramatically changes once individual HR-HPV infected basal cells lose the control over the expression of the viral genes and start to express the viral E6 and E7 genes that then may initiate their deteriorating activities and initiate the carcinogenic process. Expression of the HR-HPV type E7 protein interferes with the regulation of the G1/S phase control of the cell cycle and permits proliferation of cells even if the complete cell cycle machinery is not yet prepared to replicate the DNA of the chromosomes. This results in fixation of unrepaired mutation but also in damage of the genome by itself. In normal cells this would have been counteracted by the activation primarily of p53 mediated control mechanisms that either would have resulted in stoppage of cell cycle progression or, if the genomic damage is already too severe, induction of the cellular suicide mechanisms (apoptosis) that prevent survival and expansion of cells with damaged genomes. During the normal life cycle this never occurs since the expression of the viral early genes is restricted to terminal differentiated cells of the intermediate or superficial cell layers ( Fig. 1.14 ).
Cells that display chromosomal instability under the influence of the papillomavirus genes E6 and E7 rapidly accumulate numerous morphological alterations. First aberrant mitotic figures are seen in dividing cells, then enlarged nuclei with altered chromatin structures appear, and finally severe changes of the DNA content of the cells results in aneuploidy and anisonucleosis. All these criteria are classical features that build up the cytological classification system to score the degree of abnormalities induced by HPV infections in cervical cells.
One important aspect in the pathogenesis of HPV-induced cancers in the female anogenital tract is its typical anatomical restriction to the transformation zone of the uterine cervix. Although HR-HPV infection occur at multiple sites in the male and female anogenital tract, the epithelial cells of the transformation zone of the cervix appear to be substantially more sensitive to HPV-mediated transformation than other HPV-infected epithelial cells in the vagina, vulva, outer surface of the cervix, and particularly the epithelium of the penis, which only rarely become transformed. These peculiar epidemiological features point out important molecular differences in these different host cells. It is highly likely that these differences rely on the stringency of the molecular control that prevents the activation of the E6 and E7 oncogenes in the basal cells of the epithelium. Given the fact that the molecular features that mediate this control are an inherent part of the differentiation control system of the epithelial cells, it can be speculated that the need for epithelial stem cells at the transformation zone of the uterine cervix plays an important role. Stem cells of the transformation zone retain two potential pathways: to differentiate under the influence of high estrogen levels into a single-layer glandular epithelium, and to differentiate under the influence of low estrogen levels into a multilayer squamous epithelium. These morphological differences are regulated by the activation of different sets of genes and hence different epigenetic regulatory mechanisms that become activated in either situation. There is accumulating evidence that these epigenetic control mechanisms are also used by the papillomavirus genome to either restrict or activate the expression of their genes. The multipolar differentiation capacity of the epithelial cells at the uterine transformation zone may thus explain why these cells are so much more sensitive to transformation by HR-HPV types than many other infected cells of the human body.
Once the activation of the viral oncogenes in the basal cells of the infected epithelium has occurred and the first transformed cells that display chromosomal instability have expanded, the lesions may progress from a CIN 1 lesion (representing acute replicating HPV infections) to CIN 2 and 3 lesions (reflecting transforming HPV infections) and finally invasive carcinomas ( Fig. 1.14 ).
The knowledge now gathered on this important step of the HPV-associated transformation process offers new targets for screening and diagnosis of cervical precancer and cancer. Activation of the expression of the HR-HPV E6 and E7 genes in the basal and parabasal cells results in overexpression of a cellular protein, p16INK4a. 31 - 35 All cells that express HR-HPV proteins and retain the capacity to proliferate express extremely high levels of the p16INK4a protein; hence p16INK4a is an interesting surrogate marker for HPV transformed cells.
Based on these findings it can be expected that in the near future these new techniques will have a deep impact on revised cervical cancer screening programs.
Research over the past thirty years has thus made it possible to identify an infectious agent (HR-HPV) as cause of a major human cancer. The experimental analysis as well as epidemiological research has made it possible to clarify the role of these viruses in carcinogenesis. Detailed analysis of the expression and function of viral genes has revealed that it is not the acute infection by HPVs but rather the lack of normal cellular control functions that suppress their expression in basal cells that initiates the transforming infection and may result finally in full neoplastic transformation of individual infected cells. This research has paved the way for substantially improved diagnostic and therapeutic tools, including novel assays for cervical cancer screening as well as major global vaccination programs that aim to prevent the primary infection of humans by HPV and were shown to effectively prevent induction of HPV-induced preneoplastic or neoplastic lesions.

Concluding Remarks
The scientific analysis of cells and their components as the essential units of complex organisms has initiated a dramatic development in biomedical research and finally led to the identification of the essential biochemical basis of inheritance and diversity of life. It has initiated a paradigm shift in biology from a purely descriptive science to one that has made it possible to fully explore the construction kits of life. Genes have been recognized as the basic units of inheritance and their activity in given cells has been realized as the determinants of a distinct biological phenotype. The biological and chemical principles that control the activities of genes within cells have been explored and have revealed why and when certain cells adopt distinct functions. Since the expression of certain sets of genes also determines the shape of cells, functional irregularities and errors are frequently indicated by altered morphology. In particular early changes associated with neoplastic transformation of cells are indicated by initially discrete and later substantial and prominent morphological alterations of cells. Early detection of cancer and many other diagnostic applications in medicine are based on these fundamental concepts and have built the cornerstones of pathology and cytopathology.
As achievements in understanding genes and their functions in cellular biology and pathology have progressed, distinct markers or proteins encoded by defined genes that indicate specific alterations have been identified. In recent years this concept has been of particular importance for the area of oncology, where the (over)expression of certain proteins has been found to be associated with the initial events that finally result in cellular transformation and outgrowth of cancer cells. Neoplastic transformation of cells has been realized to be a malfunction of genes that control growth and differentiation. Basic principles of this malfunction have been elucidated. Desegregation of chromosomes during cell division, lack of genome surveillance during the replication of the genetic code, and activation of viral (e.g. papillomavirus) oncogenes in cells have been identified as causes of neoplastic transformation. Based on these fundamental concepts, functional genomics and proteomics have revealed the expression of distinct patterns of proteins in specific disease conditions. Such proteins are now usually referred to as biomarkers. The detection and evaluation of biomarkers is becoming a more and more important part of pathology and cytopathology as it makes it possible to track down morphological alterations of cells to the functional level of genes. The contribution of certain carcinogenic human papillomavirus types to carcinogenesis particularly of cells at the uterine transformation zone is a good example for this paradigm and exemplifies the diagnostic potential that these new concepts hold for the future.

References

1. Lander E.S., Linton L.M., Birren B., et al. Initial sequencing and analysis of the human genome. Nature . 2001;409(6822):860-921.
2. Pich A., Chiusa L., Margaria E. Prognostic relevance of AgNORs in tumor pathology. Micron . 2000;31(2):133-141.
3. Fischer A.H., Taysavang P., Jhiang S.M. Nuclear envelope irregularity is induced by RET/PTC during interphase. Am J Pathol . 2003;163(3):1091-1100.
4. Alirol E., Martinou J.C. Mitochondria and cancer: is there a morphological connection? Oncogene . 2006;25(34):4706-4716.
5. Gitai Z. The new bacterial cell biology: moving parts and subcellular architecture. Cell . 2005;120(5):577-586.
6. Ji J., Chen X., Leung S.Y., et al. Comprehensive analysis of the gene expression profiles in human gastric cancer cell lines. Oncogene . 2002;21(42):6549-6556.
7. Vogelstein B., Kinzler K.W. Cancer genes and the pathways they control. Nat Med . 2004;10(8):789-799.
8. Nigg E.A. Origins and consequences of centrosome aberrations in human cancers. Int J Cancer . 2006;119(12):2717-2723.
9. Kloor M., von Knebel Döberitz M., Gebert J.F. Molecular testing for microsatellite instability and its value in tumor characterization. Expert Rev Mol Diagn . 2005;5(4):599-611.
10. Woerner S.M., Kloor M., von Knebel Döberitz M., et al. Microsatellite instability in the development of DNA mismatch repair deficient tumors. Cancer Biomark . 2006;2(1-2):69-86.
11. Ghadimi B.M., Sackett D.L., Difilippantonio M.J., et al. Centrosome amplification and instability occurs exclusively in aneuploid, but not in diploid colorectal cancer cell lines, and correlates with numerical chromosomal aberrations. Genes Chromosomes Cancer . 2000;27(2):183-190.
12. Weisenberger D.J., Siegmund K.D., Campan M., et al. CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat Genet . 2006;38(7):787-793.
13. zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application. Nat Rev Cancer . 2002;2(5):342-350.
14. zur Hausen H. Condylomata acuminata and human genital cancer. Cancer Res . 1976;36(2 pt 2):794.
15. Gissmann L., zur Hausen H. Partial characterization of viral DNA from human genital warts (Condylomata acuminata). Int J Cancer . 1980;25(5):605-609.
16. Gissmann L., Diehl V., Schultz-Coulon H.J., et al. Molecular cloning and characterization of human papilloma virus DNA derived from a laryngeal papilloma. J Virol . 1982;44(1):393-400.
17. Gissmann L., Boshart M., Durst M., et al. Presence of human papillomavirus in genital tumors. J Invest Dermatol . 1984;83(1 Suppl):26s-28s.
18. Boshart M., Gissmann L., Ikenberg H., et al. A new type of papillomavirus DNA, its presence in genital cancer biopsies and in cell lines derived from cervical cancer. EMBO J . 1984;3(5):1151-1157.
19. Durst M., Gissmann L., Ikenberg H., et al. A papillomavirus DNA from a cervical carcinoma and its prevalence in cancer biopsy samples from different geographic regions. Proc Natl Acad Sci USA . 1983;80(12):3812-3815.
20. de Villiers E.M., Fauquet C., Broker T.R., et al. Classification of papillomaviruses. Virology . 2004;324(1):17-27.
21. Munoz N., Bosch F.X., de S.S., et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med . 2003;348(6):518-527.
22. Lau S., Franco E.L. Management of low-grade cervical lesions in young women. CMAJ . 2005;173(7):771-774.
23. Khan M.J., Castle P.E., Lorincz A.T., et al. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst . 2005;97(14):1072-1079.
23a. Vinokurova S., Wentzensen N., Kraus I., et al. Type-dependent integration frequency of human papillomavirus genomes in cervical lesions. Cancer Res . 2008;68(1):307-313.
24. Ho G.Y., Bierman R., Beardsley L., et al. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med . 1998;338(7):423-428.
25. Mantovani F., Banks L. The human papillomavirus E6 protein and its contribution to malignant progression. Oncogene . 2001;20(54):7874-7887.
26. Munger K., Basile J.R., Duensing S., et al. Biological activities and molecular targets of the human papillomavirus E7 oncoprotein. Oncogene . 2001;20(54):7888-7898.
27. Duensing S., Munger K. Mechanisms of genomic instability in human cancer: insights from studies with human papillomavirus oncoproteins. Int J Cancer . 2004;109(2):157-162.
28. Wentzensen N., Vinokurova S., von Knebel Döberitz M. Systematic review of genomic integration sites of human papillomavirus genomes in epithelial dysplasia and invasive cancer of the female lower genital tract. Cancer Res . 2004;64(11):3878-3884.
29. zur Hausen H. Papillomaviruses causing cancer: evasion from host-cell control in early events in carcinogenesis. J Natl Cancer Inst . 2000;92(9):690-698.
30. von Knebel Döberitz M. New markers for cervical dysplasia to visualise the genomic chaos created by aberrant oncogenic papillomavirus infections. Eur J Cancer . 2002;38(17):2229-2242.
31. Klaes R., Friedrich T., Spitkovsky D., et al. Overexpression of p16(INK4A) as a specific marker for dysplastic and neoplastic epithelial cells of the cervix uteri. Int J Cancer . 2001;92(2):276-284.
32. Klaes R., Benner A., Friedrich T., et al. p16INK4a immunohistochemistry improves interobserver agreement in the diagnosis of cervical intraepithelial neoplasia. Am J Surg Pathol . 2002;26(11):1389-1399.
33. Sano T., Oyama T., Kashiwabara K., et al. Expression status of p16 protein is associated with human papillomavirus oncogenic potential in cervical and genital lesions. Am J Pathol . 1998;153(6):1741-1748.
34. Wentzensen N., Bergeron C., Cas F., et al. Evaluation of a nuclear score for p16INK4a-stained cervical squamous cells in liquid-based cytology samples. Cancer . 2005;105(6):461-467.
35. Zhang Q., Kuhn L., Denny L.A., et al. Impact of utilizing p16INK4A immunohistochemistry on estimated performance of three cervical cancer screening tests. Int J Cancer . 2007;120(2):351-356.
CHAPTER 2 Basic Cytogenetics and the Role of Genetics in Cancer Development

Alain Verhest, Pierre Heimann

Contents
Introduction
Historical Background
Basic Knowledge of Cytogenetics
Cell Cycle
The Interphase
The Mitosis
The Meiosis
The Chromosome Structure
Methodology
The Karyotype
Fluorescent in Situ Hybridization
Comparative Genomic Hybridization (CGH)
Acquired Chromosomal Aberrations in Cancer
Introduction
Lymphomas
Sarcomas
Thyroid Carcinomas
Clinical Applications of Conventional Cytogenetics and Fish in Cytology
Introduction
FISH Strategy
Application
Concluding Remarks

Introduction
This chapter will summarize the knowledge acquired on conventional cancer cytogenetics in the second half of the last century and introduces additional applications of fluorescent in situ hybridization available for the study of cancer development and evolution.
Other indications of these techniques applied on cytology samples are also described in Chapter 36 .

Historical Background
As suspected by von Hansemann more than a century ago, cancers are associated with nuclear and mitotic anomalies in their cells.
In 1914, Boveri hypothesized his theory on somatic mutations responsible for the origin and development of malignant transformation. He stressed the acquisition of an unbalanced chromosome constitution as a cause of cancer illustrated by mitotic asymmetry and asynchrony, and foresaw the monoclonal origin of the cancer cell. It took at least 40 more years to establish the exact number of human chromosomes. The blood-culturing method became more successful than the squash method when colcemid was discovered to arrest the mitotic cycle in metaphase by poisoning the mitotic spindle and to prevent the centromeres from dividing. The erroneous adjunction of a hypotonic solution to a pellet of harvested cells was an unexpected improvement in the spread of individualized chromosomes rid of their cellular envelope, resulting in a nicer dispersal on the metaphase spread.
In 1956 Tjio and Levan accurately reported that the human somatic cell contains 46 chromosomes, including 22 pairs of autosomes and one pair of sex chromosomes; one X of maternal origin and the other chromosome—X or Y—being from the paternal source. 1 Rarely have discoveries had such impact on modern biology and medicine as the description of the 46-chromosome karyotype. The newborn cytogenetic discipline investigated simultaneously the field of inherited diseases and acquired chromosomal anomalies in cancer cells.
Trisomies of chromosome 21 in mongolism and of other autosomes or numerical variations of sex chromosomes proved their specificity and consequently their diagnostic value in congenital syndromes. In 1960 Nowell and Hungerford reported the first evidence of a chromosome anomaly specifically associated with a malignant disease, the chronic myelogenous leukemia. 2 They showed the recurrent presence in leukemic leukocytes of a deleted small chromosome that they named the Philadelphia (Ph) chromosome in reference to the city where they were working. This proof of a genetic cause in cancer was the starting point to new insights into the pathways of malignant initiation and progression.

Basic Knowledge of Cytogenetics
The human somatic cell contains two copies of each chromosome, one from paternal and the other from maternal origin. Therefore the karyotype is diploid with doubled amount of deoxyribonucleic acid (DNA) (2 n ) compared to the gametes ( n ) with a single set of 23 chromosomes.
The first step is to review the different stages of the cell cycle which are essential to the acquisition of chromosomes suitable for karyotyping.

Cell Cycle
The cell cycle is a process of successive cell divisions (mitosis) interrupted by so-called “resting” periods (interphase). Actually, the resting cell is very active metabolically with continuous molecular interactions between DNA, ribonucleic acid (RNA), and proteins.

The Interphase
The interphase is the period wherein the cell is in a nondividing state and can be at different stages: the first gap (G1) is between the last mitosis and the S-phase (phase of DNA synthesis) and the second gap (G2) is between the completion of the S-phase and the next mitosis (M). The mitotic division occupies only a short time in the cell cycle. If the cell reaches its ultimate stage of differentiation and will not divide anymore, the cell is said to be in phase G0 of the cycle. G0 applies also for those cells that have temporarily stopped dividing ( Fig. 2.1 ).

Fig. 2.1 Schematic representation of the cell cycle with the four sequential phases (see text). The cell cycle checkpoints are located at the G1/S and G2/M transitions.
During the G1 phase, the cell is metabolically active and requires many organelles for protein synthesis while acquiring the potential for the DNA-doubling process. The duration of the entire cycle depends on the time of the G1 phase, which varies according to different conditions and tissue types. G1 phase may last from only a few hours to weeks or months, depending on the mitotic rate of the tissue. The phase of DNA synthesis (chromosome replication) has a duration of approximately 8 hours. The replication is not homogeneous throughout the genome, and asynchronism of replication occurs, particularly in the synthesis of the heterochromatin composing the inactivated X chromosome.
DNA replication is achieved when all the chromosomes are duplicated in two identical sister chromatids with the consequence that the total amount of DNA is now doubled compared to the normal 2 n value of the interphase nucleus. The following phase, G2, takes about 4 hours and accumulates the cytoplasmic organelles necessary to complete the mitosis.
This step-by-step progression is controlled by a series of checkpoints which stop the process if the previous phase is not achieved. Different proteins act sequentially on the cell cycle: the cyclin-dependent kinases (CDKs), the cyclins, and the CDK inhibitors (CKIs).
Activation of kinases by cyclins positively regulates the cycle by allowing the cell to enter the successive phases. If the quality of DNA synthesis is impaired, CKIs would automatically stop the process and drive the cell to apoptosis.

The Mitosis
Although the cell cycle is a continuous process, mitosis has four distinct phases ( Fig. 2.2 ).

Fig. 2.2 The four different phases constituting the mitotic process (cytokinesis being included in the telophase).

Prophase
Condensation and fragmentation of the chromatin into chromosomes becomes evident. The nucleolus vanishes and the centrioles, replicated in G2, migrate to opposite poles of the cell. Each chromosome is still attached to the nuclear membrane and composed of a double strand of sister chromatids. A constricted area called centromere becomes apparent on the chromosomes and the nuclear membrane disintegrates.

Metaphase
The chromosomes are aligned at the equatorial plate of the mitotic spindle and attached by their centromere to the network of microtubules. Metaphase chromosomes are composed of two sister chromatids joined together by the centromere.

Anaphase
The centromeres are split into two parts and both strands of the sister chromatids are attracted to opposite pole by shortening of the spindle fibers. The chromosomes, pulled apart, are clustered at each pole of the cell.

Telophase
Telophase results in the formation of a nuclear membrane. The constriction of the cellular membrane starts the division of the cytoplasm (cytokinesis). The chromosomes progressively melt back into a chromatin network. At the end, both daughter cells have the same number of chromosomes as the maternal cell.

The Meiosis
The meiosis is a more complex process by which the gonad cell undergoes two cellular divisions.
The meiosis I follows stages similar to the mitotic division. During prophase I, each chromosome is duplicated. Chromatid exchanges occur between paired homologue chromosomes which are linked together by their sites of junction: the chiasmas. This process, called “crossing over,” results in genetic recombination, with the consequence that genomes between maternal and daughter cells will not be strictly identical. Anaphase I start with the migration of homologue chromosomes to the opposite poles of the cell without splitting of their centromere. Meiosis II arises without previous DNA synthesis and produces the longitudinal separation of the two chromatids, thereby reducing the cell to a haploid n number of 23 single-stranded chromosomes. The fecundation of the ovule by the spermatozoid will restitute the diploid value of somatic cells and provide a complete zygotic genome.

The Chromosome Structure
The chromosomes are composed of DNA and associated histone and non-histone proteins.
This combination, called chromatin, is individualized into visible chromosomes only during mitosis. The double helix of DNA described by Watson and Crick is supercoiled around protein cores in a complex structure of nucleosomes. Compacted nucleosomes constitute chromatin segments of approximately 30 nm in diameter observable in electron microscopy. Further condensation makes it optically identifiable as heterochromatin in the interphase and as chromosomes at the late prophase. An animation on cell division and chromosome structure can be found at http://www.johnkyrk.com . 3
The extremities of the chromosomes are called telomeres. They preserve the integrity of chromosomal extremities by allowing replication to occur without loss of coding sequences, but undergo repetitive shortenings themselves after each cellular division. The so-called “mitotic clock” counts the number of cell divisions that have occurred and pushes the cell to apoptosis before a critical telomeric shortening is reached. If this should occur, chromosomes would be prone to fuse end to end, giving rise to sticky ends that would favor mitotic aberrations and promote the accumulation of subsequent genetic rearrangements, possibly leading toward the first crucial steps in the development and progression of neoplasia. 4

Methodology

The Karyotype
In the 1950s and 1960s, human chromosomes were studied with Giemsa or Wright stains, making it possible for these chromosomes to be counted accurately and grouped together according to their length and the position of the centromeric constriction. The 22 pairs of autosomes and the sex chromosomes were thus classified into seven groups, A to G. The largest pairs are numbered 1 to 3 in group A. The centromere is located in the middle of chromosomes 1 and 3 and displaced in a submetacentric position in pair 2. Group B is composed of pairs 4 and 5, both with a subtelomeric centromere. Group C is the largest and is composed of medium-sized chromosomes including pairs 6 to 12 and chromosome X. Most of them are submetacentric and roughly classified by decreasing length. Group D is composed of chromosome pairs 13–15 and characterized by a distal acrocentric centromere. Group E contains the metacentric pair 16 and the submetacentric 17 and 18 sets. Chromosome pairs 19 and 20 are smaller metacentric chromosomes and constitute group F. Group G is composed of small acrocentric chromosomes arbitrarily placed in pairs 21 and 22. The small Y chromosome is included in group G.
Accurate individual classification of chromosomes was rendered possible by the banding techniques developed first by applying fluorescent quinacrine mustard on metaphase preparations. 5 This fluorescent agent reveals transverse bright bands (Q banding) of different intensities along the chromosome arms. Other procedures using trypsin digestion (which removes proteins from chromatin) and Giemsa staining yield dark G bands superimposed on the bright Q bands. This led to a very precise identification of each individual chromosome ( Fig. 2.3 ). Techniques with heat denaturation in saline solution obtained a reverse staining called R bands with optional enhancing of telomeric ends in T banding.

Fig. 2.3 G banded karyotype from a normal male showing 22 pairs of autosomal chromosomes and two X and Y sexual chromosomes
The different banding pattern for each of the 23 different chromosomes allows for a perfect pairing of homologues. The number of bands can be raised up to 800 by the high-resolution staining technique obtained on prometaphase chromosomes. The dark G bands correspond to a compact conformation of the chromatin while the clear bands are composed of rather uncoiled chromatin. The dense Q and G bands are G+C-rich and contain repetitive inactive DNA. Active genes are supposed to be in clear bands; constitutive heterochromatin is located in the pericentromeric regions as revealed by C banding and appears as chromocenters in the nondividing nucleus. Chromosome Y has a unique strong fluorescent appearance visible in the interphase nucleus as a bright dot also visible as a dark C band. With these staining methods, the chromosomes 21, already recognized in the prebanding era because of their known involvement in Down syndrome, remained classified as such, and the minute marker of CML was consequently considered as belonging to pair 22.

The Standardized Reporting
In 1971 at the International Conference in Paris, a nomenclature for banded chromosomes was adopted and extensively explained in a later revised version (ISCN 1985), codifying the way in which all possible numerical and structural chromosome anomalies should be reported. 6
On each banded chromosome pair, the upper arms are designated as p arms ( p etit, meaning “short” in French); the longer arms below are designated as q arms. Regions and bands are numbered starting from the centromere.
According to this nomenclature, the Philadelphia (Ph) chromosome was revealed by J. Rowley as being a t(9;22) translocation, hence the result of a reciprocal translocation between the q arms of chromosomes 9 and 22, with breakpoints positioned in q34 and q11 chromosomal regions, respectively 7 ( Fig. 2.4 ). Rapidly the complexity of neoplasia-associated chromosome aberrations appeared difficult to adequate with the current nomenclature. Two standing committees proposed in 1991 and 1995 new consensus guidelines to suit the description of tumor karyotypes, including fluorescent in situ hybridization (FISH) methodology. However, the new abnormalities reported with an increasing variety of FISH probes and the new confusing subtleties of ISCN 1995 accumulated a greater rate of syntax errors. The heterogeneity of the observations and the variability of the banding resolution made these ISCN nomenclatures not very practical to use for the description of cancer-associated chromosomal abnormalities, and still favors the use of personal simplified nomenclatures. 8

Fig. 2.4 G banded karyotype showing a t(9;22)(q34;q11) translocation corresponding to the so-called Philadelphia (Ph) chromosome. Arrows indicate the derivative chromosomes 9 and 22 involved in the translocation.

Karyotyping
Mitoses suitable for karyotyping are obtained more easily from lymphocyte cultures stimulated to grow by phytohemagglutinin. Cells cultures from amniotic fluid or biopsy of chorionic villi allow antenatal diagnoses. Skin fibroblasts or fetal blood samples from the umbilical cord of stillborns are also suitable. Direct examination of bone marrow or short-term cultures are the techniques of choice for hematological disorders. Short-term tissue cultures took advantage of technical improvements such as methotrexate synchronization or collagenase digestion in the analysis of lymphomas and solid tumors.
Chromosomes are counted and analyzed on slides. For decades, the better metaphase spreads were photographed, and each chromosome was manually cut out before being classified on a sheet of paper. Nowadays, on-screen karyotyping is the commonly used method for routine metaphase analysis. Once acquired by the automated capture device, metaphases can be quickly and accurately presented for chromosome assignment. The CytoVision system (Applied Imaging) used in our laboratory provides classifiers for standard banding methods ( Fig. 2.5 ).

Fig. 2.5 G-banding karyotype compared to the ideogram according to ISCN 1985.

Fluorescent in Situ Hybridization
The principle of in situ hybridization (ISH) is an uncoiling of the double DNA strand by heat denaturation followed by subsequent specific hybridization of the targeted DNA molecules with the complementary labeled DNA probe. By this procedure, ISH detects the precise location of unique DNA sequences directly on the chromosomes. The hybridized sequence is then revealed by two layers of fluorescent or chromogenic-labeled antibodies. FISH is preferred to chromogenic ISH (CISH) because of its higher sensitivity and the greater palette of artificial colors available.
Different types of probes, including the centromeric, the locus specific, and the whole chromosome probes, are described:
• Centromeric probes are the most sensitive. They bind to highly repetitive juxtacentromeric heterochromatin. Their strong signal remains very easily detectable on tissue sections. They are preferentially used for detecting gains and losses of entire chromosomes, namely aneusomies.
• Locus-specific probes are designed to detect unique sequences spanning specific genomic loci. They are used to detect specific gene amplifications, duplications, deletions, or chromosomal translocations. These latter can be revealed by fusion of colors with the use of dual-colour probes flanking chromosomal breakpoints involved in translocations.
• Whole chromosome probes, known as chromosome painting, reveal the whole chromosome except the centromeric region. They are used to identify the origin of chromosomal markers such as ring chromosomes and to refine complex chromosomal translocations.
In congenital diseases, the use of probes has advantageously circumvented the metaphase search for frequent trisomies or microdeletion syndromes. Although banding analysis remains the standard for identifying acquired chromosome abnormalities in cancer, FISH is now used as an easy and reliable technical substitute to search for well-documented specific chromosomal abnormalities in metaphase or interphase cells. Nowadays, FISH is used on a regular basis as a complementary tool to conventional cytogenetics, justifying the term “molecular cytogenetics.”

The FICTION Technique
F luorescent i mmunophenotyping and interphase c ytogenetics as a t ool for the i nvestigation o f n eoplasms (FICTION) is a combination of fluorescent immunophenotyping and in situ hybridization, making it possible to study genetic abnormalities in phenotypically selected cells. 9 The technical principle and main practical applications of this method will be discussed later in this chapter. Suffice to say it is currently used to detect recurrent chromosomal abnormalities in multiple myeloma at diagnosis. This method may be applied to any type of tumor cell displaying a specific immunophenotype but is of little value for minimal residual disease.

Multicolor Metaphase FISH
Multicolor-FISH includes mainly two different methods called spectral karyotyping (SKY) and multiplex-FISH (M-FISH). In SKY, the chromosomes are first stained with a mixture of 24 chromosome-specific painting probes; each one being labeled with a different combination of five fluorochromes. The spectral pattern of chromosomes is then classified using computer software to identify individual chromosomes. M-FISH uses a combinatorial labeling scheme with only five fluorochromes having different emission spectra. Those fluorochromes are similar to that used for SKY but the method for detecting and discriminating the different combinations of fluorescence signals is different.
Both methods are useful in characterizing complex chromosomal rearrangements and in documenting ambiguous marker or ring chromosomes 10 ( Fig. 2.6 ).

Fig. 2.6 Multicolor FISH (SKY) allowing the identification of every human chromosomal pair with an individual color using 24 different painting probes (22 pairs of autosomes plus the two sexual chromosomes). Normal chromosomes are uniform in color, whereas rearranged chromosomes will display two or more colors. This method makes it possible to detect cryptic rearrangements and marker chromosomes in complex karyotypes as demonstrated here.

Comparative Genomic Hybridization (CGH)
This method has the advantage of circumventing the need for tumor cell metaphases. Total genomic tumoral DNA is labeled in green and the normal reference DNA in red. Both differentially labeled tumor and normal DNA will be hybridized together to normal human metaphases and will compete with one another. The ratio of the fluorescent green and red intensities is measured along every chromosome, making it possible to give an overview of DNA sequence copy number changes—gains and losses—in the neoplastic cells mapped on normal chromosomes. CGH is thus able to detect amplified and deleted genomic regions harboring oncogenes or tumor suppressor genes, respectively. The limitation of this method is that it can identify DNA imbalances but not balanced chromosomal translocations (hence, without loss or gain of chromosomal material subsequent to translocation).

Acquired Chromosomal Aberrations in Cancer

Introduction
It has long been agreed that tumor cells carry chromosomal aberrations, but their causes have only recently been more deeply explored. 11, 12 Until the seventies, cytogeneticists were dealing with malignant effusions or long-term cell cultures yielding roughly recognizable multiple chromosome changes, with very large amounts of rearranged DNA in complex aneuploidies. Consequently, this situation led to disillusion in the literature seeded by a plethora of reports with confusing malignant karyotypes, suggesting to most scientists that the chromosomal rearrangements observed were just epiphenomena accompanying the process of malignancy. At that time, no method was able to show molecular changes at the gene level. Karyotype analysis based on banding techniques renewed interest in the characterization of cytogenetic abnormalities in malignant tumors. It appeared evident that nonrandom primary changes involved specific chromosome regions, and were subsequently overwhelmed by secondary more massive variations affecting randomly all chromosomes. This state of overall genomic instability developed during the malignant clonal progression.
Cytogenetic investigations focused initially on leukemias. They identified a constantly increasing number of characteristic chromosomal patterns after the Ph/CML association was detected. The FAB classification of leukemias was consequently enriched by the addition of prototypic karyotypic profiles. Beside leukemias, other cytogenetic and molecular information emerged with studies of lymphomas and sarcomas. In those tumors, relatively simple balanced rearrangements often appeared as fingerprints for a unique tumor type. These specific chromosomal abnormalities were rapidly considered as reliable diagnostic, prognostic, and predictive parameters on daily routine. The importance of chromosomal identification in the diagnosis of human leukemias, lymphomas, and mesenchymal tumors is now recognized as a component of the current subclassifications in the World Health Organization (WHO) fascicules dealing with classification of tumors of hematopoietic and lymphoid tissues and soft-tissue tumors. 12, 13
Nowadays, the well-accepted opinion is that cancer is a genetic disease with two main genetic events triggering cancer initiation: the activation or deregulation of oncogenes as a consequence of point mutation, amplification, or chromosomal translocation; and the inactivation of tumor suppressor genes due to chromosomal deletion, mutation, or epigenetic mechanisms. In malignant epithelial tumors, the prevailing view is that they do not exhibit tumor-specific genetic alteration but rather complex karyotypes with multiple abnormalities shared by carcinoma of different histological subtypes and origins. However, single and specific chromosomal translocations are encountered in some epithelial malignancies such as thyroid carcinoma, kidney carcinoma of childhood and young adult, aggressive midline carcinoma, and a surprisingly great number of prostate cancer. 14
Recurrent and specific chromosome abnormalities can be easily investigated by FISH at diagnosis. The method originally used on metaphase plates is also applicable on nondividing cells (interphase cells) provided by smears, cytospins, or paraffin-embedded tissues. It has proved to be suitable for the detection of numerical deviations on previously stained slides or fresh smears and to be feasible for improving the sensitivity of conventional cytology yielding “atypical cells” in cell suspensions. 15 As it will be illustrated below, FISH may be more sensitive than conventional cytology. FISH combined with cytology can improve the diagnostic sensitivity of detecting malignancy in bronchial brushing and washing specimens. 16 FISH has many more applications in all fields of diagnostic cancer cytology, with significant improvement in tumor classification and a critical value in selection of patients who will benefit from targeted therapies (see Chap. 36 ).
In the following sections, we will review the chromosomal abnormalities observed in lymphoma and sarcoma with their relationship to tumor development. We will mainly focus on specific aberrations that can be used as diagnostic tools in complement to cytology. In the same way, the examples of chromosomal markers in carcinoma will be limited to thyroid carcinoma. In a second part, we will discuss the applications of FISH in the field of cytology, again limiting our comments to lymphoma and sarcoma. The contribution of FISH in multiple myeloma will also be mentioned because of the novel and promising FICTION technique used to detect chromosomal abnormalities in selected nondividing plasma cells.

Lymphomas
Recurrent chromosomal abnormalities in lymphoma are mainly represented by balanced chromosomal translocations that exert their tumorigenic action by two alternative molecular mechanisms ( Fig. 2.7 ). In the first mechanism, the breakpoints on both chromosomes will occur adjacent to two genes and bring them close together but will not alter the protein produced by one of the targeted genes, mainly an oncogene. This latter is translocated close to strong promoter/enhancer elements of the other gene involved, hence the immunoglobulin ( Ig ) or T-cell receptor ( TCR ) genes. The functional consequences are constitutive activation of the oncogene through its overexpression driven by Ig or TCR enhancers. In the second mechanism, the chromosomal breakpoints occur within the coding sequence of each gene, such that the two broken genes are fused, leading to a chimeric gene translated into a new chimeric protein with dysregulated function. The first mechanism accounts for the majority of lymphoma diseases while the second molecular event predominates in sarcoma.

Fig. 2.7 Main molecular mechanisms subsequent to chromosomal translocations encountered in cancer. (A) In the first mechanism, breakpoints on both chromosomes will spare the coding sequence of the targeted genes. The translocation will lead to the juxtaposition of strong promoter/enhancers elements (blue lozenge) from one gene ( A ) with the entire intact coding sequence of another gene ( B ), leading to overexpression of this latter. In the classical example, promoter/enhancers are brought by Ig or TCR coding genes and the targeted coding sequence is oncogenes such as BCL2 or BCL1 . (B) The second mechanism is characterized by chromosomal breakpoints occurring within the coding sequence of both genes involved in the translocation, leading to a chimeric gene translated into a hybrid protein with altered function. White vertical bars denote chromosomal breakpoints.
There is an abundant literature demonstrating good correlations between chromosomal abnormalities and different lymphoma subtypes. 12 Identification of specific genetic aberrations has several meaningful implications in non-Hodgkin lymphoma's (NHL). First, it may help in accurately diagnosing NHL. For example, identification of the t(11;14) translocation makes it possible to distinguish mantle cell lymphoma (MCL) from small lymphocytic lymphoma/chronic lymphoid leukemia (SLL/CLL). The presence of the t(2;5) translocation is the characteristic genetic feature of a subgroup of anaplastic large cell lymphoma (ALCL). Second, demonstration of chromosomal translocations may help in prognostic assessment of NHL; marginal zone lymphoma of MALT type with a t(11;18) is unlikely to respond to antibiotic therapy. By contrast, the MALT-NHL negative for the t(11;18) is most often associated with Helicobacter pylori gastritis and more often responds to antibiotic therapy. The presence of the t(2;5) translocation and its consecutive anaplastic lymphoma kinase ( ALK ) overexpression in ALCL is associated with good prognosis. Third, identification of genetic abnormalities in NHL may serve as markers for staging assessment and for studies of minimal residual disease.
As Hodgkin's lymphoma does not exhibit any consistent or specific genetic abnormality detectable by cytogenetics or FISH analysis, the following topic will focus on non-Hodgkin's lymphoma. Within this last group, we will restrict our talk to NHL subtypes exhibiting characteristic chromosomal aberrations that can be used as diagnostic tools on a regular basis, and will organize this section according to the REAL/WHO morphological classification, hence from small cell lymphomas to large cell lymphomas.

Follicular Lymphoma
Follicular lymphoma (FL) is characterized by the t(14;18)(q32;q21) translocation ( Fig. 2.8 ), which juxtaposes the B-cell lymphoma/leukemia 2 ( BCL2 ) oncogene at 18q21 into the heavy chain immunoglobulin ( IgH ) gene locus at 14q32, leading to upregulated expression of the BCL2 protein. 17 BCL2 is an antiapoptotic gene, and its overexpression leads to prolonged cell survival that may make the cell more vulnerable to additional genetics events, leading to cell overgrowth and cancer. In a minority of cases, variant translocations such as t(2;18)(p11,q21) and t(18;22)(q21;q11), which relocate the BCL2 oncogene to the kappa light chain immunoglobulin ( IgL kappa ) gene locus and lambda light chain immunoglobulin ( IgL lambda ) gene locus, respectively, have also been observed.

Fig. 2.8 Karyotype of follicular lymphoma showing the balanced t(14;18)(q32;q21) translocation (arrows). The gains for chromosomes 3, 12, 15, and X as well as deletion 13q are additional abnormalities associated with clonal evolution.
This translocation t(14;18) and its variants are observed in up to 85% of FL which are mainly represented by histological grades 1, 2, and 3A. The remaining 15% cases do not exhibit a t(14;18)(q32;q21) translocation and are essentially constituted by FL grade 3B. 18, 19 Among them, a minority (~30%) exhibit BCL2 overexpression on immunohistochemistry, resulting from a non-Ig-related mechanism. The origin of this BCL2 gene overexpression is still unknown but could be due to duplication of chromosome 18 as observed in some karyotypes, or could involve other unknown mechanisms favoring BCL2 overexpression. 19 The clinical outcome of this subgroup seems to be similar to that of follicular lymphoma with t(14;18). The major subgroup (~70%) does not show any BCL2 overexpression but presents a recurrent translocation of the 3q27 chromosomal region, resulting in a disruption of the B-cell lymphoma/leukemia 6 ( BCL6 ) oncogene located at this breakpoint. This abnormality is also observed in diffuse large B-cell lymphomas (DLBCL), a feature that will be discussed later. Of interest, these 3q27+ FL grade 3B show peculiar clinicopathologic features distinct from their t(14;18)+ counterparts 20 : a stage III/IV disease as well as a bulky mass are less frequently observed, and they usually disclose a CD10 − phenotype. Finally, this genetic subgroup seems to have a better survival rate and have clinically more in common with de novo 3q27+ DLBCL. 19, 20 These findings indicate that the search for BCL2 and BCL6 rearrangement status by genetic analysis may be clinically warranted for all cases of follicular lymphoma.
Although the t(14;18) translocation is an early event and is critical for lymphomagenesis, it is by itself insufficient to produce FL. As said before, the prolonged cell survival provided by BCL2 overexpression allows the acquisition of further genetic events that contribute to the development of FL. 21 These genetic events occur as a series of chromosomal gains and losses that can be detected at diagnosis as complex and heterogeneous karyotypes. It is not the karyotypic complexity but rather the type of abnormalities exhibited that underlies the varied clinical outcome observed in FL. Recurrent cytogenetic aberrations that have been noted to correlate with a more aggressive disease include chromosomal gains such as +7, +12 or gain of 12q13-14, +18 and chromosomal losses including del 6q, del(9)(p21), and del(17)(p13), the two last aberrations corresponding to loss of tumor suppressor genes p16 and p53 . Beside a complex karyotype, 3q27/ BCL6 translocations can subsequently occur in t(14;18)+ FL, less frequently in low than in high grades, and have been shown to correlate with a risk of transformation to diffuse large B-cell lymphoma. 22

Mantle Cell Lymphoma
According to the REAL/WHO classification, the diagnosis of mantle cell lymphoma should be based on clinicomorphological but also cytogenetic or molecular features. 12 The genetic hallmark of MCL is the t(11;14)(q13;q11) translocation ( Fig. 2.9 ) that juxtaposes part of the IgH locus on chromosome 14q32 to the entire coding sequence of BCL1 oncogene, also named Cyclin D1 or PRAD1, located on chromosome 11q13. 23 BCL1 gene is thus brought under the control of an IgH enhancer, leading to overproduction of cyclin D1 protein, a mechanism similar to that observed for the BCL2 oncogene in FL. Cyclin D1 is one of the key regulators of the cell cycle, and complexes with CDK-4 and -6 in order to promote the G1/S-phase transition of the cell cycle. Increased Cyclin D1 production in MCL will dramatically induce cells to enter the S-phase and, therefore, tumor cell proliferation, by inhibiting the cell cycle inhibitory effects of the retinoblastoma (Rb) and CDK inhibitors p27kip1 proteins. 24 Concurrent disruptions of other cell cycle-associated genes contribute also to the pathogenesis of MCL. In particular, homozygous deletions of the CDK inhibitor p16 INK4 were observed in aggressive variants of MCL. p16 INK4 is an inhibitor of CDK-4 and -6 and thus maintains the Rb protein activity by preventing its phosphorylation. p16 INK4 deletion and an increased level of Cyclin D1 may therefore work together in promoting the G1/S-phase transition in MCL cells.

Fig. 2.9 Karyotype of mantle cell lymphoma displaying the t(11;14)(q13;q11) chromosomal translocation (arrows), associated with multiple additional abnormalities such as interstitial deletion involving one chromosome 13, loss of the normal chromosome 14, and marker chromosome (A). This profile is observed in aggressive cases.
The t(11;14) translocation is very specific to MCL among other B-NHL and is detected by conventional cytogenetics in 60–75% of MCL cases, but this number rises to nearly 100% with the use of FISH.
Beside the presence of t(11;14) translocation, the study of the overall cytogenetic profile brings prognostic meanings. Normal karyotype or karyotype with a single t(11;14) is associated with the typical form of MCL and is a good prognostic factor. In the majority of aggressive cases, t(11;14) is associated with a complex karyotype including numerous structural and numerical alterations of chromosomes 1, 2, 3, 9, 11, 13, 17 as well as unidentified chromosomal aberrations (markers). 25 Also, near-tetraploid karyotypes (hence ±92 chromosomes) seem to be characteristic for the blastoid variant MCL. These karyotypic features occurring in aggressive MCL cases reflect the existence of alterations in both the DNA damage response pathways and mitotic checkpoints that may constitute another important pathogenetic mechanism in this lymphoma subtype. Indeed, one of the most frequently additional cytogenetic aberrations observed in MCL is deletion in the 11q22-23 chromosomal region where the ATM (ataxia-telangiectasia mutated) gene is located. ATM gene plays a key role in genomic stability by activating gatekeeper and caretaker genes such as p53 and BCRA1 in response to DNA damage. 26 ATM inactivation in MCL is associated with a high number of chromosomal alterations, suggesting that it may, at least in part, be responsible for the chromosomal instability in these lymphomas. 26

Marginal Zone B-cell Lymphoma
Several chromosomal abnormalities are encountered in marginal zone B-cell lymphomas (MZL) and are distributed according to the three different subtypes: extranodal MZL of MALT type, nodal MZL, and splenic MZL
In MALT lymphoma, four main recurrent chromosomal translocations have been observed and demonstrate a site-specificity in terms of their incidence: t(11;18)(q21;q21), t(1;14)(p22;q32), t(14;18)(q32;q21), and the recently described t(3;14)(p14.1;q32) ( Table 2.1 ). The latter, limited to MALT lymphoma of the thyroid, skin and ocular adnexa regions, leads to the juxtaposition of the transcription factor Forkhead box-P1 ( FOXP1 ) next to the enhancer region of the IgH gene. 27 This molecular event results in FOXP1 gene overexpression but the pathogenetic relevance of this translocation is still not known.

Table 2.1 Four main recurrent chromosomal translocations observed in MALT lymphomas
The three other translocations affect a common signaling pathway, resulting in the constitutive activation of the nuclear factor-κB (NF-κB), a transcription factor which plays a major role in cellular activation, proliferation and survival. 28 The t(1;14)(p22;q32) is detected in approximately 5% of MALT lymphoma, arising in localizations such as stomach, intestine, and lung. This translocation results in overexpression of the BCL10 gene (chromosome 1p22) due to its juxtaposition with the IgH gene enhancer. The t(14;18)(q32;q21) translocation, cytogenetically identical to the t(14;18)(q32;q21) involving BCL2 gene in follicular lymphoma, is observed in more or less 20% of MALT lymphoma, especially in non-gastrointestinal localizations such as liver, lung, salivary glands, skin, and ocular adnexa. This translocation brings the mucosae-associated lymphoid tissue ( MALT1 ) gene, also involved in antigen-receptor-mediated NF-κB activation, under the control of the IgH enhancer region, with subsequent MALT1 overexpression. The t(11;18)(q21;q21) represents the most common translocation, accounting for 15–40% of cases, and is observed in stomach, intestine, and lung MALT lymphoma cases. It results in the reciprocal fusion of the API2 and MALT1 genes. API2 (cellular inhibitor of apoptosis protein 2) gene is believed to be an apoptosis inhibitor by inhibiting the biological activity of caspases 3, 7, and 9.
The pathogenesis of those three translocations sharing the same molecular pathway is beginning to be understood. 28, 29 NF-κB activation is driven by stimulation of cell-surface receptors, such as B- or T-cell receptors. In unstimulated lymphocytes, NF-κB proteins are bound with inhibitory κB (IκB) proteins and sequestered in the cytoplasm. Phosphorylation of the IκB proteins by the IκB kinase (IKK) heterodimer leads to ubiquitylation and degradation of IκB, allowing NF-κB to migrate to the nucleus and transactivate genes involved in cellular activation, proliferation and survival, and induction of effector function of lymphocytes.
In MALT lymphoma with t(1;14) translocation and BCL10 overexpression, BCL10 is able to complex with MALT1 and trigger aberrant NFκB activation without the need for upstream signaling. With the t(14;18) translocation causing MALT1 overexpression, MALT1 interacts and stabilizes BCL10, leading to its cytoplasmic accumulation. Both proteins in high cellular concentration will then synergistically favor a constitutive NF-κB activity. In t(11;18) positive MALT lymphoma, the API2-MALT1 chimeric protein activates NF-κB through self-oligomerization, and bears a gain of function when compared to wild type MALT1. This higher activation is also due to the API2 protein partner. Indeed, wild-type API2 downregulates BCL10 expression by ubiquitylation and degradation, a mechanism used to regulate BCL10 activity after antigen receptor stimulation. The API2–MALT1 protein is no longer able to ubiquitylate it and high BCL10 expression will synergistically increase API2–MALT1's intrinsic capacity for NF-κB activation, independently of any antigen-receptor activation.
Because of their specificity, the identification of these chromosomal translocations can be of interest for diagnostic purposes. They have also an immediate impact on treatment decisions, at least for two of them. Indeed, a causal relationship between H. pylori infection in the stomach and development of gastric MALT lymphoma has been clearly demonstrated, and 75% of these lymphomas can be successfully treated with appropriate antibiotics targeting H. pylori . 28 However, the presence of either the t(11;18) or t(1;14) translocation defines patients who will not respond to H. pylori eradication. At the opposite, gastric MALT lymphoma without these chromosomal translocations, sometimes carrying trisomies of chromosomes 3, 12, and 18, can be effectively treated by antibiotic treatment, at least at their early stages. However, they can progress, become H. pylori -independent and transform into high-grade tumors following the acquisition of additional genomic alterations (such as TP53 and CKN2A inactivation). Intriguingly, t(11;18) positive MALT lymphomas will rarely develop into high-grade tumors, unlike their t(1,14) counterparts. These clinical features indicate that chromosomal abnormalities in some MALT lymphoma can also serve as prognostic parameters.
In splenic marginal zone lymphoma (SMZL), cytogenetic alterations include mainly partial or complete trisomy 3, and interstitial deletion of chromosome 7q involving segments of variable size, usually centered around the 7q31q32 region ( Fig. 2.10 ). Recent gene expression profiling revealed that genes mapping to the 7q31 chromosomal region were consistently downregulated, among which three were found to be very SMZL-specific: ILF1 , Senataxin , and CD40 . 30

Fig. 2.10 Karyotype showing an interstitial deletion of chromosome 7q as observed in splenic marginal zone lymphoma . In the present case, the deletion involves the 7q22q32 chromosomal segment.
Nodal marginal zone lymphoma (NMZL) is a very rare disease. However, local regional lymph node of MALT lymphoma is virtually indistinguishable from NMZL, requiring clinical information and, in some respect, cytogenetic data to diagnose it. NMZL is characterized by frequent trisomies of chromosomes 3, 7, and 18, but the characteristic translocations of MALT lymphoma are never seen. 12

Small Lymphocytic Lymphoma
The histology, immunophenotypic and cytogenetic features of small lymphocytic lymphoma are indistinguishable from the more common CLL. 12 Chromosomal aberrations observed in SLL include thus trisomy 12, 11q, and 17p deletions—all of them being poor-risk cytogenetic parameters—and a 13q14 deletion which is considered as a marker of good prognosis. A t(14;19)(q32;q13) translocation occurs infrequently in SLL and juxtaposes the BCL3 gene located on chromosome 19 next to the enhancer region of the Ig-heavy-chain gene, leading to BCL3 overexpression. When present, it confers a more aggressive behavior. 31

Lymphoplasmacytic Lymphoma
Lymphoplasmacytic lymphoma (LPL) is a rather uncommon entity but its diagnosis remains challenging for most pathologists. Cytogenetic investigations had previously considered the t(9;14)(p13;q32)—juxtaposing the PAX5 transcription factor with the Ig-heavy-chain gene enhancer—as characteristic of LPL, but more recent studies question the accuracy of this association. Firstly, no PAX5 rearrangement was detected in a series of 13 LPL. 32 Secondly, PAX5/IgH rearrangement was observed in other types of lymphoma including T-cell-rich B-cell lymphoma, post-transplantation diffuse large B-cell lymphoma, and some cases of SMZL. 33

Diffuse Large B-cell Lymphoma
Diffuse large B-cell lymphoma is a very heterogeneous clinicopathologic entity, displaying numerous and disparate chromosomal aberrations. In this section, we will only focus on the most frequent cytogenetic aberrations observed in DLBCL, hence chromosomal translocations involving BCL6 , BCL2 and C-MYC oncogenes.
The translocations involving the 3q27 chromosomal region are the most characteristic and frequent cytogenetic aberrations, detected in 30 to 40% of DLBCL 12 ( Figs 2.11 A and 2.11 B). The 3q27 breakpoint involves the BCL6 gene, which is required for germinal center (GC) formation and the B-cell immune response. The gene partners of the BCL6 chromosomal translocations are multiple. They most often involve the Ig-heavy- or -light-chain (κ and λ) genes on chromosome bands 14q32, 2p11 and 22q11, but more than 20 non- Ig partners have also been described, a phenomenon termed “promiscuous translocation”. 34 Whatever the partner is, the chromosomal translocation brings the entire coding sequence of BCL6 under the control of a replaced promoter that will cause its deregulated expression during B-cell differentiation.

Fig. 2.11 (A) Karyotype of a diffuse large B-cell lymphoma exhibiting a characteristic t(3;14)(q27;q32) chromosomal translocation (arrows). Other abnormalities such as additional material of unknown origin attached to the 1p36 chromosomal region of one chromosome 1, to the 6p24p25 chromosomal regions of both chromosomes 6, and deletion of the 7p21 segment of one chromosome 7 are additional aberrations reflecting clonal evolution. (B) Metaphase FISH with the use of a dual-color (green and red) break-apart probe specific to the BCL6 gene. The yellow signal (juxtaposition of green and red colors) identifies the normal BCL6 gene, whereas splitting of the green and red signals indicates a disruption of the other BCL6 gene subsequent to the t(3;14) chromosomal translocation.
BCL6 plays a key role in the generation of a germinal center by B cells. It encodes a transcriptional repressor protein that downregulates the expression of the B-lymphocyte-induced maturation protein 1 ( BLIMP1 ) gene necessary for plasma cell differentiation, and also the expression of p27kip1, cyclin D2, and P53 which control the cell cycle, apoptosis, DNA repair, and maintenance of genomic stability. 35 In a normal situation, BCL6 expression is tightly regulated during B-cell ontogenesis, being restricted to B cells in the GC. In contrast, the heterologous Ig and non- Ig promoters exhibit a broader spectrum of activity in B-cell ontogenetic stages and will prevent BCL6 downregulation in post-GC cells. A block in the normal downregulation of BCL6 might thus favor differentiation arrest, continuous cell proliferation, survival, and genetic instability, all of which allowing neoplastic transformation. Indeed, the 3q27/ BCL6 rearrangement is sufficient in itself to produce lymphoma as demonstrated by transgenic mice studies. In addition and independently of BCL6 translocations, point mutations and small deletions of BCL6 have been reported in approximately 70% of DLBCL, leading also to its deregulated expression. 35
The clinical relevance of BCL6 gene translocations has been initially a subject of controversy with studies reporting improved survival in patients with BCL6 translocation, and other failing to show any statistically significant impact of such rearrangements on the clinical outcome of DLBCL. 36 More recently, a cDNA microarray analysis demonstrated that DLBCL patients with the germinal center B-cell-like (GCB) gene expression profile had a better overall survival than those with the activated B-cell-like (ABC) expression pattern. 37 As BCL6 is a marker of the GCB-type signature, its mRNA and protein levels were correlated to clinical outcome of DLBCL patients: high-level expression of BCL6 was associated with significantly longer overall survival and shown to be a predictor of a favorable treatment outcome in cases of DLBCL. 36
In some cases, 3q27/BCL6 translocation coexists with other translocations in a single clone, including t(14;18)(q32;21) and t(8;14)(q24;q32), involving BCL2 and c-MYC oncogenes, respectively. This coexistence of two to three chromosomal translocations seems not necessarily to have a significant impact on the clinical features. 38 Finally, it must be added that around 20% of DLBCL exhibit a t(14,18)(q32;q21) similar to that associated with follicular lymphoma and mutually exclusive of BCL6 rearrangements.

Burkitt's Lymphoma
Burkitt's lymphoma (BL) and its leukemic equivalent, the L3 variant of acute lymphoblastic leukemia, are characterized in nearly 90% of cases by a reciprocal chromosomal translocation that juxtaposes the c-Myc oncogene (chromosome 8q24) to one of the immunoglobulin genes located on chromosome 14q32 ( IgH ), chromosome 22q11 ( Igλ ), or chromosome 2p12 ( Igκ ) ( Fig. 2.12 ). All three chromosomal translocations lead to overexpression of the c-Myc gene product. C -Myc gene is a transcription factor that regulates a very large number of genes through heterodimerization with the partner protein Max. 39 The genes targeted by the c-Myc/Max heterodimer complexes are involved in cell proliferation, differentiation, and apoptosis. Such global transcriptional regulatory function may explain why c-Myc overexpression is sufficient in itself to promote lymphoma diseases as demonstrated in transgenic mice studies.

Fig. 2.12 Karyotype showing a t(8;14)(q24;q32) chromosomal translocation (arrows) characteristic of Burkitt's lymphoma (or ALL L3). Segmental duplication of chromosome 1q and loss of chromosome 17p are recurrent additional chromosome aberrations in this type of lymphoma.
The so-called “Burkitt-like” form is characterized by three cytogenetic categories: one with an 8q24/ c-MYC translocation, a second with associated 8q24/ c-MYC and 18q21/ BCL2 translocations, and a third with miscellaneous rearrangements, frequently including an 18q21/ BCL2 chromosomal translocation.
Recurrent chromosome aberrations associated with the 8q24 translocations include duplications of the 1q21q25 chromosomal region, 6q11q14 and 17p chromosomal deletions, and trisomies for chromosomes 7, 8, 12, and 18. A recent cytogenetic and CGH study on BL has demonstrated that the presence of abnormalities on chromosome 1q (demonstrated either by cytogenetics or by CGH) and gains of 7q (ascertained only by CGH) were associated with adverse prognosis. 40

Anaplastic Large Cell Lymphoma
Anaplastic large cell lymphoma is a CD30+ T-cell NHL that can be divided in two majors groups according to the WHO classification: (1) systemic nodal ALCL and (2) primary cutaneous ALCL. As this latter group does not exhibit specific chromosomal alteration, it will not be pursued further in this review. In this section, we will only focus on systemic nodal ALCL, more particularly on anaplastic lymphoma kinase-positive ALCL where a characteristic t(2;5)(p23;q35) translocation is observed in approximately 60% of cases. This translocation fuses the nucleophosmin ( NPM ) gene on chromosome 5q35 to the ALK gene on chromosome 2p23, leading to the NPM-ALK chimeric gene. 41 It is present in approximately 75% of ALCL with ALK gene rearrangement. In the remaining approximately 25% of cases, 2p23/ALK locus translocates with various partner genes. 41 The common molecular features of all ALK rearrangements is the fusion of the ALK tyrosine kinase domain to the 5′ region of partners which provide a strong promoter and most likely an oligomerization motif allowing constitutive activation and aberrant expression of the ALK kinase.
ALK gene encodes for a receptor tyrosine kinase normally expressed in fetal and mature nervous systems but not in lymphoid cells. As any receptor tyrosine kinase and in normal situation, ALK protein will activate signaling pathway and cell cycle after oligomerization induced by binding with its ligand. In ALK rearrangements, the partner gene brings to ALK the ability to self-associate in a ligand-independent fashion, leading to its constitutive activation. In addition, the gene partner brings a strong promoter, driving illegitimate and high levels of ALK receptor tyrosine kinase fusion gene expression in lymphoid cells. The functional consequence is to exaggerate and dysregulate otherwise normal downstream signals which will promote cell growth and inhibit apoptosis. 41 Clearly, ALK activation is a critical step in the development of ALCL of T cell origin. As ALK gene is not expressed in normal lymphoid cells, the immunodetection of ALK protein in a lymphoid tumor represents a highly sensitive test for identification of lymphoma with ALK rearrangement, correlating in nearly 100% of cases with the presence of such abnormality.
Regardless of other clinical and biological prognostic parameters, the outcome for patients with ALK-positive ALCL is significantly better than that for patients with ALK-negative ALCL with the 5-year survival rates ranging between 79 and 88% and 28 and 40%, respectively. 42 Additional information on lymphomas is found in Chapter 24 Lymph Nodes and Flow Cytometry.

Sarcomas
Although sarcomas are relatively rare neoplasms in adulthood, they represent the most frequent malignant tumors in childhood and young adults. Abundant genetic studies have revealed that a significant number of sarcoma are associated with specific chromosomal abnormalities (mainly chromosomal translocations) that can be used as practical diagnostic markers in histological equivocal cases. 13, 14, 43, 44 A typical example is the so-called “small round blue cell” undifferentiated pattern shared by disparate tumor entities such as embryonal or alveolar rhabdomyosarcoma, Ewing's sarcoma, neuroblastoma, and lymphoma.
Two major genetic groups distinguishable at the cytogenetic level are observed in sarcomas. One group is characterized by a near-diploid karyotype with a single or few chromosomal abnormalities, whereas the second exhibits complex karyotype with numerous aberrations that reflect severe disturbance in genomic stability. Sarcoma with genetic abnormalities not detectable by conventional cytogenetics and/or FISH means—such as GIST and its specific c-KIT mutation—will not be discussed in this section.

Sarcomas with Single Karyotypic Abnormalities
This group is characterized by karyotype harboring single and tumor-specific chromosomal translocations ( Table 2.2 ). Most of these translocations lead to fusion genes encoding aberrant transcription factors but a small subset creates aberrant chimeric genes related to growth-factor signaling pathway.
Table 2.2 Translocations associated with sarcomas Translocation Genes Type of fusion gene EWING'S SARZCOMA t(11;22)(q24;q12) EWSR1-FLI1 Transcription factor t(21;22)(q22;q12) EWSR1-ERG Transcription factor t(7;22)(p22;q12) EWSR1-ETV1 Transcription factor t(17;22)(q21;q12) EWSR1-ETV4 Transcription factor t(2;22)(q33;q12) EWSR1-FEV Transcription factor CLEAR-CELL SARCOMA t(12;22)(q13;q12) EWSR-ATF1 Transcription factor DESMOPLASTIC SMALL ROUND-CELL TUMOR t(11;22)(p13;q12) EWSR-WT1 Transcription factor MYXOID CHONDROSARCOMA t(9;22)(q22-31;q11-12) EWSR-NR4A3 Transcription factor MYXOID LIPOSARCOMA t(2;16)(q13;p11) FUS-DDIT3 Transcription factor t(12;22)(q13;q12) EWSR1-DDIT3 Transcription factor ALVEOLAR RHABDOMYOSARCOMA t(2;13)(q35;q14) PAX3-FKHR Transcription factor t(1;13)(p36;q14) PAX7-FKHR Transcription factor SYNOVIAL SARCOMA t(X;18)(p11;q11) SYT-SSX Transcription factor DERMATOFIBROSARCOMA PROTUBERANS t(17;22)(q22;q13) COL1A1-PDGFB Growth factor CONGENITAL FIBROSARCOMA t(12;15)(p13;q25) ETV6-NTRK3 Transcription factor-receptor INFLAMMATORY MYOFIBROBLASTIC TUMOR 2p23 rearrangements TMP3-ALK; TMP4-ALK Growth factor-receptor ALVEOLAR SOFT-PART SARCOMA t(X;17)(p11.2;q25) ASPL-TFE3 Transcription factor
The Ewing's family of tumors, which includes Ewing's sarcoma and primitive neuroectodermal tumor (ES/PNET), are characterized by a t(11;22)(q24;q12) translocation leading to the EWSR1-FLI1 fusion gene and observed in nearly 90% of cases of ES/PNET ( Fig. 2.13 ). The remaining cases show alternative chromosomal translocations fusing the EWSR1 gene (chromosome 22q12) with partner genes other than FLI1 and that belong to the same ETS family of transcription factors. EWSR1 gene is also involved in chromosomal translocations arising in several other tumoral entities such as the intra-abdominal desmoplastic small round-cell tumor (DSRCT), myxoid chondrosarcoma, and clear cell sarcoma. However, EWSR1 gene fused in each case with gene partners not encountered in the Ewing's family of tumors, giving rise to specific fusion genes suitable for diagnostic purposes. 13, 43, 44

Fig. 2.13 Karyotype of an Ewing's tumor with the characteristic t(11;22)(q24;q12) chromosomal translocation (arrows) leading to the EWS-FLI1 fusion gene. Secondary recurrent chromosomal abnormalities such as monosomies 6 and 15 and trisomies 2 and 14 are also observed.
Some new data indicate that soft-tissue tumors can no longer be classified only on basis of their site of origin but also according to their genetic aberrations. 43, 45 Congenital fibrosarcoma and mesoblastic nephroma were thought to be unrelated tumors until cytogenetic analysis revealed a common aberration, hence the t(12;15)(p13;q25) translocation with subjacent ETV6-NTRK3 fusion gene, indicating that they are simply the same tumoral entity that develops in different locations. Another similar example is illustrated by the t(X;17)(p11.2;q25) translocation shared by the alveolar soft-part sarcoma (ASPS) and a cytogenetic subset of childhood papillary renal cell carcinoma (PRCC). 46 Although this translocation is cytogenetically unbalanced in ASPS and balanced in PRCC, it gives rise at the molecular level to the same ASPL-TFE3 fusion transcript in both tumoral types. Therefore, some fusion genes can exert their oncogenic properties in more than one target cell type and seems not to play any role in cell differentiation. On the other hand, in vitro experiments showed that fusion proteins such as EWS-FLI1 contribute to the phenotypic features of ES/PNET by subverting the differentiation program of its neural crest precursor cell to a less differentiated and more proliferative state. In synovial sarcoma, the SYT gene on chromosome 18q11 can fuse with various members of the SSX cluster located on chromosome Xp11 ( Fig. 2.14 ). The SSX2 translocation partner is more likely observed in monophasic synovial sarcoma, whereas SSX1 is much often associated with the biphasic forms, indicating that this latter gene partner may drive epithelial differentiation in synovial sarcoma. Finally, some data support the hypothesis that the gene fusion occurs in an already established lineage that imposes constraints such that the target cell selects the fusion gene. In contrast, other observations suggest that this fusion will modulate the phenotypic features of the undifferentiated precursor harboring this fusion gene. 45

Fig. 2.14 Karyotype of a synovial sarcoma with the specific t(X;18)(p11;q11) chromosomal translocation (arrows). Losses and gains of other chromosomes represent additional secondary changes.
Sarcoma-associated chromosomal translocations and/or their respective fusion genes may have some prognostic impacts. 14, 43 In Ewing's sarcoma, several molecular variants are observed in the EWS-FLI1 fusion gene due to various breakpoint junctions. The most common, designated type 1 (linking exon 7 of EWS with exon 6 of FLI1 ) is associated with a better prognosis than other variants. The SYT-SSX fusion type in synovial sarcoma appears to be a significant prognostic factor since patients with the SYT-SSX2 variant have an improved overall survival when compared with SYT-SSX1 positive patients, independent of the histological type. Patients with metastatic alveolar rhabdomyosarcoma having the PAX7-FKHR fusion gene show a substantially better prognosis than those with the PAX3-FKHR translocation. These variations in behavior could be due to subtle differences in the biochemical activities of the variant fusion proteins, with a better prognosis associated with variants having a less transcriptional activity.
The close association between specific translocations and distinct sarcoma types indicates that they are early events in tumorigenesis but their exact role in tumor development remains often difficult to assess. In the small subset of translocations with aberrant chimeric genes related to growth-factor signaling pathways (see Table 2.2 ), the pathogenesis arises through cell cycle activation although this is probably not sufficient per se to induce full transformation. The great majority of chromosomal translocations in sarcoma involve transcription factors without obvious putative oncogenic properties at first sight. Transcription factors are proteins that directly interact with the DNA strand of their target genes, and regulate the expression of these genes by binding their promotor regions upstream of RNA transcription sites. A translocation will lead to aberrant gene fusion composed of the DNA (or RNA)-binding domain of a transcription factor fused with the transactivation domain of another transcription factor. The functional consequence is that the transcriptional activity of the latter will be deviated toward downstream genes targeted by the DNA-binding domain provided by the transcription factor partner. Moreover, most of these chimeric proteins show enhanced transcriptional activity compared with their constitutive normal protein, providing eventually a gain of function mechanism. It is thus believed that these phenomena lead to dysregulation of gene expression, accounting for the tumoral properties of fusion genes in sarcoma.
This general opinion can be illustrated by the t(2;13)(q35;q14) and t(1;13)(p36;q14) translocations arising in alveolar rhabdomyosarcoma and corresponding to the PAX3/FKHR and PAX7/FKHR fusion genes, respectively. 43, 47 Both translocations fuse the DNA-binding domain of PAX3 or PAX7 to the transactivation domain of FKHR. PAX genes activate myogenesis, and fork head in rhabdomyosarcoma ( FKHR ) is though to have pro-apoptotic activities. The resulting PAX-FKHR fusion gene is a highly potent activator with a transcriptional activity 10 to 100 times as high as that brought by wild-type PAX gene. This enhanced transcriptional activity is further amplified by mechanisms of PAX-FKHR fusion genes amplification. As the DNA-binding of FKHR is lost in the PAX-FKHR fusion, any DNA-binding specificity of the fusion gene is directed by the PAX sequence, leading to dysregulated expression of downstream target genes of PAX genes. Consequently, PAX3/FKHR will be able, firstly, to inhibit cellular apoptosis through a PAX3 target gene, the anti-apoptotic protein BCL-XL and, secondly, to activate c-MET , PDGFαR , or c-RET oncogene, downstream targets of PAX3 involved in migration and proliferation of myogenic precursors.
Other sarcoma-associated fusion genes have been shown to get tumoral properties by activating growth factor receptors. MET oncogene has been recently shown to be a direct transcriptional target of the ASPL-TFE3 fusion gene. Induction of MET by ASPL-TFE3 results in strong MET autophosphorylation and activation of downstream signaling in the presence of hepatocyte growth factor. 48
Another question that remains a matter of debate is whether these chromosomal translocations are sufficient for neoplastic transformation. 14, 43 Although expression of certain gene fusions can induce sarcoma in primary mesenchymal progenitor cells, secondary mutations are likely to be required for full malignancy as observed in the context of hematological disorders. Loss of tumor suppression genes expression (such as P16 and RB ) is observed in more than 50% of various sarcoma. Activation of common growth-factor pathways not directly due to chromosomal translocation is described in sarcomas including the insulin-like growth factor 1 (IGF1) pathway in alveolar RMS, the platelet-derived growth factor receptor (PDGFR) in DSRCT, and the c-KIT receptor pathway in Ewing's tumors. Parallel to the situation observed in childhood leukemia, 49 it is possible that some chromosomal translocations associated with childhood tumors arise during fetal development, leading to a “pre-malignant state” preceding the sarcomatous transformation induced by additional genetic aberrations.

Sarcomas With Complex Karyotypes
This group of sarcoma does not exhibit any specific and recurrent chromosomal translocation but rather complex karyotypes with multiple numerical and structural aberrations characteristic of severe genetic and chromosomal instability 43 ( Table 2.3 ) ( Fig. 2.15 ). The underlying genetic mechanisms frequently include alterations in cell-cycle genes such as P53 , INK4A , and RB1 as well as genes directly involved in DNA-repair pathways. Oncogene amplifications occur in cytogenetically complex sarcoma. MDM2 and MYCN gene amplifications are well-known examples. MDM2 amplification is observed in liposarcoma (other than myxoid) and malignant fibrous histiocytomas. MDM2 is a p53 inhibitor and its amplification will lead to inability of p53 to induce apoptosis in cells with DNA damage, which, in turn, will induce genomic instability. MYCN amplification ( Fig. 2.16 ) is used as a genetic parameter for better therapeutic stratification of patients suffering from neuroblastoma, one of the most frequent malignant tumors in childhood. 50 MYCN is a member of the MYC family of proto-oncogenes which are transcription factors promoting cell proliferation and inhibiting terminal differentiation. In view of its function, MYCN is involved in the genesis of a wide range of cancers including neuroblastoma, small cell lung carcinoma, some cases of medullary thyroid carcinoma, retinoblastoma, and breast cancers. A forced expression of MYCN in central nervous system cells in mouse leads to the development of a subgroup of neuroblastomas, indicating that it is sufficient for malignant transformation. Additional information on sarcomas is found in Chapter 18 .
Table 2.3 Sarcomas with complex karyotypes Type of sarcoma Fibrosarcoma (other than congenital) Leiomyosarcoma Malignant fi brous histiocytoma Osteosarcoma Chondrosarcoma (types other than extraskeletal myxoid) Liposarcoma (types other than myxoid) Embryonal rhabdomyosarcoma Malignant peripheral nerve-sheath tumour Angiosarcoma Neuroblastoma a
a Neuroblastoma is quoted in this table as it belongs to the “small round blue cell tumours” group.

Fig. 2.15 Typical example of a complex karyotype as observed in embryonal rhabdomyosarcoma showing multiple numerical and structural abnormalities. The latter (isochromosome 17q and two chromosome markers) are marked with arrows.

Fig. 2.16 Interphase FISH demonstrating amplification of the MYCN oncogene (red signals) in a case of neuroblastoma. The two green spots correspond to centromeric probes specific for the chromosome 2 and used as control for diploid status assesment of the analyzed cell.

Thyroid Carcinomas
Among epithelial malignancies, two histological types of thyroid carcinoma, namely the papillary and follicular thyroid carcinoma, deserve to be mentioned as they exhibit specific genetic aberrations that represent reliable diagnostic parameters.

Papillary Thyroid Carcinoma
Papillary thyroid carcinoma (PTC) is characterized by rearrangements of the RET oncogene, a receptor tyrosine kinase (RTK) gene located on chromosomal region 10q11.2. These activating rearrangements, called RET/PTC, are caused by either paracentric inversion of chromosome 10 or balanced translocations involving chromosome 10 and various chromosome partners 51 ( Table 2.4 ). The molecular consequences are fusion of the tyrosine kinase domain of RET with the 5′ part of the various gene partners with subsequent release of the extracellular ligand-binding and juxtamembrane domains of RET receptor. As the juxtamembrane domain negatively regulates RET mitogenic signaling, its deletion contributes to RET/PTC activation, which is further enhanced by dimerization potential brought by the gene partner. 52 This leads to ligand-independent activation of the RET kinase, signaling pathway stimulation and cell-cycle activation; a well-known oncogenic process in tumoral cells harboring RTK rearrangements. As part of its oncogenic effect, RET/PTC directly modulates genes involved in inflammation/invasion of the cell such as various cytokines (GM-CSF, M-CS, IL6, etc.), chemokines (CCL2, CXCL12, etc.), and chemokine receptors (CXCR4). The induction of an inflammatory-type reaction may explain the chronic inflammatory reaction observed in this type of cancer. 52

Table 2.4 Characteristics of different types of RET/PTC rearrangement in papillary thyroid carcinoma
The prevalence of RET/PTC in papillary thyroid carcinoma is highly variable (0–87%), depending on age of patient, geographic regions, and sensitivities of the detection methods used (polymerase chain reaction versus FISH), particularly if the rearrangement is present only in a small proportion of tumor cells or if the RET/PTC transcripts is expressed at low levels. The average prevalence is 20–30% in sporadic adult cases and rises to 45–60% among tumors from children and young adults. It is higher (50–80%) in papillary carcinoma associated with radiation exposure, and it is thought that the close association between RET/PTC translocations and irradiation is due to spatial proximity of the participating chromosomal loci in the nuclei of thyroid cells, providing a structural basis for radiation-induced illegitimate recombination of the genes. 14 Most studies concur that RET/PTC rearrangements are rare or absent in benign adenomas, and not observed in other types of thyroid carcinomas. They are more frequent in PTC exhibiting a classic architecture and in microcarcinomas. 52 Among the different variants of RET/PTC translocations, RET/PTC1 and 3 are the most frequent, accounting for more than 90% of all rearrangements. 53
A small subset of PTC (around 10%) is characterized by rearrangement of the NTRK gene, another receptor tyrosine kinase, located on chromosome 1q22 and encoding one of the receptors for the nerve growth factor. NTRK gene activation is due to chromosome 1 inversions or balanced translocations between chromosome 1 and 10, resulting in fusion of the NTRK tyrosine kinase domain to 5′-end sequences from at least three different genes: tropomyosin ( TPM3 ) or TPR gene, both on chromosome 1, and TFG gene located on chromosome 3. 51, 53

Follicular Thyroid Carcinoma
Follicular thyroid carcinomas are characterized by PPARγ ( peroxisome proliferator-activated receptor γ ) gene rearrangements in 25–50% of cases, mainly under the form of a distinctive t(2;3)(q13;p25) chromosomal translocation. 53 This translocation leads to the fusion of the PAX8 gene ( paired box gene 8 ) with PPARγ gene, resulting in a fusion protein designed PPFP. PAX8 is a transcription factor expressed at high levels in thyrocytes and necessary for normal thyroid development. PPARγ encodes a nuclear hormone receptor transcription factor whose activity is related to adipocyte differentiation, lipid and carbohydrate metabolism, and cellular proliferation and differentiation. PPFP is thought to exert its oncogenic properties through a mechanism in which it acts as a dominant-negative inhibitor of wild-type PPARγ . This results in inhibition of apoptosis and promotion of proliferation as well as anchorage-independent growth of thyroid follicular cells. PPARγ has mainly been observed in low-stage follicular carcinomas with vascular invasion and has been identified at apparent lower frequency in adenomas.

Clinical Applications of Conventional Cytogenetics and fish in Cytology

Introduction
Cytological assessment of a fine-needle aspiration (FNA) specimen remains the first-line morphological investigation of any suspected mass but cytomorphology alone—hence, without tissue architecture—is not always sufficient for a definitive diagnosis. 54, 55 For example, small-to-intermediate cell lymphomas such as MCL, FL, or MZL can show overlapping cytomorphologic features with one another as well as with reactive lymph node hyperplasia. Limitations of FNA are also encountered in soft-tissue neoplasms, especially in the diagnostic management of small round-cell tumors. Most of the diagnostic problems can be solved with the help of immunocytochemistry but limitations can be encountered mainly due to immunophenotypic heterogeneity among small B-NHL subtypes. 56 For examples, the intensity of CD10 expression in FL has been shown to be variable, and even negative in some cases. 57 MCL and SLL can be distinguished by differences in CD23 expression but CD23 can be weakly expressed in both subtypes. 58 CD5 expression may not systematically be used as a diagnostic criterion between MCL and SLL, and some FL can also exhibit a CD5 positivity. 59 It is thus necessary that FNA examination be supplemented with ancillary methods such as karyotype, FISH, or polymerase chain reaction (PCR). Conventional cytogenetics allows complete karyotype analysis and, as such, remains the historic gold standard by which everything is started. However, it is a cumbersome and time-consuming procedure requiring adequate fresh tissue and special cell culture techniques. PCR and interphase FISH (I-FISH) methods are more practical in that they can bypass the need of cell culture. They have both their own advantages and disadvantages, and must be considered as complementary rather than competing with one another. It is therefore not surprising that both have been included in a combined diagnostic algorithm proposed in the literature. 60 - 62 However, I-FISH remains a less sophisticated laboratory technique than PCR and offers a greater qualitative sensitivity in studies of tumor-associated chromosomal abnormalities as will be illustrated later. This technique is advantageous for FNA specimens because it requires only a few cells. It allows also retention of cellular morphology, which permits simultaneous evaluation of morphology and chromosomal alterations. Moreover, recent studies have demonstrated the feasibility of FISH on Papanicolaou-stained archival cytology slides, highlighting the good flexibility of such method. 63 - 66 These advantages probably explain why FISH is becoming more and more popular in cytopathology laboratories.

FISH Strategy
Interphase FISH requires simple material such as cytospins from FNA specimens. Cytospin is an optimal preparation for I-FISH because the monolayer allows excellent hybridization results. Cytospin preparation can be made by Ficoll-Hypaque gradient-separation technique and then fixed in methanol-glacial acetic acid (3:1) for 20 minutes at −20°C. The slides will be then air-dried and stored at −20°C prior current FISH procedure. Specimen handling is thus very simple, but it is critical to avoid delays in specimen processing in order to prevent possible degradation of the target DNA and subsequent poor hybridization results. 67 Subsequent FISH steps can be then easily performed without further manipulation of the samples, with the use of commercially available kit sets including the premixed probes, and according to the protocol recommended by the manufacturer. At least, 200 nonoverlapping and intact nuclei per case and at least two different areas on the same slide should be scored before giving a result. The great advantage of working on an interphase cell can nevertheless be a source of interpretative pitfalls in that random chromosome colocalizations occur not infrequently in normal nuclei and can mimic chromosomal translocations. Although most of the commercial probes have been designed to limit the risk of false-positive profiles, it remains critical to determine the frequency of such false-positive cells in order to define a cutoff level. Normal lymphocyte nuclei can be used as negative control to assess hybridization efficiency, and the cutoff level for positivity should be set at the mean (%) ± 3 standard deviations. Beside this pitfall, other good practice recommendations are needed and must be known by the user. Such guidelines are detailed in an excellent overview recently published that we highly recommend to the reader. 68
The commercial probes are usually several hundred kilobases in length and yield large, bright and easily detectable signals. They are currently available to detect many of the relevant chromosomal abnormalities described in the previous section and are known to be highly sensitive. 69 For detection of chromosomal translocations, three different kinds of probes are available, including the dual-fusion probes, the single-fusion extra-signal probes, and the break-apart probes, all being dual-color probes ( Figs 2.17 A and 2.17 B). Dual-fusion and extra-signal probe sets are made of two differentially labeled (green and red) DNA segments, each of these segments identifying one of the chromosomal loci involved in the translocation. For the dual-fusion probes, an abnormal pattern will be represented by one red and one green signal (representing the normal homolog) and by two fusion or colocalization signals corresponding to the chromosomal translocation and its reciprocal (“2F,1R,1G” pattern). Typical examples are probes designed to detect lymphoma-associated chimeric genes subjacent to translocation such as the BCL2-IgH or BCL1-IgH in follicular or mantle cell lymphoma, respectively ( Fig. 2.18 A). Such probes make it possible to significantly reduce the risk of false positives as the possibility that two overlapping signals are due to random spatial proximity of the participating chromosomal loci remains very low. The abnormal pattern for extra-signal translocation probes will be represented by a single fusion (corresponding to one derivative chromosome) plus a small extra signal representing the residual portion of one of the loci involved in the translocation. Again, the probability that such pattern is observed in a normal nucleus is very low. A well-known example is the probe used to detect the BCR-ABL chimeric gene in chronic myeloid leukemia. Such a probe has not been designed for detection of recurrent chromosomal translocations in lymphoma or sarcoma and will not be illustrated here. Dual-color break-apart probes are made of differentially labeled (green and red) DNA segments located on either side of a breakpoint cluster region within a target gene. The separation of green and red signals indicates break between the 5′ and 3′ regions of the rearranged gene. In normal cells, the two probes colocalize to produce two yellow fusion signals (corresponding to two copies of nonrearranged genes), whereas in the case of translocation involving one of the two genes, one of the fusion signal splits, resulting in a characteristic 1 red–1 green–1 yellow fusion (“1R1G1F”) signal pattern. The break-apart strategy offers the advantage of detecting in a single experiment all recurrent rearrangement of a gene involved in translocations with different gene partners. A typical example is the EWSR1 gene which can fuse with no less than nine different gene partners ( Fig. 2.18 B).

Fig. 2.17 Schematic representation of two different types of dual-color probes, the dual fusion and break-apart probes. (A) Left: Dual-fusion probes are composed of two differentially labeled (green and red) DNA segments, each of these segments identifying one of the genes/loci involved in the chromosomal translocation. The probes are usually several kilobases in length and extend largely on both sides of the gene of interest. Right: a normal pattern will show two red and two green spots, whereas a cell harboring a chromosomal translocation will demonstrate two fusion or colocalization signals corresponding to the chromosomal translocation and its reciprocal; the red and green spots indicate the two remaining normal chromosomes (“2F,1R,1G” pattern). (B) Left: break-apart probes are made of two differentially labeled (green and red) DNA segments flanking the breakpoint cluster region of a gene involved in chromosomal translocations. Right: a normal cell will show two yellow fusion signals corresponding to two copies of a normal gene. The disruption of one of these two copies subsequently to a chromosomal translocation will lead to split of one yellow signal into two red and green signals (“1R,1G,1F” pattern).

Fig. 2.18 (A) Interphase FISH of follicular lymphoma cells with the use of a dual-fusion BCL2-IgH probe. The two fusion/colocalization signals indicate the existence of a BCL2-IgH oncogene and its reciprocal while the green and red signals correspond to the remaining normal IgH and BCL2 genes respectively. (B) Interphase FISH with the use of a dual-fusion break-apart probe specific to the EWSR1 locus. The lower nucleus shows a normal pattern, whereas the upper one displays split of one EWSR1 gene copy as it can be observed in Ewing's tumors.
Interphase FISH is also able to identify submicroscopic chromosomal deletions as well as numerical chromosomal abnormalities such as trisomy or monosomy. The probes used to detect entire chromosomal gains or losses are juxtacentromeric alphoid DNA sequences while submicroscopic deletions will be identified with locus-specific probes. To ensure the quality of hybridization (mainly the hybridization properties of the tumor cells being analyzed), a control probe, labeled with a different fluorophore and identifying any other chromosome, will be cohybridized with the probe of interest. For detection of microdeletion, the control probe will also serve to identify the chromosome harboring the deleted region. Examples are trisomy 3 ( Fig. 2.19 ) and deletion of chromosome 7q in marginal zone lymphoma.

Fig. 2.19 Interphase FISH using centromeric probes for chromosomes 7 (green) and 18 (red). Both cells show three signals for each probe indicative of trisomies 7 and 18 as observed in nodal marginal zone lymphoma.

Application

Lymphomas
Studies demonstrating the feasibility and diagnostic utility of FISH in FNA specimens have focused on the most frequent lymphoma such as FL and, to a certain extent, diffuse large B-cell lymphoma. Although less common, MCL has also been a subject of interest because of the clinical relevance and difficulties to differentiate it cytologically from other small cell NHLs. Among the latter, small lymphocytic lymphoma may be difficult to diagnose when it presents as isolated lymphadenopathy.
As mentioned earlier, FISH and PCR remain complementary methods for detecting predictable chromosomal abnormalities in lymphoma, but comparative studies on specimens such as tissue imprints, cytospins, or smears have demonstrated a higher qualitative sensitivity of I-FISH. In follicular lymphoma, the detection rate of the t(14;18) translocation with PCR was 70% at best, whereas a positive result could be achieved in around 90% of cases with FISH. 65, 66, 70 - 73 The low detection rate encountered with the PCR technology is due to mutation involving primer binding sequences and to the fact that the current PCR method applicable in routine use is not able to detect breakpoints outside the known major breakpoint region (MBR) and minor cluster region (mcr). A similar situation is encountered in MCL where the sensitivity of FISH analysis for the direct detection of the t(11;14) translocation largely exceeds DNA-PCR methods; the detection rate reaching nearly 100% according to FISH studies, 64, 69, 72, 74, 75 while it falls in the range of 40% with the second method. 74, 60 The lower qualitative sensitivity offered by DNA-PCR is mainly due to the wider variation of BCL1 gene breakpoints that are difficult to span with primers. FISH analysis circumvents these limitations by using IgH/BCL2 and IgH/BCL1 dual-fusion probes covering the entire BCL2 and BCL1 gene, respectively. Moreover, these results highlight the greater applicability of FISH since all known BCL2 and BCL1 breakpoints can be covered and detected with a single-probe set. Among other small-cell lymphoproliferative disorders, SLL/CLLs are characterized by recurrent chromosomal abnormalities such as trisomy 12 or interstitial deletion involving 13q14 chromosomal region. Interphase FISH can easily detect these aberrations 63, 76 and, together with negative results for BCL2 and BCL1 rearrangements, can help in assessing an accurate diagnosis of SLL/CLL. In FNA specimens displaying monomorphous lymphoid population composed of medium-sized or large cells, a proper diagnosis of Burkitt, diffuse large B-cell, or anaplastic lymphoma can be easily reached with the use of specific break-apart probes targeting C-MYC , BCL6 , or ALK gene, respectively. 72, 77 Beside the detection of lymphoma-associated specific translocations, atypical patterns revealed by interphase FISH can help in better classifying lymphoma. For example, a current IgH/BCL1 dual-fusion probe can identify hypotetraploid profiles with extra copies of BCL1 signals as observed in blastoid variants of MCL. 74, 78
A recent study aimed at comparing the utility of I-FISH and flow cytometry immunophenotyping (FCM) in a series of FL and DLBCL. 71 They found that detection of t(14;18) by FISH was a slightly more sensitive (85%) diagnostic marker than identification of the typical CD19+/CD10+ immunophenotype profile by FCM (75%). FISH appeared to be more sensitive because it could also detect FLs with an atypical CD19 + /CD10 - pattern. In the same study, a BCL2 gene rearrangement was detected by FISH in 29% of DLBCL cases, whereas FCM was able to identify a CD10+ monoclonal population in only 23% of such lymphoma cases. I-FISH appeared thus to be slightly more sensitive than FCM in identifying germinal center B-like DLBCL.

Sarcomas
Most (if not all) chromosomal translocations described in soft-tissue sarcomas (STS) are detectable by I-FISH. This method is thus particularly useful in diagnostically difficult cases such as small blue cells tumors. Several studies aimed at comparing the efficiency of both reverse transcriptase (RT)-PCR and FISH techniques for a molecular diagnosis in sarcoma. 79, 80 Both methods were complementary and had their own advantages and disadvantages in terms of specificity and qualitative and quantitative sensitivity. However, the FISH break-apart approach appears to be very practical in that the use of a single break-apart probe can recognize each specific translocation such as the t(X;18) in synovial sarcoma, the t(2;13) or t(1;13) in alveolar rhabdomyosarcoma, and the t(12;16) in myxoid liposarcoma. The potential disadvantage of such an approach would be its inability to distinguish Ewing/PNET from other sarcomatous types harboring EWS gene rearrangements (see Table 2.2 ) since the partner gene is not detected. In most cases, these neoplasms are nevertheless distinguishable from each other on the basis of clinical data or immunocytochemical differences.
Several studies have demonstrated the usefulness of cytogenetics 81, 82 or I-FISH 83 - 85 as an adjunct in making a definitive diagnosis of sarcoma by FNA. As our knowledge about the specific chromosomal abnormalities associated with sarcoma is constantly increasing, there is good hope that I-FISH will allow accurate diagnosis on more cases investigated by FNA, obviating open surgical biopsy preceding therapy.

Multiple Myeloma
Multiple myeloma (MM) is characterized by numerous chromosomal abnormalities which have been shown to significantly impact survival in patients with such disease. 86 The most relevant alterations include hyperdiploidy, monosomy 13/deletion 13q14, deletion 17p, t(11;14)(q13;q32), and t(4;14)(p16;q32)translocations, giving rise to the BCL1-IgH and FGFR3/MMSET-IgH chimeric genes, respectively. Hyperdiploidy is associated with a favorable prognosis but all other abnormalities represent unfavorable parameters, among which the t(4;14) and deletion17p appear to be the most important. Their identifications have implications for the design of risk-adapted treatment strategies. Historically, testing for abnormalities in MM was based on conventional chromosomal analysis performed on bone marrow, but results were often falsely normal since the actively normal myeloid cells were analyzed rather than the monoclonal plasma cells, which infrequently enter mitosis. Standard FISH studies were thus employed to detect the classical abnormalities associated with MM. Again, erroneously normal results were most often obtained since this method is not able to distinguish between normal cells and small clones of monoclonal plasma cells. A novel FICTION method, which is a combination of fluorescent immunophenotyping and in situ hybridization, has thus been developed. 9 First, antibodies against the cytoplasmic immunoglobulins λ or κ are applied in order to specifically identify the plasma cells thanks to their cytoplasmic fluorescence. Second, FISH probes will be hybridized to all cell types, but only specifically target plasma cells will be analyzed. This FICTION method is thus capable of detecting chromosomal abnormalities in bone marrow specimens even when few plasma cells are present ( Fig. 2.20 A and 2.20 B).

Fig. 2.20 Illustration of the FICTION method on multiple myeloma cells. The plasma cells are detected with the use of antibodies directed against the cytoplasmic immunoglobulins λ or κ. These antibodies are colored with fluorochrome AMCA (blue color). (A) Two plasma cells with deletion of the p53 locus (chromosome 17p13) demonstrated by the absence of one red signal, whereas the existence of both chromosomes 17 is confirmed by a specific chromosome 17 centromeric probe (green signals). (B) Plasma cell harboring the FGFR3/MMSET-IgH fusion genes corresponding to the t(4;14)(p16;q32) translocation. The two yellow fusion signals are due to red and green signals relocating next to each other, and indicate the FGFR3/MMSET-IgH fusion gene and its reciprocal. The green and red signals correspond to remaining normal IgH and FGFR3/MMSET genes respectively.

Concluding Remarks
Over the past two decades, conventional cytogenetics has made it possible to identify nearly all chromosomal abnormalities associated with specific histological subtypes of lymphoproliferative disorders and soft-tissue tumors. These chromosomal aberrations made it possible, in a second step, to pinpoint the underlying oncogenes and to study the pathogenesis of tumors bearing such abnormalities. In addition to their role in fundamental research, these alterations rapidly appeared to be powerful diagnostic and prognostic parameters relevant to use on a regular basis. The constant emergence of commercial probes yielding large, bright, and easily detectable signals made the FISH method a reliable tool for detecting specific chromosomal abnormalities on nondividing cells provided by cytology specimens such as smears, cytospin, or liquid-based samples. At present time, there is enough evidence in the specialized literature demonstrating that I-FISH, in conjunction with other ancillary tools such as immunocytochemistry and molecular biology, constitutes a suitable complementary approach in the cytological diagnosis of cancers detailed in this chapter.

References

1. Tjio J.H., Levan A. The chromosome number of man. Hereditas . 1956;42:1-6.
2. Nowell P.C., Hungerford D.A. A minute chromosome in human chronic granulocytic leukaemia. Science . 1960;132:1497.
3. Cell biology Animation. Available at: http://www.johnkyrk.com
4. Stewénius Y., Gorunova L., Jonson T., et al. Structural and numerical chromosome changes in colon cancer develop through telomere-mediated anaphase bridges, not through mitotic multipolarity. Proc Natl Acad Sci USA . 2005;102(15):5541-5546.
5. Caspersson T., Zech L., Johansson C., et al. Identification of human chromosomes by DNA-binding fluorescent agents. Chromosoma . 1970;30:215-227.
6. An International System for Human Cytogenetic Nomenclature (ISCN 1985). Report of the Standing Committee on Human Cytogenetic Nomenclature. Birth Defects Orig Artic Ser . 1985;21(1):1-117.
7. Rowley J.D. Letter: A new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and Giemsa staining. Nature . 1973;243(5405):290-293.
8. Mascarello J.T., Cooley L.D., Davison K., et al. Problems with ISCN FISH Nomenclature make it not practical for use in clinical test reports or cytogenetic databases. Genet Med . 2003;5(5):370-377.
9. A combined cytoplasmic immunoglobulin staining and FISH method (cIg FISH) to detect prognostic genetic anomalies in multiple myeloma and related plasma cell proliferative disorders. January 2005. URL: Available at: http://www.mayomedicallaboratories.com/media/articles/communique/mc2831-0105.pdf
10. McNeil N, Montagna C, Difilippantonio MJ, et al. Comparative cancer cytogenetics. Atlas Genet Cytogenet Oncol Haematol 2003. Available at: http://atlasgeneticsoncology.org/Deep/ComparCancerCytogID20011.html
11. Mitelman F. Cancer cytogenetics update 2005. Atlas Cytogenet Oncol Haematol 2005. Available at: http://atlasgeneticsoncology.org/Deep/CancerCytogenet2005ID20050.html
12. Jaffe E.S., Harris N.L., Stein H., et al, editors. World Health Organization. Classification of Tumours. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon: IAC Press, 2001.
13. Fletcher C.D.M., Unni K., Mertens K. World Health Organization. Classification of Tumours. Pathology and Genetics of Tumors of Soft-Tissue and Bone . Lyon: IAC Press; 2002.
14. Mitelman F., Johansson B., Mertens F. The impact of translocations and gene fusions on cancer causation. Nat Rev Cancer . 2007;7(4):233-245. Epub 2007 Mar 15
15. Cajulis R.S., Yu G.H., Gokaslan S.T., et al. Modified interphase cytogenetics technique as an adjunct in the analysis of atypical cells in body fluids. Diagn Cytopathol . 1997;16(4):331-335.
16. Halling K.C., Rickman O.B., Kipp B.R., et al. A comparison of cytology and fluorescence in situ hybridization for the detection of lung cancer in bronchoscopic specimens. Chest . 2006;130(3):694-701.
17. Tsujimoto Y., Cossman J., Jaffe E., et al. Involvement of the bcl-2 gene in human follicular lymphoma. Science . 1985;228(4706):1440-1443.
18. Horsman D.E., Okamoto I., Ludkovski O., et al. Follicular lymphoma lacking the t(14;18)(q32;q21): identification of two disease subtypes. Br J Haematol . 2003;120(3):424-433.
19. Bosga-Bouwer A.G., van Imhoff G.W., Boonstra R., et al. Follicular lymphoma grade 3B includes 3 cytogenetically defined subgroups with primary t(14;18), 3q27, or other translocations: t(14;18) and 3q27 are mutually exclusive. Blood . 2003;101(3):1149-1154.
20. Jardin F., Gaulard P., Buchonnet G., et al. Follicular lymphoma without t(14;18) and with BCL-6 rearrangement: a lymphoma subtype with distinct pathological, molecular and clinical characteristics. Leukemia . 2002;16(11):2309-2317.
21. Sehn L.H. Optimal use of prognostic factors in non-Hodgkin lymphoma. Hematology Am Soc Hematol Educ Program . 2006:295-302.
22. Akasaka T., Lossos I.S., Levy R. BCL6 gene translocation in follicular lymphoma: a harbinger of eventual transformation to diffuse aggressive lymphoma. Blood . 2003;102(4):1443-1448.
23. Remstein E.D., Kurtin P.J., Buño I., et al. Diagnostic utility of fluorescence in situ hybridization in mantle-cell lymphoma. Br J Haematol . 2000;110(4):856-862.
24. Fernàndez V., Hartmann E., Ott G., et al. Pathogenesis of mantle-cell lymphoma: all oncogenic roads lead to dysregulation of cell cycle and DNA damage response pathways. J Clin Oncol . 2005;23(26):6364-6369.
25. Woroniecka R., Pienkowska-Grela B., Grygalewicz B., et al. Significance of chromosomal markers in the diagnosis of mantle cell lymphoma (MCL). J Appl Genet . 2002;43(4):545-553.
26. Zhou B.B., Elledge S.J. The DNA damage response: putting checkpoints in perspective. Nature . 2000;408(6811):433-439.
27. Streubel B., Vinatzer U., Lamprecht A., et al. T(3;14)(p14.1;q32) involving IGH and FOXP1 is a novel recurrent chromosomal aberration in MALT lymphoma. Leukemia . 2005;19(4):652-658.
28. Isaacson P.G., Du M.Q. MALT lymphoma: from morphology to molecules. Nat Rev Cancer . 2004;4(8):644-653.
29. Bertoni F., Zucca E. Delving deeper into MALT lymphoma biology. J Clin Invest . 2006;116(1):22-26.
30. Ruiz-Ballesteros E., Mollejo M., Rodriguez A., et al. Splenic marginal zone lymphoma: proposal of new diagnostic and prognostic markers identified after tissue and cDNA microarray analysis. Blood . 2005;106(5):1831-1838.
31. Michaux L., Dierlamm J., Wlodarska I., et al. t(14;19)/BCL3 rearrangements in lymphoproliferative disorders: a review of 23 cases. Cancer Genet Cytogenet . 1997;94(1):36-43.
32. George T.I., Wrede J.E., Bangs C.D., et al. Low-grade B-cell lymphomas with plasmacytic differentiation lack PAX5 gene rearrangements. J Mol Diagn . 2005;7(3):346-351.
33. Poppe B., De Paepe P., Michaux L., et al. PAX5/IGH rearrangement is a recurrent finding in a subset of aggressive B-NHL with complex chromosomal rearrangements. Genes Chromosomes Cancer . 2005;44(2):218-223.
34. Akasaka H., Akasaka T., Kurata M., et al. Molecular anatomy of BCL6 translocations revealed by long-distance polymerase chain reaction-based assays. Cancer Res . 2000;60(9):2335-2341.
35. Lossos I.S. Molecular pathogenesis of diffuse large B-cell lymphoma. J Clin Oncol . 2005;23(26):6351-6357.
36. Lossos I.S., Morgensztern D. Prognostic biomarkers in diffuse large B-cell lymphoma. J Clin Oncol . 2006;24(6):995-1007.
37. Alizadeh A.A., Eisen M.B., Davis R.E., et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature . 2000;403(6769):503-511.
38. Ohno H. Pathogenetic and clinical implications of non-immunoglobulin; BCL6 translocations in B-cell non-Hodgkin's lymphoma. J Clin Exp Hematop . 2006;46(2):43-53.
39. Li Z., Van Calcar S., Qu C., et al. A global transcriptional regulatory role for c-Myc in Burkitt's lymphoma cells. Proc Natl Acad Sci USA . 2003;100(14):8164-8169.
40. García J.L., Hernandez J.M., Gutiérrez N.C., et al. Abnormalities on 1q and 7q are associated with poor outcome in sporadic Burkitt's lymphoma. A cytogenetic and comparative genomic hybridization study. Leukemia . 2003;17(10):2016-2024.
41. Kutock J.L., Aster J.C. Molecular biology of anaplastic lymphoma kinase-positive anaplastic large-cell lymphoma. J Clin Oncol . 2002;20(17):3691-3702.
42. ten Berge R.L., Oudejans J.J., Ossenkopele G.J., et al. ALK-negative systemic anaplastic large cell lymphoma: differential diagnostic and prognostic aspects—a review. J Pathol . 2003;200(1):4-15.
43. Helman L.J., Meltzer P. Mechanisms of sarcoma development. Nat Rev Cancer . 2003;3(9):685-694.
44. Lazar A., Abruzzo L.V., Pollock R.E., et al. Molecular diagnosis of sarcomas: chromosomal translocations in sarcomas. Arch Pathol Lab Med . 2006;130(8):1199-1207.
45. Barr F.G., Zhang P.J. The impact of genetics on sarcoma diagnosis: an evolving science. Clin Cancer Res . 2006;12(18):5256-5257.
46. Argani P., Antonescu C.R., Illei P.B., et al. Primary renal neoplasms with the ASPL-TFE3 gene fusion of alveolar soft part sarcoma: a distinctive tumor entity previously included among renal cell carcinomas of children and adolescents. Am J Pathol . 2001;159(1):179-192.
47. Parham D.M., Ellison D.A. Rhabdomyosarcomas in adults and children: an update. Arch Pathol Lab Med . 2006;130(10):1454-1465.
48. Tsuda M., Davis I.J., Argani P., et al. TFE3 fusions activate MET signaling by transcriptional up-regulation, defining another class of tumors as candidates for therapeutic MET inhibition. Cancer Res . 2007;67(3):919-929.
49. Greaves M.F., Wiemels J. Origins of chromosome translocations in childhood leukaemia. Nat Rev Cancer . 2003;3(9):639-649.
50. Vandesompele J., Baudis M., De Preter K., et al. Unequivocal delineation of clinicogenetic subgroups and development of a new model for improved outcome prediction in neuroblastoma. J Clin Oncol . 2005;23(10):2280-2299.
51. Pierotti M.A. Chromosomal rearrangements in thyroid carcinomas: a recombination or death dilemma. Cancer Lett . 2001;166(1):1-7.
52. Santoro M., Mellilo R.M., Fusco A. RET/PTC activation in papillary thyroid carcinoma: European Journal of Endocrinology Prize Lecture. Eur J Endocrinol . 2006;155(5):645-653.
53. Kondo T., Ezzat S., Asa S.L. Pathogenetic mechanisms in thyroid follicular-cell neoplasia. Nat Rev Cancer . 2006;6(4):292-306.
54. Dass D.K. Value and limitations of fine-needle aspiration cytology in diagnosis and classification of lymphomas: A review. Diagn Cytopathol . 1999;21(4):240-249.
55. Soderlund V., Tani E., Skoog L., et al. Diagnosis of skeletal lymphoma and myeloma by radiology and fine needle aspiration cytology. Cytopathology . 2001;12(3):157-167.
56. Dey P., Amir T., Al Jassar A., et al. Combined applications of fine needle aspiration cytology and flow cytometric immunophenotyping for diagnosis and classification of non Hodgkin lymphoma. Cytojournal . 2006;3:24.
57. Kaleem Z., White G., Vollmer R.T. Critical analysis and diagnostic usefulness of limited immunophenotyping of B-cell non-Hodgkin lymphomas by flow cytometry. Am J Clin Pathol . 2001;115(1):136-142.
58. Gong J.Z., Lagoo A.S., Peters D., et al. Value of CD23 determination by flow cytometry in differentiating mantle cell lymphoma from chronic lymphocytic leukemia/small lymphocytic lymphoma. Am J Clin Pathol . 2001;116(6):893-897.
59. Barry T.S., Jaffe E.S., Kingma D.W., et al. CD5+ follicular lymphoma: a clinicopathologic study of three cases. Am J Clin Pathol . 2002;118(4):589-598.
60. Safley A.M., Buckley P.J., Creager A.J., et al. The value of fluorescence in situ hybridization and polymerase chain reaction in the diagnosis of B-cell non-Hodgkin lymphoma by fine-needle aspiration. Arch Pathol Lab Med . 2004;128(12):1395-1403.
61. Kocjan G. Best Practice No 185. Cytological and molecular diagnosis of lymphoma. J Clin Pathol . 2005;58(6):561-567.
62. Dey P. Role of ancillary techniques in diagnosing and subclassifying non-Hodgkin's lymphomas on fine needle aspiration cytology. Cytopathology . 2006;17(5):275-287.
63. Liu K., Dodd L.G., Powers J.A., et al. Detection of trisomy 12 by fluorescence in situ hybridization on archival cytopathologic material in chronic lymphocytic leukemia/small lymphocytic lymphoma. Acta Cytol . 2000;44(3):368-374.
64. Bentz J.S., Rowe L.R., Anderson S.R., et al. Rapid detection of the t(11;14) translocation in mantle cell lymphoma by interphase fluorescence in situ hybridization on archival cytopathologic material. Cancer . 2004;102(2):124-131.
65. Kishimoto K., Kitamura T., Fujita K., et al. Cytologic differential diagnosis of follicular lymphoma grades 1 and 2 from reactive follicular hyperplasia: cytologic features of fine-needle aspiration smears with Pap stain and fluorescence in situ hybridization analysis to detect t(14;18)(q32;q21) chromosomal translocation. Diagn Cytopathol . 2006;34(1):11-17.
66. Richmond J., Bryant R., Trotman W., et al. FISH detection of t(14;18) in follicular lymphoma on Papanicolaou-stained archival cytology slides. Cancer . 2006;108(3):198-204.
67. Jiang F., Katz R.L. Use of interphase fluorescence in situ hybridization as a powerful diagnostic tool in cytology. Diagn Mol Pathol . 2002;11(1):47-57.
68. Wolff D.J., Bagg A., Cooley L.D., et al. Guidance for fluorescence in situ hybridization testing in hematologic disorders. J Mol Diagn . 2007;9(2):134-143.
69. Frater J.L., Tsiftsakis E.K., His E.D., et al. Use of novel t(11;14) and t(14;18) dual-fusion fluorescence in situ hybridization probes in the differential diagnosis of lymphomas of small lymphocytes. Diagn Mol Pathol . 2001;10(4):214-222.
70. Alkan S., Lehman C., Sarago C., et al. Polymerase chain reaction detection of immunoglobulin gene rearrangement and bcl-2 translocation in archival glass slides of cytologic material. Diagn Mol Pathol . 1995;4(1):25-31.
71. Gong Y., Caraway N., Gu J., et al. Evaluation of interphase fluorescence in situ hybridization for the t(14;18)(q32;q21) translocation in the diagnosis of follicular lymphoma on fine-needle aspirates: a comparison with flow cytometry immunophenotyping. Cancer . 2003;99(6):385-393.
72. Buno I., Nava P., Alvarez-Doval A., et al. Lymphoma associated chromosomal abnormalities can easily be detected by FISH on tissue imprints. An underused diagnostic alternative. J Clin Pathol . 2005;58(6):629-633.
73. Belaud-Rotureau M.A., Parrens M., Carrere N., et al. Interphase fluorescence in situ hybridization is more sensitive than BIOMED-2 polymerase chain reaction protocol in detecting IGH-BCL2 rearrangement in both fixed and frozen lymph node with follicular lymphoma. Human Pathol . 2007;38(2):365-372.
74. Belaud-Rotureau M.A., Parrens M., Dubus P., et al. A comparative analysis of FISH, RT-PCR, PCR, and immunohistochemistry for the diagnosis of mantle cell lymphomas. Mod Pathol . 2002;15(5):517-525.
75. Caraway N.P., Gu J., Lin P., et al. The utility of interphase fluorescence in situ hybridization for the detection of the translocation t(11;14)(q13;q32) in the diagnosis of mantle cell lymphoma on fine-needle aspiration specimens. Cancer . 2005;105(2):110-118.
76. Caraway N.P., Du Y., Zhang H.-Z., et al. Numeric chromosomal abnormalities in small lymphocytic and transformed large cell lymphomas detected by fluorescence in situ hybridization of fine-needle aspiration biopsies. Cancer . 2000;90(2):126-132.
77. Shin H.J., Thorson P., Gu J., et al. Detection of a subset of CD30+ anaplastic large cell lymphoma by interphase fluorescence in situ hybridization. Diagn Cytopathol . 2003;29(2):61-66.
78. Parrens M., Belaud-Rotureau M.A., Fitoussi O., et al. Blastoid and common variants of mantle cell lymphoma exhibit distinct immunophenotypic and interphase FISH features. Histopathology . 2006;48(4):353-362.
79. Bridge R.S., Rajaram V., Dehner L.P., et al. Molecular diagnosis of Ewing sarcoma/primitive neuroectodermal tumor in routinely processed tissue: a comparison of two FISH strategies and RT-PCR in malignant round cell tumors. Mod Pathol . 2006;19(1):1-8.
80. Jambhekar N.A., Bagwan I.N., Ghule P., et al. Comparative analysis of routine histology, immunohistochemistry, reverse transcriptase polymerase chain reaction, and fluorescence in situ hybridization in diagnosis of Ewing family of tumors. Arch Pathol Lab Med . 2006;130(12):1813-1818.
81. Udayakumar A.M., Sundareshan T.S., Goud T.M., et al. Cytogenetic characterization of Ewing tumors using fine needle aspiration samples. A 10-year experience and review of the literature. Cancer Genet Cytogenet . 2001;127(1):42-48.
82. Kilpatrick S.E., Bergman S., Pettenati M.J., et al. The usefulness of cytogenetic analysis in fine needle aspirates for the histologic subtyping of sarcomas. Mod Pathol . 2006;19(6):815-819.
83. Hummel P., Yang G.C., Kumar A., et al. PNET-like features of synovial sarcoma of the lung: a pitfall in the cytologic diagnosis of soft-tissue tumors. Diagn Cytopathol . 2001;24(4):283-288.
84. Sápi Z., Antal I., Pápai Z., et al. Diagnosis of soft tissue tumors by fine-needle aspiration with combined cytopathology and ancillary techniques. Diagn Cytopathol . 2002;26(4):232-242.
85. Das K., Mirani N., Hameed M., et al. Fine-needle aspiration cytology of alveolar rhabdomyosarcoma utilizing ThinPrep liquid-based sample and cytospin preparations: a case confirmed by FKHR break apart rearrangement by FISH probe. Diagn Cytopathol . 2006;34(10):704-706.
86. Stewart A.K., Bergsagel P.L., Dispenzieri A., et al. A practical guide to defining high-risk myeloma for clinical trials, patient counseling and choice of therapy. Leukemia . 2007;21(3):529-534.
CHAPTER 3 Cytologic Screening Programs

Luiz M. Collaço, Lucilia Zardo

Contents
Principles of screening
Cervical Cancer and Screening
Cervical Cancer Incidence and Mortality Worldwide
Efficacy of Screening
Design of Screening Programs
Features of Successful Screening Programs
Limitations of Screening Programs
Screening Programs and HPV Vaccine
Screening Programs and HPV DNA Test
The Role of Laboratory in Screening Programs
Early Detection of Cancer in Other Sites
New Developments in Cytological Screening
Liquid-Based Cytology (LBC)
Automated Cytology
Concluding Remarks

Principles of Screening
Screening of diseases gained significance in medicine at the end of the nineteenth century, when public health authorities emphasized the importance of screening methods for certain diseases. An example is the radiological screening of immigrants, searching for infectious diseases such as tuberculosis in the USA. 1
The idea of screening for early detection of cancer was accepted in the 1920s after the development of exfoliative cytological techniques initiated through the work of Babes 2 and Papanicolaou. 3 In 1941 George Papanicolaou demonstrated a test for the early detection of cervical cancer, contributing toward the creation of screening programs. 4, 5 Prevention and early diagnosis are major factors in reducing morbidity and mortality resulting from neoplasia. 6
Screening of diseases presumes a test or examination that can detect the existence of a particular disease in a high-risk population, asymptomatic or with minimum symptoms of the disease. Systematic screening of diseases requires a series of elements with the objective of decreasing mortality from a particular disease. For this reason the World Health Organization lists certain principles to guide the screening systems: 7
1. The condition to be evaluated must be an important cause of morbidity or mortality.
2. The natural history of the disease must be known as well as forms of intervention in the pre-clinical stage or with the disease installed.
3. The test used for screening must have a high level of sensitivity and specificity.
4. The test to be applied must be low risk, with good acceptability by the target population and the scientific community.
5. In the case of positive tests, diagnostic methods for confirming screening finds should be possible.
6. The test must be shown to be efficient in reducing morbidity or mortality caused by the disease.
It therefore follows that screening of a particular disease requires a precise test, easy to do, at a low cost, and the capability of detecting the presence of a lesion. In principle this is not a test for a definitive diagnosis, although it can in some situations serve to indicate subsequent therapy.

Cervical Cancer and Screening
Cancer of the uterine cervix is an important cause of morbidity and mortality among women worldwide and a leading public health problem. It is the second most common cancer in women, but the most common in developing countries. 8
Because of the phases that precede the lesion in the natural progress of invasive cervical cancer, and because they can be easily discovered and treated, the disease is well suited to screening programs. The Papanicolaou test is an established method for examining the cells collected from the cervix to determine whether they show signs of pre-neoplastic differentiation.
Cytologic screening programs have led to a large decline in cervical cancer incidence and mortality in developed countries. However, cervical cancer remains largely uncontrolled in high-risk developing countries because of ineffective or no screening. 9 Approximately 85% of new cases of cervical cancer (estimated at 493,000 worldwide) and deaths from cervical cancer affect women in developing countries each year. 10
Cervical cytology, originally perceived useful in the detection of pre-invasive disease and not just for identification of invasive cervical cancer, came to be seen as a technique destined to prevent cervical cancer.
In the 1960s, its use spread among developed countries; meanwhile the concept that invasive squamous cell carcinoma of the cervix arises from a spectrum of intraepithelial precursor lesions appeared. This concept changed with the evolution of scientific knowledge on the central role of human papillomavirus (HPV) in pathogenesis of cervical cancer and its precursor lesions. 11 Although this morphology-based model of a continuum has now been supplanted by a more discrete theory of multistage carcinogenesis, the cervical intraepithelial neoplastic scale still merits consideration as the current basis of clinical management. 12
Cervical cancer screening is an example of success in the prevention of cancer. Unfortunately the majority of women who develop cervical cancer live in countries where there is a lack of infrastructure to support the organization and management of programs, or where other obstacles such as social and cultural questions make their participation difficult. Permanent efforts to find new and more effective strategies will be necessary to expand the access and participation of these women, optimizing resources and modifying the mortality statistics for the disease, mainly in these areas.

Cervical Cancer Incidence and Mortality Worldwide
Currently cervical cancer is potentially curable, but still continues to be the second most frequent cause of death by neoplasia in women and the survival rate in 5 years varies from 44 to 66%. 13
The highest incidence occurs in Latin America, the Caribbean, Africa (tropical sub-Sahara), and South and Southeast Asia 8 ( Fig. 3.1 ). Around 80% of the cases occur in developing countries and just 20% in developed countries. Socioeconomic and cultural aspects are a factor in this unequal distribution of this neoplasia around the world. However, a preponderant factor in the areas of low incidence is the level of information from the feminine population regarding the disease and the continual screening of this population. On the other hand, in developing countries, the low level of awareness of the problem, the lack of interest of the sanitary authorities, and the use of opportunist screening favors the continuance of this unfavorable situation and indicates the urgent need for the public health authorities to find a solution.

Fig. 3.1 Age-standardized (world) incidence rates of cervical cancer 2002.
Reproduced with permission of Parkin et al. Vaccine 2006;24(Suppl 3):12.
An important number of risk factors for cervical carcinoma have been identified and can therefore be controlled, avoiding the progress of pre-neoplastic lesions. These factors are early start to sexual activity, multiple partners, the number of partners a man has, infection by oncogenics HPV, precarious genital hygiene, and smoking.
Histologically the largest number of cases is of squamous cell carcinoma; however, the incidence of cervical adenocarcinoma has gradually increased over the past decades, particularly in young women, where it has doubled. 14 A larger number of adenocarcinomas are being identified, either by control of cervical cancer in developed countries or by association with HPV infection, above all the type 18.
Programs applied in Scandinavian countries and in Canada demonstrate that with continuous screening, it is possible to reduce mortality from cervical cancer by almost 75%. However, the reduction of the mortality rate is necessarily related to the real efforts by doctors and population, the frequency and quality of the specimen collection, the examination and diagnostic analysis, adequate communication between the specialists, and the efficacy of the system for management of the patients. 15

Efficacy of Screening
The efficacy of cytological screening for cervical cancer depends on the possibility of modifying the course of the disease through identification of women with high-degree precursor lesions and invasive initial lesions. With this it is possible to distinguish the woman apparently not affected from the woman who could have the disease.
Even though the efficacy of cytology screening has never been proven through randomized trials, it is generally agreed that the marked reduction in the incidence and mortality from cervical cancer before and after the introduction of screening programs in a variety of developed countries has been interpreted as strong non-experimental support for organized cervical cancer screening programs. 17, 18
The best known studies are those that compare incidence and mortality in Iceland and in the four Nordic countries 19 - 21 ( Fig. 3.2 ). Before screening was installed in Iceland, mortality had been on the increase but fell 50% in the period of 10 years from introducing screening. In the Nordic countries, the decline in cumulative incidence rates over a 15-year period, between 1966–70 and 1981–5, was related to the coverage and extent of the organized programs. In Norway, where only 5% of the population had been screened opportunistically, the incidence rates fell by 20% in comparison to Finland, with a national population-based program, where incidence fell by 65%.

Fig. 3.2 Incidence and mortality rates of cervical cancer in the Nordic countries, 1958–97 (mortality available up to 1996). Whole female population, adjusted for age to the world standard population (Laara et al. (1987); Engeland et al. (1993); Hristova and Hakama (1997); Parkin et al. (1997); Moller et al. (2002); EUROCIM (European Network of Cancer Registries) database). 21, 74 - 77 Reproduced with permission of IARC—International Agency for Research on Cancer. 12
In a study of invasive cervical cancer in British Columbia, approximately half of the new cases diagnosed had no previous cytology or the last examination had been made more than 5 years ago. 22
Two important parameters traditionally used to measure the validity of screening tests are sensitivity and the specificity. The sensitivity means the percentage of positive cases reported as being positive. It relates to the ability of disease detection and it can be calculated using the formula

The specificity means the percentage of negative cases reported as being negative. It relates to the ability of disease exclusion and it can be calculated using the formula

A third criterion is the positive predictive value that measures the probability of the disease to be present in the patients whose test was reported as positive, and it can be calculated using the formula

Glandular lesions are much less frequent than those originating from squamous epithelium and the diagnosis of the intraepithelial forms is the principal objective of the screening programs. In relation to the prevention of cervical adenocarcinoma, the Papanicolaou test is potentially a powerful weapon, but in comparison to the diagnosis of squamous lesions, the diagnosis of cervical adenocarcinoma in situ has shown a significantly higher rate of false-negatives, not being so effective in the prevention of invasive glandular lesions. 23
In 2004 a working group at the International Agency for Research on Cancer (IARC) of the World Health Organization (WHO) met to evaluate the efficacy of prevention of cervical cancer in reducing mortality caused by the disease. They concluded that the programs of prevention based on the Papanicolaou test continue being the mainstay for prevention of this type of cancer throughout the world, there being sufficient evidence that screening of cervical cancer diminishes mortality caused by the disease. 24
Despite the knowledge of the efficacy of cytopathologic tests in contributing to the reduction of cervical cancer through organized programs by their characteristics of simplicity, acceptability, and low cost, studies have shown major variations in the estimates of sensitivity and specificity of the test. A meta-analysis to estimate the accuracy of the Pap test in which data from 59 studies were combined reported estimates of sensitivity and specificity ranging from 11 to 99% and 14 to 97%, respectively. 25 A systematic review reported sensitivity and specificity ranging from 30 to 87% and 86 to 100%, respectively. 26
Sensitivity and specificity are important parameters for the evaluation of the accuracy of the screening test. However, it is important to bear in mind that the efficacy of screening is not restricted to the performance of the test used. Special emphasis must be given to the need to develop organized programs that have a systemic approach, that are well integrated into the existing health system, and which consider social, cultural, and economic aspects. A meta-analysis of social inequality and the risk of cervical cancer in 57 studies revealed that both cervical infection with HPV and a lack of access to adequate cervical cancer screening and treatment services are likely to be important in explaining the large cervical cancer incidence rates observed in different socioeconomic groups. 27 An estimated 100% increased risk of invasive cervical cancer was found for women in low social class categories when compared with women in high social class categories.
In an analogous way, this difference occurs in developing countries and those developed where the inequality in the access is the consequence of the inequality in the quality of the services. Past failures of cervical screening in developing countries are attributable to failures in program quality, rather than to technological limitations of the screening test. 28, 29
It is very important to evaluate the efficacy of the programs in reducing the disease, and also whether the screening approach chosen is cost effective, before considering extending implementation to large populations. A screening program is justified if a prior diagnosis that permits a cost-effective and measurable reduction of the disease is made.

Design of Screening Programs
Cervical cancer can be avoided when there is an early diagnosis of the precursor lesions, without local or systemic compromise. The implementation of a systematic program of prevention of gynecological cancer among women in British Columbia in 1949 reduced both the incidence and the mortality caused by this neoplasia. Among the methods available for early detection of cervical cancer, exfoliative cytology, or the Pap test, is recommended worldwide for mass screening, because the efficacy in the detection of premalignant lesions, associated with the social role of the method, permits minimization of costs with curative medicine. 1, 4, 6 - 8 ,13 ,15 ,30
From that stated above, a routine of procedures essential to the success of a program of prevention may be obtained. The basic integrated actions include: (1) care with collection, (2) processing of the smears, (3) screening and interpretation of the specimens, (4) follow-up of the patients, and (5) quality control. 31
1. Care with collection —The majority of false-negatives arise from problems with collection of specimens, and for this reason this stage should be systemized and there should be training and recycling of the personnel responsible for taking the samples. The smears must be well identified, slim, uniform, and without contaminants, and contain samples from the transformation zone, where in the majority of cases the cervical cancer develops. There should be a minimum of blood, mucus, or other obscuring material such as lubricating gel. It is also important at this moment to adequately fix the material so as not to compromise subsequent stages.
2. Processing the specimens —In general prevention programs cover a large number of tests, so laboratories should have guidance regarding the systemizing of the processing and the recording of a large volume of specimens. One of the characteristics of the Pap test is that it consists of various stages. Each stage should be monitored so as to minimize the possibility of error. The condition on arrival of the slides, and the number of slides per case, must be verified. Special care should be taken with the flow of the tests, with adequate numbering and balanced coloration with control of the number of cases colored in each set. The end product of this stage will be fundamental to a good result with the rest.
3. Screening and interpretation of the specimens —The screening should be done in as little time as possible, depending on the basic requirements of each program, by trained and qualified personnel. Care should also be taken with excess workloads for cytopathologists and cytotechnicians, and also with refresher courses and recycling. The report on the tests should be systemized and use a unique nomenclature, of which all involved in the preparation and interpretation of the results should be fully aware.
4. Follow-up of patients —The prevention program should include reference, contra-reference, and active search services. The mere detection of the lesions will not determine the impact on the natural history of the disease. For this reason the treatment of lesions in a pre-invasive stage is fundamental. Outpatient treatment centers for the more simple cases, and others for more complex cases should be integrated into a service network for the program. Mechanisms for finding and managing patients who did not return after the initial test and who show alterations are fundamental for a prevention program to function well.
5. Quality control —Quality is fundamental in gynecological cytopathology. One of the greatest problems in mass cytology is the false-negative cases. Cytopathology labs must have mechanisms for internal quality control with the objective of avoiding false-negative and false-positive tests. These mechanisms should include measures relating to the screening and interpretation of the specimens, a review of 10% of the cases seen by the cytotechnician, grouping the technicians according to hierarchy. External quality control must be included in the design of the prevention program, conducted by an accredited entity and with interlab action with the objective of guaranteeing the homogeneity and quality of the laboratory procedures. 32 It should also function as a detector of eventual problems and could indicate a need for redirectioning continuous education efforts within the program. For additional information on quality assurance in cytopathology see Chapter 4 .
Cancer screening may be offered to a population either as an organized program or opportunistically, or as some combination of the two. Opportunistic screening is spontaneous and initiated either by the individual or healthcare provider during routine healthcare encounters. It is often associated with low coverage of people at high risk and excessive repetition of procedures at frequent intervals, high costs, and a small benefit at the population level. Systematic or organized screening programs refer to planned and concerted public health application of early detection and treatment in defined populations, operating under precise protocols and guidelines. 33
Some countries with organized screening programs can reduce the incidence of cervical cancer by up to around 80% in areas with high-quality screening, good coverage, and a reliable follow-up. Organized programs with systematic call-up, recall, follow-up, and vigilant systems have shown more expressive effects with less resources than less organized programs. 34 Various alternative screening strategies are being researched for developing countries, although the challenge in less-developed countries is surpassed by the complex array of problems that go far beyond the introduction of simplified technology. 35

Features of Successful Screening Programs
The success of cervical cancer screening is shown by its ability to reduce the incidence of cervical cancer and the resulting mortality, in a cost-effective way. To be successful it is fundamental that the program is organized and broad-based, developing along the line of care for cervical cancer. All the stages involved in the finding of the women, the collection of material for the cytological test, transport and processing of the slides, identification of lesions, and finally the delivery of results, treatment, and follow-up of the women with alterations should happen in sequence, synchronized and with the highest quality. Any failure in one of these stages can compromise the impact of the screening on the health of the population. The following are some aspects of successful screening programs:
• Government policy : Planning within a governmental policy and national planning for cancer control. This includes the definition of the age range of the population to be prioritized and the frequency (interval) of screening, apart from production of instructions to guide the process, including recommendations regarding nomenclature and the therapeutic action for the lesions identified.
• Coverage : Measures to guarantee good coverage, with special attention to identification of women in the target population. Education of these women regarding cervical cancer screening can contribute to increased attendance and confidence in the procedure, apart from facilitating understanding of the results of the cytological test.
• Integrated system : The different levels of healthcare in the program should be integrated like a network, with the capacity to ensure continuity of the care within the different levels.
• Health professionals : Good results can be achieved by educating and training the health professionals, improving the attention given to the women, the quality of samples collected, the quality of the screening and the results of the tests, and also the research and follow-up of the patients with lesions needing treatment.
• Quality of the diagnosis : Efficient and high-quality laboratory service, which should preferably be centralized; quality control of cytology reading.
• Infrastructure of health services : Adapting the services to give the treatment needed, with the capacity for attending to the planned demand, in relation both to equipment, installations, and material and to the human resources available. It is of fundamental importance to guarantee the supply and the accessibility of the health services.
• Information system : An integrated information system linking the different elements of the program, permitting identification of each woman and the exchange of management information, and monitoring and referring women with results showing alterations to the respective health services, with a view to ensuring that these patients receive appropriate diagnosis and treatment, should be achieved.
• Indicators : Monitoring and evaluating cervical cancer prevention programs is essential for effective, efficient planning and service organization, as well as for patient management. Indicators created to evaluate performance at the different stages of the program should be monitored regularly, using information generated preferably through the routine information system. 25
• Leadership : Leadership, management skills, attention to linkages at all levels of the program, and budgeting skills are essential. 8

Limitations of Screening Programs
Limitations of prevention programs can be related to different factors, such as errors and failures in the program as well as socioeconomic and cultural problems.

Errors and Failures in the Program
The first limitation refers to the Pap test. Although it is the most effective screening test in oncology, it shows failures with low sensitivity where the false-negatives vary between 3 and 13%, and high specificity, with false-positives less than 5%.
Achieving the ideal coverage is another problem that limits screening programs. For a prevention program to diminish cervical cancer mortality it must achieve a coverage estimated at around 80% of the women, so this must be the target. For this the community must be mobilized and informed in order to make the women realize the causes and consequences of cervical cancer and so submit to the tests. The warnings should be spread through all means of communication including explanatory folders and pamphlets. Lack of knowledge becomes one of the main allies of the inefficiency of screening programs. Strategies must be established to encourage regular participation of women in the program and the return of women with abnormal results. Around 29% do not return after taking the test. 36 Some measures that can help are assistance with transport, slides/films, and personal letters with folders. It is very important to individualize the incentives according to the socioeconomic level of the patients.
Other causes of failures in the programs include inadequate collection of material to be examined, errors of interpretation in the cytopathology lab, absence of adequate follow-up, and failures in the treatment of precursor lesions. It becomes fundamental to keep the multidisciplinary teams who work with prevention stimulated and up to date, aware of the importance of each stage and its role in achieving the final result, which is to avoid death by cervical cancer.

Socioeconomic and Cultural Problems
Limitations that extrapolate the limits of the program and for this reason can be considered as extrinsic also occur. In this group are political limitations, as mainly in developing countries other questions could be prioritized and oppose the prevention of cervical cancer just as with survival, infectious/contagious diseases such as tuberculosis or AIDS, vaccination, or the fight against hunger and bad nutrition.
Sociocultural and behavioral questions can also be a negative influence in the running of programs and should be detected and minimized. This group includes the level of education of the women that could influence directly actions and response to the programs, religious aspects, and the decisive influence of the companion on the woman, contributing to diminish women's adhesion to the screening programs.
Finally it is important to those countries or regions installing screening programs to pay attention to the following points 37 :
1. Try to screen the largest possible number of women—with at least one test during their lifetime.
2. Focus on the workup for those women who had a bad diagnosis in the primary screening such as high-degree lesions or invasive carcinomas.
3. Design the programs remembering that it will not be possible to detect all the carcinomas.
4. Try to make the women, medical staff, politicians, and legal authorities understand that the Pap test is not a perfect test, and that there are advantages and disadvantages in its use. Because of this they should make an effort to optimize results and even add methods to help the prevention program achieve its targets.

Screening Programs and HPV Vaccine
HPV is a sexually transmitted infection recognized as necessary for the development of cervical cancer, and this strong association is currently seen as causal. 11 Oncogenic types of HPV DNA are detected in virtually all cervical cancers and recognition of this crucial role has stimulated the investigation and development of HPV vaccines in both prophylactic and therapeutic settings. The natural history of this cancer offers two opportunities for intervention to interrupt the course of the disease: primary and secondary prevention. Primary prevention refers to measures to prevent infection by HPV using vaccines. Secondary prevention, which is detection and treatment of precursor cervical cancer lesions, traditionally uses cytopathology as a screening test.
The prophylactic HPV vaccines are prepared from empty viral capsids called virus-like particles (VLPs) composed of the capsid proteins L1 and L2. These particles do not contain viral genetic material and thus are unable to multiply, which means they are non-infectious. Early studies showed that L1 protein has the intrinsic capacity to assemble into empty capsid-like structures whose immunogenicity is similar to that of infectious virions. 38 Studies of HPV16 L1 VLP vaccine have indicated that they were well tolerated and highly immunogenic, generating high levels of antibodies against HPV16. 39
Two studies provided convincing evidence that the prophylactic vaccine is effective in preventing new and persistent infections of the genital tract with the two types of HPV most commonly associated with cervical cancer and its precursors.
The first published randomized, clinical trial of prophylactic vaccines developed to determine whether an HPV16 L1 VLP vaccine could prevent HPV16 infection in women reported that the vaccine had a 100% efficacy in the vaccine group when compared to the placebo group. In this study, women who received the vaccine had the titer of HPV16 antibodies 58.7 times as high as the titer among women with serologic evidence of natural HPV16 infection. 40
A second study on a randomized, double-blind, controlled trial to assess the efficacy, safety, and immunogenicity of a bivalent HPV16/18 L1 VLP vaccine showed that the vaccine was effective in prevention of incident and persistent cervical infections with these two virus types, and associated cytological abnormalities and precancerous lesions. This study reported that the vaccine efficacy was 91.6% against incident infection and 100% against persistent infection with HPV16/18. 41
So far there exist two vaccines with types HPV16 and 18, more commonly associated with cervical cancer, which protect against both new and persistent infections. One of the vaccines also includes types 6 and 11, with a protector effect against genital warts. 42 With this it is hoped to prevent around 70% of cervical cancer cases worldwide among women who have never been exposed to the high-risk types of virus. Bearing in mind that the prevalence of types 16 and 18 varies from country to country, it is believed that the vaccine that would include the seven types of most common HPV in the world (16, 18, 45, 31, 33, 52, 58) would be able to prevent about 87% of all cases, with small regional variations. 43 However, the addition of multiple new types of VLP in one unique vaccine could present technical obstacles for the manufacturers. 44
Although the results of studies with vaccine are promising, some questions still need to be answered for an efficient implementation. The first of them refers to viral types other than 16 and 18 strongly associated with cervical cancer. 18 Screening and treatment services will still be required, because the vaccines only prevent about 70% of cervical cancer cases and because it will be years, if not decades, before we see the full benefit of vaccination in terms of a reduction in the incidence of cervical cancer. 45 Efforts to develop multivalent vaccines that could control most HPV infections associated with cervical disease will be necessary.
Furthermore, duration of the antibody response and protection from vaccination remains to be determined and a long-term follow-up of vaccinated women is required before the impact can be fully identified. 18 Trial data for both vaccines suggest they offer a minimum of 4–5 years efficacy, of close to 100%, in preventing persistent infection by the vaccine genotypes. 46 There are no predictions regarding prevention of cancer among women who have already experienced an infection. The vaccines are not designed to treat people who have already been infected with these genotypes. Moreover, it is not known whether effective coverage against some genotypes could favor the emergence of more pathogenic types. 42
Some subjects are particularly relevant for programs, and service delivery strategies still remain under study. There are several open issues as well as the performance in Africa where chronic malnutrition, HIV, and other infections diseases may compromise the immune response, the cross-protection, vaccine compatibility, and the safety and efficacy in specific populations, such as pregnant women and immunocompromised patients. There is still a lack of data on infants and young children. 47
While the development of a prophylactic HPV vaccine may be the ultimate solution to the prevention of cervical cancer, it is unlikely that the vaccine will be widely available in low-resource settings within the next decade. In the meantime large numbers of women already infected with risk types of HPV remain at risk of developing cervical cancer. 18 To prevent the disease in women already infected or in those infected with other types of virals not included in the vaccines, it will be necessary to maintain and improve cervical cancer control actions, principally in developing countries where the programs are less effective.
To measure the clinical benefits and cost-effectiveness of HPV vaccination, a computer-based model of various cancer prevention policies has been developed. The most effective strategy was one that combined vaccination and triennial conventional cytologic screening strategies. This approach would reduce the absolute lifetime risk of cervical cancer by 94% compared with no intervention, in a cost-effective manner. 48 HPV vaccine will be an additional tool in the strategies for reducing morbidity and mortality from cervical cancer but will not replace screening and early treatment, and will be a component of a comprehensive strategy with the long-term goal of eliminating cervical cancer. 44

Screening Programs and HPV DNA Test
Cervical-vaginal cytology continues to be the method of choice for mass prevention of cervical cancer. However, there are discussions regarding its limitations with regard to sensitivity, specificity, and ability to reproduce. 49
One of the first applications of the HPV test in clinical practice has been in women referred for a colposcopy after an abnormal pap smear. 50 The combination of methods has been proposed in an attempt to improve the sensibility of the Pap test. Among these the association of cytology with the molecular test for HPV using hybrid capture (HC) has been highlighted. This technology presents high sensitivity for high-degree intraepithelial lesions and age-dependent specificity. In young women the specificity of HC-2 is lower than cytology and in women over 35 the specificity is equal to cytology. Studies show that the woman with negative HC-2 for HPV with normal cytology is at low risk for developing cervical cancer in the next 10 years. 50 Other studies have shown an improvement in sensitivity in cervical intraepithelial neoplasia (CIN) 3 cases.
The FDA approved HC-2 for HPV DNA as an assistant method for cytology in women over the age of 30. 24 Apart from the improvement in sensitivity in the detection of lesions the use of HC will bring about a larger spacing of time between screenings and virtual reduction in the number of consultations in a screening program.
However, some studies have shown evidence that the HPV test could be potentially superior in screening when compared with cytology. 51, 52

The Role of Laboratory in Screening Programs
The laboratory can make an important contribution to the structuring and organization of cervical screening programs based on the Papanicolaou test. It should be included in the global planning, with a logical structure of hierarchy, regionalization, and above all integration with the healthcare system.
Data collected through the laboratory, apart from producing epidemiological and administrative information on the results of the tests, permit the creation of indicators to help in monitoring and evaluating, not just the quality of the laboratory activities, but also the quality of programming, thus helping to generate pertinent and useful investigations which contribute toward the improvement of effectiveness and efficiency of the program. 53 Using laboratory data it is possible to achieve some of the required guarantees in cervical screening programs, such as to confirm that women in the target population are being screened and are receiving appropriate management, and to confirm the target geographic group coverage ( Table 3.1 ). 29

Table 3.1 Confirming that women in target demographic groups are screened
Monitoring of the different steps in the program can provide valuable information for the identification of problems and planning of respective measures for improvement. Care in the setting of parameters and indicators is fundamental in the monitoring of each stage of the program, in order to take into account local differences. The indicators can provide useful data both for the laboratory and for the local and/or regional manager in programming suitable action.
The lab, when integrated into a screening program, should have among its objectives top quality production, training, and updating of personnel and the guarantee of a secure place of work, where risk factors are under control and the environment is protected. 53
In the area of laboratory governance, it is possible to contribute toward an atmosphere of monitoring and evaluation, which helps in decision-making and guarantees attention to quality. The system of internal monitoring of laboratory quality includes a set of actions, which should be developed and disseminated in a coordinated way, involving the various stages in the work process, from collecting a sample to issuing the report. The system aims to accompany and evaluate the cyto- and histopathologic diagnostic procedures in the laboratories, thus helping to determine areas where improvements can be planned and implemented, and also evaluate the impact of these actions and the incorporation of new practices.
In some situations, the indicators measured in the lab can show evidence of problems in previous stages, as for example in the collection of material or in the use of an inadequate fixer. At these times, it is the role of the lab, acting as an integral part of the network, to inform the health units that send the material for examination to improve quality in a team effort, assisting in the planning and implementation of corrective measures and improvements.
It is recommended that laboratories have a system that permits monitoring of test quality, establishing evaluation criteria and maintaining records of the results found. This is a simple and low-cost measure that reflects technical advances of the staff involved, improving the relationship between the clinic and the lab, and in the last analysis improving care to the patients.

Early Detection of Cancer in Other Sites
As mentioned at the beginning of this chapter, screening refers to the use of simple tests across a healthy population in order to identify individuals who have disease, but do not yet have symptoms. Early diagnosis, a concept different from screening, refers to the detection of early clinical stages of disease in symptomatic subjects. 54
Evaluations of the potential of the Pap test to be a practical screening test for endometrial cancer have shown both sensitivity and positive predictive value too low. It is likely that the early detection of some endometrial cancers are mainly incidental. However, abnormal endometrial cells on the Pap test may be markers for increased risk, especially when they are present in the secretory phase of the menstrual period, or when the patient is postmenopausal. 55 When unexpected bleeding occurs, evaluation becomes diagnostic rather than screening and the woman should undergo an endometrial biopsy and/or other diagnostic tests as appropriate. 56
Prospective studies of lung cancer screening have not demonstrated persuasively that screening for lung cancer with sputum cytology in combination with chest radiography saves lives. 56, 57 Although none of the studies showed fewer deaths in the experimental group than in the control group, none of the studies compared disease outcome in a group offered screening with a group strictly not invited to, or encouraged to, have screening. Sputum cytology was believed to have potential for the early detection of lung cancer, but showed little added advantage over chest X-ray in the NCI cooperative trials, and was not associated with any reduction in deaths from lung cancer. 56 However, sputum cytology was effective in identifying roentgenographically occult lung carcinoma in its early and occult stages, particularly in patients at high risk for this disease, where lung carcinoma was suspected on the basis of symptoms, smoking, or airflow obstruction. 58
Esophageal cancer has a very poor prognosis, mainly because most tumors are asymptomatic and do not receive medical attention until they are unresectable. Early detection is the key to the treatment. At the current time, screening and surveillance for esophageal cancer is still controversial and screening is not recommended except for very selected subgroups. The presence of risk factors such as long-term tobacco or alcohol use, achalasia, or squamous head and neck cancers may identify selected groups for screening. 59
Another indication is in high-incidence populations, such as those found in northern China where esophageal cancer is endemic. 60 Since the late 1950s, Chinese scientists have performed many studies to develop early detection strategies for this disease. The principal early detection technique developed was esophageal balloon cytology. With this technique it was possible to collect exfoliated cells and scrape the surface of the esophageal mucosa. Although the comparison of cytological diagnoses with concurrent histological findings showed low (14–36%) sensitivities for the cytological detection of biopsy-proven cancers, cytologic screening could decrease the mortality by facilitating early detection of the disease. 61 Positive cytology must be verified by endoscopy and biopsy.
Urine cytology is an inappropriate test for screening the general population. Because of the low prevalence of bladder cancer, the positive predictive value of the tests is low. There is inadequate evidence to determine whether screening for bladder and other urothelial cancers would have any impact on mortality. 62
Urinary cytology can be helpful in the follow-up of patients exposed to carcinogenic agents by contributing to early diagnosis of urinary tract tumors. By far the greatest known environmental risk factor in the general population is tobacco, especially cigarette smoking, with individuals who smoke having a four- to sevenfold increased risk of developing bladder cancer compared to individuals who have never smoked. 62, 63 Bladder cancer was first linked to an occupational exposure. An increased risk of bladder cancer has been identified for a variety of occupations including chemical and rubber workers, leather workers, painters, dye workers, truck drivers, and garage and gas station workers. 64, 65, 66
Urine cytology is primarily used for detection of cancer in patients like those exposed to industrial chemicals and metals, cigarette smokers, and those with schistosomiasis, associated with increased risks of bladder cancer. Furthermore, it is also used for diagnosis of symptomatic patients and follow-up of patients with a history of urinary tract neoplasia.
The use of cytology as a screening test for cancer has demonstrated discouraging results in sites other than uterine cervix. In asymptomatic persons it can be more harmful than beneficial for the possibility of false-positive results, leading to unnecessary expense and morbidity from follow-up procedures. However, in a general way, it has had an important role in the diagnosis of early clinical stages of disease when already symptoms are proven or when there has been clinical suspicion. Early detection of cancer greatly increases the chances for successful treatment, education being one of the most important elements for recognizing possible warning signs and taking prompt action that will lead to early diagnosis. Increased awareness of possible warning signs of cancer can have a great impact on the disease.

New Developments in Cytological Screening

Liquid-Based Cytology (LBC)
New technology for alternative and complementary forms of screening alterations in the cervix has recently been proposed, and one of these is known as liquid-based cytology (LBC). In this method, the cervical cells are immersed in a conserving liquid before being fixed on the slide, avoiding desiccation and reducing the quantity of obscuring material. Liquid cytology can be prepared by manual or automated methods. LBC methods have been used routinely in laboratories in the majority of developed countries, whereas developing nations more frequently use manual liquid cytology methods. 67 - 69 Although the cost is higher, various studies have shown the advantages of using LBC. Special attention has been given to the use of residual material in the vial, which can be used for:
(a) Preparation of additional slides;
(b) Molecular testing of infectious agents;
(c) DNA cytometry; and
(d) DNA ploidy analysis.
An additional sample can be potentially useful in clarifying diagnoses in cases showing undetermined nuclear atypia, questionable graduation of the lesion, excess or scarcity of cells, blood, or exudates. 70, 71
Since their introduction and approval as a method of detection of lesions on the cervix, the various LBC methods have been the object of diverse studies that emphasize the lower number of interpretative errors, 70 more effective diagnosis of lesions, and greater speed of analysis due to the ease of reading the slides. 72 Emphasized also is the reduction of unsatisfactory cases due to collection of samples and the increase in the number of cases diagnosed with low- or high-degree lesions. Because of these findings the FDA approved LBC in 1996 as a screening method, and today in the USA around 80% of cervical cytology tests use the liquid base. 24
However, despite the benefits shown by liquid cytology, studies point out the need to evaluate the cost–benefit of using one or other type of cytology, as the cost of LBC is still high when compared with the conventional method. A recently published review showed that the principal automated techniques, used in the USA, have a questionable cost-effectiveness ratio. 70 Scotland adopted the method and the UK's National Health Service (NHS) is recommending the introduction of the method in the screening program.

Automated Cytology
Automation of cytology has been studied for many years with the purpose of introducing methods that reduce errors caused by human fatigue, and that can detect lesions when the sample contains a lesser number of abnormal cells. Automated methods available can be used with conventional cytology or with LBC. Despite technological development and the emergence of automatic apparatus, studies have shown that automated screening would not improve the outcome of cervical cytology. For additional information on automated systems see Chapter 34 .
One of the advantages of this methodology is the possibility of testing a large number of cases with a minimum possibility of error. 73 Its association with LBC would also provide a still reduced number of unsatisfactory cases. Nevertheless the high cost of equipment and the implementation of the technology makes its use difficult principally in developing nations.

Concluding Remarks
Cytologic screening is an important method for certain diseases, especially cervical cancer, and an example of successful prevention of this disease. The majority of cervical cancer occurs in developing countries. The success of cervical cancer screening is shown by its ability to reduce the incidence of cervical cancer and the resulting mortality. The integration of procedures is essential for a successful screening program. Recently new technologies for alternative and complementary forms of screening such as liquid-based cytology and automated cytology have been proposed. A combination of methods has been proposed in an attempt to improve the sensibility of the Pap test. Among these, the association of cytology with the molecular test for HPV using hybrid capture has been highlighted. Automated cytology may be used for the purpose of reducing human errors caused by human fatigue, and to detect lesions with a lesser number of abnormal cells in the sample. HPV vaccine will be an additional tool in the strategies to reduce morbidity and mortality from cervical cancer and will be a component of a comprehensive strategy with the long-term goal of eliminating the disease. Cytologic screening can also be performed in selected high-risk populations for lung, esophageal, and bladder cancer.

References

1. Anderson G.H., Flynn K.J., Hickey L.A., et al. A comprehensive internal quality control system for a large cytology laboratory. Acta Cytol . 1997;31:895-899.
2. Babes A. Diagnostic du cancer du col utérin par les frottis. PresseMedicale . 1928;36:451-454.
3. Papanicolaou GN. New cancer diagnosis. In: Proceedings of the Third Race Betterment Conference January 1928; Battle Creek, Michigan , Race Betterment Foundation, 528-534.
4. Papanicolaou G.N., Traut H.F. The diagnostic value of vaginal smears in carcinoma of the uterus. Am J Obstet Gynecol . 1941;42:193-206.
5. Papanicolaou G.N. Atlas of Exfoliative Cytology. Boston: Commonweath Fund University Press, 1954.
6. Lazcano-Ponce E., Alonso P., Ruiz-Moreno J.A., et al. Recomendations for cervical cancer screening programs in developing countries. The need for equity and technological development. Salud Publica de Mexico . 2003;45:449-462.
7. World Health Organization. Cytological screening in the control of cervical cancer: Technical guidelines. Geneva 2 . 1988.
8. World Health Organization. Cervical cancer screening in developing countries: report of a WHO consultation. Genebra . 2002:1-75.
9. Sankaranarayanan R., Budukh A.M., Rajkumar R. Effective screening programmes for cervical cancer in low- and middle-income developing countries. Bull World Health Organ . 2001;79(suppl):954-962.
10. Stewart B.W., Kleihues P., editors. World Cancer Report 2003. Lyon: IARC Press, 2003.
11. Walboomers J.M., Jacobs M.V., Manos M.M., et al. Human papillomavirus is a necessary cause of invasive cancer worldwide. J Pathol . 1999;189:12-19.
12. International Agency for Research on Cancer. IARC Handbooks of Cancer Prevention: Cervical Cancer Screening . Lyon: IARC; 2005. 1-57
13. Richart R.M. Task Force Two: Reasons for Regional I Differences in Prevalence of Cervical Lesions. Chicago: International Conference on the Fight Against Cervical Cancer, 2000.
14. Waddell C. Glandular prediction: the pride and the prejudice. Cytopathology . 2004;15:177-180.
15. Koss L.G. The Papanicolaou test for cervical cancer detection: a triumph and a tragedy. JAMA . 1989;261:737-743.
16. Parkin D.M., Bray F. The burden of HPV-related cancers. Vaccine . 2006;24(Suppl 3):11-25.
17. Miller A.B., Anderson G., Brisson J., et al. Report of a National Workshop on Screening for Cancer of the Cervix. Can Med Assoc J . 1991;145(10):1301-1325.
18. Denny L. The prevention of cervical cancer in developing countries. BJOG . 2005;112:1204-1212.
19. Johannesson G., Geirsson G., Day N. The effect of mass screening in Iceland, 1965-1974, on the incidence and mortality of cervical carcinoma. Int J Cancer . 1978;21:418-425.
20. Hakama M. Trends in the incidence of cervical cancer in the Nordic countries. In: In:Magnus K., editor. Trends in Cancer Incidence . New York: Hemisphere Press; 1982:279-292.
21. Laara E., Day N.E., Hakama M. Trends in mortality from cervical cancer in the Nordic countries: association with organised screening programmes. Lancet . 1987;1:1247-1249.
22. Anderson G.H., Benedet J.L., LeRiche J.C. Invasive cancer of the cervix in British Columbia: A review of demographic and screening histories of 437 cases seen from 1985-1988. Obstet Gynecol . 1992;80(1):1-4.
23. Krane J.F., Granter S.R., Trask C.E., et al. Papanicolaou smear sensitivity for the detection of adenocarcinoma of the cervix: a study of 49 cases. Cancer . 2001;93:8-15.
24. International Agency for Research on Cancer (IARC). IARC confirms efficacy of cervix cancer screening for women 25–65 in reducing mortality. Press release no. 151. Lyon; 2004 [cited 2004 May 3]. Acessed on 2007 Mar 31 Available from: http://www.iarc.fr/ENG/Press_Releases/Summary.pdf
25. Fahey M.T., Irwig L., Macaskill P. Meta-analysis of Pap test accuracy. Am J Epidemiol . 1995;141:680-689.
26. Nanda K., McCrory D.C., Myers E.R., et al. Accuracy of the Papanicolaou test in screening for and follow-up of cervical cytologic abnormalities: a systematic review. Ann Intern Med . 2000;132:810-819.
27. Parikh S., Brennan P., Boffetta P. Meta-analysis of social inequality and the risk of cervical cancer. Int J Cancer . 2003;105:687-691.
28. Miller A.B. Quality assurance in screening strategies. Review. Virus Res . 2002;89(2):295-299.
29. Suba E.J. Crossing the quality chasm: a requirement for successful cervical cancer prevention in developing countries. Clin Lab Med . 2004;24:945-963.
30. Guidozzi F. Obstet screening for cervical cancer. Gynecol Surv . 1996;51:247-252.
31. Anderson G.H., Boyes D.A., Benedet J.L., et al. Organisation and results of cervical cytology screening programme in British Columbia, 1955-1985. BMJ . 1988;296:975-978.
32. Collaco L.M., de Noronha L., Pinheiro D.L., et al. Quality assurance in cervical screening of a high risk population: a study of 65,753 reviewed cases in Parana Screening Program, Brazil. Diagn Cytopathol . 2005;33:441-448.
33. Sankaranarayanan R., Thara S., Sharma A., et al. Accuracy of conventional cytology: Results from a multicentre screening study in India. J Med Screen . 2004;11:77-84.
34. Sankaranarayanan R., Gaffikin T.L., Jacob M., et al. A critical assessment of screening methods for cervical neoplasia. Int J Gynaecol Obstet . 2005;89(suppl 2):S4-S12.
35. Zeferino L.C., Derchain S.F. Cervical cancer in the developing world. Best practice research. Clin Obstet Gynec . 2006;20:339-354.
36. Marcus A.C., Crane L.A., Kaplan C.P., et al. Improving adherence to screening follow-up among women with abnormal Pap smears: results from a large clinic-based trial of three intervention strategies. Med Care . 1992;30:216-230.
37. Garner D. Potencial limitations of screening programs. Task Force 13. International Consensus Conference on the Fight against Cervical Cancer . International Academy of Cytology. 2000.
38. Kirnbauer R., Booy F., Cheng N., et al. Papillomavirus L1 major capsid protein self-assembles into virus-like particles that are highly immunogenic. Proc Natl Acad Sci USA . 1992;89:12,180-12,184.
39. Harro C.D., Pang Y.Y., Roden R.B., et al. Safety and immunogenicity trial in adult volunteers of a human papillomavirus 16 L1 virus-like particle vaccine. J Natl Cancer Inst . 2001;93:284-292.
40. Koutsky L.A., Ault K.A., Wheeler C.M., et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med . 2002;347:1645-1651.
41. Harper D.M., Franco E.L., Wheeler C., et al. Efficacy of a bivalent L1 virus-like particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: a randomised controlled trial. Lancet . 2004;364:1757-1765.
42. Villa L.L., Costa R.L., Petta C.A., et al. Prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in young women: a randomized double-blind placebo controlled multicentre phase II efficacy trial. Lancet Oncol . 2005;6:271-278.
43. Muñoz N., Bosch F.X., Castellsagué X., et al. Against which human papillomavirus types shall we vaccinate and screen? The international perspective. Inter J Cancer . 2004;111:278-285.
44. WHO. Report of the consultation on human papillomavirus vaccines. 2005; Geneva; 1-34. Acessed on 2007 Mar 31. Available from: http://www.who.int/vaccines-documents/DocsPDF05/816screen.pdf
45. Parry J.Vaccinating against cervical cancer. Bull World Health Organ. 2007;2:85.
46. Harper D.M., Franco E.L., Wheeler C., et al. Sustained efficacy up to 4–5 years of a bivalent L1 virus-like particle vaccine against human papillomavirus types 16 and 18: follow-up from a randomised control trial. Lancet . 2006;367:1247-1255.
47. Pollack A.E., Balkin M., Edouard L., et al. Ensuring access to HPV vaccines through integrated services: a reproductive health perspective. Bull World Health Organ. 2007;1:85.
48. Goldie S.J., Kohli M., Grima D., et al. Projected clinical benefits and cost-effectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst . 2004;96:604-615.
49. Wright T.C.Jr., Cox J.T., Massad L.S. 2001 consensus guidelines for the management of women with cervical intraepithelial neoplasia. Am J Obstet Gynecol . 2003;1:295-304.
50. Bosch FX: Cervical cancer epidemiology and rationale for human papilomavirus (HPV) related preventive strategies. Twenty-first International Conference and Clinical Workshop Papilomavirus , 2004.
51. Franco E.L., Ferenckzy A. Is HPV testing with cytological triage a more logical approach in cervical cancer screening? Lancet Oncol . 2006;7:527-529.
52. Girianelli V.R., Thuler L.C., Szklo M., et al. Comparison of human papillomavirus DNA tests, liquid-based cytology and conventional cytology for the early detection of cervix uteri cancer. Eur J Cancer Prev . 2006;15:504-510.
53. Diehl I.S., Prado R., Magna R.M. Organización Panamericana de la Salud. Manual de Procedimientos del Laboratório de Citologia. Red Panamericana de Citologia y Tamizaje . 2002:1-84. Red-PAC
54. World Health Organization. Screening and early detection of cancer. Accessed on 2007 Mar 31. Available from: http://www.who.int/cancer/detection/en/
55. Gondos B., King E.B. Significance of endometrial cells in cervicovaginal smears. Ann Clin Lab Sci . 1977;7(6):486-490.
56. Smith R.A., von Eschenbach A.C., Wender R., et al. American Cancer Society guidelines for the early detection of cancer: update of early detection guidelines for prostate, colorectal, and endometrial cancers. Also: update 2001—testing for early lung cancer detection. CA Cancer J Clin . 2001;51:38-75.
57. Fontana R.S., Sanderson D.R., Woolner L.B., et al. Screening for lung cancer. A critique of the Mayo Lung Project. Cancer . 1991;67(4 Suppl):1155-1164.
58. Petty T.L. The early identification of lung carcinoma by sputum cytology. Cancer . 2000;89(11 Suppl):2461-2464.
59. Wang K.K., Wongkeesong M., Buttar N.S. American Gastroenterological Association medical position statement: role of the gastroenterologist in the management of esophageal carcinoma. Gastroenterology . 2005;128:1468-1470.
60. Dong Z., Tang P., Li L., Wang G. The strategy for esophageal cancer control in high-risk areas of China. Jpn J Clin Oncol . 2002;32(suppl 1):S10-S12.
61. Dawsey S., Shen Q., Nieberg R.K., et al. Studies of esophageal balloon cytology in Linxian, China. CA Epidem Biomar Prev . 1997;6:121-130.
62. National Cancer Institute. U.S. National Institutes of Health. Screening for bladder and other urothelial cancers. Acessed on 2007 Mar 31. Available at: http://www.cancer.gov/cancertopics/pdq/screening/bladder/HealthProfessional/page2
63. Burch J.D., Rohan T.E., Howe G.R., et al. Risk of bladder cancer by source and type of tobacco exposure: a case-control study. Intern J Cancer . 1989;44(4):622-628.
64. Cole P., Hoover R., Friedell G.H. Occupation and cancer of the lower urinary tract. Cancer . 1972;29(5):1250-1260.
65. Brownson R.C., Chang J.C., Davis J.R. Occupation, smoking, and alcohol in epidemiology of bladder cancer. Am J Pub Health . 1987;77:1296-1300.
66. Anton-Culver H., Lee-Feldstein A., Taylor T.H. Occupation and bladder cancer risk. Am J Epidem . 1992;136:89-94.
67. Richard K., Dziura B., Hornish A. Cell block preparation as a diagnostic technique complementary to fluid-based monolayer cervicovaginal specimens. Acta Cytol . 1999;43(1):69-73.
68. Diaz-Rosario L.A., Kabawat S.E. Cell block preparation by inverted filter sedimentation is useful in the differential diagnosis of atypical glandular cells of undetermined significance in ThinPrep specimens. Cancer . 2000;90(5):265-272.
69. Massarani-Wafai R., Bakhos R., Wojcik E.M., et al. Evaluation of cellular residue in the ThinPrep PreservCyt vial. Diagn Cytopathol . 2000;23(3):208-212.
70. Longatto Filho A., Pereira S.M., Di Loreto C., et al. DCS liquid-based system is more effective than conventional smears to diagnosis of cervical lesions: study in high-risk population with biopsy-based confirmation. Gynecol Oncol . 2005;97(2):497-500.
71. Dahl M.B., Ejersbo D., Holund B. Cause and follow-up of inadequate cervical smears in the county of Funen. Ugeskr Laeger . 2002;164(37):4280-4283.
72. Alves V.A., Bibbo M., Schmitt F.C., et al. Comparison of manual and automated methods of liquid-based cytology. A morphologic study. Acta Cytol . 2004;48(2):187-193.
73. Nieminen P. Task force nine: Automated prescreening and rescreening. International Conference on the Fight against Cervical Cancer, Chicago . 2000.
74. Engeland A., Haldorsen T., Tretli S., et al. Prediction of cancer incidence in the Nordic countries up to the years 2000 and 2010. A collaborative study of the five Nordic Cancer Registries. APMIS . 1993;38(Suppl):1-124.
75. Hristova L., Hakama M. Effect of screening for cancer in the Nordic countries on deaths, cost and quality of life up to the year 2017. Acta Oncol . 1997;36(Suppl 9):1-60.
76. Parkin D.M., Whelan S.L., Ferlay J., et al, editors. Cancer Incidence in Five Continents, vol. VII. No. 143. Lyon: IARC: IARC Scientific Publications, 1997.
77. Moller B., Fekjaer H., Hakulinen T., et al. Prediction of cancer incidence in Nordic countries up to the year 2020. Eur J Cancer Prev . 2002;11(Suppl 1):S1-S96.
CHAPTER 4 Diagnostic Quality Assurance in Cytopathology

Marluce Bibbo, Catherine M. Keebler

Contents
Introduction
Quality Assurance Measures
Laboratory Directors
Cytotechnologists
Physical Laboratory Facilities
Safety Precautions
Equipment
Specimen Collection
Preparation, Fixation, and Staining Procedures
Slide Evaluation Workload
Cytologic Terminology
Laboratory Records, Logs, and Files
Internal Quality Assurance Mechanisms
CLIA 88 Update
Rapid Re-evaluation
Computer-Assisted Quality Assurance Mechanisms
External Quality Assurance Mechanisms
Continuing Education Practices
Concluding Remarks

Introduction
Cytopathologists are concerned about and committed to quality assurance and quality control in their laboratories. These practices include, among others, the use of intralaboratory and extradepartmental consultations, case reviews, correlation of cytologic and histopathologic specimens, hierarchic review of cytopathology, and review of completed diagnostic reports. Most of the quality assurance techniques are well described. 1 - 18 ,20 ,21
In the past, formal organization and mandatory documentation of these quality assurance efforts may have been limited or deficient. Formal rules may be difficult to apply, because laboratories and screening programs vary depending on the volume and type of cytodiagnostic material received and on the size and experience of its staff members. Even though the detailed design of a quality assurance program emanates from the cytopathology laboratory director, basic quality control and quality assurance principles of structure, organization, documentation, and systematic review must be in place.
The enactment of the Clinical Laboratory Improvement Amendment of 1988 by the US Department of Health and Human Services, 22 the convening of two national conferences on cytologic quality assurance by the Centers for Disease Control, Atlanta, Georgia, 23 - 25 the publication of the Quality Assurance Manual by the College of American Pathologists (1988), 26 the American Medical Association Committee Report on the subject, 27 the publication of the Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology by the Tutorials of Cytology, 28 and the proposal by the National Cancer Institute Workshop on the Bethesda System of reporting cytologic findings, 29 as well as editorials, letters to the editors, 30 and general public interest led to the assurance of high diagnostic standards and fostered intense activities in the quality assurance and quality control sectors. 3, 24, 28, 31 - 52

Quality Assurance Measures
Cytopathology is a practice of medicine and represents a medical consultation, in both gynecologic and nongynecologic anatomic sites. The basic principles of quality assurance apply to all types of cytologic specimens.
The following represents several minimum quality assurance stipulations to which most cytopathologists will probably agree.

Laboratory Directors
The laboratory should be directed by a legally qualified physician with a specialist qualification in pathology, including special training and expertise in cytopathology. In a case in which the current laboratory director or co-director (associate cytopathologist) does not have board qualification in pathology but has had special training in cytopathology, this situation may be approved under a “grandfather clause.” The director or designated medical professional is responsible for proper performance and reporting of all tests done in the cytopathology laboratory. The director or designated cytopathologist should be physically present in the laboratory to direct the staff, be available for consultations, review all reactive and abnormal gynecologic cytology samples, review fine needle aspiration samples, and review all nongynecologic samples. In addition, a supervisor or senior cytotechnologist should be assigned to review 10% of the negative cases, including high-risk cases as designated by cytologic findings, clinical findings, and patient histories. This procedure should help to detect any discrepancies in interpretation that may occur prior to issuing the final cytologic report.
In addition the director should develop a quality assurance plan, a manual of laboratory policies/procedures, and ensure the written policies and procedures reflect actual laboratory practice. Issues and problems identified through the quality management process need to be addressed and resolved. Gathering of laboratory statistics is best accomplished by collection of monthly reports during quality assurance meetings presided over by the director.

Cytotechnologists
Cytotechnologists should meet one of the following requirements:
(1) Be certified as a cytotechnologist by either the American Society of Clinical Pathologists or the US Department of Health, Education and Welfare; or
(2) Previously have been admitted to the practice of cytotechnology by existing regulations under a grandfather clause.

Physical Laboratory Facilities
The laboratory should be clean, well lighted, adequately ventilated, and functionally arranged so as to minimize problems in specimen handling, evaluation, and reporting. The area for specimen preparation and handling should be separate from the area where specimens are evaluated and reported. Formaldehyde and xylene (if in use) should be carefully monitored due to the possible presence of hazardous vapor concentrations.

Safety Precautions
Laboratory personnel must be protected against hazards (chemical, electric, fire, infections, or others) by using well-ventilated hoods and biologic safety hoods for handling potentially infectious material. Fire precautions should be posted and tested. Each employee should participate in fire drills and should know the location of the fire extinguisher, blankets, emergency fire alarm, and exits. Safety shower, eye wash stations, and procedures to follow in case of chemical spills or splash to the body should be posted and readily visible in the laboratory.

Equipment
An adequate number of binocular microscopes of good quality and proper working order must be available. Laboratory instruments and equipment should be under periodic maintenance to monitor and ensure malfunctions do not adversely affect analytical results. A sample of slides from slide preparation instruments including liquid-based technology and cytocentrifuge or filtration methods should be routinely reviewed microscopically for technical acceptability.

Specimen Collection
Cytologic specimens should be accepted and examined only if requested by a licensed medical practitioner and collected in accordance with instructions regarding recommended collection techniques. The cytopathology laboratory should inform the originator of the sample if the specimens are “unsatisfactory” and detail adequacy qualifiers such as presence or absence of a transformation zone component or obscuring factors in “satisfactory samples.” 29, 53 - 56

Preparation, Fixation, and Staining Procedures
The specimens must be identified with the patient’s name and/or a unique identifier and must be accompanied by a requisition form with the requesting physician’s name, address, date of specimen collection, specimen source, and appropriate clinical information about the patient. When the specimen arrives in the laboratory the laboratory staff affix an accession number or bar code label on each slide for further identification. The laboratory should have written criteria for rejecting specimens. Fixation while the specimen is still wet is recommended for conventional cell samples and rinsing of spatula and brush in preservative solution (kit provided by manufacturers) for liquid-based specimens. The Papanicolaou staining procedure is strongly suggested for most cytologic samples, unless additional staining procedures are warranted. Staining solutions and chemicals used in the cytopathology laboratory should be labeled with the time of preparation, purchase, or both. Staining solutions should be filtered regularly to avoid contamination and should be covered when not in use. Effective measures to prevent cross-contamination between gynecologic and nongynecologic specimens during the staining process must be used. Separate runs followed by filtration or changing of solutions or a separate staining setup is recommended.

Slide Evaluation Workload
Regulations as to the number of specimens a cytotechnologist may evaluate in a 24-hour period are currently set at 100 slides per an 8-hour day. This regulation may not do justice to the various conditions that influence the quality of the slide evaluation performance. The percentage of atypical cases evaluated versus the percentage of negative cases in varying populations as well as screening of nongynecologic specimens should be considered when workloads are established. This regulation ensures that the number and type of cytologic samples evaluated do not, through fatigue, adversely affect the cytotechnologist’s performance. Some slides are easier and less fatiguing to evaluate and some cytotechnologists are more experienced than others. Other activities in which the cytotechnologist participates, such as participation on the fine needle aspiration service and quality control activities, should appropriately reduce their workload in the evaluation of cell samples. The interpretation of the cytotechnologist should become a permanent record and available for future review.

Cytologic Terminology
The vaginal/ectocervical/endocervical cytology sample should be interpreted preferably by using the Bethesda System. 56 The nongynecologic material should be interpreted in medical terms, i.e., conform and correspond to diagnostic reporting systems in histopathology.

Laboratory Records, Logs, and Files
Each specimen should be recorded and a sequential accession number assigned together with the name of the patient and the originator of the sample. Test records must be retained for at least 2 years. The negative gynecologic cell samples should be retained on file for a minimum of 5 years and negative fine needle aspirates for 10 years or indefinitely if they exhibit abnormal features. The modern cytopathology laboratory should use a computerized file system. Such a system permits laboratory professionals to have information on all previous cytologic or histologic reports on a given patient available when the new cell sample is being evaluated. Modern computerized data collection and retrieval systems are also essential for continuing quality control and assurance mechanisms. A record of workload and diagnostic performance should be maintained as a part of the personnel data file for each cytotechnologist.

Internal Quality Assurance Mechanisms

CLIA 88 Update
The standard requirements of CLIA 88 affecting the cytopathology laboratory in the United States have been published in the federal register. 57 Recent updates of CLIA 88 have added more rules and regulations to the already highly regulated field of gynecologic cytology. 58 In addition to the following:
1. Prospective 10% review of negative gynecologic cell samples including high-risk cases may be performed, prior to reporting patient results, by the pathologist, supervisor, or cytotechnologist who has had 3 years of continuous cytology experience within the past 10-year period.
2. Retrospective re-evaluation for current high-grade squamous intraepithelial lesions (HSIL) or cancer cases with a review of negative specimens obtained within the previous 5 years should be performed when clinically significant discrepancies are found that will affect current patient care; notification of physician and an amended report are required.
3. Cytology/histology correlation for gynecologic cases with a diagnosis of HSIL, adenocarcinoma, or other malignant neoplasm should be carried out with causes of discrepancies, such as sampling or diagnostic errors on biopsy or cytologic cell samples, documented.
new rules require:
1. An evaluation of the case reviews of each individual against the laboratory’s overall statistical values, documentation of discrepancies including reasons for deviation, and, if appropriate, corrective action taken.
2. A workload limit based on individual’s performance, every 6 months, based on review of 10% negative cases and comparison of individual’s interpretation with the pathologist should be established. The maximum workload limit is 100 slides in 24 hours. Pathologists who perform primary screening are not required to include tissue pathology slides and previously examined cytology slides in the 100-slide workload limit.
3. The pathologist should confirm all nongynecologic cases and interpret each gynecologic slide as reactive or with any abnormality, including atypical squamous cells of undetermined significance (ASCUS), atypical glandular cells (AGC), low-grade squamous intraepithelial lesion (LSIL), high-grade squamous intraepithelial lesion, and carcinoma. The report needs to be signed by the pathologist to reflect review or if a computer report is generated it must reflect an electronic signature authorized by the pathologist.
4. Reports with narrative descriptive nomenclature for all results are recommended. Corrected reports must include the basis for correction.
5. Records of initial examinations and all rescreening results must be documented.
6. Annual statistics should include total gynecologic and nongynecologic cases, by specimen type, and diagnosis including unsatisfactory, cytology/histology correlation discrepancies, and cases reclassified from normal to LSIL or higher.
7. When performing evaluations using automated and semiautomated screening devices, the laboratory must follow manufacturers’ instructions for preanalytic, analytic, and postanalytic phases of testing, as applicable, and meet the applicable requirements.
8. Enrollment for proficiency testing (PT) as required by Centers for Medicare and Medicaid Services. For more details see section on External Quality Assurance Mechanisms.

Rapid Re-evaluation
Rapid rescreening (RR) of cervical smears for internal quality control has been advocated by a number of authors. 59 - 62 In a meta-analysis of 14 studies by Arbyn and Schenck, 61 RR was a more effective quality control method than full rescreening of a 10% random sample. Although RR has received great support as the quality control method of choice in some countries, the cost-effectiveness of its potential advantage is still unknown.

Computer-Assisted Quality Assurance Mechanisms
With the approval by the Food and Drug Administration (FDA) of two automated instruments for quality assurance rescreening of gynecologic cytology, namely, the Focal Point System by Becton Dickinson/TriPath and the ThinPrep Imaging System (TIS) by Cytyc Corporation, automated screening of gynecologic cytologic samples for quality assurance has become a practical reality.
Detailed proposed specifications for automated instruments have been published, 63 and reviews of the instruments and the concepts can be found in Chapter 34 , Automation of Pap Smears, and in the Compendium on the Computerized Cytology and Histology Laboratory 64 and in the Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . 65
The development of automated cytologic devices had to pass two milestones: the application of the system as a quality assurance or quality control instrument, and the application of automated devices for primary screening of gynecologic cell samples. 66, 67
The AutoPap System, now named Focal Point System, showed superior sensitivity and specificity for detection of abnormal slides with ASCUS and more severe lesions when compared to current practice. 68 Other investigators had similar results. 69 - 71 The Focal Point System is a more effective way of detecting errors within a laboratory and reduces the false-negative rate (FNR) by greater than 25% according to Renshaw et al. 72
The ThinPrep Imaging System was statistically more sensitive than manual evaluation for detection of ASCUS or more severe lesions and statistically equivalent for LSIL and HSIL diagnoses. 73 According to Biscotti et al. 74 the sensitivity of this system equals or exceeds the sensitivity of manual screening without adversely affecting specificity. Significant increase in SIL detection was demonstrated with the TIS by Dziura et al. 75 A recent study by Bolger et al. 76 showed sensitivity and specificity of the imager equivalent to that of primary manual screening.
These devices may improve the rate of false-negative evaluations—but at a price. In other words, in the quality control mode the question is not whether the use of either instrument may be harmful to the patient but whether the cost–benefit ratio is such that supplemental funds will be made available by third-party payers to support this additional quality control procedure. Currently several third-party payers have reimbursed for these procedures and current procedural terminology (CPT) codes are available for billing purposes. For additional information see Chapter 34 .
Before automated devices can be effectively applied in primary screening, one will have to consider other major factors that preclude successful reduction of prevalence and incidence of gynecologic precancerous and malignant lesions. To obtain maximum results for a population-screening program, we need to educate women to obtain periodic gynecologic examinations, to teach medical personnel to take and fix the specimens appropriately, to make sure laboratory personnel follow proper staining procedures, and to ensure the patient will make herself readily available for required repeat cytologic examinations, colposcopy, and/or biopsies. After a certain time of intensive cytologic evaluation, prevalent cancers and high-grade precursors will have disappeared (the “sweeping” effect of an effective population-screening program). What remains are incident low-grade precursors and “interval” cancers and “interval” high-grade precursors that in fact are missed positives (i.e., false-negatives), if one does not believe in the very rapidly growing malignant lesion.
We should also re-emphasize to cytotechnologists that their services are urgently needed in the future, even with approval of automated devices for routine application, because the systems either work interactively under cytotechnologic and cytopathologic guidance or produce alarm messages on selected cases that in fact do not contain atypia (false-positives), which will take extensive effort and time by a human re-evaluator to locate and override the alarm message.
The development of a fully automated diagnostic cytology and histology system, i.e., complete without professional interaction and involvement, was and remains a daydream that is neither feasible nor desirable. However, interactive systems as aids in quality assurance mechanisms and improvement of productivity are constructive developments and major technologic accomplishments.

External Quality Assurance Mechanisms
External quality assurance mechanisms with peer review by professional organizations or by state or federal governmental bodies are currently being implemented. The Center for Medicare and Medicaid Services (CMS) has approved two proficiency tests by the American Society for Clinical Pathology (ASCP) and College of American Pathologists (CAP). Some states, e.g., New York and Maryland, have had a testing program for cytopathology laboratories enacted and operational for many years. 42, 77 Any external program will be welcomed by the high-quality laboratory, but there may be problems in the funding and execution of an unbiased, objective, and reproducible testing system. 19, 34, 38, 39, 44, 46, 47, 51, 78 - 80 ,82 Members of the Cytopathology Educational and Technology Consortium have recommended modifications of the rules regarding proficiency testing including testing interval, utilizing validated and monitored slides, and changing the grading system. 81

Continuing Education Practices
The laboratory director should conduct continuing educational activities within the laboratory; provide up-to-date reference materials, such as cytopathology textbooks, compendia on clinical cytology, cytologic journals, visual image teaching slide sets, cytology websites, and CDs. The staff should be encouraged to participate in ongoing educational events, such as local, regional, national, and international cytology meetings and tutorials.

Concluding Remarks
The described quality assurance stipulations represent minimum quality assurance measures to which most laboratories adhere. In the United States laboratories are also governed by the described standard requirements of CLIA 88. External testing programs are a welcome component of a laboratory quality assurance but there may be problems in the execution of an unbiased, objective, and reproducible evaluation system. A number of professional organizations are working with the federal government in the United States to improve the current CLIA 88 cytology proficiency testing. Unacceptable errors do occur in the cytology laboratory. Automated systems and molecular HPV testing are available and may help to reduce the false-negative rate of the Pap test. Continuing education practices and a creative learning environment for the cytotechnologist and cytopathologist are necessary to improve the diagnostic results for the patient.

References

1. Allen K.Quality control. Tirol J.H., editor. ASCT Cytopathology Quality Assurance Guide, vol I. Raleigh, NC: ASCT, 1992.
2. Anderson G.H., Flynn K.J., Hickey L.A., et al. A comprehensive internal quality control system for a large cytology laboratory. Acta Cytol . 1997;31:895-910.
3. Bonfiglio T.A. Quality assurance in cytopathology: Recommendations and ongoing quality assurance activities of the American Society of Clinical Pathologists. Acta Cytol . 1989;33:431-433.
4. Erozan Y.S. Quality control in cytopathology. Clin Lab Med . 1986;6:707-713.
5. Gay J.D., Donaldson L.D., Goellner J.R. False-negative results in cervical cytologic studies. Acta Cytol . 1985;29:1043-1046.
6. Hindman W.M. A proposal for quality control in gynecologic cytology. Acta Cytol . 1987;31:384-385.
7. Inhorn S.L., Clarke E. A statewide proficiency testing program in cytology. Acta Cytol . 1971;15:351-356.
8. Inhorn S.L., Shalkham J.E. Quality assurance in cytology: Monitoring cytotechnologist performance. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:231-237.
9. Kaminsky F.C., Burke R.J., Haberle K.R., et al. Rescreening policies in cervical cytology and their effect on detecting the truly positive patient. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:73-78.
10. Keebler C.M. A quality assurance program for the cytopathology laboratory. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:207-213.
11. Krieger P., Naryshkin S. Random rescreening as a quality assurance tool in cytology. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:109-114.
12. Melamed M.R. Quality control in cytology laboratories. Gynecol Oncol . 1981;12:206-211.
13. Melamed M.R., Inhorn S.L., Kachenmeister L.A., et al. Cytology. In: Inhorn S.L., editor. Quality Assurance Practices for Health Laboratories . Washington, DC: American Public Health Association; 1978:509-544.
14. Mitchell H., Medley G., Drake M. Quality control measures for cervical cytology laboratories. Acta Cytol . 1988;32:288-292.
15. Patten F.W. Organization of the laboratory. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:214-219.
16. Saigo P.E., Gatscha R.M., Warren, G.P. Cytology staff: quality assurance: one laboratory’s experience. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:148-150.
17. Schenck U., Reuter B. Workload in the cytology laboratory. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:223-230.
18. Vooijs G.P., van Aspert van Erp A.J.M., van’t Hoff Grootenboer B.E., et al. Parameters of quality control in cervical cytodiagnosis. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:95-108.
19. Vooijs G.P., van Aspert van Erp A.J.M., van’t Hoff Grootenboer B.E., et al. Intra- and interobserver variability. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:15-27.
20. Wied G.L. Quality control standards for laboratories and cytotechnology registration. Acta Cytol . 1970;14:557.
21. Wied (moderator) G.L. Symposium on Quality Assurance. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:297-318.
22. US Department of Health and Human Services. Clinical Laboratory Improvement Amendment,1988.
23. Centers for Disease Control. Proceedings of two conferences on the state of the art in quality control measures for diagnostic cytology laboratories. Atlanta, GA, March 12, 1988, September 1, 1988.
24. Paris A.I. Conference on the state of the art in quality control measures for diagnostic cytology laboratories: Background and introduction. Acta Cytol . 1989;33:423-426.
25. Solomon D. Introduction to the proceedings of the conference on the state of the art in quality control measures for diagnostic cytology laboratories. Acta Cytol . 1989;33:427-430.
26. College of American Pathologists. Surgical Pathology/Cytopathology Quality Assurance Manual. Skokie, IL: College of American Pathologists, 1988.
27. American Medical Association Committee Report of the Council on Scientific Affairs. Quality Control in Cervical Cytology: The Papanicolaou Smear. Chicago: AMA, 1989.
28. Vooijs G.P.Quality assurance and quality control in clinical cytology: Symposium by correspondence (moderator) Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology. Chicago: Tutorials of Cytology, 1995, 319-389.
29. National Cancer Institute Workshop on Terminology. The Bethesda System. JAMA . 1989;262:931-934.
30. Wied G.L. Quality assurance measures in cytopathology [editorial and letters to the editor]. Acta Cytol . 1988;32:913-939.
31. Ashton P.R. American Society for Cytotechnology quality assurance survey data: Summary report. Acta Cytol . 1989;33:451-454.
32. Clark A.H. Current manpower pool in cytotechnology. Acta Cytol . 1989;33:455-459.
33. Collins C.L. The challenge of cytology proficiency testing. Lab Med . 1994;25:219-220.
34. DeBoy J.M., Jarboe B.R. Maryland’s cytology proficiency testing: Positive impacts, national applicability. Lab Med . 1994;25:232-233.
35. Gardner N.M. In-house quality assurance program in a state cytology laboratory. Acta Cytol . 1989;33:487-488.
36. Greenberg R.S., Chow W.H., Liff J.M. Recent trends in the epidemiology of cervical neoplasia. Acta Cytol . 1989;33:463-470.
37. Gupta P.K., Erozan Y.S. Cytopathology laboratory accreditation, with special reference to the American Society of Cytology programs. Acta Cytol . 1989;33:443-447.
38. Inhorn S.L. Proficiency testing in Wisconsin from Medicare to CLIA '88. Lab Med . 1994;25:226-229.
39. Kraemer B.B. Quality assurance activities of the College of American Pathologists. Acta Cytol . 1989;33:434-435.
40. Lipa M., Fletcher A., Chen V., et al. Guidelines for quality assurance programs in cytopathology. Canad Soc Cytol Bull . 1989;5:2.
41. Luff R.D., Lobritz R.W. Diagnostic performance of United States Air Force cytology laboratories: An eight-year experience in central monitoring. Acta Cytol . 1989;33:475-478.
42. Newton L.E. Cytopathology proficiency testing in New York State: The first 25 years. Lab Med . 1994;25:230-231.
43. Parker J.E. Education and training for cytopathologists: Its role in quality assurance. Acta Cytol . 1989;33:448-450.
44. Penner D.W. An overview of the College of American Pathologists; program in surgical pathology and cytopathology: Data summary of diagnostic performance in cervical cytopathology. Acta Cytol . 1989;33:439-442.
45. Rube I.F. Experience in managing a large-scale rescreening of Papanicolaou smears and the pros and cons of measuring proficiency with visual and written examinations. Acta Cytol . 1989;33:479-483.
46. Somrak T.M. Cytology proficiency testing: The challenge of implementation. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:246-250.
47. Somrak T.M. Quality assurance and risk management issues in the cytopathology laboratory. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:220-222.
48. Steiner C. Cervical cancer screening from the public health perspective. Acta Cytol . 1989;33:471-474.
49. Taylor W.R., Nadel M.R., Smith R.A., et al. Cervical cancer screening and demonstration projects to identify barriers to preventing cervical cancer mortality. Acta Cytol . 1989;33:460-462.
50. Taylor W.R., Plott A.E., Cheek S.W., et al. Epilogue: Establishment of the center for chronic disease prevention and health promotion. Acta Cytol . 1989;33:489-490.
51. Thompson D.W. Canadian experience in cytology proficiency testing. Acta Cytol . 1989;33:484-486.
52. Vooijs G.P. Quality assurance and control in cytology: Literature search. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:390-399.
53. Gilbert F.E., Hicklin M.D., Inhorn S.L., et al. Conclusions of a study group on standards of adequacy of cytologic examinations of the female genital tract. Acta Cytol . 1973;17:559-561.
54. Solomon D., Henry M. Specimen adequacy. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:90-94.
55. Vooijs G.P., Elias A., van der Graaf Y., et al. The influence of sample takers on the cellular composition of cervical smears. Acta Cytol . 1986;30:251-257.
56. Solomon D., Davey D., Kurman R., et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA . 2001;287:2114-2119.
57. Rules and Regulations—Cytology—Federal Register February 28 1992, 57:40
58. Rules and Regulations—Cytology—Federal Register January 24 2003, 68:16
59. Faraker C.A. Partial rescreening of all negative smears. An improved method of quality assurance in laboratories undertaking cervical screening. Cytopathology . 1993;4:47-50.
60. Baker A., Melcher D.H. Rapid cervical cytology screening. Cytopathology . 2001;2:299-301.
61. Arbyn M., Schenck U. Detection of false negative Pap smears by rapid reviewing. a metaanalysis. Acta Cytol . 2000;44(6):949-957.
62. Collaco L.M., Noronha L.D., Bleggi-Torres L.F., et al. Quality control in cervical cancer screening: Brazilian Experience. Acta Cytol . 2005;49:694-696.
63. Specifications of automated cytodiagnostic systems, proposed by the International Academy of Cytology. Analyt Quant Cytol . 1984;6:146.
64. Wied G.L., Bartels P.H., Rosenthal D.L., et al, editors. Compendium on the Computerized Cytology and Histology Laboratory. Chicago: Tutorials of Cytology (International Academy of Cytology), 1994.
65. Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology. Chicago: Tutorials of Cytology, 1995.
66. Grohs H.K., Husain O.A.N., editors. Automated Cervical Cancer Screening. New York: Igaku-Shoin, 1994.
67. Hutchinson M.L., Zahniser D.J. Cervical cytology automation. In: Marchevsky A.M., Bartels P.H., editors. Image Analysis . New York: Raven; 1994:309-334.
68. Wilbur D.C., Prey M.U., Miller W.M., et al. The AutoPap system for primary screening in cervical cytology: Comparing the results of a prospective, intended-use study with routine manual practice. Acta Cytol . 1998;42:214-220.
69. Bibbo M., Hawthorne C. Does use of the AutoPap assisted primary screener improve cytologic diagnosis? Acta Cytol . 1999;43:23-26.
70. Bibbo M., Hawthorne C. Performance of the AutoPap primary screening system at Jefferson University Hospital. Acta Cytol . 1999;43:27-29.
71. Marshall C.J., Rowe L., Bentz J.S. Improved quality-control detection of false-negative Pap smears using the AutoPap 300 QC system. Diagn Cytopathol . 1999;20:170-174.
72. Renshaw A.A., Lezon K.M., Wilbur D.C. The human false-negative rate of rescreening Pap tests. Measured in a two-arm prospective clinical trial. Cancer . 2001;93(2):106-110.
73. Dawson A.E. Can we change the way we screen? The ThinPrep Imaging System®: Clinical trial data and early experience. Cancer Cytopathol . 2004;102:340-344.
74. Biscotti C.V., Dawson A.E., Dziura B., et al. Assisted primary screening using the automated ThinPrep Imaging System. Am J Clin Pathol . 2005;123(2):281-287.
75. Dziura B., Quinn S., Richard K. Performance on an imaging system vs. manual screening in the detection of squamous intraepithelial lesions of the uterine cervix. Acta Cytol . 2006;50:309-311.
76. Bolger N., Heffron C., Regan I., et al. Implementations and evaluation of a new automated interactive image analysis system for cervical cytology. Acta Cytol . 2006;50:483-491.
77. Collins D.N., Kaufman W., Clinton W. Quality evaluation of cytology laboratories in New York State. Acta Cytol . 1974;18:404-413.
78. Lunz M.E. Validity and reliability of competency testing. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:241-243.
79. Plott A.E. Cytology proficiency testing research and development at the Centers for Disease Control, 1967–1992. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:244-245.
80. Rosenthal D.L. Criteria and standards for cytology proficiency testing. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:238-240.
81. Scientific issues related to the cytology proficiencytesting regulations. Cytopathology Education and Technology Consortium. Acta Cytol . 2006;50:191-197.
82. Bonfiglio T.A., Nayar R., Abdul-Karim F.W. ASCP sets high standards for regulatory gynecologic cytology proficiency testing. Lab Med . 2006;37:733-735.
Chapter 5 Evaluation of the Sample in Smears and Liquid-Based Preparations

Marluce Bibbo, Joseph F. Nasuti

Contents
Introduction
Cervicovaginal Cytology
Specimen Type
Patient Identification
Clinical Information
Microscopic Evaluation
Nongynecologic Cytology
Specimen Type
Specimen Cross-Contamination
Specimen Mishandling
Concluding Remarks

Introduction
Several factors play a role in the evaluation of the cellular sample. The method of sample collection and fixation, the laboratory procedure to process the sample, and the integration of the morphologic features observed in the sample with the clinical information may affect the quality of interpretation/diagnosis reached.
The ultimate goal in specimen processing is to preserve, as much as possible, in vitro or in vivo aspects of the sample obtained. This includes the original size, shape, and texture of the cytoplasmic and nuclear components present. Among the desired results of a well-preserved cellular specimen is the ability to accurately assess quantitative and semiquantitative criteria including hyperchromasia, nuclear/cytoplasmic ratio, and chromatin patterns. The recognition of fine nuclear details such as grooves, notches, inclusions, and pseudo-inclusions are essential for the definitive diagnosis of certain subtypes of thyroid, breast, urinary, and soft-tissue tumors with concomitant prognostic implications. Equally important is maximizing the number of true tissue fragments. In contrast to the screening nature of the Pap test, where an interpretation of the sample is followed by a biopsy to establish the final diagnosis, fine-needle aspiration (FNA) is used to obtain a definitive diagnosis and as more targeted therapies are developed, expectations on the diagnostic performance of cytopathologists will increase. 1

Cervicovaginal Cytology

Specimen Type
Two types of specimen are available for cervicovaginal cytology: smear for the conventional Pap (CP) and liquid-based preparation (LBP), which emerged as an alternative sampling and preparation method in the 1990s. 2 In the United States ThinPrep and SurePath are the first two LBP approved by the Food and Drug Administration. In several countries manual methods of LBP such as DNA Citoliq, Cyto-Screen, PapSpin, and Autocyte Manual are available. The overall quality of most LBP is surprisingly good because cell preservation is enhanced in contrast to conventional smears, which may have thick and thin areas or air-drying artifacts. 3, 4

Patient Identification
It is important to match the name of the patients on the smears or LBP vials with the names in the requisition form, to prevent mix-ups. The use of an automated processing system improves the accuracy of patient identification and ensures patient chain of custody with bar-coded labels and etched numbers on the glass slides.

Clinical Information
Clinical information needs to be integrated with the sample interpretation. Minimal information required is patient age, date of last menstrual period, history of previous abnormal Paps, the latter being more often available in the laboratory computer system. Age and menstrual status are particularly important for the interpretation of endometrial cells. A history of previous malignancies of the female genital tract will alert the laboratory of the possibility of a recurrent disease or changes secondary to treatment.

Microscopic Evaluation
In the United States most laboratories follow the 2001 Bethesda System for reporting cervicovaginal cytology. 5 Other countries utilize national reporting systems. In most laboratory settings the evaluation of gynecologic specimens is performed by both cytotechnologists, who screen all samples, and cytopathogists who are responsible for the final interpretation of all abnormal cases .

Sample Adequacy Assessment
Over the years there has been considerable debate about what constitutes an adequate sample. In the 2001 Bethesda conference this issue was addressed and specific guidelines for assessing sample adequacy emerged. Applying the minimum squamous cellularity criteria a cellularity of 10,000 to 12,000 squamous cells is considered adequate for conventional Paps. There is no need to count squamous cells but rather estimate the cellularity based on density cell patterns in reference images. In samples with cytolysis, atrophy, and cell clustering, when cell adequacy is borderline, professional judgment and hierarchical review is recommended. The presence or absence of endocervical cells/transformation zone component does not affect sample adequacy but should be mentioned as a quality factor. Specimens with more than 75% of squamous cells obscured should be termed unsatisfactory assuming that no abnormal cells are present. For liquid-based preparations 5,000 squamous cells is considered adequate cellularity and estimates are reached by counting the number of cells in ten high-power fields across the main diameter of the preparation. Using a 40× objective and FN20 eyepiece the number of cells in each field should be 3 for the ThinPrep and 7 for the SurePath for adequate cellularity. 6
Unsatisfactory sample rates are variable. Centrifugation LBP have a lower rate of unsatisfactory samples than filtration-based preparations because in the latter red blood cells or inflammatory cells may plug the filter pores. Reprocessing is recommended to lower the unsatisfactory rate. 7, 8
A meta-analysis of prospective studies comparing cytologic diagnosis and sample adequacy showed LBP improved sample adequacy and equal or superior results in diagnosing premalignant cervical lesions when compared with conventional Papanicolaou test. 9 In a comparison study of automated versus manual LBP few differences were found in the sample adequacy and cellular presentation: less uniform distribution of cells and more artifacts were noted in the manual methods such as DNA Citoliq and AutoCyte Prep compared with the ThinPrep method; sample adequacy and overall quality of all LBP were surprisingly good. 3

Interpretation
Detailed criteria for the interpretation of gynecologic samples are described in Chapters 6 – 11 of this book. The following highlights some observations in LBP. 6, 10 - 14
For LBP fixed in ethanol-based fixative the interpretation is closer to conventional smears. When LBP are concentrated in smaller areas, with three-dimensional cell clusters above the plane of squamous cells, focusing may be required more often.
There are many similarities in the evaluation of LBP and conventional smears but there are differences:
• Cell size;
• Pattern; and
• Background.
Because samples are collected in a liquid-based fixative solution, the presentation of the cellular material appears different, concentrated and evenly distributed ( Fig. 5.1 ). Cells are smaller, dispersed, and single, although cell clusters are also present. Cells in solution tend to round up and wet fixation enhances cytoplasm and nuclear morphology. Hyperchromasia as observed in abnormal cells on conventional smears is not always present in liquid-based preparations, especially in methanol-based fixatives, and lack of hyperchromasia may render the interpretation of high-grade squamous intraepithelial lesion more difficult. Variations in nuclear size and shape and especially appreciation of nuclear contours play an important role in the evaluation on LBP.

Fig. 5.1 Normal squamous and endocervical cells appear evenly distributed. SurePath (Papanicolaou x LP).
The background is generally clean and debris more clumped. Blood, mucus, and inflammation are rarely obscuring, and inflammatory cells tend to cling to epithelial cells. Tumor diathesis has a “ratty” appearance but this type of background can also be observed with cytolytic or inflammatory patterns.
Some cytologic entities have key features in LBP:
Key features of atrophy
• Fewer bare nuclei;
• Flat sheets of parabasal cells; and
• Preserved nuclear polarity ( Fig. 5.2 ).
Key features of trichomonads
• Smaller;
• Difficult to visualize; and
• More visible nuclei, eosinophilic granules, or flagella.
Key features of lymphocytic cervicitis
• Lymphoid cells in clusters; and
• Confused with endometrial cells ( Fig. 5.3 ).
Key features of repair
• Cohesive cell groups;
• More rounded cytoplasmic borders;
• Less streaming; and
• Prominent nucleoli.
Key features of metaplastic cells
• Often single, smaller, and rounder;
• Confused with HSIL; and
• Paler chromatin pattern.
Key features of low-grade squamous intraepithelial lesions (LSIL)
• Large cell size;
• Koilocytosis;
• Multinucleation; and
• Nuclei show decreased hyperchromasia ( Fig. 5.4 ).
Key features of high-grade squamous intraepithelial lesion (HSIL)
• Fewer abnormal cells;
• Single cells more common than sheets;
• Syncytial aggregates;
• High nuclear cytoplasmic ratio; and
• Nuclear membrane irregularities ( Figs. 5.5 , 5.6 ).
Key features of squamous cell carcinoma
• Single cells and syncytial aggregates;
• Pleomorphic and less hyperchromatic nuclei;
• Rare nucleoli; and
• Diathesis surround cells or cling to malignant cells ( Figs. 5.7 , 5.8 ).
Key features of endocervical adenocarcinoma in situ
• Strips of cells with pseudostratification;
• Nuclear crowding; and
• Subtle appearance of feathering and rosettes.
Key features of endometrial cells
• Tight or lose cell clusters;
• Vacuolated cytoplasm;
• Enhanced nuclear detail; and
• Confused with low-grade endometrial adenocarcinoma ( Fig. 5.9 ).
Key features of endometrial adenocarcinoma
• Papillary configurations;
• 3D groups; and
• Less prominent tumor diathesis ( Fig. 5.10 ).

Fig. 5.2 Flat sheet of parabasal cells with preserved nuclear polarity in atrophy. ThinPrep (Papanicolaou x MP).

Fig. 5.3 Lymphoid cells appear in aggregates in chronic lymphocytic cervicitis. ThinPrep (Papanicolaou x MP).

Fig. 5.4 Koilocytes showing perinuclear cavitation and binucleation in LSIL. ThinPrep (Papanicolaou x MP).

Fig. 5.5 Single cells with high n/c ratio and hyperchromasia in HSIL . SurePath (Papanicolaou x MP).

Fig. 5.6 Immature cells showing irregular nuclear outlines in HSIL . ThinPrep (Papanicolaou x MP).

Fig. 5.7 Syncytium of malignant cells in squamous cell carcinoma with variation in nuclear size and shape, irregular chromatin distribution and clinging diathesis. ThinPrep (Papanicolaou x HP).

Fig. 5.8 Pleomorphic cells and clean background in squamous cell carcinoma . SurePath (Papanicolaou x HP).

Fig. 5.9 Normal endometrial cells in tight cluster and enhanced nuclear detail. ThinPrep (Papanicolaou x MP).

Fig. 5.10 Cluster of malignant cells and granular tumor diathesis in endometrial adenocarcinoma . ThinPrep (Papanicolaou x MP).

Nongynecologic Cytology

Specimen Type
The three major processing modalities for nongynecologic specimens consist of direct smear , filtration (Millipore, SurePath, and ThinPrep) and cytocentrifugation-based preparations (Cytospin). All three techniques share the capacity to archive a portion of the specimen for the application of special and immunocytochemical stains. Each modality subjects the cellular material to different degrees of physical forces and chemical influences. This will result in certain artifact profiles that can affect cytomorphologic interpretation. True tissue fragments with architectural features similar to those of histologic specimens are present in quality direct smears ( Fig. 5.11 ). The use of Rapid Pap and/or Diff Quik stains in conjunction with the direct smears technique allows for on-site cellular adequacy of FNA. The importance of on-site specimen evaluation cannot be overstated since it has been shown conclusively to significantly reduce the inadequacy rate. 15 The preservation of diagnostically important true tissue fragments and the three-dimensional microtopography of the Millipore cellulose filter produces an increased depth of field and subsequent exquisite cytologic detail ( Fig. 5.12 ). The ThinPrep technique consistently produces the truest monolayer, thus minimizing the obscuring effects of background elements and cellular clumping ( Fig. 5.13 ). SurePath and Cytospin preparations also present excellent cytomorphology ( Figs. 5.14 , 5.15 ).

Fig. 5.11 Follicular architecture in follicular thyroid adenoma FNA. Direct smear (Diff Quik x MP).

Fig. 5.12 Nuclear grooves, chromatin clearing, margination and three-dimensional effect in papillary thyroid carcinoma FNA. Millipore filter (Papanicolaou x HP).

Fig. 5.13 True monolayer of benign urothelial cells in clear background in voided urine. ThinPrep (Papanicolaou x MP).

Fig. 5.14 True tissue fragment of papillary urothelial carcinoma in voided urine. SurePath (Papanicolaou x MP).

Fig. 5.15 Metastatic ovarian carcinoma in peritoneal fluid showing clean background and cells with good nuclear detail. Cytospin (Papanicolaou x MP).
When sufficient cellularity and technical support are available cell blocks can provide much in the way of additional diagnostic information. Included among the possibilities are architectural and staining properties that approach what is seen in conventional paraffin-embedded hematoxylin and eosin (H and E) stained surgical biopsies . In this manner cell blocks often provide an immeasurable level of comfort to many pathologists more accustomed to conventional histologic diagnoses. In particular types of FNA specimens such as salivary gland cystic lesions and thyroid nodules conventionally prepared cells blocks may provide key information to clinch a definitive diagnosis . A common example is in distinguishing chronic sialadenitis from Warthin's tumor in which characteristic large papillary oncocytic lymphoepithelial tissue fragments are best appreciated and often only appear in the cell block pellet derived from needle rinses ( Fig. 5.16 ). 16 Similarly, cell blocks have been reported to be extremely helpful in discerning malignant papillary tissue fragments containing fibrovascular cores from benign papillary tissue fragments that for the most part lack fibrovascular cores. 17 Also the characteristic Orphan Annie-eyed nuclear chromatin pattern (felt by some to be a useful reproducible histologic artifact) is best seen in cytology specimens that have been conventionally processed into formalin-fixed paraffin-embedded cell blocks. 18 - 21 In a study by Sanchez and Selvaggi additional morphologic information derived from cell blocks was found to be diagnostically contributory in 31% thyroid FNAs . 22 In addition to its unique ability to provide essential histologic clues cell blocks offer in theory and often in practice the opportunity to perform the same battery of ancillary studies as conventionally processed surgical biopsy tissue including molecular analysis and electron microscopy. 23, 24

Fig. 5.16 Warthin's tumor in parotid FNA. Cell block (H&E x MP).
A few published reports document the fact that immunocytochemistry can be performed successfully on Cytospin, ThinPrep, Millipore filter, and direct smear slides. 25 - 28 Options are somewhat limited in terms of the number and types of antibodies available due to the typically sparse and delicate nature of the cellular material present in these preparations. Many articles, on the other hand, describe the diagnostic usefulness of a variety of immunohistochemical stains on cell block material. 29 - 38 In particular, paraffin-embedded cell blocks have proven efficacious by helping to resolve benign reactive lesions from both primary and metastatic tumors ( Fig. 5.17 ). Such common dilemmas may often require a panel of immunohistochemical stains. See Chapter 35 for additional information.

Fig. 5.17 Metastatic lung adenocarcinoma in pleural fluid. Cell block (TTF1 immunostain positivity x MP).

Comparison of Nongynecologic Processing Techniques
There are many articles found in the contemporary literature which compare the efficacy of the processing modalities based on costs, cellular yield, unsatisfactory rates, artifacts, and diagnostic accuracy. Most deal with specific types of specimens such as urine, pleural fluid, pancreatic/biliary tract, soft tissue, breast, and thyroid FNA. The recommendations of the various authors differ somewhat in favoring one processing technique over the other. A thorough review of the relevant literature, however, leads one to the conclusion that direct smears, cytocentrifugation, and filtration techniques are worthy of routine use with comparable diagnostic and cost parameters for most nongynecologic specimens. In the end the decision to utilize one or more of the processing techniques depends upon weighting the service demands against the financial and labor resources available in a particular laboratory.
The following highlights some observations in the literature concerning these processing modalities:
Cellular yield was found to be superior by the ThinPrep method which retained three times as many cells as cytocentrifugation. 39 There were no statistically significant differences in unsatisfactory rates, sensitivity, specificity, or positive predictive value in both FNA and body cavity fluid specimens processed by ThinPrep and direct smears methods. 40 Overall technical quality was reported to be improved by ThinPrep processing when compared to direct smears on split FNA specimens due to cleaner background and better monolayer formation. 41, 42 Some authors, however, advised caution to avoid diagnostic errors when interpreting ThinPrep slides due to the increased incidence of following cytologic artifacts ( Table 5.1 ): disruption of tissue fragments, formation of cell clusters, aggregation of lymphocytes, cellular shrinkage, attenuation of nuclear details, and exaggeration nucleolar prominence. 42 In comparison to SurePath processed specimens ThinPrep slides demonstrated increased cellular shrinkage , flattening, and fragmentation of large cellular sheets and nuclear chromatin patterns were reportedly more difficult to evaluate. 43 SurePath slides were also found to contain larger branched three-dimensional tissue fragments. 43
Table 5.1 Artifacts more commonly seen with nongynecologic LBP Cellularity Nuclear Background Cellular shrinkage Increased naked nuclei Loss of adipose tissue, stroma, mucin, and colloid Disruption of tissue fragments Decreased nuclear chromatin details Colloid that appears as dense droplets Flattening and fragmentation of large cellular sheets Attenuation of nuclear grooves and pseudo inclusions Aggregation of lymphocytes Formation of cell clusters Exaggerated nucleolar prominence   Artificially increased single epithelial cells    
Separate studies involving FNAs of thyroid nodules, breast and salivary gland lesions, and pancreatic and soft-tissue tumors reported somewhat conflicting results in terms of unsatisfactory rates, quality of nuclear details, diagnostic accuracy, and relative prevalence of artifacts when ThinPrep-processed slides were compared to direct smears. 44 - 53 Among the artifacts attributed to the ThinPrep method for FNA specimens are the inability to assess cellularity of individual passes; diminished/distorted extracellular and stromal elements such as mucin, stroma, adipose tissue, and colloid that also appeared as dense droplets ( Fig. 5.18 ); crowded tight tissue clusters with loss of cellular preservation; increased cellular and tissue fragment disruption; artificially increased single epithelial cells; numerous naked nuclei; pronounced nucleoli; decreased nuclear details; and attenuation of nuclear grooves and pseudo-inclusions. Authors did note that significantly more conventional smears were limited by air-drying artifact. Additionally, ThinPrep slides had greater cellularity, improved nuclear detail, and more easily recognizable myoepithelial cells relative to direct smears ( Fig. 5.19 ). An added benefit of greater suitability for immunoperoxidase staining was also documented for ThinPrep processing. 48

Fig. 5.18 Dense colloid droplet in thyroid FNA in goiter. ThinPrep (Papanicolaou x MP).

Fig. 5.19 Benign ductal cells in association with myoepithelial cells . ThinPrep (Papanicolaou x MP).
Specific examples in which artifact-associated ThinPrep-processed FNA slides compromised diagnostic accuracy were cited: Four out of 21 fibroadenomas were correctly diagnosed on ThinPrep-processed breast FNA due to artificially increased single ductal epithelial cells and a lack of background stroma. 47 Three benign pancreatic lesions were interpreted as atypical/suspicious due to presence of single atypical cells with distinct nucleoli, and one mucinous pancreatic neoplasm was incorrectly diagnosed due to lack of background mucin. 51
The diagnostic value for pleural fluid specimens of ThinPrep versus Cytospin was compared by examining a large spectrum of cytologic features that would distinguish malignant mesothelioma (e.g., peripheral cytoplasmic skirt, bubbly cytoplasm, cyanophilic cytoplasm, and scalloped border of cell balls) from pulmonary adenocarcinoma (e.g., two-cell population, inspissated cytoplasmic material, cytoplasmic vacuole, angulated and indented nuclei, and smooth border of cell balls). 54 Based on statistical analysis most cytologic features examined in this study can be seen in both preparation techniques. The authors therefore concluded that ThinPrep preparation of pleural effusions does not appear to provide additional diagnostic value when compared to Cytospin preparation.
A comparative analysis of urine specimens processed by both ThinPrep and Cytospin techniques found that the cytomorphology and screening time were comparable with both techniques. 55 However, in cases of transitional cell carcinoma Cytospin was superior in terms of better preservation of architectural features and produced less artificial empty spaces and air-drying artifact ( Fig. 5.20 ). A contrasting study compared cytocentrifugation and ThinPrep techniques for cost efficiency including wages, investments in instrumentation and consumables, overall cytomorphologic quality, and suitability for molecular studies. 56 Based on their examination of 224 split urine samples the authors reported that cytocentrifugation with disposable chambers resulted in a global cytomorphologic quality superior to that of ThinPrep. In addition utilizing a 200-specimen per month calculation a greater cost efficiency was achieved with cytocentrifugation than with ThinPrep. In a similar subsequent study the same authors compared cytomorphologic quality of urine specimens prepared by ThinPrep, direct smears, Cytocentrifugation, and ThinPrep and Millipore filtration. 57

Fig. 5.20 High-grade urothelial carcinoma . Cytospin (Papanicolaou x MP).
The conclusions of the study were as follows:
• Direct smears show good overall results;
• Cytocentrifugation with reusable chambers should be avoided;
• Cytocentrifugation with disposable chambers (Cytofunnels or Megafunnel chambers) gives excellent results; and
• Millipore filtration followed by blotting should be avoided due to its poor global quality.
Contrasting results were reported when voided urine specimens processed via Millipore filter cytosieve technique were examined. 58 True tissue fragments consisting of either flat sheets or three-dimensional structures were significantly more common in voided urine specimens with follow-up biopsies of TCC than in negative biopsies. A considerably lower rate of tissue fragments was reported when voided urine specimens were processed utilizing cytocentrifugation with no statistically significant correlation found between the incidence of cell groups in voided urine specimens and the presence of TCC in follow-up biopsies. 59 It has been suggested that the reason for these discrepant findings is due to the stronger, more disruptive centripetal force imposed on true tissue fragments by cytocentrifugation relative to the weaker, gentler forces of gravity and suction encountered by the Millipore filter cytosieve technique. 58

Specimen Cross-Contamination
The potentially catastrophic problem of cross contamination of cytology specimens can occur with any processing modality or in any stage of the process from fixation to coverslipping ( Fig. 5.21 ). Fortunately steps can be taken in the specimen processing and staining to minimize its rate occurrence in cytopreparatory laboratories. 60, 61

Fig. 5.21 Cross-contamination from ovarian carcinoma (see Fig.5.15 ) in pancreatic cyst FNA. Cell block (H&E x MP).

Specimen Processing

A. All test requisitions, specimen cups, tubes, and slides must be identified with their own unique accession number before processing.
B. Laboratory technicians must carefully aspirate sample into pipette tip, making sure no sample gets sucked into the pipette barrel. If this happens the pipette is immediately removed from production. The barrel is cleaned with alcohol and water, dried, and tested for contamination before being placed back into production.
C. Pipette tips are changed for every sample.
D. All cytochambers and holders, if not disposable, are soaked for at least 30 minutes and then washed in the dishwasher and dried before being used the following afternoon.
E. All staining solutions must be either filtered or changed after every staining run.
F. All microscopists shall inform the supervisor of suspected contamination. Immediate corrective action shall follow with appropriate documentation of each occurrence.

Specimen Staining

A. All stains and staining solutions must be checked, filtered, and changed daily with documentation supporting this. Discard all water rinses after usage.
B. Special staining is to be performed in separate Coplin jars with all reagents made fresh daily.
C. Hacker coverslipping xylene wells must be filtered after each use.
D. Known positive cases should be stained separately in Coplin jars or stained in the regular staining dishes provided that the regular staining setup is then discarded of all solutions.
E. Any suspected floater (atypical cell contaminant seen on a different focal plane found on a slide) identified on a slide should be brought to the attention of the supervisor. For nongynecologic specimens, it is recommended that additional slides be made with the leftover sediment. Again no staining should continue until all stains, dishes, and jars are either filtered or changed and cleaned.
F. If contamination is suspected and no residual specimen sediment is available, consideration must be given to cancellation of the test. If the test is canceled a phone call must be placed to the requesting clinician and a hard copy of the report with appropriate explanation shall be issued. All inquires will be documented in the reconciled report logbook.
It is also important to remember that the possibility of cross-contamination is not limited to the cytopreparatory laboratory. This is particularly true when cell blocks are processed along with surgical pathology tissue utilizing the same formalin baths and automated processors. Contamination due to shedding of malignant cells and tissue can occur from cell block to histology tissue block and vice versa . To minimize this type of cross-contamination the following steps should be taken:
1. The histology laboratory should be notified when cell blocks are requested from cytology specimens likely to shed tumor cells such as ascites fluid from a known ovarian carcinoma patient ( Fig. 5.21 ).
2. Cassettes for cell blocks should be placed in separate formalin baths from histology tissue cassettes.
3. Cassettes for cell blocks whenever possible should have their own separate runs in the automated processors without surgical pathology tissue cassettes.

Specimen Mishandling
Once errors in handling the specimens are detected, a root cause analysis should take place to help identify what, how, and why the error happened. Understanding why a mistake occurred is the key to develop effective quality control measures to prevent it from occurring again. A review of skill-based activities, if appropriate, to ensure appropriate level of hands-on training should be provided in addition to the training development process to ensure adequate guidance. For errors related to sample interpretation see Chapter 4 , Diagnostic Quality Assurance in Cytopathology .

Concluding Remarks
The parameters for evaluation of smears and liquid-based preparations have been described in this chapter. As we have seen, several factors play an important role in the evaluation of the cellular sample and for optimal results good fixation and processing techniques, availability of clinical information, and expertise in interpretation are required. Criteria for interpretation of gynecologic and nongynecologic samples are described in all chapters on diagnostic cytology in this book but observations on liquid-based preparations were highlighted here to show differences in cell size, pattern background, and artifacts. A comparison of nongynecologic processing modalities and specimen cross-contamination as well as procedures to prevent the problem were also presented. The use of adjunct techniques in diagnostic cytology is already an important component of the specimen evaluation. In the future the diagnostic potential of cytology will increase with development of new fluorescent in situ hybridization (FISH) probes and more assays to identify genetic alterations that serve as therapeutic targets in addition to the use of microarrays, allowing a global and integrative approach to diagnosis.

References

1. Bibbo M., Akerman M., Alves V.A.F., et al. How technology is reshaping the practice of nongynecologic cytology. Acta Cytol . 2007;51:123-152.
2. Austin M.R., Ramzy I. Increased detection of epithelial cell abnormalities by liquid-based gynecologic cytology preparations. A review. Acta Cytol . 1998;42:178-184.
3. Alves V.A.F., Bibbo M., Schmitt F.C.L., et al. Comparison of manual and automated methods of liquid-based cytology: A morphologic study. Acta Cytol . 2004;48:187-193.
4. Mintzer M.P., Curtis P., Resnick J.C., et al. The effect of the quality of Papanicolaou smears on the detection of cytologic abnormalities. Cancer (Cancer Cytopathol) . 1998;87:113-117.
5. Solomon D., Davey D., Kurman R., et al. The 2001 Bethesda System: Terminology for reporting results of cervical cytology. JAMA . 2001;287:2114-2119.
6. Solomon D., Nayar R. The Bethesda System for Reporting Cervical Cytology , 2nd edn. New York: Springer-Verlag; 2004.
7. Hellman D.J., Bolick D.R. Improving the adequacy of the ThinPrep Pap Test ∗ . Reference Pathology Services, Murray, Utah. Acta Cytol . 1998;42:1241.
8. Vassilakos P., Cossali D., Albe X., et al. Efficacy of monolayer preparations for cervical cytology: Emphasis on suboptimal specimens. Acta Cytol . 1996;40:496-500.
9. Bernstein S.J., Sanchez-Ramos L., Ndubisi B. Liquid-based cervical cytologic smears study and conventional Papanicolaou smears: A metaanalysis of prospective studies comparing cytologic diagnosis and sample adequacy. Am J Obstet Gynecol . 2001;185:308-317.
10. Douglas K.L. ThinPrep Morphology Reference Manual. Cytic Corp, 1994.
11. ThinPrep Pap Test Morphology Reference Atlas. Cytic Corp, 2004.
12. AutoCyte Prep Reference Manual. TriPath Imaging, 1998.
13. Wilbur D.C., Dubeshter B., Angel C., et al. Use of Thin-layer preparations or gynecologic smears with emphasis on the cytomorphology of high-grade intraepithelial lesions and carcinomas. Diagn Cytopathol . 1996;14(3):201-211.
14. Wood M.D., Horst J.A., Bibbo M. Weeding atypical glandular cell look-alikes from the true atypical lesions in liquid-based pap tests: A review. Diagn Cytopathol . 2007;35:12-17.
15. Nasuti J.F., Gupta P.K., Bloch Z.W. Diagnostic value and cost-effectiveness of on-site evaluation of fine needle aspiration specimens. Review of 5,688 cases. Diagn Cytopathol . 2002;27:1-4.
16. Nasuti J.F., Yu G.H., Gupta P.K. Fine needle aspiration of cystic parotid gland lesions. An institutional review of 46 cases with histologic correlation. Cancer Cytpathol . 2006;90:111-116.
17. LiVolsi A.L. Surgical Pathology of the Thyroid. vol 22. 1990 WB Saunders Philadelphia Major Problems in Pathology
18. Carcangiu M.L., Zampi G., Rosai J. Papillary thyroid carcinoma. A study of its many morphologic expressions and clinical correlates. Pathol Annu . 1985;20:1-44.
19. Deligeorgi-Politi H. Nuclear crease as a cytodiagnostic feature of papillary thyroid carcinoma in fine needle aspiration biopsies. Diagn Cytopathol . 1987;3:307-310.
20. Vickery A.L. Thyroid papillary carcinoma. Pathologic and philosophic controversies. Am J Surg Pathol . 1983;7:797-807.
21. Vickery A.L., Carcangiu M., Johannessen J.V., et al. Papillary carcinoma. Sem Diagn Pathol . 1985;2:90-100.
22. Sanchez N., Selvaggi S.M. Utility of cell blocks in the diagnosis of thyroid aspirates. Diagn Cytopathol . 2006;34:89-92.
23. Schoedel K.E., Finkelstein S.D., Swalsky P.A. Molecular profiling of primary and metastatic neoplasms in lung using material obtained by fine needle aspiration. A report of two cases. Diagn Cytopathol . 2004;30:342-346.
24. Young N.A., Naryshkin S., Katz S.M. Diagnostic value of electron microscopy on paraffin-emdedded cytologic material. Diagn Cytopathol . 1993;9:282-290.
25. Fetsch P.A., Simsir A., Brosky K., et al. Comparison of three commonly used cytologic preparations in effusion immunocytochemistry. Diagn Cytpathol . 2002;26:61-66.
26. Lozano M.D., Panizio A., Toledo G.R., et al. Immunocytochemistry in the differential diagnosis of serous effusions. A comparative evaluation of eight monoclonal antibodies in Papanicolaoy stained smears. Cancer . 2001;93:68-72.
27. Fadda G., Rossi E.D., Mule A., et al. Diagnostic efficacy of immunocytochemistry on fine needle aspiration biopsies processed by thin-layer cytology. Acta Cytol . 2006;50:129-135.
28. Baloch Z.W., Cobb C., Roberts S., et al. Millipore filter cell block preparation. An alternative to cell block in non-gynecologic specimens of limited cellularity. Diagn Cytopathol . 1999;20:389-392.
29. Chandan V.S., Faquin W.C., Wilbur D.C., et al. The utility of GLUT-1 immunolocalization in cell blocks. An adjunct to the fine needle aspiration diagnosis of cystiv squamous lesions of the head and neck. Cancer . 2006;108:124-128.
30. Chhieng D.C., Benson E., Eltoum I., et al. MUC1 and MUC2 expression in pancreatic ductal carcinoma obtained by fine needle aspiration. Cancer . 2003;99:365-371.
31. Morgan R.L., DeYoung B.R., McGaughy V.R., et al. MOC-31 aids in the differentiation between adenocarcinoma and reactive mesothelial cells. Cancer . 1999;87:390-394.
32. Fetsch P.A., Abati A., Hijazi Y.M. Utility of the antibodies CA19-9, HMBE-1 and thrombomodulin in the diagnosis of malignant mesothelioma and adenocarcinoma in cytology. Cancer . 1998;84:101-108.
33. Levine P.H., Joutovsky A., Cangiarella J., et al. CDX-2 expression in pulmonary fine needle aspiration specimens. A useful adjunct for the diagnosis of metastatic colorectal adenocarcinoma. Diagn Cytopatol . 2006;34:191-195.
34. Saad R.S., Essig D.L., Silverman J.F., et al. Diagnostic utility of CDX-2 expression in separating metastatic gastrointestinal adenocarcinoma in fine needle aspiration cytology using cell blocks. Cancer . 2004;102:168-173.
35. Bedrossian C.W., Davilia R.M., Merenda G. Immunocytochemical evaluation of liver fine needle aspirations. Arch Pathol Lab Med . 1998;113:1225-1230.
36. Yoder M., Zimmerman R.L., Bibbo M. Two-color immunostaining of liver fine needle aspiration biopsies with CD34 and carcinoembryonic antigen. Potential utilization in the diagnosis of primary hepatocellular carcinoma versus metastatic tumor. Anal Quant Cytol Histol . 2004;26:61-64.
37. Lin F., Abdallah H., Meschter S. Diagnostic utility of CD10 in differentiating hepatocellular carcinoma from metastatic carcinoma in fine needle aspiration biopsy of the liver. Diagn Cytopathol . 2004;30:92-97.
38. Shin S.J., Hoda R.S., Ying L., et al. Diagnostic utility of the monoclonal antibody A103 in fine needle aspiration biopsies of the adrenal. Am J Clin Pathol . 2000;113:295-302.
39. Papillo J.L., Lapin D. Cell yield verses cytocentrifuge. Acta Cytol . 1994;38:33-36.
40. Leung C.S., Chiu B., Bell V. Comparison of ThinPrep and conventional preparations. Nongynecologic cytology evaluation. Diagn Cytopathol . 1997;16:368-371.
41. Dey P., Luthra U.K., George J., et al. Camparison of ThinPrep and conventional preparations on fine needle aspiration cytology material. Acta Cytol . 2000;44:46-50.
42. Michael C.W., Hunter B. Interpretation of fine-needle aspirates processed by the ThinPrep technique. Cytologic artifacts and diagnostic pitfalls. Diagn Cytopathol . 2000;23:6-13.
43. Micheal C.W., McConnel J., Pecott J., et al. Comparison of ThinPrep and TriPath PREP liquid-based preparations in nongynecologic specimens. A pilot study. Diagn Cytopathol . 2001;25:177-184.
44. Biscotti C.V., Hollow J.A., Toddy S.M., et al. ThinPrep versus conventional smear cytologic preparations in the analysis of thyroid fine needle aspiration specimens. Am J Clin Pathol . 1995;104:150-153.
45. Afify A.M., Liu J., Al-Khafaji B.M. Cytologic artifacts and pitfalls of fine needle aspirations using ThinPrep. A comparative retrospective review. Cancer . 2001;93:179-186.
46. Malle D., Valeri R.M., Pazaitou-Panajiotou K., et al. Our experience from the use of ThinPrep technique in thyroid fine needle aspiration. Acta Cytol . 2006;50:23-27.
47. Perez-Reyes N., Mulford D.K., Ruthowski M.A., et al. Breast fine needle aspiration. A comparison of thin-layer and conventional preparation. Am J Clin Pathol . 1994;102:108-110.
48. Biscotti C.V., Shorie J.H., Gramlich T.L., et al. ThinPrep versus conventional smear cytologic preparations in analyzing fine needle aspiration specimens from palpable breast masses. Diagn Cytopathol . 1999;21:137-141.
49. Bedard Y.C., Pollett A.F. Breast fine needle aspiration. A comparison of ThinPrep and conventional smears. Am J Clin Pathol . 1999;113:312-314.
50. Al-Khafaji B.M., Afify A.M. Salivary gland fine needle aspiration using the ThinPrep technique. Diagnostic accuracy, cytologic artifacts and pitfalls. Acta Cytol . 2001;45:567-574.
51. de Luna R., Eloubeidi M.A., Sheffield M.V., et al. Comparison of ThinPrep and conventional preparations in pancreatic fine needle aspiration biospy. Diagn Cytopathol . 2004;30:71-76.
52. Siddiqui M.T., Gokaslan S.T., Saboorian M.H., et al. Split sample comparison of ThinPrep and conventional smears in endoscopic retrograde cholangiopancreatography-guided pancreatic fine needle aspirations. Diagn Cytopathol . 2005;32:70-75.
53. Das K., Hameed M., Heller D., et al. Liquid-based versus conventional smears in fine needle aspiration of bone and soft tissue tumors. Acta Cytol . 2003;47:197-201.
54. Ylagan L.R., Zhai J. The value of ThinPrep and cytospin preparation in pleural effusion cytologic diagnosis of mesothelioma and adenocarcinoma. Diagn Cytopathol . 2005;32:137-144.
55. Nassar H., Ali-Fehmi R., Madan S. Use of ThinPrep monolayer technique and cytospin preparation in urine cytology. A comparative analysis. Diagn Cytopathol . 2003;28:115-118.
56. Piaton E., Huntin K., Faynel J., et al. Cost efficiency analysis of modern cytocentrifugation methods versus liquid based (Cytcy ThinPrep) processing of urinary sample. J Clin Pathol . 2004;57:1208-1212.
57. Piaton E., Faynel J., Huntin K., et al. Conventional liquid-based techniques versus Cytyc ThinPrep processing of urinary samples. A qualitative approach. BMC Clin Pathol . 2005;5:9-14.
58. Nasuti J.F., Fleisher S.R., Gupta P.K. Significance of tissue fragments in voided urine specimens. Acta Cytol . 2001;45:147-152.
59. Goldstein M.L., Whitman T., Renshaw A.A. Significance of cell groups in voided urine. Acta Cytol . 1998;42:290-294.
60. Holmquist D.H., Keebler C.M. Cytopreparatorytechniques. In: Keebler C.M., Somrak T.M., editors. The Manual of Cytotechnology . 7th edn. Chicago: American Society of Clinical Pathologists Press; 1993:411-448.
61. Bales E.B., Durfee G.R.Principles of operation of a cytopathology laboratory. Koss L.G., editor. Diagnostic Cytology and Its Histopathologic Bases. vol 2. Philadelphia:JB Lippincott, 1992, 1451-1531 4th edn
Part II
DIAGNOSTIC CYTOLOGY
CHAPTER 6 The Bethesda System for Reporting Cervical Cytology

Ritu Nayar, David C. Wilbur, Diane Solomon

Contents
The Bethesda System: Historical Perspective
The 2001 Bethesda System
Report Format
Specimen Adequacy
Bethesda 2001 Specimen Adequacy Categories
Squamous Cellularity
Quality Indicators
Management Guidelines
Impact on Laboratory Practice
General Categorization
Interpretation/Result
Negative for Intraepithelial Lesion or Malignancy (NILM)
Endometrial Cells
Epithelial Cell Abnormalities: Squamous Cell
Epithelial Cell Abnormalities: Glandular Cell
Educational Notes/Suggestions
Ancillary Testing
Automated Review
Interobserver Reproducibility in Cervical Cytology
The Bethesda System and Reporting Anal-Rectal Cytology
Concluding Remarks

The Bethesda System: Historical Perspective
Terminology forms the basis for effective communication between the laboratory and clinician. The clinician is expected to provide relevant patient information to the laboratory. It is the laboratory's responsibility to report results using terminology that clearly conveys the diagnostic interpretation of the morphologic findings. The use of a uniform diagnostic terminology facilitates communication by establishing a common language that, in theory, does not vary significantly from cytologist to cytologist or laboratory to laboratory. However, terminology is not static over time; rather, it evolves in parallel with increased understanding of the pathogenesis and biology of disease. The framework, therefore, must be flexible enough to incorporate advances in scientific knowledge without creating undue confusion or complexity.
In 1988, the National Cancer Institute sponsored an open workshop—including cytotechnologists, pathologists, clinicians, and representatives of professional organizations—to develop a uniform descriptive terminology for cervical/vaginal cytologic interpretation. The format that emerged became known as The Bethesda System (TBS). 1
Approximately two years later, a second meeting was convened to critique and refine the terminology based on experience with the use of the system in actual laboratory practice. Minor modifications were incorporated into the 1991 Bethesda System that streamlined the terminology. 2 In addition, an ad hoc committee developed criteria for specimen adequacy and Bethesda interpretive categories, culminating in the first TBS atlas that outlined and illustrated the morphologic features. 3 By the mid- to late 1990s, there was a significant penetration of the Bethesda System into cytopathology practice with approximately 90% of laboratories in the United States using the Bethesda terminology for reporting of cervical/vaginal cytology. 4
Among all the changes introduced by the implementation of TBS terminology into practice, none was more controversial than the category of atypical squamous cells (ASCUS). At that time, the majority of abnormal Pap tests reported annually in the United States, approximately 2.5 million, were interpreted as ASCUS and had highly variable management at considerable cost to the healthcare system. Another 1.2 million were interpreted as low-grade squamous intraepithelial lesion (LSIL). 5 In an effort to determine the best management strategy (effective as well as cost-effective) for women with these equivocal and low-grade abnormalities, the National Cancer Institute (NCI) sponsored the ASCUS/LSIL Triage Study (ALTS), which was completed in 2001. 6 This study, as detailed below, has allowed for a data-driven approach to management of these prevalent cervical cytologic abnormalities.
From its inception, the fundamental aim of the Bethesda System has been to communicate clinically relevant information from the laboratory to the patient's healthcare provider, using uniform, reasonably reproducible terminology which reflects the most current understanding of the biology of cervical neoplasia. Advances in the understanding of the biology of cervical cancer, results from clinical trials, the introduction of liquid-based cytology, human papillomavirus (HPV) testing, and automated screening devices for cervical cytology led to the decision to convene the third Bethesda workshop in April 2001.

The 2001 Bethesda System

TBS 2001 Process
Approximately eight months prior to the workshop, nine forum groups consisting of 6 to 10 individuals with a breadth of expertise in the area of cervical cancer, were organized under the sponsorship of the NCI to formulate draft recommendations. Internet bulletin boards were open to the worldwide cytology community for six months during the pre-conference process of review and discussion. Over 1000 comments were considered in revising the pre-workshop drafts. The 2001 Bethesda workshop was co-sponsored by 44 international professional organizations and attended by over 400 individuals, including pathologists, cytotechnologists, gynecologists, attorneys, patient advocates, and other healthcare workers involved in women's health initiatives. The revised draft recommendations were presented by each forum group and after open discussions and voting by all participants, the 2001 Bethesda consensus terminology was finalized and published in 2002. 7
Following the Bethesda workshop, the American Society for Colposcopy and Cervical Pathology (ASCCP) held a comparable consensus workshop on patient management in September 2001. This was also preceded by an Internet discussion, and resulted in the development of evidence-based management guidelines for abnormal cervical cytology corresponding to the 2001 Bethesda reporting format. 8 The ASCCP management guidelines were subsequently updated at a consensus conference held in September 2006. 9
After the initial publication of the Bethesda System 2001 terminology ( Table 6.1 ), the NCI approached the American Society of Cytopathology (ASC) to collaborate on publication of the second edition of the Bethesda Atlas 10 and the development of an accompanying Bethesda System educational website. 11 Images chosen for the atlas and website underwent an extensive selection/validation process, and included classic as well as morphologically difficult and “borderline” images, illustrated on both conventional and liquid-based preparations. A subset of images chosen for the Bethesda atlas were used to assess interobserver reproducibility in gynecologic cytology—the details of this Bethesda interobserver reproducibility project are described below. 12
Table 6.1 The 2001 Bethesda System SPECIMEN TYPE Indicate conventional smear vs liquid-based preparation vs other ADEQUACY OF THE SPECIMEN
• Satisfactory for evaluation (describe presence or absence of endocervical/transformation zone component and any other quality indicators, e.g., partially obscuring blood, inflammation)
• Unsatisfactory for evaluation … (Specify reason)
– Specimen rejected/not processed (specify reason)
– Specimen processed and examined, but unsatisfactory for evaluation of epithelial abnormality because of (specify reason) GENERAL CATEGORIZATION (OPTIONAL)
• Negative for intraepithelial lesion or malignancy
• Other
• Epithelial cell abnormality: see Interpretation/Result (specify squamous or glandular as appropriate ) INTERPRETATION/RESULT Negative for intraepithelial lesion or malignancy (when there is no cellular evidence of neoplasia, state this in the General Categorization above and/or in the Interpretation/Result section of the report—whether or not there are organisms or other non-neoplastic findings) ORGANISMS
• Trichomonas vaginalis
• Fungal organisms morphologically consistent with Candida spp.
• Shift in flora suggestive of bacterial vaginosis• Bacteria morphologically consistent with Actinomyces spp.
• Cellular changes associated with herpes simplex virus OTHER NON-NEOPLASTIC FINDINGS (OPTIONAL TO REPORT; LIST NOT INCLUSIVE)
• Reactive cellular changes associated with:
– Inflammation (includes typical repair)
– Radiation
– Intrauterine contraceptive device (IUD)
• Glandular cells status posthysterectomy
• Atrophy Other
• Endometrial cells (in a woman > 40 years of age) (specify if “negative for squamous intraepithelial lesion”) Epithelial cell abnormalities SQUAMOUS CELL
• Atypical squamous cells
– of undetermined significance (ASC-US)
– cannot exclude HSIL (ASC-H)
• Low-grade squamous intraepithelial lesion (LSIL) (encompassing: HPV†/mild dysplasia/CIN 1)
• High-grade squamous intraepithelial lesion (HSIL) (encompassing: moderate and severe dysplasia, CIS/CIN 2 and CIN 3)
– with features suspicious for invasion
• Squamous cell carcinoma GLANDULAR CELL
• Atypical
– endocervical cells (NOS or specify in comments)
– endometrial cells (NOS or specify in comments)
– glandular cells (NOS or specify in comments)
• Atypical
– Endocervical cells, favor neoplastic
– Glandular cells, favor neoplastic
• Endocervical adenocarcinoma in situ
• Adenocarcinoma
– endocervical
– endometrial
– extrauterine adenocarcinoma
– not otherwise specified (NOS) Other malignant neoplasms (specify) ANCILLARY TESTING Provide a brief description of the test method(s) and report result so that it is easily understood by the clinician. AUTOMATED REVIEW If case is examined by automated device, specify device and result. EDUCATIONAL NOTES AND SUGGESTIONS (OPTIONAL) Suggestions should be concise and consistent with follow-up guidelines published by professional organizations (references to relevant publications may be included).

Report Format
The basic structure of TBS includes three elements, based on communication needs germane, but not limited, to gynecologic cytology: (1) statement of specimen adequacy, (2) general categorization, and (3) descriptive terminology. The specimen type—conventional smear, liquid-based preparation, or other—should also be stated in the report ( Table 6.1 ).

Specimen Adequacy
Reporting of adequacy was an important quality assurance measure introduced by the Bethesda System. The 1988 Bethesda System incorporated a classification of three categories of specimen adequacy—satisfactory, less than optimal, and unsatisfactory—into the format of the report but did not outline specific morphologic criteria for evaluation of adequacy. Participants at the 1991 Second Workshop, and others in the cytopathology community, 13, 14 voiced the need for developing consensus guidelines. In response, following the second workshop, a Criteria Committee formulated the definitions for adequacy based on a combination of experience and review of an admittedly sparse scientific database. Three categories—“satisfactory,” “satisfactory but limited by…,” and “unsatisfactory”—based on estimates of overall squamous cellularity, assessment of the transformation zone component, and the presence/extent of obscuring or limiting factors, were suggested in an initial attempt to develop a more standardized approach to the evaluation of adequacy. 15 It was emphasized that the indicated percentages should be used as general ranges, not strict numerical cutoffs and that patient-related clinical factors and previous cytologic findings should always be taken into consideration.

Bethesda 2001 Specimen Adequacy Categories
In 2001, substantial changes were made to the adequacy component of TBS. The previously used borderline adequacy category of “less than optimal” (1988)/“satisfactory but limited by…” (1991) was deleted in order to provide the clinician a clearer and more reproducible indication of the adequacy of the specimen. 16 The classification recommended in TBS 2001 is either as “satisfactory” or “unsatisfactory”:
• Satisfactory 10 —Satisfactory for evaluation (describe presence or absence of endocervical/transformation zone component and any other quality indicators, e.g., partially obscuring blood, inflammation, etc.). For “satisfactory” specimens, including information on transformation zone sampling and other adequacy qualifiers (obscuring elements, poor preservation, etc.) encourages specimen takers to pay greater attention to specimen procurement and handling. Any factors that compromise specimen quality can be mentioned in a note.
• Unsatisfactory 10 —For unsatisfactory specimens the report should indicate whether the laboratory processed/evaluated the slide. Suggested wording is:
– Rejected specimen—Specimen rejected/not processed because ( specify reason: specimen not labeled, broken slide, etc. ).
– Fully evaluated, unsatisfactory specimen—Specimen processed and examined, but unsatisfactory for evaluation of epithelial abnormality because of ( specify reason: inadequate squamous component, obscuring blood, etc. ).
Additional comments/recommendations may be made as deemed appropriate.
While unsatisfactory specimens which are processed and evaluated are not suitable for excluding an intraepithelial lesion or malignancy, the presence of endometrial cells in a women 40 years or older, or the presence of organisms, can be reported in this context, since this information may prove to be clinically relevant for patient management.
As in prior Bethesda adequacy guidelines, if abnormal cells are detected, the specimen cannot be categorized as “unsatisfactory.”

Squamous Cellularity
TBS 1991 required that well-preserved and well-visualized squamous epithelial cells should cover more than 10% of the slide surface. In order to address adequacy on conventional as well as liquid-based preparations, and to improve interobserver reproducibility, TBS 2001 went further to provide numerical estimates of what constitutes adequacy for squamous cellularity in cervical cytology preparations.
• Conventional smears : An adequate conventional preparation should have a minimum of approximately 8,000–12,000 well-preserved and well-visualized squamous cells. This minimum cell count should be estimated , not counted. The count includes both nucleated mature and metaplastic squamous cells. The percentage of hypocellular areas, if present, should be estimated and the fields counted should reflect this proportion. The Bethesda atlas 10 and website 11 provide “reference images” of known cellularity at low (4×) magnification as a resource for cytologists to compare with the specimen being assessed.
• Liquid-based preparations (LBP): An adequate LBP should have an estimated minimum of at least 5,000 well-visualized, well-preserved squamous cells. 10, 17 Estimation of cellularity is suggested in borderline cases by performing representative field counts. A minimum of 10 fields, usually at 40×, are assessed along a diameter that includes the center of the preparation. The average number of squamous cells per field is thus estimated. One preliminary study suggested that LBPs containing 5,000–20,000 squamous cells should be considered as “borderline” cellularity. 18
The reader is referred to the Bethesda atlas for cellularity tables and figures. 10 These guidelines may change in the future based on additional data.
The TBS numeric criteria for cellularity may not be applicable to vaginal specimens, cases with extensive cytolysis, cell clustering, and some cases of atrophy. Cytologists should utilize clinical information and their best judgment when interpreting such cases. At present there are no published studies specifically addressing the relationship between low cellularity and false-negative rate.

Quality Indicators

Patient/Specimen Identification and Technical Interpretability
Correct specimen identification is essential for evaluation and is required by the Clinical Laboratory Improvement Amendments of 1988 (CLIA '88). In addition to ensuring that the specimen corresponds to the correct patient, proper identification allows the laboratory to locate prior records and slides from the patient that may influence the current evaluation.
The cellular material must be well fixed and unobscured for interpretation. Minimal data regarding how obscuring factors affect the interpretive reliability of a cervical specimen are available. In order to be considered obscuring, the epithelial cell morphology must be uninterpretable. For example, although most cervical samples contain inflammatory cells, moderate numbers do not generally obscure the nuclei of squamous cells. Even a large amount of inflammation or blood may be acceptable if it is spread thinly such that the intermixed epithelial cells can be easily visualized. In general, specimens with more than 75% of epithelial cells obscured are considered unsatisfactory. A variety of factors may compromise visualization of the cells. Heavy inflammation, blood, and extensive cytolysis are patient-related and independent of the sample taker. However, presence of air-drying, thick uneven smears, or lubricant is often inversely correlated with the skill and experience of the clinician. With liquid-based preparations a number of these factors become less significant; however, appropriate collection/rinsing techniques need to be utilized. Clinicians who repeatedly obtain technically poor quality specimens may benefit from constructive feedback provided in the written report, by telephone, or in a summary format comparing adequacy rates of peer group clinicians.

Clinical Information
Providing pertinent clinical information may increase the sensitivity and reliability of the evaluation by directing attention to a clinical question or by clarifying otherwise uncertain cytologic findings. At a minimum, age and date of last menstrual period should be provided. Absence of this information does not, however, preclude evaluation; therefore, the specimen may remain categorized as “satisfactory” in these circumstances.

Sampling of the Transformation Zone
Presence of endocervical or squamous metaplastic cells forms the microscopic basis for the assumption that the transformation zone has been sampled. The numeric criterion for a transformation zone component, at least 10 well preserved endocervical or squamous metaplastic cells, did not change from TBS 1991; however, due to the widespread utilization of liquid-based preparations, single endocervical/metaplastic cells are acceptable and clusters are no longer required. This definition applies to specimens from both premenopausal and postmenopausal women having a cervix. In the situation of marked atrophy, where metaplastic and endocervical cells often cannot be distinguished from parabasal cells, the laboratory has the option of making a comment regarding the difficulty in assessing the transformation zone component. Patient factors, such as location of the transformation zone, age, pregnancy, and previous therapy, may limit the clinician's ability to obtain an endocervical sample despite optimal collection technique.
Numerous cross-sectional studies have demonstrated that smears with endocervical and/or metaplastic cells have a significantly higher frequency and higher grade of squamous epithelial abnormality detected than do smears without such cells. 19 - 22 Paradoxically, short-term longitudinal studies of women whose initial negative smears lacked an endocervical component have shown no increase in abnormalities on repeat, satisfactory smears (as might be expected if the initial smears had a higher false-negative rate). 23, 24
Based on the above studies, TBS 2001 does not require the presence of a transformation zone component to categorize a specimen as satisfactory—adequate squamous cellularity is the only criterion. The absence of a transformation zone component is considered to be a quality indicator. With the reported increase in endocervical carcinoma, 25, 26 the importance of the transformation zone component may undergo further evaluation in the future, in order to ensure optimized screening performance in the setting of endocervical neoplasia.

Management Guidelines
The ASCCP has published management guidelines for Pap test specimen adequacy and quality indicators. 9 The preferred management for unsatisfactory Pap tests is a repeat test within a short interval of 2–4 months. Unsatisfactory cases are unreliable for detection of an epithelial abnormality; furthermore a longitudinal study found that unsatisfactory Pap tests are more often from high-risk patients and have significantly more SIL/cancer on follow-up than patients with satisfactory index Paps. 27 The guidelines suggest a repeat Pap test in 12 months for most women who lack a transformation zone component or whose cytology is partially obscured unless there is a history of prior adequate/negative Pap tests. Indications for considering earlier repeats are also outlined and depend on additional patient risk factors. 28 For quality assurance, it may be prudent to have the pathologist review unsatisfactory cases prior to final sign-out because of the clinical implications of such a report and the association of obscuring blood/inflammation with invasive cancers. 29

Impact on Laboratory Practice
The incorporation of specimen adequacy as an integral part of the cervical cytology report has been acknowledged as one of the most important contributions of TBS. The impact on laboratory practice has been dramatic. Surveys conducted by CAP revealed that in 1990 only 35% of responding laboratories routinely reported specimen adequacy; by 1992, this figure increased to 85%. 30 A 1991 CAP survey found that most laboratories reported unsatisfactory specimen rates of 0.5–1.0%. By the year 2003, a CAP survey assessing Bethesda implementation and reporting rates 31 found that 73.6% of responding laboratories had eliminated the use of the “satisfactory but limited by…” category and the 2001 TBS minimum squamous cellularity criteria had been adopted by 85.3%. Experts had predicted an increase in unsatisfactory rates with the use of TBS 2001 criteria. While some studies have indeed reported this (up to a tenfold increase on conventional smears 32 ), the CAP survey 31 and other reports from the United States 33 and Europe 34 did not show an increase in the unsatisfactory rate after conversion to TBS 2001 adequacy criteria. Possibilities suggested by the authors include improved sampling and preparation methods, related predominantly to liquid-based methodology, or, alternatively, lack of attention to the TBS 2001 criteria. 31

General Categorization
The general categorization is a clerical device to aid clinicians and their office staff in triaging patients/prioritizing cases for review and to assist laboratories in compiling statistical information.
There are 3 headings used under the general category:
1. “Negative for intraepithelial lesion or malignancy” for specimens in which an epithelial abnormality is not identified. Organisms and other benign/reactive cellular changes can be reported in the Interpretation under this category.
2. “Other” may be utilized for cases in which there is no clear cytologic abnormality but the findings may warrant follow-up/investigation based on patient risk, for example endometrial cells in a woman 40 years of age or older.
3. “Epithelial cell abnormality” may be utilized for squamous or glandular epithelial abnormalities. Specify as far as possible which type is present.
If more than one diagnostic entity is present—for example, an infectious process and an epithelial abnormality—the specimen should be categorized according to the most clinically significant lesion; in this example, epithelial cell abnormality. However, the general category should not replace narrative (descriptive) terminology for communicating the interpretation/result. Some laboratories also extend the concept of a general or summary categorization to nongynecologic specimens.

Interpretation/Result
The prior Bethesda system use of the term “diagnosis” was replaced by “interpretation/result” in 2001. Workshop participants felt that cervical cytology is a screening, not a diagnostic test, that provides the clinician with information on morphologic findings that need to be integrated with the patient's other clinical findings for a final diagnosis and subsequent management. 7

Negative for Intraepithelial Lesion or Malignancy (NILM)
The Bethesda 2001 category of NILM is used to report non-neoplastic findings in the absence of an intraepithelial lesion or malignancy. This term is used both as a general categorization and as an interpretation and incorporates the reporting of organisms and other non-neoplastic findings such as reactive cellular changes ( Table 6.1 ). The NILM category replaces the two prior Bethesda categories of “within normal limits” (WNL) and “benign cellular changes” (BCC). The basis of this change was to clearly communicate to the physician that despite any other “benign” changes reported, the Pap test is “negative” or without evidence of cervical intraepithelial neoplasia or malignancy.
Clearly, the main purpose of cervical cytology screening is the detection of cervical squamous cell carcinoma and its precursors; however, reporting the findings of organisms or reactive conditions can make an important contribution to patient care. This documentation can facilitate patient triage, provide clinical-cytologic correlation, and focus attention on cytomorphologic criteria during microscopic screening and interpretation of cervical cytology.
The category of “infections” was changed to “organisms” in TBS 2001 since the presence of some organisms represents colonization rather than a clinically significant infection. Excellent specificity and reproducibility can be achieved for the cytopathologic interpretation of fungal elements, Trichomonas vaginalis , Actinomyces , and herpes simplex virus, by application of reproducible morphologic criteria. The interpretation of Chlamydia spp. is not listed in TBS because of the acknowledged low diagnostic accuracy of routine cytology for this organism and because of the availability of other, more accurate detection methods. TBS lists the organisms that should be reported; however, the laboratory is advised to discuss the relevance of reporting organisms and other non-neoplastic findings with their clinicians and come to a decision about what to report under the NILM category.
Cells manifest reactive morphologic changes in response to a variety of traumatic insults such as infection, inflammation, and radiation. Reparative processes, radiation, atrophy, and intrauterine contraceptive devices are examples of entities that induce cellular changes that may mimic intraepithelial lesions or even cancer. Severe reactive/reparative changes are difficult to distinguish from neoplastic changes and such interpretations are well known to have lower reproducibility than classic repair. 35 It is, however, important to recognize benign reactive features in order to avoid overinterpretation and resulting false-positive interpretations. A CAP report indicates that reparative changes tend to be easier to recognize on LBP, yielding less false positives than on conventional smears. 36
Keratotic cellular changes—hyperkeratosis, parakeratosis, and dyskeratosis—are descriptive terms that do not clearly communicate a diagnostic interpretation and are not included in TBS. The classification of such changes as benign/reactive or dysplastic should be based on the cytoplasmic and nuclear alterations present and reported under the appropriate general category/interpretation.
Occasionally, benign-appearing glandular cells may be seen in post-hysterectomy patients that can have a wide variety of sources, including adenosis, metaplasia, and prolapse of the remaining fallopian tube after a simple hysterectomy. 37, 38 This finding can be communicated to the clinician under the NILM category and per current ASCCP guidelines does not require further follow-up. 9 Other non-neoplastic changes that may be reported under NILM include atrophy and tubal metaplasia.
Details regarding the morphology of these entities are discussed elsewhere in this book.

Endometrial Cells
TBS 1991 recommended that benign-appearing endometrial cells in postmenopausal women be reported as an “epithelial cell abnormality” based on the increased risk for endometrial adenocarcinoma (6%) and endometrial hyperplasia (12%) on a meta-analysis. 3, 39, 40
In TBS 2001, a new category was included to report the presence of benign-appearing endometrial cells in women aged 40 years or older. 7, 10 The basis for including this new category in TBS 2001 was twofold: (a) review of the published literature showed an exceedingly low rate of significant lesions in anyone less than 40 years of age, and (b) pathologists may lack clinical information on menstrual dates/menopausal status, hormone therapy/tamoxifen, abnormal bleeding, and other endometrial carcinoma risk factors. It is important to include in the interpretation whether the cytology is “negative for squamous intraepithelial lesion.”
Only exfoliated, intact endometrial cells should be reported under the “other” category. As described in Bethesda 2001, the exfoliated groups of endometrial cells may be of epithelial and/or stromal origin; morphological distinction of these two cell types is usually not possible. Directly sampled lower uterine segment or abraded stromal cells/histiocytes, when present alone, should not be reported under this category. Atypical endometrial cells should be reported as an epithelial glandular cell abnormality. 1, 10
The prevalence of benign-appearing endometrial cells cervical in Pap tests from women aged 40 years or older is difficult to assess due to differences in study designs, but has been estimated to range from 1–3/100 to 1/1600 or less. 39 After adoption of TBS 2001, there have been many reports in the cytology literature that have shown minimal risk associated with this interpretation, especially in premenopausal women. 40 This TBS category has been controversial for clinicians and initially resulted in an increase in endometrial biopsies.
It may be useful to add an educational note to this interpretation in order to clearly communicate to clinicians that this interpretation has an increased risk of neoplasia, but the risk is low, especially in premenopausal women and those without endometrial carcinoma risk factors, and that clinical correlation with other risk factors and symptoms is necessary. Examples of educational notes for this interpretation can be found in the second edition of the Bethesda atlas. 10
The 2006 ASCCP guidelines provide additional guidance and suggest that for asymptomatic women who are documented by clinical history to be premenopausal, with benign appearing endometrial cells, endometrial stromal cells, or histiocytes; no further evaluation is required. For documented postmenopausal women with endometrial cells, on the other hand, endometrial assessment is suggested, regardless of symptoms. 9

Epithelial Cell Abnormalities: Squamous Cell
Squamous intraepithelial lesion (SIL) encompasses the morphologic spectrum of noninvasive squamous epithelial abnormalities associated with HPV infection. Since the Bethesda System was introduced in 1988, this spectrum has always been divided into low-grade (LSIL) and high-grade (HSIL) categories. LSIL encompasses changes referred to as “HPV effect,” “koilocytosis,” and mild dysplasia/cervical intraepithelial neoplasia (CIN 1). HSIL includes moderate dysplasia (CIN 2) and severe dysplasia/carcinoma in situ (CIN 3). The basis for this bipartite classification of SIL in TBS is based on the principles that this division (a) better reflects natural history and clinical management and (b) has better intra- and interobserver reproducibility than does a three-tiered reporting system.

Atypical Squamous Cells (ASC)
The term ASCUS was initially introduced into the earliest version of the Bethesda System to reflect the reality and limitations of light microscopy in classifying borderline cytologic changes. The use of multiple ASCUS qualifiers such as “not otherwise specified” (NOS), “favor reactive,” and “favor SIL/dysplasia” led to overuse of this category and by 1996, ASCUS interpretations accounted for a mean of 5.2% of all cervical cytology reports in the United States. 41 ASCUS interpretations caused dilemmas for clinicians due to the lack of standardized follow-up and variability of outcomes.
With advances in the understanding of the biology of HPV infections and results from various natural history studies, 42 as well as from the NCI ALTS trial, 6 the focus of cervical cancer screening has shifted from detecting and treating any CIN to focusing on treating high-grade CIN. Based on this concept, in TBS 2001, the term ASCUS was replaced by ASC, which has a narrower definition and only two qualifiers: atypical squamous cells of undetermined significance (ASC-US) and atypical squamous cells, cannot exclude HSIL (ASC-H). 7 A subclassification was aimed at having greater clinical utility by clearly separating equivocal findings into those that are worrisome for HSIL in distinction from other types of ASC. As a general guide, the majority of ASC interpretations should fall into the ASC-US qualifier (90–95%) with only 5–10% into the ASC-H category. 7, 10
ASC is not a single biologic or interpretive entity: it encompasses a spectrum of cellular changes reflecting a variety of pathologic processes that for one reason or another cannot be more definitively categorized. Specifically, ASC should be used for changes suggestive of SIL, that are either quantitatively or qualitatively insufficient for a definitive interpretation. For a cell to be classified as ASC, it should show squamous differentiation, an increase in nuclear cytoplasmic ratio, and minimal nuclear changes. 10 In each case of ASC, the cytopathologist must consider the summation of the morphologic abnormalities in terms of quantity and severity within the context of the clinical information provided.

Atypical Squamous Cells of Undetermined Significance (ASC-US)
Most often, ASC-US involves noninflammatory changes in squamous cells with mature, superficial/intermediate-type cytoplasm. Nuclear enlargement is approximately two-and-a-half to three times the area of a normal intermediate squamous nucleus, but the chromatin remains evenly distributed without significant hyperchromasia. Nuclear outlines are smooth and regular, although there may be variation in nuclear size. The differential diagnosis is usually between a reactive change versus LSIL but the change(s) quantitatively or qualitatively fall short of establishing a definitive interpretation of LSIL. Round or ovoid cells that resemble large metaplastic or small intermediate cells may also be classified as ASC-US. In liquid-based preparations, the cells may appear smaller and rounder compared to conventional smears. The cells in question should always be compared to “normal”-appearing intermediate cells on the same slide. In distinguishing reactive changes, cells that demonstrate pale round nuclei and even chromatin distribution favor an interpretation as NILM rather than ASC.

Atypical Squamous Cells, Cannot Exclude High-Grade Squamous Intraepithelial Lesion (ASC-H)
The ASC-H category is useful for changes suggestive of, but fall short of a definite interpretation of HSIL. The differential includes HSIL and mimics of HSIL. A variety of patterns can be recognized:
1. Small cells with a high nuclear to cytoplasmic ratio or “atypical (immature) metaplasia. Nuclear abnormalities such as abnormal shapes, hyperchromasia, and chromatin irregularity favor HSIL over benign metaplasia.”
2. Crowded sheet pattern or so-called hyperchromatic cell groups. Dense cytoplasm, polygonal cell shape, and distinct cell borders favor squamous over endocervical cells. This cell pattern includes a broad differential from normal (atrophy, endometrial cells) to neoplastic (endocervical adenocarcinoma, HSIL, or HSIL involving glands) changes.
3. Atypical cells in the setting of atrophy, atypia seen following radiation therapy, poorly preserved endometrial cells or histiocytes, and intrauterine device users may all show cellular changes that are difficult to distinguish from HSIL. In such situations, a designation as ASC-H may be appropriate.

Laboratory Reporting of ASC
Subsequent to the publication and dissemination of TBS 1988/1991, many clinicians felt overwhelmed by ASCUS interpretations in their patient practices. This phenomenon was not limited to the United States or to TBS; greatly increased rates of minor degrees of abnormality have been observed in countries that do not use TBS. 43 The reasons underlying this real or perceived ASCUS explosion were twofold: (1) The constant specter of medical–legal litigation has lowered the threshold for diagnosis of cellular abnormalities in many laboratories; 9, 31 and (2) atypical cases historically may have been camouflaged in vague terms such as “inflammatory atypia,” “benign atypia,” “borderline HPV,” and “koilocytotic atypia.” The aggregation of all such equivocal cases under one heading highlighted the subjective, interpretative nature of cytopathologic diagnosis, something long understood by laboratorians but not always recognized by clinicians. Some contend that in TBS 1991, ASCUS merely replaced the old Pap Class 2 or “inflammatory atypia” designations. However, a study by Sidawy and Tabbara demonstrated that by using criteria similar to those outlined above, almost two-thirds of 88 smears previously interpreted as “inflammatory atypia” could be reclassified as reactive; only 3 out of 57 cases (5%) had CIN (all low grade) on follow-up colposcopic biopsies. In contrast, among the smears that fulfilled ASCUS criteria, 61% correlated with colposcopic biopsies positive for CIN. 44
Laboratory rates of ASC will vary depending on the patient population, the diagnostic criteria used, and the experience and skill of the microscopist(s). If used appropriately, ASC should be an infrequent designation employed only when cellular changes elude a more definitive interpretation. Although there is no “correct” percentage rate of ASC, benchmarks were provided when ASCUS was introduced in the Bethesda terminology. In a low-risk population, it was suggested that the rate of ASCUS should be less than 5%. For laboratories that serve high-risk populations (e.g., sexually transmitted disease clinics or colposcopy clinics), the rate of ASCUS could be higher, but by 1991 guidelines should not exceed two to three times the rate of SIL; thus the ratio of ASCUS/SIL suggested was in the range of 2–3:1. A 1993 CAP survey focusing on laboratory utilization of ASCUS found that 86% of responding laboratories used the term ASCUS and the median ASCUS rate was 2.8%, with 90% of laboratories reporting rates of less than 9%. The median ASCUS/SIL ratio was 1.7; for 90% of laboratories, the calculated ratio was less than 3.6. 41 In 2003, a follow-up CAP survey on Bethesda 2001 implementation and reporting rates showed a decrease in the average ASC/SIL ratio (from a median of 2.0 in 1996 to 1.4 in 2002). This can be explained by increased LSIL detection on LBP and also possibly by using Bethesda 2001 criteria more stringently. 31
Sherman and colleagues, in a study correlating cytopathologic diagnoses with detection of HPV DNA, also found that use of TBS criteria reduced the percentage of inconclusive “atypical” smears. Overall, a consistent relationship between high-risk HPV detection and TBS diagnostic categories was evident. High-risk HPV types were detected in 10% of negative smears, 30% of ASCUS, and 60% of SIL specimens. Based on these data, the authors proposed using high-risk HPV testing as an objective quality assurance measure to assess the performance of a cytopathology laboratory. 45 These results were substantiated by ALTS. 5, 6
It is well established that ASC-US is one of the least reproducible cytologic interpretations. 12, 46 Various quality assurance monitors may be utilized to evaluate the laboratory's utilization of ASC. These include the following:
1. Correlation of ASC-US cases with high-risk HPV positivity rates; results from ALTS indicate that this should be in the range of 40–60%, or in essence that ASC-US is a 50–50 proposition between SIL (usually LSIL) and cellular changes unrelated to HPV; 47, 48
2. Correlation of ASC cases with results of colposcopically directed biopsy;
3. Review of ASC cases by a second cytopathologist; and
4. Calculation of ASC/SIL ratio.
The ASC-HPV+/ASC ratio closely mirrors the ASC/SIL ratio. However, the ASC-HPV+/ASC ratio offers the additional advantage of identification of aberrant trends where ASC and SIL are both being misinterpreted, which may allow ASC/SIL ratios to remain within “acceptable” ranges despite the erroneous trend. 49 After implementation of LBPs, many laboratories have reported an increase in SIL rates over the increase in ASC rates, such that lower ASC/SIL ratios are being seen in many laboratories. 31 The prior 2–3:1 suggested ratio for ASC/SIL may undergo revision as future benchmarking results are gathered.

Clinical Management of ASC
Women with ASC have a low prevalence of invasive cancer, estimated at 0.1–0.2%. 50 The prevalence of CIN 2/3 is substantially higher in women with ASC-H (37–40%) 51, 52 than in those with ASC-US (11.6%). 51 ASC-US/high-risk HPV-positive cases over 2-years follow-up in ALTS have the same cumulative risk of CIN 2/3, about 27–28%, as a cytologic LSIL. 47 In contrast women in the ALTS who were ASC-US/high-risk HPV-negative showed a very low (1.4%) absolute risk of subsequently detected CIN 3 or worse and no cancers were detected in the two-year study period, similar to women with negative cytology in the absence of HPV testing. 53
The ASCCP consensus guidelines 8, 9 for ASC follow-up have seen widespread penetration into US practice. For ASC-US, HPV DNA testing, repeat cytological testing, and colposcopy are all acceptable methods for managing women over 20 years of age. When liquid-based cytology is used, reflex oncogenic or high-risk HPV DNA testing is the preferred management approach. ASC-US/high-risk HPV positive women should be managed in a fashion similar to those with LSIL. In adolescents (20 years and younger), follow-up with annual cytology is suggested due to the high prevalence of HPV in this age group and the low risk of persistence. 9
ASC-H on the other hand needs more aggressive follow-up, with colposcopic evaluation at the first interpretation. HPV testing is not recommended for triage of ASC-H. However, if a CIN 2/3 lesion is not identified at colposcopy, follow-up with either HPV testing at 12 months or cytology testing at 6 and 12 months is recommended. 9

Squamous Intraepithelial Lesions (SIL)
In TBS, LSIL and HSIL encompass the spectrum of precursors to squamous carcinoma of the cervix. Unlike CIN and dysplasia classifications that maintain HPV as a separate diagnostic category, low-grade SIL incorporates changes of HPV as well as mild dysplasia/CIN 1. High-grade SIL includes moderate dysplasia/CIN 2, severe dysplasia/CIN 3, and carcinoma in situ/CIN 3. Cytologists are of course free to append degrees of dysplasia or CIN classifications to a SIL interpretation.

Conceptual Basis for Two-Tiered Terminology of SIL
Previous terminology classifications—degrees of dysplasia and grades of CIN 1 to 3—have emphasized the morphologic continuum of squamous lesions that was thought to reflect a continuous process in the development of cervical cancer. Natural history studies 42 and HPV research have since established that HPV infection is a necessary cause for cervical carcinogenesis; 55, 56 however, even most oncogenic or high-risk HPV types cause transient low-risk lesions that regress, and cervical carcinoma develops in a small subset of persistent/progressive HPV infections. 57 It is estimated that approximately 70% of cervical cancers are associated with HPV 16 or 18. 58
The two-tiered LSIL/HSIL Bethesda approach attempts to morphologically distinguish minor from significant lesions; however, morphology is an imperfect reflection of biologic potential. Low-grade lesions, particularly those that persist, may progress or be associated with the development of high-grade lesions, and some high-grade lesions may regress. 57 Some have questioned setting TBS division of LSIL and HSIL at the breakpoint of CIN 1–2, or mild/moderate dysplasia, arguing that some CIN 2/moderate dysplasias should be considered low-grade lesions. However, because some CIN 2 lesions represent high-grade disease processes, conservatism dictated its inclusion into the more severe TBS category to ensure maximal sensitivity of the process.

Morphologic Features
An interpretation of LSIL based on cellular changes associated with HPV requires nuclear as well as cytoplasmic abnormalities. Nuclear changes may include enlargement with hyperchromasia or pyknosis, and chromatin smudging and wrinkling of nuclear contours. Cytoplasmic changes consist of a well-defined perinuclear cavity, associated with peripheral thickening of the cytoplasm or cytoplasmic orangeophilia, and rounding of cellular contours. Specimens with subtle changes that fall short of definitive LSIL may be categorized as ASC-US. Cytoplasmic vacuolization (pseudokoilocytosis) alone, in the absence of any nuclear atypia, is considered a benign change and should not be classified as LSIL or ASC-US.
Intraepithelial precursors of squamous cell carcinoma present a spectrum of morphologic changes within which one is able, in most cases, to classify lesions as LSIL or HSIL; however, occasional “borderline” cases occur. In the CAP Interlaboratory Comparison Program in Cervicovaginal Cytology (PAP), the discrepant rate between low- and high-grade lesions ranged from 9.8 to 15% for cytotechnologist, pathologist, laboratory, and all responses. 59 Cytology and histology may also be discrepant; 15–25% of women with LSIL cytology are found to have histologic CIN 2/3 on further work-up. 48, 59 Features that favor a high-grade lesion include increased numbers of abnormal cells, higher nucleus to cytoplasmic ratios, greater irregularities in the outline of the nuclear envelope and nuclear chromatin distribution, and increased number of chromocenters. The appearance of the cytoplasm may also assist in determining whether a borderline case is low- or high-grade SIL. LSIL changes typically involve “mature,” intermediate, or superficial type cytoplasm with well-defined polygonal borders. Cells of HSIL have a more immature type of cytoplasm, either delicate and lacy or dense/metaplastic, with rounded cell borders. Lesions previously termed “pleomorphic dysplasia,” “keratinizing dysplasia,” or “atypical condyloma,” which are composed of single cells or clusters of cells with enlarged hyperchromatic nuclei and abundant but abnormally keratinized cytoplasm, are always considered HSIL.

HSIL, Cannot Exclude Invasion
In rare cases of HSIL, invasive carcinoma may be difficult to exclude. Examples include atypical keratinized cells without diathesis/necrosis in the background or cases in which the background is suspicious but malignant cells are not seen. 10 This terminology may be used in such cases to communicate the increased concern to the clinician.

Squamous Cell Carcinoma
Squamous cell carcinoma is defined as an invasive malignant tumor with squamous differentiation. TBS does not subdivide squamous cell carcinoma into keratinizing and non-keratinizing types, although the atlas does discuss the morphology separately. On LBPs, tumor diathesis may be more difficult to recognize; and as such, in the United States a trend toward undercalling squamous cell carcinoma as HSIL, especially on LBP, has been noted. 60

Management of SIL
Low-grade SIL. Based on natural history studies of HPV infection, it is clear that the majority of cytologically detected LSIL regress within an average of two years. 42 After implementation of liquid-based cervical cytology, there has been a steady increase in the rate of LSIL in the United States—in 2003 the median rate was 2.4%. 31 Anecdotal experiences suggest that this has further increased with the use of location guided screening. In ALTS, the HPV positivity rate in LSIL was 83%; a meta-analysis published in 2006 reported a 76.6% positivity rate. 61 Thus HPV DNA testing is not suggested for initial triage of LSIL. Initial colposcopy identifies prevalent CIN 2 or greater in 18% of women with LSIL; subsequent follow-up over 2 years identified another approximately 10% CIN 2/3, irrespective of whether the initial colposcopy was negative or showed histologic CIN 1. 54, 62
Colposcopy is recommended for managing LSIL; exceptions include adolescents, postmenopausal, and pregnant women. If no lesion is identified or colposcopy is unsatisfactory, endocervical sampling is recommended for non-pregnant women. If CIN 2/3 is not detected, post-colposcopically, either HPV testing at 12 months or repeat cytology at 6 and 12 months is suggested. In adolescents with LSIL, initial colposcopy and/or HPV testing is not recommended; they should be followed by annual cytologic testing. Further details can be found in the ASCCP management guidelines. 9
High-grade SIL. The median percentile reporting rate of HSIL in the United States is estimated at 0.5%, 31 and approximately 2% of women with HSIL cytology have invasive carcinoma. 63 Follow-up of cytologic HSIL carries a significant risk of a CIN 2/3—a single colposcopy identifies 53–66% of prevalent CIN 2/3; and CIN 2/3 is found in 84–97% of women who proceed to a loop electrosurgical procedure (LEEP). 64 Thus both colposcopy and LEEP are acceptable for management of cytologic HSIL. 9

Epithelial Cell Abnormalities: Glandular Cell

Background
Cervical cytology is primarily a screening test for cervical squamous intraepithelial lesions and squamous cell carcinoma; cytology may have lower sensitivity for detection of glandular lesions has lower sensitivity due to limitations in sampling and interpretation.

Reporting Glandular Cells in TBS 2001
In TBS 2001, the term atypical glandular cells of undetermined significance (AGUS) has been eliminated to avoid confusion, particularly among clinical staff, with ASC-US. Abnormal glandular cells should be subclassified when possible as endocervical or endometrial; otherwise the generic term “atypical glandular cells” should be used. It is also advisable to use the qualifiers “not otherwise specified” or “favor neoplastic” for endocervical and glandular cells to convey the level of concern about any abnormality identified. The qualifier “favor reactive” from TBS 1991 has been eliminated as follow-up studies show that results are similar to those in the NOS category, and as such, this qualifier provides no useful predictive value. Atypical endometrial cells are not further qualified due to difficulty in doing so and lack of reproducibility of the morphologic criteria. Adenocarcinoma in situ is a separate interpretive entity in TBS 2001, having been well described and shown to have moderately good reproducibilty since the 1991 TBS version. 65

Atypical Glandular Cells (AGC)
As with its squamous ASC counterpart, this designation applies to glandular cells that demonstrate changes beyond those encountered in benign reactive processes, yet which are insufficient for an interpretation of in situ or invasive adenocarcinoma. This interpretation should be further qualified, where possible, to indicate whether the cells are thought to be of endocervical or endometrial origin. This category includes a broad morphologic spectrum ranging from atypical-appearing, reactive processes all the way to adenocarcinoma in situ (AIS). Therefore, lesions falling into this category should be further subclassified, if possible, according to whether a neoplastic process is favored or the changes are non-specific (NOS). Specific comments may be added to the interpretation if pertinent clinical findings and/or history are available and relevant (polyps, IUD, etc.).

Atypical Endocervical Cells, NOS
Endocervical cells can show a variety of changes associated with benign/reactive processes in the endocervical canal. Reactive endocervical cells can show some pleomorphism of cell size as well as nuclear enlargement, multinucleation, and prominent nucleoli; however, there is usually a honeycomb or sheet-like pattern and nuclei remain round and the chromatin bland. Such changes are usually recognized as NILM and not included in the AGC category. Cells that show cytologic changes beyond those recognized easily as reactive such as significant nuclear enlargement/crowding, hyperchromasia, loss of mucin, and loss of polarity should be considered for inclusion in the atypical endocervical cells, NOS category. Such changes may be seen in conditions such as tubal metaplasia, radiation therapy, endocervical polyps, and microglandular hyperplasia, and in IUD users, but also in neoplastic conditions in a small percentage of cases. This category therefore includes changes that are in excess of those attributable to a reactive/reparative condition but which fall short of those seen in glandular neoplasia.

Atypical Endocervical Cells, Favor Neoplastic
These cells are characterized by cellular strips and rosettes demonstrating elongated, overlapping nuclei with moderately coarse chromatin and hyperchromasia. The peripheral border of the glandular clusters may be “feathered,” with protruding nuclei, in contrast to the smooth communal border typical of glandular fragments. In LBPs cells are more rounded and three-dimensional. Cellular changes, while suspicious for in situ or invasive adenocarcinoma, are quantitatively or qualitatively insufficient for an outright interpretation as such.

Atypical Endometrial Cells
These are usually small groups of cells with slightly enlarged nuclei, and variable prominence of nucleoli and nuclear hyperchromasia. Their distinction from cytologically benign endometrial cells is based primarily on the criterion of increased nuclear size. When dealing with LBPs, it is important to keep in mind that menstrual/shed endometrium is often well preserved and may show nuclear size and shape pleomorphism and the presence of nucleoli. The differential of atypical endometrial cells is broad and may include endometrial polyps, endometritis, IUD associated changes, hyperplasia, and carcinoma.

Endocervical Adenocarcinoma (In Situ and Invasive)
Endocervical AIS is a high-grade endocervical neoplastic lesion that cytologically demonstrates nuclear enlargement, hyperchromasia, stratification, and mitotic activity. Invasive carcinoma overlaps cytologically with AIS, but may show features of invasion, including prominent nucleoli and tumor diathesis. The possibility of a coexisting squamous lesion should always be carefully assessed when a glandular lesion is detected, due to the high rate of coexistence of SIL in cases with AIS. 66 - 68

Endometrial Adenocarcinoma
The cytologic features are directly related to the histologic grade of the tumor, with well-differentiated cases yielding malignant cells with minimal atypia and poorly differentiated tumors being obviously malignant. Tumor diathesis is often difficult to appreciate, particularly in LBP. In general endometrial lesions yield fewer cells than do directly sampled endocervical lesions.

Extrauterine Adenocarcinoma
A clean background and tumors whose cytologic features are not characteristic of uterine/cervical tumors should raise the possibility of metastasis. Diathesis is usually not seen unless there is direct extension from the rectum or bladder with associated tissue destruction.

Diagnostic Difficulties
Criteria indicating invasion—tumor diathesis and macronucleoli—may be absent in the majority of well-differentiated, early adenocarcinomas. It also can be difficult to differentiate SIL with gland involvement from AIS. HSIL/CIS involving endocervical glands may yield round cell clusters with smooth peripheral contours showing group polarity and “columnar” shape of individual cells, thus mimicking a glandular abnormality. 69 Additionally, SIL and AIS may coexist in up to 50% of cases, and at conization, a high proportion of AIS specimens demonstrate concurrent SIL. 68
Benign entities such as tubal metaplasia, directly sampled lower uterine segment (LUS) endometrial cells, and cervical endometriosis may all morphologically mimic AIS. Fragments of tubal metaplasia may demonstrate crowded sheets of glandular cells with enlarged nuclei as well as cell fragments with nuclear palisading and nuclear overlap, mimicking some of the morphologic features of AIS. 70 However, rosette formation is uncommon in tubal metaplasia, and the nuclear chromatin tends to be more finely granular. The most helpful findings though, when present, are cytoplasmic terminal bars and cilia.
Directly sampled endometrial tissue may mimic AGC or glandular neoplasia. Inadvertent sampling of the LUS may occur because of closer approximation of the LUS to the cervical os following cone biopsy 71 or with aggressive use of endocervical brushes. In contrast to spontaneously exfoliated endometrial cells, which typically shed as tight ball-like clusters, direct brushing of endometrial tissue yields large cellular fragments. These fragments often recapitulate their native three-dimensional architecture with branching tubular glands enmeshed in stroma composed of round to spindle-shaped cells. 72 Glandular cells show crowding with overlapping round nuclei and scant cytoplasm. Peripheral palisading may be evident. The low power recognition of branching glands and glandular-stromal complexes is an important clue to avoid confusion with AGC or glandular neoplasia.
Conventional smears with a diagnosis of adenocarcinoma consistently identified correctly by CAP interlaboratory glass slide program participants were significantly more likely to have more abnormal cells, larger abnormal cells, larger nuclei, marked atypia, and hyperchromasia than cases that performed poorly. 73 Glandular lesions have a slightly different morphology on LBPs; specifically the cells may be flatter, feathering less prominent, and diathesis more difficult to appreciate. 10 Details are discussed elsewhere in this book. However, as for squamous lesions, there have been reports showing increased detection of glandular abnormalities on LBPs compared to conventional smears. 68, 74, 75

Management
Atypical glandular cells are estimated to be reported in only 0.2% of cervical cytology tests in the United States. 31 While AGC may be associated with benign and reactive conditions such as endocervical/endometrial polyps, it is clear from several studies that AGC is a “high-risk” interpretation compared to ASC; the reported rate of neoplasia in follow-up of AGC ranges from 9 to 38%. 67, 68, 76, 77
Due to the high risk of a significant lesion associated with a cytologic interpretation of AGC, colposcopy with endocervical sampling is recommended for women with all subcategories of AGC and AIS. In women over 35 years of age, additional endometrial sampling is recommended. Endometrial sampling is also recommended for women under the age of 35 with clinical indications suggesting that they may be at risk for neoplastic endometrial lesions, such as unexplained vaginal bleeding or conditions suggesting chronic anovulation. In women with atypical endometrial cells, both endometrial and endocervical sampling should be done initially. In 2006 ASCCP suggested that while HPV testing alone is not appropriate for initial triage of any subcategory of AGC or AIS, HPV DNA testing at the time of colposcopy is preferred in women with atypical endocervical, endometrial, or glandular cells NOS, and the results should be utilized in overall patient management. 9

Educational Notes/Suggestions
The use of educational notes/comments is optional. If these are used by the pathologist/laboratory, it is suggested that they be concise, be phrased in the form of a suggestion, not a directive, and be substantiated by published guidelines from professional organizations. 7 Examples can be found in the second edition of the Bethesda atlas. 10

Ancillary Testing
If ancillary testing, such as high-risk HPV, has been performed, whether the report is issued concurrently with the cervical cytology result or as an addendum/separate report will depend on the laboratory's information system, turnaround time for such testing, and clinical expectations. The methodology utilized for the ancillary test should be specified. Suggestions for reporting of molecular tests are provided in the Bethesda atlas. 10

Automated Review
For cervical cytology preparations that undergo computer-only or computer-assisted review, the type of instrument used and any result should be included in the report. In addition, if there was no “human” review of the slide, this should be made clear in the report.

Interobserver Reproducibility in Cervical Cytology
In an effort to improve standardization, clarity, and reproducibility of cervical cytology reporting, the second edition of the Bethesda atlas 10 emphasized more detailed morphologic criteria and had many more images, which were complimented by additional images on the Bethesda website. 11 In addition, as part of the ASC–NCI Bethesda Project, a web based interobserver reproducibility study was designed to gauge cervical cytology reproducibility prior to publication of the atlas and website. A range of classic and borderline images (77) were included for interpretation; approximately 651 cytotechnologists and pathologists worldwide participated in the study. It was apparent from the results that the morphology presented was more important in classifying images correctly than were professional or academic degrees, or other variables assessed. In this study, exact agreement with the TBS panel was relatively low (57%), although agreement was 84.1% at the threshold of distinguishing NILM from non-negative. Participants achieved a higher sensitivity for correctly classifying high-grade squamous lesions than that for high-grade glandular lesions. The details of this study have been published 12 and all the images and associated histograms of participants' responses are available for review on the Bethesda website. 11

The Bethesda System and Reporting Anal-Rectal Cytology
Anal cancer is considered an appropriate target for cytologic screening in selected high-risk populations. The anatomic commonality of the anal–rectal canal and the cervical mucosa is reflected in that both have a transformation zone. HPV is a common risk/etiologic factor for cancers of the anus and cervix and subsequently the morphology of cytology samples from both sites is comparable. It follows that sampling devices, preparation techniques, and morphologic interpretation using the Bethesda system terminology utilized for cervical cytology can readily be applied for anal–rectal cytology screening.
Adequacy criteria for anal–rectal cytology are based, at present, on limited personal experiences. As a guide, minimum adequacy cellularity should be in the range of 2000–3000 nucleated squamous cells for conventional smears and for LBP samples 1–2 nucleated cells/high-power field for ThinPrep (20 mm diameter) preparations and 3-6 nucleated squamous cells/high-power field for SurePath (13 mm diameter) preparations. 10
Normal elements seen in anal–rectal specimens include nucleated, anucleate, and metaplastic squamous cells, rectal columnar cells, fecal matter, and mucus. A comment should be included in the report about the presence of a transformation zone component. Cytomorphologic criteria are quite similar to those utilized for cervical cytologic interpretation; however, there is a higher incidence of poor preservation, cellular degeneration, and cytoplasmic keratinization/parakeratosis, and classic koilocytes are less frequently identified. 10, 78
When targeting high-risk groups, the rate of epithelial abnormalities noted is far higher than that reported for cervical cytology. 78, 79 Reports from the United States suggest that anal–rectal cytology screening is sensitive but has low specificity for predicting the grade of the lesion, with a tendency to under-represent the grade of squamous abnormality. While it has been shown that screening high-risk patients by cytology is effective, the present impediments to the success of early detection of anal cancer by this method include limited clinical expertise and means for the subsequent treatment/follow-up of these patients, and the high risk of complications associated with excisional procedures at this site.

Concluding Remarks
Cervical cytology has seen many changes since its introduction in the 1960s—liquid-based sampling techniques, automated preparation, computer-assisted screening, HPV DNA testing, and more recently dual testing, which combines cytology screening with HPV testing. HPV vaccines entered the market in 2006, and are likely to further decrease the incidence of invasive squamous cell carcinoma of the cervix and its precursor lesions. Cervical cancer screening guidelines have undergone significant changes after implementation of LBP and HPV testing and with advances in the understanding of cervical neoplasia. 80 It is predicted that the number of cervical cytology tests performed will decrease significantly in the future if there is compliance with these guidelines. 81
The Bethesda System has met the goals that were conceived at the time of its implementation in 1988. It has seen successful penetration into cervical cytology reporting worldwide, allowing laboratories to use consistent terminology in conveying results to clinicians and thus enabling comparison of studies across many countries and health care systems. The use of the Bethesda ASC-US terminology prompted the NCI-sponsored ALTS trial, the results of which have significantly impacted the management of equivocal and low-grade cervical cytologic abnormalities. The Bethesda terminology has been updated twice—in 1991 and 2001 since its inception in 1988 in order to keep pace with the advances in our understanding of cervical cancer and evolving technologies in cervical cancer screening and prevention. The Bethesda System also provided the basis for the ASCCP to develop consensus guidelines for management of cervical cytologic abnormalities as defined by TBS. This process of re-evaluation and revision will continue in the future in order to provide the most accurate, reproducible, and relevant terminology. Optimal communication and ultimately patient care outcomes will therefore ensue.

References

1. National Cancer Institute Workshop. The 1988 Bethesda System for reporting cervical/vaginal cytologic diagnoses. JAMA . 1989;262:931-934.
2. National Cancer Institute Workshop. The Bethesda System for reporting cervical/vaginal cytologic diagnoses. Acta Cytol . 1993;37:115-124.
3. Kurman R.J., Solomon D. The Bethesda System for Reporting Cervical/Vaginal Cytologic Diagnoses: Definitions, Criteria and Explanatory Notes for Terminology and Specimen Adequacy . New York: Springer-Verlag; 1994.
4. Davey D., Woodhouse S., Styer P., et al. Atypical epithelial cells and specimen adequacy: current laboratory practices of the participants in the College of American Pathologists interlaboratory comparison program in cervicovaginal cytology. Arch Pathol Lab Med . 2000;124:203-211.
5. ASCUS LSIL Triage Study (ALTS) Group. Results of a randomized trial on management of cytology interpretations of atypical squamous cells of undetermined significance. Am J Obstet Gynecol . 2003;188:1383-1392.
6. Schiffman M., Adrianza M.E. ASCUS-LSIL Triage Study. Design, methods and characteristics of trial participants. Acta Cytol . 2000;44(5):726-742.
7. Solomon D., Davey D., Kurman R., et al. for the Forum Group Members and the Bethesda 2001 Workshop. The 2001 Bethesda System—terminology for reporting results of cervical cytology. JAMA . 2002;287:2114-2119.
8. Wright T.C.Jr., Cox J.T., Massad L.S., et al. ASCCP-Sponsored Consensus Conference. 2001 Consensus Guidelines for the management of women with cervical cytological abnormalities. JAMA . 2002;287:2120-2129.
9. ASCCP consensus guidelines. Am J Obstet Gynecol . 2007;197(4):346-355.
10. Solomon D., Nayar R., editors. The Bethesda System for Reporting Cervical Cytology, 2nd edn, New York: Springer, 2004.
11. Bethesda System Web Atlas. Available at the American Society of Cytopathology website: http://www.cytopathology.org
12. Sherman M.E., DasGupta A., Schiffman M., et al. Bethesda Interobserver Reproducibility Study (BIRST). Cancer Cytopathol . 2007;111:15-25.
13. Bottles K., Reiter R.C., Steiner A.L., et al. Problems encountered with The Bethesda System: The University of Iowa experience. Obstet Gynecol . 1991;78:410-414.
14. Kline T.S., Solomon D. Guidelines for specimen adequacy: A plea for workable definitions. Diagn Cytopathol . 1991;7:1-2.
15. Solomon D., Henry M. Specimen adequacy. In: Wied G.L., Keebler C.M., Rosenthal D.L., et al, editors. Compendium on Quality Assurance, Proficiency Testing and Workload Limitations in Clinical Cytology . Chicago: Tutorials of Cytology; 1995:90-94.
16. Birdsong G. Pap smear adequacy: is our understanding satisfactory … or limited? Diagn Cytopathol . 2001;24:79-81.
17. Studeman K.D., Ioffe O.B., Puszkiewicz J., et al. Effect of cellularity on the sensitivity of detecting squamous lesions in liquid-based cervical cytology. Acta Cytol . 2003;47(4):605-610.
18. Bolick D.R., Kerr J., Stanely B.E., et al. Effect of cellularity in the detection rates of high grade and low grade squamous intraepithelial lesions. Acta Cytol . 2002;46:922-923. (abstract)
19. Bos A.B., van Ballegooijen M., van den Akker-van Marle E., et al. Endocervical status is not predictive of the incidence of cervical cancer in the years after negative smears. Am J Clin Pathol . 2001;115:851-855.
20. Elias A., Linthorst G., Bekker B., et al. The significance of endocervical cells in the diagnosis of cervical epithelial changes. Acta Cytol . 1983;27:225-229.
21. Mauney M., Eide D., Sotham J. Rates of condyloma and dysplasia in Papanicolaou smears with and without endocervical cells. Diagn Cytopathol . 1990;6:18-21.
22. Mitchell H., Medley G. Influence of endocervical status on the cytologic prediction of cervical intraepithelial neoplasia. Acta Cytol . 1992;36:875-880.
23. Mitchell H., Medley G. Longitudinal study of women with negative cervical smears according to endocervical status. Lancet . 1991;337:265-268.
24. Mitchell H.S. Longitudinal analysis of histologic high grade disease after negative cervical cytology according to endocervical status. Cancer Cytopathol . 2001;93:237-240.
25. Smith H.O., Tiffany M.F., Qualls C.R., et al. The rising incidence adenocarcinoma relative to squamous cell carcinoma of the uterine cervix in the United States: a 24-year population based study. Gynecol Oncol . 2000;108:397.
26. Zheng T., Holford T.R., Ma Z., et al. The continuing increase in adenocarcinoma of the uterine cervix: a birth control phenomenon. Int J Epidemiol . 1996;25:252-258.
27. Ransdell J.S., Davey D.D., Zaleski S., et al. Clinicopathologic correlation of the unsatisfactory Pap smear. Cancer Cytopathol . 1997;81:139-143.
28. Davey D., Austin M., Birdsong G., et al. ASCCP patient management guidelines: Pap test adequacy and quality indicators. J Low Genit Tract Dis . 2002;6:195-199.
29. Nielsen M.L., Davey D.D., Kline T.S. Specimen adequacy evaluation in gynecologic cytopathology. Diagn Cytopathol . 1993;9:394-403.
30. Davey D.D., Woodhouse S., Styer P., et al. Atypical epithelial cells and specimen adequacy. Arch Pathol Lab Med . 2000;124(2):203-211.
31. Davey D.D., Neal M.H., Wilbur D.C., et al. Bethesda 2001 Implementation and reporting rates: 2003—practices of participants in the College of American Pathologists Interlaboratory Comparison Program in cervicovaginal cytology. Arch Pathol Lab Med . 2004:;28(11):1224-1229.
32. Fidda N., Miron J., Rodgers W.H., et al. Impact of the new Bethesda system 2001 on specimen adequacy of cervicovaginal smears. Diagn Cytopathol . 2004;30:235-239.
33. Quddus M.R., Sung C.J., Eklund C.M., et al. ASC:SIL ratio following implementation of 2001 Bethesda system. Diagn Cytopathol . 2004;30(4):240-242.
34. Prandi S., Beccati D., Aloysio G.D., et al. Applicability of the Bethesda system 2001 to a public health setting. Cancer Cytopathol . 2006;108(5):271-276.
35. Colgan T.J., Woodhouse S.L., Styer P.E., et al. Reparative changes and the false-positive/false-negative Papanicolaou test. Arch Pathol Lab Med . 2001;125(1):134-140.
36. Snyder T.M., Renshaw A.A., Styer P.E., et alfor the Cytopathology Resource Committee. Altered recognition of reparative changes in ThinPrep specimens in the College of American Pathologists Gynecologic Cytology Program. Arch Pathol Lab Med . 2005;129(7):861-865.
37. Ponder T.B., Easley K.O., Davila R.M. Glandular cells in vaginal smears from posthysterectomy patients. Acta Cytol . 1997;41:1701-1704.
38. Bwetra C. Columnar cells in posthysterectomy vaginal smears. Diagn Cytopathol . 1992;8(4):342-345.
39. Greenspan D.L., Cardillo M., Davey D.D., et al. Endometrial cells in cervical cytology: review of cytologic features and clinical assessment. J Low Gen Dis . 2006;10(2):111-122.
40. Browne T.J., Genest D.R., Cibas E.S. The clinical significance of benign-appearing endometrial cells on a Papanicolaou test in women 40 years or older. Am J Clin Pathol . 2005;124(6):834-837.
41. Interlaboratory Comparison Program in Cervicovaginal Cytology. 1993 PAP supplemental questionnaire on Laboratory Practice: ASCUS. College of American Pathologists, 1994.
42. Ho G.Y.F., Bierman R., Beardsle L., et al. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med . 1998;338:423-428.
43. Raffle A.E., Alden B., Mackenzie E.F.D. Detection rates for abnormal cervical smears: What are we screening for? Lancet . 1995;345:1469-1473.
44. Sidawy M.K., Tabbara S.O. Reactive change and atypical squamous cells of undetermined significance in Papanicolaou smears: A cytohistologic correlation. Diagn Cytopathol . 1993;9:423-429.
45. Sherman M.E., Schiffman M.H., Lorincz A.T., et al. Toward objective quality assurance in cervical cytopathology. Am J Clin Pathol . 1994;102:182-187.
46. Stoler M.H., Schiffman M. Atypical Squamous Cells of Undetermined Significance-Low-grade Squamous Intraepithelial Lesion Triage Study (ALTS) Group. Interobserver reproducibility of cervical cytologic and histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study. JAMA . 2001;285(11):1500-1505.
47. The ALTS Group. Results of a randomized trial on the management of cytology interpretations of atypical squamous cells of undetermined significance. Am J Obstet Gynecol . 2003;188:1383-1392.
48. Stoler M.H. New Bethesda terminology and evidence based management guidelines for cervical cytology findings. JAMA . 2002;287(16):2140-2141.
49. Ko V., Nanji S., Tambouret R., et al. Testing for HPV as an objective measure for quality assurance in cervical cytology. Cancer Cytopath . 2007;111:67-73.
50. Jones B.A., Novis D.A. Follow-up of abnormal gynecologic cytology: a College of American Pathologists Q-probes study of 16132 cases from 306 laboratories. Arch Pathol Lab Med . 2000;124:665-671.
51. Sherman M.E., Solomon D., Schiffman M. Qualification of ASCUS: a comparison of equivocal LSIL and Equivocal HSIL cervical cytology in the ASCUS LSIL Triage Study. Am J Clin Pathol . 2001;116:386-394.
52. Sordon M., Dilworth H.P., Ronnett B.M. Atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion. Diagnostic performance, human papillomavirus testing, and follow-up results. Cancer Cyopathol . 2006;108:32-38.
53. Safaeian M., Solomon D., Wacholder S., et al. Risk of precancer and follow-up management strategies for women with human papillomavirus-negative atypical squamous cells of undetermined significance. Obstet Gynecol . 2007;109:1325-1331.
54. Cox J.T., Solomon D., Schiffman M. Prospective follow up suggests similar risk of subsequent CIN 2 or 3 among women with CIN 1 or negative colposcopy and directed biopsy. Am J Obstet Gynecol . 2003;188:1406-1412.
55. Nubia Muñoz N., Bosch F.X., Sanjosé S.D., et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med . 2003;348(6):518-527.
56. Lorincz A.T., Reid R., Jenson A.B., et al. Human papillomavirus infection of the cervix: Relative risk associations of 15 common anogenital types. Obstet Gynecol . 1992;79:328-337.
57. Wright T.C., Schiffman M. Adding a test for human papillomavirus DNA to cervical cancer screening. N Engl J Med . 2003;348(6):489-490.
58. Khan M.J., Castle P.E., Lorincz A.T., et al. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst . 2005;97:1072-1079.
59. Sherry L., Woodhouse S.L., Stastny J.F., et al. Interobserver variability in subclassification of squamous intraepithelial lesions: Results of the College of American Pathologists Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med . 1999;123(11):1079-1084.
60. Renshaw A.A., Henry M.R., Birdsong G.G., et al. Cytologic features of squamous cell carcinoma in conventional smears: Comparison of cases that performed poorly with those that performed well in the College of American Pathologists Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med . 2004;129(9):1097-1099.
61. Arbyn M., Sasieni P., Meijer C.J., et al. Clinical applications of HPV testing: A summary of meta-analyses. Vaccine . 2006;24(Suppl 3):S78-S89.
62. Guido R., Solomon D., Schiffman M., Burke L. Comparison of management strategies for women diagnosed as CIN 1 or less, postcolposcopic evaluation: Data from the ASCUS and LSIL Triage Study (ALTS), a Multicenter Randomized Trial. J Low Genit Tract Dis . 2002;6:176.
63. Jones B.A., Davey D.D. Quality management in gynecologic cytology using interlaboratory comparison. Arch Pathol Lab Med . 2000;124:672-681.
64. Evans M.F., Adamson C.S., Papillo J.L., et al. Distribution of human papillomavirus types in ThinPrep Papanicolaou tests classified according to the Bethesda 2001 terminology and correlations with patient age and biopsy outcomes. Cancer . 2006;106(5):1054-1064.
65. Renshaw A.A., Mody D.R., Lozano L.R., et al. Detection of adenocarcinoma in situ of the cervix in Papanicolaou tests: comparison of diagnostic accuracy with other high-grade lesions. Arch Pathol Lab Med . 2004;128:53-157.
66. Eddy G.L., Strumpf K.B., Wojtowycz M.A., et al. Biopsy findings in five hundred thirty-one patients with atypical glandular cells of uncertain significance as defined by the Bethesda System. Am J Obstet Gynecol . 1997;177:1188-1195.
67. Burja I.T., Thompson S.K., Sawyer W.L., et al. Atypical glandular cells of undetermined significance on cervical smears. A study with cytohistologic correlation. Acta Cytol . 1999;43:351-356.
68. Diaz-Montes T.P., Farinola M.A., Zahurak M.L., et al. Clinical utility of atypical glandular cells (AGC) classification: cytohistologic comparison and relationship to HPV results. Gynecol Oncol . 2007;14(2):366-371.
69. Selvaggi S.M. Cytologic features of squamous cell carcinoma in situ involving endocervical glands in endocervical cytobrush specimens. Acta Cytol . 1994;38:687-692.
70. Novotny D.B., Maygarden S.J., Johnson D.E., et al. Tubal metaplasia. A frequent potential pitfall in the cytologic diagnosis of endocervical glandular dysplasia. Acta Cytol . 1992;36:1-10.
71. Lee K. Atypical glandular cells in cervical smears from women who have undergone cone biopsy. Acta Cytol . 1993;37:705-709.
72. de Peralta-Venturino M.N., Purslow M.J., et al. Endometrial cells of the “lower uterine segment” (LUS) in cervical smears obtained by endocervical brushings: A source of potential diagnostic pitfall. Diagn Cytopathol . 1995;12:263-271.
73. Renshaw A.A., Schwartz M.R., Wang E., et al. Cytologic features of adenocarcinoma, not otherwise specified, in conventional smears: comparison of cases that performed poorly with those that performed well in the College of American Pathologists Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med . 2006;130(1):23-26.
74. Austin R.M., Ramzy I. Increased detection of epithelial cell abnormalities by liquid-based gynecologic cytology preparations. A review of accumulated data. Acta Cytol . 1998;42(1):178-184.
75. Ashfaq R., Gibbons D., Vela C., et al. ThinPrep Pap test. Accuracy for glandular disease. Acta Cytol . 1999;43(1):81-85.
76. DeSimone C.P., Day M.E., Tovar M.M., et al. Rate of pathology from glandular cells classified by the Bethesda system 2001 nomenclature. Obstet Gynecol . 2006;107:1285-1291.
77. Chhieng D.C., Gallaspy S., Yang H., et al. Women with atypical glandular cells: a long term follow-up study in a high risk population. Am J Clin Pathol . 2004;122:575-579.
78. Arian S., Walts A.E., Thomas P., et al. The anal Pap smear: cytomorphology of squamous intraepithelial lesions. Cytojournal . 2006:2-4.
79. Nayar R., Contreras N., Luzadeer R., et al. Screening for anal cancer: cytologic findings and follow-up. Cancer Cytopathol . 2006;108:423. (abstract)
80. Saslow D., Runowicz C.D., Solomon D., et al. American Cancer Guideline for the early detection of cervical neoplasia and cancer. CA Cancer J Clin . 2002;52:342.
81. Solomon D., Breen N., McNeel T. Cervical cancer screening rates in the United States and the potential impact of implementation of screening guidelines. CA Cancer J Clin . 2007;57:105-111.
CHAPTER 7 Microbiology, Inflammation, and Viral infections

Prabodh K. Gupta, Cindy McGrath

Contents
Introduction
Vaginal Microbiology
General Features
Background Changes
Cellular Changes
Infections of the Female Genital Tract
Bacterial Infections
Viral Infections
Chlamydial Infection
Fungal Infections
Parasitic Infections
Concluding Remarks

Introduction
The lower female genital tract includes the vulva, the vagina, the cervix, and the uterine cavity. It is in direct communication with the external environment, and prone to various noninfectious and infectious inflammatory reactions. Although most of these infections remain confined locally; they can progress and the organisms may ascend to the fallopian tubes and the ovaries. Occasionally, microbial infections from the lower genital tract can disseminate via the peritoneal space or the hematogenous or lymphatic routes.
It must be appreciated that while a large number of women harboring genital infections may remain asymptomatic, vaginal infection can produce a number of clinical symptoms. Increased vaginal secretions, along with sloughed vaginal epithelial cells, other infective organisms, bacteria, and inflammatory cells, constitute acute vulvovaginitis. Inappropriate use of over-the-counter medications, personal hygiene, tight clothing, impermeable panty hose (panty hose vulvitis), reaction to various laundry detergents, and washed clothes may contribute to the symptoms of vulvovaginitis. 1 Often personal and social reasons delay medical intervention.
Key features of vulvovaginitis
• Increased vaginal secretions;
• Sloughed vaginal epithelial cells;
• Bacteria;
• Infectious organisms; and
• Inflammatory cells.
Vaginitis accounts for nearly 6 million visits to the healthcare providers per year with an annual cost of over a billion dollars to society. 1 Vulvovaginal irritation, itching, pain, ulceration with bleeding, dyspareunia, and warty growths are some of the common presenting features of vaginal infections. Most often, a woman may remain minimally symptomatic and may not seek medical help for various personal and social reasons.

Vaginal Microbiology
Among healthy women, the vaginal milieu is polymicrobial and contains a large number and variety of aerobic as well as obligate and facultative anaerobic organisms. 2 The most frequently recovered bacteria include lactobacilli, Streptococcus viridans , and Staphylococcus epidermidis ; none of which cause symptoms. Bacteroides and Gardnerella vaginalis may be culturable from 20 and 30%, and 50% of the asymptomatic women, respectively. Staphylococcus occurs infrequently in the healthy vaginal flora. Table 7.1 lists the microorganisms that can be commonly recovered from vaginal specimens.
Table 7.1 Common microbial organisms in the vaginal flora Lactobacilli Bacteroides species Diphtheroids Peptococcus species Staphylococcus species Peptostreptococcus species Streptococcus species Fusobacterium species Enterobacter (not group A) Clostridium species Gardnerella vaginalis Bifidobacterium
Pregnancy, besides causing a growth of lactobacillary flora, does not appear to affect the microbial composition of the vagina significantly. Estrogenic hormones and similar substances help in the epithelial maturation of the vagina and support the growth of an extraordinary number of microbes. Transplacental hormonal exchanges also influence the vaginal epithelium of the newborn infant. Bacterial composition and an adult type of microenvironment may occur in a newborn female infant.
Menarche and menopausal changes also affect the bacterial makeup of the vagina. Hormone or hormone-like medications, contraceptives, intrauterine contraceptive devices (IUDs), barrier diaphragms, pessaries, and other similar substances and contraception may directly or indirectly influence the microbial balance of the lower genital tract. 3 Common factors influencing the vaginal microbial flora are presented in Table 7.2 . It must be appreciated that the vaginal flora is in a dynamic state physiologically and in health. It contains a large number of organisms that, under poorly understood conditions, may become pathogenic and cause disease.
Table 7.2 Common factors influencing vaginal microbial flora Physiologic Diseases and drugs Local factors Parturition Hepatic disorders Infections Pregnancy Hormonal imbalance IUD Menstruation Metabolic diseases Pessary Menopause Erosion and infections Diaphragm   Oral contraceptives Vaginal douche   Hormonal mimic drugs Surgery   Antibodies Trauma     Abortion     Sexual exposure

General Features
A number of general cytological features represent the various effects of infective processes. These include the changes listed in Table 7.3 . Specific cytological changes frequently are associated with certain infections and are described in their respective areas. Only some of these responses may occur under specific inflammatory conditions.
Table 7.3 Cytologic features of vaginopancervical smears in infective processes General Cellular degenerative changes Cellular reactive Background changes Nuclear Hyperplasia and repair Acute and chronic inflammation and cellular obscuring Cytoplasmic Degenerative Fresh and old blood   Metaplasia Cytolysis   Parakeratosis Cell distribution changes   Hyperkeratosis     Pseudoparakeratosis     Multinucleation     Histiocytic proliferation     Dysplasia

Background Changes

Inflammation and Cellular Obscuring
Overgrowth of microbes in the vaginal milieu may result in obscuring of morphologic details in the smear ( Fig. 7.1 ). In such smears, numerous polymorphonuclear leukocytes occur, often interspersed with a large number of histiocytes. Excessive bacterial growth may also contribute to cellular obscuring. It must be realized that no meaningful evaluation of cellular change may be possible on such smears. In all such cases, it is almost mandatory that appropriate therapy be initiated and a repeat smear examined before an opinion is rendered. Atrophic epithelium of the vagina is particularly prone to inflammatory changes ( Fig. 7.2 ) that may mimic atypia. In such cases local hormonal application generally helps in proper interpretation of cells. In specimens with overwhelming inflammatory exudates, it may not be possible to differentiate vaginitis from cervicitis and endocervicitis. Inflammatory exudates and vaginal microbial flora is considerably altered in the liquid-based gynecological slides (LBGS). 4
Postmenopausal atrophy and atypia:
• Important to treat with local estrogens;
• Repeat cytologic examination after 6 weeks; and
• Immediate colposcopy not recommended.

Fig. 7.1 Heavy, acute inflammatory exudate. (A) Vaginopancervical smear (Papanicolaou × LP). (B) Note the reduced background inflammation in LBGS (Papanicolaou × LP)

Fig. 7.2 Atrophic smear with inflammatory background and some “atypical” cells (arrows). (A) Vaginopancervical smear. (B) Same patient after topical estrogen application. Note the cellular maturation and obvious “atypical” cells (arrow). LBGS (Papanicolaou × MP).
Cellular specimens from specific areas, e.g. the vaginal, cervical, and endocervical smear, 5 or vulvar or lateral vaginal wall scrapings, may reveal inflammatory response in one or more preparations that can help localize the infective process within the lower genital tract. A specific cervical inflammation with predominant lymphohistiocytic reaction may occur in follicular cervicitis (discussed later). Granulomatous reaction may occur in the presence of foreign bodies (e.g. suture material, surgical clips, IUD) or specific infections such as tuberculosis.

Bleeding
In infectious processes, both fresh and old blood may be observed. Postmenopausal women with atrophic, thin vaginal mucosa may bleed more easily. Similar changes may occur in Trichomonas vaginalis infection, which produces the typical “strawberry” cervical lesion. Whereas the fresh bleeding is recognizable without much difficulty, old bleeding, observed as fibrin, should be differentiated from mucus. In direct smear preparations, fibrin threads are uniformly thick and reveal nodal formations at the points of intersections of interlacing threads. Sometimes, hemosiderin pigment may be observed within the macrophages and extracellularly. Sometimes, old hemorrhage may contain hematoidin crystals that appear as “cockleburs,” as described by Hollander and Gupta 6 ( Fig. 7.3 ).

Fig. 7.3 Hematoidin crystals , also known as cockleburs, are seen associated with macrophage response. (A) Vaginopancervical smear (Papanicolaou × MP). (B) Hematoidin crystals LBGS. Note the clean background and lack of obvious macrophages (Papanicolaou × MP).
Cytolysis, the process of cellular degeneration due to bacterial overgrowth, commonly affects the intermediate squamous epithelial cells. The process is believed to be glycogen dependent. Late menstrual cycle and pregnancy as well as hormonal contraceptives often cause lactobacilli overgrowth. Pale staining, vesicular nuclei with little or no cytoplasm of the intermediate cells predominate in such smears. Numerous lactobacilli may occur interspersed with the remaining cellular remnants ( Fig. 7.4 ).

Fig. 7.4 Lactobacilli . Vaginopancervical smear (Papanicolaou × HP).

Cellular Changes
In the cervicovaginal smears, intermediate and superficial cells occur predominantly among healthy and normally menstruating women. Heavy inflammation frequently causes an exfoliation of parabasal cells in the vaginal smear. Although often present in postmenopausal women and in the immediate postpartum period, occurrence of these three types of cells in the premenopausal age group should be carefully evaluated. Caution need to be exercised in rendering hormonal evaluation in smears with excessive inflammation. The parabasal cells may exfoliate from ulceration of the squamous epithelium of the vagina and the ectocervix.
Under the persistent effect of various microbial infections and inflammatory reactions, both squamous and columnar epithelial cells may undergo degenerative changes. Almost all these changes are nonspecific, but their identification helps in the proper interpretation of more serious cellular alterations. This is critical because most degenerative changes may be accompanied by concurrent regenerative, reactive, and metaplastic changes. An extremely heterogenous group of cellular changes that includes reactive, degenerative, metaplastic, and neoplastic features is termed atypical squamous cells (ASC). Morphological features of these cells overlap and generally are nonspecific, meaning that they cannot be precisely separated between neoplastic and non-neoplastic changes. Additional studies are often necessary for further characterization of these changes (discussed elsewhere).

Degeneration
In inflammatory states cytoplasm of the squamous and columnar cells may be completely or partially disintegrated. However, the major changes are observed within the nuclei. These have been detailed by Frost. 7 Briefly; the nuclei may become compact, dense, and pyknotic with loss of all chromatin details ( Fig. 7.5 ). Such nuclei may have a distinct circumferential cytoplasmic clearing or hollow, causing a perinuclear “halo” ( Fig. 7.6 ), often seen in association with Trichomonas and other infections. Nuclei undergoing degenerative changes frequently lose the sharp details of their nuclear envelope, the chromatin, and the interphase. The nuclear chromatin may clump irregularly or appear beaded along the nuclear margins. They may become blurry and opaque. Other changes include nuclear swelling, with partial or total disintegration of the nuclear envelope, karyorrhexis, and karyopyknosis ( Fig. 7.7 ).

Fig. 7.5 Nuclear degeneration. Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.6 Perinuclear clearing or halos. Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.7 Nuclear degeneration among endocervical cells infected with adenovirus infection. Vaginopancervical smear (Papanicolaou × HP).

Degenerative, Regenerative (Repair), and Metaplastic Changes in Inflammation
The epithelial cells of the lower genital tract (ectocervix, endocervix, and transformation zone), under the influence of persistent irritation (infectious and non-infectious) and repair, undergo morphologic changes commonly referred to as metaplasia. These essentially reflect a benign process of tissue repair. An overestimation of reparative changes is a most common error in interpretive cytology. Geirsson and co-workers 8 referred to these as atypical reparative changes (ARC). The majority of these changes are believed to be derived from the columnar and squamous epithelia, but reserve or pluripotential cells may be involved in the genesis of the metaplasia. It must be appreciated that there is a continuum of changes observed in the healing phase. Cytomorphologic changes, for convenience, are grouped under the term metaplasia. Some of these morphologic changes can be indistinguishable from “early” intraepithelial dysplastic alterations, observed within both the squamous and the glandular epithelia. These cellular features are grouped as “atypical squamous or glandular cells of undetermined significance” (ASC, AGC) (discussed elsewhere).
The earliest discernible changes are referred to as presquamous metaplasia: columnar differentiation phase, or type I ARC of the cervical columnar epithelium. Under the effect of chronic irritation and repair processes, the surface cells of the columnar epithelium continue to mature, and there is a proliferation of the basal or reserve cells. These small, undifferentiated cells commonly occur in small tissue fragments. They have high nucleus-to-cytoplasmic (N/C) ratios and prominent nucleoli ( Fig. 7.8 ). The nuclear chromatin is fine and uniformly distributed, and the nuclear membrane is well delineated and thin. These may have an inflammatory background. The cells of subluminal origin (progenitor cells) can often be mistaken for undifferentiated neoplasia.

Fig. 7.8 Squamous metaplasia (atypical reactive cells, ARC I). LBGS (Papanicolaou × MP).
As the changes evolve, the subluminal cells differentiate from the germinal layers upward. These changes reflect the immature squamous metaplastic epithelium and have been referred to as ARC II. The cells may appear to be columnar and have excessive goblet cell proliferation and mucus production. Signet-ring forms may be recognized. These cells can have numerous macronucleoli, coarse chromatin, and modest but pale cytoplasm. If not carefully examined, the changes can be mistaken for a neoplastic lesion of the endocervix ( Fig. 7.9 ). Presence of heavy acute inflammation generally helps in the correct interpretation of these changes. In the LBGS, inflammation is less obvious and these cells often occur as small tissue fragments and may be reported as AGC.

Fig. 7.9 Squamous metaplastic changes with columnar cell hyperplasia (ARC II). LBGS (Papanicolaou × MP).
As the changes progress from the subluminal, via the presquamous, to the keratinizing stratified squamous phase , the cells become oval or polygonal with sharp borders and dense cytoplasm. They lie in sheets and reveal no obvious cilia and mucus. Intercellular bridges may be seen at times. These cells with their metaplastic changes have been called ARC type III. The nuclear changes may be reactive or degenerative with pyknosis. These cells may be mistaken for squamous cell carcinoma ( Fig. 7.10 ).

Fig. 7.10 Metaplastic changes revealing keratinizing stratified squamous metaplasia (ARC III). LBGS (Papanicolaou × MP).

Squamous Epithelium
When stressed, as by a chronic irritation or an infective injury, squamous epithelium responds in a number of ways. These changes essentially represent alteration of functional differentiation of the affected cells and are mostly cytoplasmic in nature. Proper identification of these cytoplasmic features is necessary, as they may mask a more serious underlying disease process. These changes—hyperkeratosis, parakeratosis, basal cell hyperplasia, pseudoparakeratosis, and dyskeratosis—have been discussed elsewhere in this book. They reflect abnormalities of maturation with normal keratin formation in cells that normally do not reveal these changes.
Squamous epithelial cell changes
• Hyperkeratosis;
• Parakeratosis;
• Basal cell hyperplasia;
• Pseudoparakeratosis; and
• Dyskeratosis.
Dyskeratosis is mentioned here because of its relationship with viral infections and developing cancer. This represents an abnormality of the squamous cells in which the cytoplasmic maturation is altered. The affected cells reveal premature, hypermature, or atypical keratinization. It is a common occurrence in the presence of chronic infections, such as those caused by human papillomavirus (HPV). The cytomorphologic features are further detailed in the appropriate sections.

Endocervical Columnar Epithelium
In addition to the squamous metaplasia discussed earlier, endocervical cells may undergo other morphologic changes including columnar cell hyperplasia and hyperplastic polyp formation.

Columnar Cell Hyperplasia
Endocervical cells frequently enlarge and produce excessive mucus. Such changes occur with chronic irritation of the endocervical canal, such as among women using IUDs 8 or hormonal contraceptives, and who are exposed to certain infections; these are discussed separately.

Hyperplastic Polyp
Hyperplastic endocervical columnar cells may proliferate to produce finger-like epithelial processes—polyps. As described by Ramzy 9 and Frost, 10 these polyps are three-dimensional structures with three distinct planes—a floor or base composed of a sheet of polygonal cells, a middle plane that makes up the sides of the polyp, and a top or surface layer that, like the base, is also a sheet of polygonal cells. In the center of the polyp, a connective tissue core that contains fibroblasts, collagen, and capillary vessels, may be recognizable.

Tubal Metaplasia
Generally, endocervical epithelium may contain ciliated columnar cells only in small numbers (5–10%). Under conditions of chronic irritation, sheets of ciliated cells representing tubal metaplasia may occur in the specimens collected from the transformation zone and the adjacent endocervix. These cells may show pseudostratification and atypia and can be a common source of misinterpretation as neoplastic cells. 11 Tubal metaplastic cells may stain positively with p16 INK 4A antibodies, 12 which may lead to confusion with neoplastic processes.
Columnar epithelial cell changes
• Hyperplasia
• Polyp formation
• Squamous metaplasia
• Tubal metaplasia

Endometrium
Heavy acute inflammation, with pronounced reactive, degenerative, and metaplastic changes may be observed in these cells during the later half of menstrual bleeding, following instrumentation, in the postpartum period, and in association with the usage of an IUD. Retained gestational products and foreign bodies may result in extensive squamous metaplasia, multinucleated giant cell reaction, and calcification. Some of these changes are discussed later in this chapter.

Infections of the Female Genital Tract

Bacterial Infections
Bacteria most commonly infect the female genital tract. Bibbo and Wied 13 reported nonspecific organisms including mixed bacteria and coccobacilli in nearly 20% of patients. Among children these infections occur commonly and may be hormonally dependent. Vaginal or vaginopancervical smears often reveal a number of bacilli and coccid organisms (mixed infections) as detailed by Wied and Bibbo. 14 These organisms, although diffusely scattered, may occur in clumps and as microcolonies. Appropriate microbiologic isolation techniques are necessary for specific species identification but is generally not considered necessary for clinical management of the disease.
Bacterial vaginitis (nonspecific vaginitis) was first described by Gardner and Dukes. 15 They stated, “Any woman whose ovarian activity is normal and who has a gray, homogenous, malodorous vaginal discharge with a pH of 5.0 to 5.5 that yields no Trichomonads is likely to have Haemophilus vaginalis vaginitis.” It is also known as nonspecific vaginosis/vaginitis so named by Blackwell and Barlow, 16 or bacterial vaginosis (BV), a term used by the International Agency for Research on Cancer. 17 This is the most common cause for the clinical entity of bacterial vaginosis. Instead of making a specific diagnosis of G. vaginalis infection, reporting of a “shift in the bacterial flora” is the current term used to describe the organism variously named as Haemophilus vaginalis and Corynebacterium vaginalis . Gardner and Dukes 18 first described these organisms. Regarding the etiology of BV, the statement by Fredricks and Marrazzo that “BV probably results from infection with complex communities of bacteria that consist of metabolically interdependent (syntrophic) species” appears true. 19
Morphologically, the organisms are Gram-negative or Gram-variable, are 0.1–0.8 nm in diameter, and appear bacillary or coccobacillary. The microbe, although it shares many characteristics with Corynebacterium , is catalase-negative and is now classified separately. Petersdorf and colleagues 20 and Ledger and associates 21 found that as many as 40–50% of women may have vaginal infection with G. vaginalis and be asymptomatic. Among symptomatic women, leucorrhoea and pruritus with inflamed vaginal mucosa and occasional punctate hemorrhages are commonly observed. Increased growth and concentration of these organisms may not denote pathogenicity. 17 It is believed that patients with pure G. vaginalis infection are asymptomatic when the vaginal pH is less than 4.5. Secondary organisms interplay with G. vaginalis and alter this synergistic relationship. A raised pH over 4.5 (5.0 to 6.5) and an interaction with various bacteroides and peptococci may produce clinical disease. Recently, considerable interest has been exhibited in the study of BV. Molecular identification of the associated bacteria has revealed the presence of three bacteria in the Clostridiales order. These are named as bacterial vaginosis-associated bacterium (BVAB1, BVAB2, and BVAB3). BVAB are considered highly specific for BV infection. 22
Patients with a high pH of the vagina have a vaginal discharge with a distinct fishy odor. When the pH is further raised by potassium hydroxide (KOH), this odor is manifest in the “whiff test.” 23 Such preparations of vaginal reactions and KOH, when examined microscopically, have the diagnostic “clue cells” ( Fig. 7.11 ). These refer to normal polygonal squamous cells having thin, transparent cytoplasm covered by tiny coccobacillary forms of G. vaginalis . Edges of the “infected” cells reveal the BV changes. The cell borders may be indistinct and on a different plane of focus. Similar clue cells are observed in the fixed and stained Papanicolaou preparations ( Fig. 7.12 ). A variable amount of acute inflammation may be present in the background. Mere complete or partial covering of the squamous epithelial cells by the organisms ( Fig. 7.13 ) or their sticking to the cellular margins ( Fig. 7.14 ) per se should not be considered diagnostic for G. vaginalis . To be diagnostic, clue cells should have bacterial organisms not only covering the surfaces of the affected cells but also spreading beyond the margins of the squamous cells. In LBGS preparations, organisms appear in a clean background ( Fig. 7.15 ). Detection of BV is reported to be considerably less in LBGS-based preparations than in conventional slides. 4 This may not be an entirely true observation. A high degree of diagnostic accuracy exists in cytologic detection of clue cells and culture confirmation for G. vaginalis . Schnadig and co-workers 24 cultured G. vaginalis in nearly 90% of the cases that contained clue cells. This infection is believed to be sexually transmissible, and an accurate diagnosis is necessary.

Fig. 7.11 “Clue cells” phase contrast. Vaginopancervical smear (unstained × MP).

Fig. 7.12 Gardnerella vaginalis (BV) infection. Vaginopancervical smear (Papanicolaou × LP).

Fig. 7.13 Partial obliteration of the squamous cell by the coccobacillary organisms. Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.14 Organisms sticking to squamous cell. Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.15 “Clue cells” in LBGS (Papanicolaou × LP).

Micrococcus Vaginitis (Toxic Shock)
This entity is now rarely observed in current practice. This group of microbes includes a large number of Gram-positive coccoid organisms commonly observed in female genital tract smears, and Gram-negative diplococci. Staphylococcus aureus may be recovered from the vagina in about 5% of normal women. These organisms frequently cause vaginitis and vaginal discharge and may produce toxic shock syndrome. This association was documented by Shands and co-workers in 1980. 25 These organisms characteristically occur singly and can be seen within the polymorphonuclear leukocytes or other infected epithelial cells. In vaginal smears, occasionally fragments of tampon fibers may be observed ( Fig. 7.16 ). However, the finding of coccoid organisms or tampon fibers in the vaginal smear does not have any correlation with the clinical occurrence of toxic shock syndrome.

Fig. 7.16 Tampon fibers. Occasionally, these may have a core center that may contain red blood cells. Vaginopancervical smear (Papanicolaou × LP).

Lactobacillus Vaginitis (Cytolytic Vaginosis)
Lactobacilli are a heterogeneous group of organisms normally present in the vaginal flora. They occur in abundance in the late luteal phase and in pregnancy, prefer an acid environment, and are common among women using hormonal preparations (contraceptives and replacements) and in the premenarchal and menopausal age groups. They are Gram-positive, immobile, non-spore-forming anaerobes or facultative anaerobes. Certain species may be aerobic in their growth characteristics. In the presence of lactobacilli, glycogen-rich intermediate cells are often lysed. Smears in such cases show cellular crowding, cytolysis with cytoplasmic debris, and numerous bare nuclei occurring in a predominantly bacillary background. False clue cells can be reported in these cases as the lactobacilli adhere to the edges of squamous cells. Lactobacilli may be observed in up to 50% of healthy women depending on the day of the menstrual cycle. In the symptomatic population, the observed figure may be lower, about 20%. It is debatable whether pure lactobacilli (an unlikely occurrence) produce vaginitis, although vaginal discharge and leucorrhoea may occur as a result of excessive cytolysis.

Gonococcus Vaginitis
These Gram-negative diplococci cause abundant, purulent vaginal exudates. The infection affects the urethra and the perivaginal glands. On the surface of squamous cells, these organisms occur as bean-shaped diplococci. The gonococci are better observed in the air-dried areas of the smears, such as the edges of the smear. This is an uncommon occurrence in properly prepared LBGS. Within the air-dried distended polymorphonuclear leukocytes, diplococci may be present in large numbers ( Fig. 7.17 ). Gonococcus vaginitis is a venereal infection with important social and medical implications. Although it is detectable cytologically, we do not advise rendition of such a diagnosis on cytologic examination of Papanicolaou stained smears alone; they may be indistinguishable from other cocci organisms, phagocytosed debris, or Chlamydia organisms.

Fig. 7.17 Gonococcal organisms. These reveal diplococci within the polymorphonuclear leukocytes, and on the surface. Vaginopancervical smear (Papanicolaou × OI).

Curved Anaerobic Bacterial Vaginitis
These motile, anaerobic, rod-shaped organisms resemble Wolinella and have been recognized as a cause of nonspecific vaginitis by Hjelm and colleagues. 2, 26, 27 In the Papanicolaou stained smears, these bacteria cannot be easily diagnosed but are better detected in wet mount preparations. Clinically, the presentation is of nonspecific vaginitis.

Vaginal Lactobacillosis
A recently recognized clinical picture has been reported among women who have used antifungal local medications for genital Candida infection for a prolonged period of time, generally more than 20 months. These result in the proliferation of giant lactobacilli accompanying often the yeast forms of Candida organisms. A correct morphologic recognition of this condition is important for specific treatment with appropriate antibiotics 28, 29 ( Fig. 7.18 ).

Fig. 7.18 Vaginal lactobacillosis. This picture reveals numerous “giant” lactobacilli along with some yeast forms of Candida (Papanicolaou × HP).

Foreign-Body Vaginitis
A forgotten tampon is the most common cause of this type of vaginitis, in which there is a secondary overgrowth of anaerobic organisms. The tampons may irritate and ulcerate the vaginal wall and ectocervix. Occasionally, fragments of tampons can be observed in vaginal smears. Their presence is not diagnostic of vaginitis. Heavy acute inflammation, mucus, and foreign-body giant cells may be observed.

Allergic and Acute Vaginitis
Numerous eosinophils may occur in the cervical samples obtained from women with vaginal discharge. Generally, the causes are noninfectious and associated with an allergic reaction to vaginal douche, contraceptives, or various items of clothing 30 ( Fig. 7.19 ).

Fig. 7.19 Allergic vaginitis. Note the numerous eosinophils in this preparation. Cervicovaginal smear (Papanicolaou × MP).

Desquamative Inflammatory Vaginitis (DIV)
This clinical condition is noninfectious in nature and may result from a number of blister-forming disorders including pemphigus vulgaris, lichen planus, and pemphigoid. 31
Pemphigus vulgaris may exfoliate parabasal size cells that have extremely prominent single or multiple nucleoli, pale chromatin, and features of reactive cells. Mitosis may be observed. Nuclear and cytoplasmic changes can simulate squamous cell carcinoma or atypical endocervical or metaplastic cells ( Figs. 7.20 , 7.21 ). 32

Fig. 7.20 Pemphigus vulgaris. In (B) cells show metaplastic changes. These features can be confused with neoplastic as well as viral changes. Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.21 Pemphigus vulgaris, cervix. Tissue biopsy reveals cellular changes similar to those seen in the smear in Fig. 7.20 . Cervical biopsy (H&E × MP).

Granuloma Inguinale
Gram-negative, encapsulated coccobacillary organisms called Calymmatobacterium granulomatis cause this venereally transmitted infection. The infection produces large, ulcerated lesions that histologically reveal inflammatory granulation tissue and numerous macrophages. These macrophages are easily identifiable in ethanol-fixed Papanicolaou stained smears. They are plump and swollen and have a lobulated cytoplasm. Within the cytoplasm, a large number of coccobacillary (1–2 μm) structures (Donovan bodies) are seen. These are safety-pin shaped with terminal or polar thickening of the cell walls ( Fig. 7.22 ). The organisms stain faintly with hematoxylin and eosin (H + E) dyes. They can be stained with Romanowsky or silver stains. Varying degrees of acute inflammation are commonly observed in the smears. The infection is more common in the tropics, and the incidence is high in India and New Guinea. Although reported, an association of granuloma inguinale and squamous cell carcinoma is controversial. It is true that the infection may cause extremely bizarre pseudoepitheliomatous hyperplasia of the squamous epithelium that can mimic neoplasm.

Fig. 7.22 Donovan bodies. A single macrophage reveals the cytoplasmic lobules containing numerous safety-pin-shaped bacillary structures. Vaginopancervical smear (Papanicolaou × OI).

Tuberculosis (Granulomatous Cervicitis)
This is a disease of the tropics and is usually secondary to extragenital, most often pulmonary, tubercular infection. Involvement is more common in the fallopian tubes and the endometrium and is thus difficult to detect cytologically. Angrish and Verma 33a reported a number of cases of cervical tuberculosis that were detected cytologically. The cervical smears reveal large aggregates of epithelioid cells. These appear as pale, cyanophilic cells in a syncytial formation with indistinct and arborizing borders and vesicular, oval nuclei ( Fig. 7.23 ). Intermixed with these one may occasionally observe Langhans-type multinucleated giant cells ( Fig. 7.24 ). These cells may contain as many as 20 to 30 peripherally arranged vesicular nuclei. A variable number of lymphocytes may be present in the background. Secondary infection is common in these ulcerated lesions, and heavy, acute inflammatory exudates may be present. A cytologic diagnosis of granulomatous disease, probably tuberculosis, can be suggested under appropriate clinical and cytologic settings.

Fig. 7.23 Epithelioid cells. Notice the syncytial formation of cells with ill-defined margins. Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.24 Cervical tuberculosis (H&E × LP).

Malacoplakia
Malacoplakia is a rare disorder that may affect the cervix. We have observed two cases occurring in postmenopausal women with atrophic smears and persistent vaginal discharge. 33 Numerous macrophages with the characteristic intracytoplasmic, laminated inclusions (Michaelis-Gutmann bodies) may be observed ( Fig. 7.25 ). They can be stained for calcium salts including calcium phosphates and carbonates by histochemical techniques such as Von Kossa's method. 34

Fig. 7.25 Michaelis-Gutmann bodies. Intracytoplasmic laminated structures from a case of malacoplakia of the cervix. Vaginopancervical smear (Papanicolaou × HP).

Langerhans Cell Histiocytosis
This rare disease of unknown etiology may involve the lower female genital tract and the endometrium. It has been included here because the lesions can both clinically and cytologically may be indistinguishable from inflammatory or neoplastic processes. It can occur as “pure” genital tract or part of the generalized systemic disease. Cervical cytology may contain atypical histiocytic cells ( Fig. 7.26 ), numerous macrophages with intranuclear grooves, eosinophils, and an occasional multinucleated giant cell ( Fig. 7.27 ). This diagnosis may be considered in the presence of intranuclear grooves in the macrophages, and eosinophils in the smear. The exact morphologic features vary from the stage of the disease. Immunohistochemical stains S100 and CD1a as well as ultrastructural demonstration of Birbeck granules are helpful diagnostically. 35, 36

Fig. 7.26 Langerhans cell histiocytosis. Note numerous histiocytic cells with intranuclear grooves (arrowhead) and an occasional eosinophil (arrow). Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.27 Langerhans cell histiocytosis. Note atypical histiocytic cells and numerous eosinophils. This smear was obtained 14 months after that in Fig. 7.26 . Vaginopancervical smear (Papanicolaou × HP).

Actinomyces
These organisms belong to the order of higher bacteria that also include Mycobacteriaceae and Streptomycetaceae . There are three common species of Actinomyces — A. israelii , A. bovis , and A. naeslundii . These bacteria are nonmotile, non-spore-forming, and anaerobic or facultative anaerobes. They are Gram-positive and occur in filamentous and diphtheroid forms. With the recent increase in IUD usage, genital Actinomyces infection appears to also be increasing in prevalence. It is, however, true that the new device designs and the judicious usage make the clinical disease less likely. 37
Actinomyces occur commonly within the tonsillar crypts, tartar of teeth, and the alimentary tracts. Actinomyces do not occur as commensals in the vaginal flora. In the female genital tract, ascending infection is the most common mode of occurrence of clinical disease; however, rarely, hematogenous and lymphatic spread, or dissemination of infection from the alimentary tract or other distant sources, may occur. Ascending infection occurs in the presence of intrauterine or intravaginal contraceptives, IUDs of various types being the most common. Vaginal pessaries, surgical clamps, and foreign bodies, including forgotten tampons, all have been associated with vaginal Actinomyces . Among untreated women, clinical disease may be manifest for as much as 12 months after the removal of the Actinomyces -associated IUD.
Gupta has reviewed the subject and the relationship of Actinomyces with clinical female genital tract disease. 38 It is appropriate to say that nearly 10% of women using an IUD may develop vaginal Actinomyces infection at some stage. If such users have symptoms of lower genital tract infection such as pelvic pain, vaginal discharge, bleeding, fever, or lower abdominal tenderness, approximately one-quarter of these women may have genital Actinomyces infection. Of the women using an IUD and being admitted to the hospital for clinically suspected pelvic inflammatory disease, about 40% may harbor the organism in the lower genital tract. 39, 40 Dissemination of the infection to distant sites has been documented by de la Monte and co-workers, 41 and by Hager and Majmudar. 42

Cytomorphology of Actinomyces
In close proximity to the calcified and mineralized fragments of a disintegrating IUD, the Actinomyces organism can be detected in Papanicolaou stained vaginal smears. Typically, the organisms appear as spidery, amorphous clumps that are darker in the center ( Fig. 7.28 A). Morphologic features of the Actinomyces colonies are more distinct in LBGS-based slides ( Fig. 7.28 B). These aggregates of Actinomyces in the cervicovaginal smears have been referred to as “Gupta bodies” by Hager and Majmudar. 42 Upon careful examination, numerous filamentous organisms with acute angle branching patterns are recognizable in these clumps ( Fig. 7.29 ). They can be uniformly thick and beaded. The filaments generally extend to the outer limits of the dark clumps. Only a few delicate, branching filamentous forms may occur scattered randomly in the smear. In Papanicolaou stained smears, calcified filamentous forms that may not be stainable by antigen antibody techniques, club forms, or the Splendore Hoeppli phenomenon may be identified. Typical sulfur granules may be observed in smears obtained from symptomatic patients ( Fig. 7.30 ). These per se are not diagnostic of Actinomyces , and proper morphologic identification of the filamentous forms is necessary in all cases. Gupta and co-workers have detailed various other morphologic forms. 38, 43

Fig. 7.28 Actinomyces . (A) Vaginopancervical smear (Papanicolaou × LP). (B) LBGS (Papanicolaou × LP).

Fig. 7.29 Colonies of Actinomyces. (A) Note the numerous filamentous structures radiating from the center. Vaginopancervical smear (Papanicolaou × HP). (B) Higher magnification of the colonies of Actinomyces (Papanicolaou × HP).

Fig. 7.30 Sulfur granule. In the center are radiating filamentous structures of Actinomyces organisms. Vaginopancervical smear (Papanicolaou × MP).
Actinomyces organisms can be stained with modified Gram, periodic acid-Schiff (PAS), and silver stains. A definitive species diagnosis requires specific antigen antibody reaction using an immunoenzymatic or immunofluorescence or bacterial culture procedures. 44
A number of organisms, including Candida , dermatophytes, and Nocardia , along with bacterial aggregates, and foreign substances such as sulfa drug crystals and contraceptive creams, may resemble Actinomyces organisms. Hematoidin crystals described by Hollander and Gupta 6 have a resemblance to sulfur granules. The differential diagnosis of Actinomyces as seen in the vaginopancervical smear is presented in Table 7.4 .
Table 7.4 Differential diagnosis of Actinomyces in vaginal smears Other organisms Candida, Aspergillus, Nocardia, Penicillium, Trichophyton, Leptotrichia , lactobacilli Miscellaneous structures Filamentous structures: Fibrin, mucus, sulfa crystals, cotton and synthetic fibers   Nonfilamentous structures: Contraceptive cream, bacterial clumps, hematoxylin pigment, spermatozoa, hematoidin, foreign material (spores, pollen, douche ingredients)
We believe that genital Actinomyces is an exogenous infection. Orogenital contact may be an important mode of acquiring the genital Actinomyces infection. The “tail” of the IUD most likely acts as a carrier for the ascent of the organisms. The tissue damage produced by the body and edges of the IUD causes a change in the oxygen reduction potential and alteration in the microbial milieu of the lower genital tract. The changed environment is conducive to the growth of these organisms. Actinomyces has been observed with all types of IUDs, including currently marketed models. Infection is more common with devices with polyfilamentous thread and with angular forms.
Key features of genital Actinomyces
• Always associated with an IUD or a foreign body;
• May cause no symptoms;
• Occur as dark, woolly clumps (Gupta bodies);
• Parallel filaments, branching at acute angle;
• Difficult to culture; and
• May be confirmed by special stains.
Occasionally, Actinomyces may occur in association with “black yeast,” a fungus Aureobasidium pullulans , commonly found in areas with poor hygienic conditions. It has large, dark-fruiting bodies ( Fig. 7.31 ). As reported by de Moraes-Ruehsen and associates, Entamoeba gingivalis , a protozoan of the oral cavity, may be found in association with Actinomyces in vaginal specimens 45 ( Fig. 7.32 ). An orogenital route of this Actinomyces infection is a distinct possibility. These nonpathogenic protozoa should be distinguished from Entamoeba histolytica that occur in the alimentary tract and which may also cause lower genital tract infection.

Fig. 7.31 Aureobasidium pullulans . These black yeast organisms can vary in color from light yellow, gold-brown, to black. Vaginopancervical smear (Papanicolaou × HP).
Reproduced with permission from Gupta PK: Intrauterine contraceptive device: Vaginal cytology, Pathologic changes, and their clinical implications. Acta Cytol 1982;26:571-613.

Fig. 7.32 Entamoeba gingivalis. (A) Vaginopancervical smear (Papanicolaou × HP). (B) LBGS (Papanicolaou × HP).

IUD-Associated Cellular Changes
In addition to the alterations in the microbial environment and Actinomyces infection, usage of the IUD is associated with cellular changes occurring in the various genital tract epithelia, as early as 10–12 weeks after an IUD insertion. These result from chronic irritation by the IUD tail and the body affecting the adjacent tissues within the endocervix and the uterine cavity. It is important to recognize these morphologic features as they can mimic and be confused with dysplastic and neoplastic cellular changes of the squamous, metaplastic, endocervical, and endometrial epithelia. These changes appear more pronounced in LBGS and interpretation can be problematic especially in the paucity of an inflammatory background. There is no definite evidence for the association of squamous dysplastic changes and IUD usage. Squamous cell changes are essentially reactive and reparative in nature. These occur in about 40% of women using IUDs. DNA analysis of IUD-associated cellular changes does not reveal any aneuploidy.
The morphological picture is further complicated by interplay among the reactive-proplastic and degenerative-retroplasticchanges occurring over a prolonged period and affected by polymicrobial and physiologic factors.
Endocervical columnar cells may become hyperplastic with large papillary tissue fragment formations. Bibbo and co-workers 46 and Gupta and colleagues 47 have systematically reviewed these changes. Columnar cell hyperplastic changes should be distinguished from adenocarcinoma ( Fig. 7.33 ). They may mimic papillary tumors of ovarian or endometrial origin. Single cells can be extremely bizarre and resemble neoplasia. Cells may show large cytoplasmic vacuoles referred to as “bubble gum” cells. The presence of heavy inflammation and degenerative changes helps diagnostically. The salient features of these cellular changes are summarized in Table 7.5 . The presence of psammoma or calcified bodies among IUD users is not an indication of neoplasm.

Fig. 7.33 “Bubble-gum” cells occurring in a patient with an IUD. (A) Vaginopancervical smear (Papanicolaou × HP). (B) IUD-associated glandular cells LBGS (Papanicolaou × MP).
Table 7.5 Comparison of IUD-associated columnar-type cells and adenocarcinoma cells Feature IUD columnar cells Tumor cells Tumor diathesis Absent Present Distribution Endocervical Random Inflammation Present Variable Cellular degeneration Present Absent| “Bubble gum” cytoplasm Present Absent Bare nuclei Absent Present Cellular preservation Poor Good Atypical histiocytic cells Absent Present
Another cell type, best described as indeterminate cell changes or “IUD cells,” probably arises from the endometrial surface. Such conclusions are supported by the work of Gupta and co-workers. 47 These cells with a high nucleus-to-cytoplasmic ratio should be distinguished from the third type of cell described by Graham 48 and from in situ carcinoma (HSIL, CIN III) cells. Nuclear degeneration, the presence of nucleoli, and a hiatus between normal and abnormal cells help differentiate these cells from true neoplastic cells ( Fig. 7.34 ). Table 7.6 summarizes the salient features of these cells. Occasionally, the endometrial-type reactive cells and the IUD cells may occur together.

Fig. 7.34 “IUD cells.” These high nucleocytoplasmic (N/C) ratio cells appear to be of endometrial origin. They frequently show multinucleation, nuclear degeneration, and nucleoli. If not carefully examined, these can be easily mistaken for cervical intraepithelial neoplasm. (A) Vaginopancervical smear (Papanicolaou × HP). (B) IUD cells in LBGS (Papanicolaou × OI).
Table 7.6 Comparison of IUD cells and cervical intraepithelial neoplasia (CIN/HSIL) cells Feature IUD cell CIN/HSIL cell Distribution Endocervical Endocervical Tissue fragments Rare Common (LBGS) Inflammation Present Absent Cellular degeneration Present Absent Preservation Poor Good Cellular hiatus Present Absent Nucleoli Present Absent Multinucleation Present Absent IUD columnar cells Present Absent
Binucleated and multinucleated giant forms and psammoma body formation are other findings that may be observed in the presence of the IUD and Actinomyces . These develop from endometrial surface changes. Extensive squamous metaplasia of the endometrial surface may occur in some cases as the result of prolonged endometritis accompanying the IUD.
Key features of IUD-associated cellular changes
• Bubble gum cells;
• IUD cells;
• Metaplastic cells;
• Mesenchymal proliferation;
• Multinucleation; and
• Psammoma body formation.

Leptotrichia buccalis
These microbes, also known as just Leptotrichia or Leptothrix , are Gram-negative, non-spore-forming anaerobic organisms. They occur in the oral and vaginal cavities as very thin, segmented, large, filamentous structures. Occasionally, branching may be observed ( Fig. 7.35 ). Morphologically they may be indistinguishable from certain forms of Doederlein's bacillus. Most frequently (75–80%), cases of Leptotrichia have concomitant T. vaginalis infection. Numerous other infective organisms, including Candida and G. vaginalis , may occur in the presence of L. buccalis infection.

Fig. 7.35 Leptotrichia buccalis. Vaginopancervical smear (Papanicolaou × HP).
Bibbo and Wied 13 made an investigative study on the prevalence of Leptotrichia in cervicovaginal smears. They observed Leptotrichia organisms in 75% cases with trichomonads, 1.5% with Doederlein's bacillus, and about 1% among patients with fungal or BV infection. Nearly half (47%) of the 1,000 patients studied were oral contraceptive users. Pregnancy and menopause were other physiologic features, followed by the postpartum state, that were often associated with the presence of L. buccalis in cervical smears. Sometimes acute inflammatory changes may be observed in the presence of Leptotrichia .

Mycoplasma
These are the smallest known organisms capable of growing in cell-free media. Jones and Davson documented a correlation between the occurrence of a “dirty” smear and mycoplasma, 61 and Mardh and co-workers confirmed these findings and reported the occurrence of coccoid organisms both on the surface and in between the squamous epithelial cells in dirty smears in cases of Mycoplasma . 49 Such features appear to have limited practical value. 48a ,49

Follicular Cervicitis
Also referred to as lymphocytic cervicitis, this is a specific type of cervical, and sometimes vaginal, lesion in which the predominant feature is the occurrence of lymphoid follicles in the subepithelial areas. When examined cytologically, numerous mature and reactive lymphoid cells and germinal macrophages (tingible bodies) are seen ( Fig. 7.36 A). 50 Cellular changes are uncommonly recognized in LBGS ( Fig. 7.36 B) and are often more problematic to interpret. Small lymphoid cells appear in aggregates especially in LBGS and do not clearly reveal tingible bodies. 51 Lack of mucus and small size of the lymphocytes appear to contribute to these morphologic changes. At times, a capillary from the germinal center of the lymphoid follicle may be scraped and observed in the smear ( Fig. 7.37 ). There is evidence that nearly 50% of the cases of follicular cervicitis are associated with Chlamydia infection. 52, 53 Follicular cervicitis is not uncommonly seen in postmenopausal atrophic smears. The precise pathogenesis of this condition in this age group is not well understood. These cells should be distinguished for malignant lymphomas, histocytes, endometrial cells, and metastatic tumor cells.
Key features of follicular cervicitis
• Must identify tingible-body macrophage;
• Difficult to diagnose in LBGS due to lymphocytic dispersion; and
• Cells should be distinguished from:
– Lymphoma;
– Metastatic tumor cells;
– Endometrial cells; and
– Histiocytes.

Fig. 7.36 (A) Follicular cervicitis. Vaginopancervical smear (Papanicolaou × OI). (B) Follicular cervicitis, LBGS. Diagnosis is difficult to make. Tingible macrophages (arrow) and lymphocytes may be observed (Papanicolaou × HP).

Fig. 7.37 Follicular cervicitis. Note the occurrence of a germinal follicle and a capillary in this picture. Vaginopancervical smear (Papanicolaou × LP).

Viral Infections
Diseases caused by these intracellular organisms are among the most common in the human body and include a most heterogeneous group of clinical conditions. Although some of the viral infections have been affecting humanity for thousands of years, changes in society, social habits, medical practice, and advances in diagnostic capabilities have resulted in a great many new viral diseases. Even though smallpox has been eradicated from the world, with antibiotics and immunosuppressive therapies, numerous dormant viral infections have become manifest.
Being intracellular by nature, viruses co-opt cellular metabolic processes in their replication cycles. In addition to the nature of the affected tissues, the virus, general and local immune responses, and particular enzymatic derangements are important in determining the cytomorphologic changes and the nature of the tissue injury or injuries. In some common viral infections, these cellular changes may be quite typical and considered of diagnostic significance.

General Features of Viral Infection

Inclusion Formation
Inclusions are discrete, dense, homogeneous, round, or oval intracellular structures consisting of viral particles in a matrix, and generally represent a stage in the replication of the virus. They do not occur in all viral infections, their formation depending upon a particular agent and on the affected tissue. Certain inclusions are typical and diagnostic. Inclusions may be observed within the nuclei, the cytoplasm, or both.

Hydropic or Ballooning Degeneration
This particular cellular change is often an effect of organelle membrane damage caused by the virus. Certain degenerative changes precede or accompany the development of inclusion bodies and are often used in the diagnostic evaluation of cellular changes ( Fig. 7.38 ).

Fig. 7.38 Hydropic degeneration. Vaginopancervical smear (Papanicolaou × HP).

Necrosis
Viruses may cause coagulative necrosis and characteristic cytoplasmic changes. Most often, the cytoplasm becomes opaque and thickened and loses its transparency and crispness. Nuclear degenerative changes with karyolysis and karyorrhexis may occur ( Fig. 7.39 ). Only ghost forms of the infected cells may remain.

Fig. 7.39 Nuclear degeneration. LBGS (Papanicolaou × HP).

Giant Cell Formation
Alterations in the membrane composition of the infected cells contribute to the fusion of cells to produce syncytial and giant forms. Sometimes nuclear inclusions may occur within the multinucleated forms.

Cellular Proliferation
Transient cellular proliferation is commonly seen in viral infections. These changes may be extreme and mimic dyskaryosis and neoplastic forms ( Fig. 7.40 ).

Fig. 7.40 (A) Atypical cellular proliferation, LBGS (Papanicolaou × MP). (B) “Atypical” cellular changes in cervical herpes, LBGS (Papanicolaou × HP).

Cellular Cohesion
In certain viral infections, the initial step is attachment to the host cell; viral proteins (antireceptors) adsorb to the cell surface furnished with appropriate receptors. The interaction may alter not only the surface of the infected cell but also the structure of the virion. Although not fully understood, viral detachment and readsorption perhaps contribute to cell clumps or plaque formation.

Cytoskeleton Changes
Cytoplasmic and nuclear changes frequently occur not as a result of the damage caused by the virus, but rather as a result of specific reorganization of the cellular or skeletal elements necessary for its growth. Alteration in intermediate keratin filaments and microtubules, and cellular metabolism contribute to the formation of ciliocytophthoria (CCP) seen in certain viral infections. It should be distinguished from detached ciliary tufts (DCT) described by Hollander and Gupta that may be observed in lower genital tract smears in the absence of a viral infection. 6, 54 In vivo hemadsorption observed occasionally may be a related phenomenon ( Fig. 7.41 ).

Fig. 7.41 Hemadsorption. This patient had herpes infection at the time when these and many other similar cells were seen. Vaginopancervical smear (Papanicolaou × HP).

Oncogenesis
Both in vitro and in vivo neoplastic transformation of viral-infected cells may occur. Numerous DNA viruses and a group of retroviruses are capable of neoplastic transformation. These commonly manifest as dyskaryosis and atypical nuclear alterations.
Quite often, in the presence of florid viral infection, no discernible morphologic changes may occur in the infected cells and tissues.
The previously mentioned cellular manifestations may or may not be reflected in all cytologic preparations and in the presence of all viral infections.

Specific Infections
Specific viral infections commonly observed in the female genital tract include herpes infection: Herpes is a Greek word meaning “to creep.” It is believed that this word was used in relation to certain clinical features of an infection that eventually was found to be related to the particular DNA virus. There are at least six different viruses in this group causing disease in humans. These are:
• Herpes simplex virus, type 1 and type 2 (HSV 1, HSV 2);
• Cytomegalovirus (CMV);
• Varicella-zoster virus;
• Epstein-Barr (EB) virus; and
• Lymphoma-associated viruses.
In the cytologic preparations from the female genital tract herpes, CMV and varicella may be detected.

Herpes Simplex Virus
Distinction between HSV 1 and HSV 2 was made based on serologic studies by Schneweis. 55 Most people acquire antibodies to HSV 1 during the first 2 years of their life. Herpetic vulvovaginitis or stomatitis due to HSV 1 may occur at the time of initial infection, generally in infancy or early adolescence. Infection is mostly asymptomatic, or it may be accompanied by upper respiratory tract or ocular symptoms. Morphologically, HSV 1 and HSV 2 cellular changes appear identical.
Although congenital or neonatal transmission may occur, HSV 2 generally occurs after puberty and the onset of sexual activity. Cutaneous lesions, commonly vesicles, tend to occur in the same area repeatedly; the interval between successive eruptions varies considerably even in the same individual. Stress, menses, and other unrelated ailments may precipitate an eruption in an otherwise healthy person. Following the initial infection, the virus remains dormant in the sacral (S2 through S4) dorsal root ganglia in the spinal cord. McDougall 56 has documented its presence in the spinal cord.
Recently, there has been an increase in the occurrence of HSV 2 cases; it is generally attributed to changed sexual and social habits. Using seroepidemiologic data, an association of HSV 2 and cervical cancer has been reported by Kessler and others. 55, 57 - 63 The precise role of HSV 2 in the development of human cervical cancer is far from resolved; evidence, however, accepts HPV as the most important causative infection.

Herpes Simplex Genitalis Virus, Type 2
For nearly 2500 years, people have used the word herpes in medical literature. Corey has briefly discussed the history of genital herpes. 64 John Astruc first described genital herpes in 1736 in the French literature. More than 100 cases of “herpes progenitalis” were reported in the late nineteenth century. Lipschutz established experimental transmission of herpes in human beings in 1921. 65 He concluded that there were differences between oral (HSV 1) and genital herpes (HSV 2) infections.
HSV 2 infection is one of the most common sexually transmitted genital infections; more than 300,000 new cases are recorded in the United States annually. The prevalence of infection varies depending on the group studied. Although in general populations the incidence of infection is not well established, genital HSV infection was diagnosed among 4.2% of those attending the Sexually Transmitted Disease Clinic in Seattle, Washington, in 1980. Women presenting at student health services have been found to have HSV 2 infection about seven to ten times more commonly than gonorrhea. Data from the Centers for Disease Control and Prevention (CDC) suggest that the prevalence of HSV 2 is increasing and that the infection is occurring in social groups that previously did not have the disease.
Primary infection may be asymptomatic or accompanied by severe constitutional symptoms. Commonly, fever, headache, and myalgia occur before the appearance of mucocutaneous lesions. Visible lesions appear between 2 and 7 days following exposure to the virus. Local pain and itching, dysuria, vaginal discharge, and inguinal lymphadenopathy may be present. The lesions are painful and often multiple. Large ulcerations that start as papules or vesicles spread rapidly. They form pustules that coalesce and break down. Unless complicated by secondary infection, these ulcers heal in 5 to 10 days with reepithelialization. Residual scarring is uncommon. Systemic symptoms and inguinal lymphadenopathy occur mainly in primary HSV 2 infection.
The cytologic diagnosis of HSV infection is important and must be made on well-preserved cells that have typical diagnostic features and have not been altered by air-drying, fixation, or inflammation. Such a diagnosis may determine proper management of patients, especially pregnant women with genital ulcerations. An HSV diagnosis, with its social and medical implications, should be rendered only when unequivocal evidence is present.
In addition to diagnosis in the standard Pap test, direct sampling of visible lesions can be performed. Such smears should be prepared from the edge and bed of the ulceration, not from the contents of the vesiculae. The latter generally contain serosanguineous material with acute inflammatory cells, eosinophils, and some macrophages. Although use of air-dried smears and their examination after Romanowsky stain (Tzanck preparation) 66 have been advocated, we do not recommend this for genital lesion diagnoses. Heavy inflammation, cellular obscuring, and degeneration often make interpretation difficult and may severely compromise the diagnostic value of air-dried smears. Cellular samples obtained from the cleared ulcer beds should be immediately fixed in 95% ethanol and examined after Papanicolaou staining.
The virus may infect the immature squamous, metaplastic, and endocervical columnar cells. Initially, the changes are proplastic and somewhat nonspecific. The infected cells can occur singly, in groups, and in tissue fragments. There is cytomegaly and karyomegaly, and the nucleocytoplasmic ratio is not much altered. These cells demonstrate a combination of reactive (proplastic) and degenerative (retroplastic) changes. The nuclei of the infected cells show changes in the chromatin structure consisting of hydropic or ballooning degeneration. The chromatin material becomes extremely finely divided and is uniformly dispersed in the nuclear sap. The chromatin–parachromatin interphase is obliterated, and nuclei assume a faintly hematoxylinophilic, homogenized appearance. Some chromatin material may be matted against the inner leaf of the nuclear envelope, which may appear uniformly thick and conspicuous. The altered nuclear morphology is commonly referred to as ground glass, bland, gelatinous, glassy, or opaque. In some cases the redistribution of chromatin may result in a beaded appearance of the nuclear margins. Nucleoli may be present and conspicuous, may have associated chromatin, and may not appear typically bright acidophilic. Although the nucleoli generally remain round or oval, sometimes irregular shapes may be observed.
In the later stages of HSV infection, the cells undergo the effects of viral replication and DNA integration. The cells may assume multinucleation, which is observed in nearly 80% of the smears from cases of genital HSV infection. The infected nuclei may have the same homogeneous chromatin pattern described previously. The nuclei appear tightly packed within the cells and reveal distinct internuclear molding ( Fig. 7.42 ). At times they may be overlapping and not molding. Large and single intranuclear inclusions appear within these nuclei. The nuclear inclusions are generally round or oval. They can be angulated and sharp ( Fig. 7.43 ). They lack structure and are densely eosinophilic. Depending on the staining procedure employed, they may appear cherry red. A clear zone, or halo, which separates it from the nuclear membrane, surrounds the intranuclear inclusion. Most often the halo is as clear as the background of the slide. Sometimes it may retain delicate, homogeneous, diffuse hematoxylinophilia. Small, inconspicuous chromatin granules can occur in the peri-inclusion halo. Inclusions may occur in infected single cells observed in nearly one-third of the cases of HSV 2 infection. Intranuclear inclusions may not be present in all of the nuclei within the multinucleated giant cells.

Fig. 7.42 “Early” herpes genitalis infection. Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.43 Herpes genitalis infection. Vaginopancervical smear (Papanicolaou × HP).
The cytoplasm in the infected cells at this early stage of HSV infection is dense. It may lose its transparent appearance and become opaque. Often it stains bright cyanophilic.
HSV-infected cells can become atypical; the enlarged cells may assume bizarre shapes ( Fig. 7.44 ). They may be hyperchromatic or degenerated and may be misinterpreted as tumor cells. The cytoplasm may show changes of the cytoskeletal structure and become dense or opaque uniformly or focally. The latter may represent keratohyaline material. Degenerative vacuoles, the ectoplasmic–endoplasmic differentiation with spiral fibrils of Erbeth, as described by Patten, 67 may be present between the two zones. The fibrillary apparatus of Herxheimer appears as delicate, uniformly thin spirals that originate at the nucleus, travel down the cytoplasm, and may be observed in cells with squamous differentiation features.
Key features of herpes genitalis infection
• Multinucleated giant cells not diagnostic;
• Nuclear changes critical for diagnosis; and
• Primary and secondary infection cannot be distinguished morphologically.

Fig. 7.44 Bizarre columnar cells in a smear with herpes genitalis infection. LBGS (Papanicolaou × HP).
Multinucleated giant cells per se do not establish the diagnosis of HSV infection. Proper nuclear features and inclusions must be identified for such diagnosis. Virus-infected cells or virocytes should be distinguished from other giant cells such as trophoblasts, foreign-body giant cells following extraneous intervention with foreign bodies or surgery, nonspecific giant cells seen in postmenopausal smears, and reactive multinucleated cells found in cases of cervicitis.
Some workers have described morphologic differences that may distinguish between primary and recurrent herpes. Paucity of intranuclear inclusions and occurrence of chromatin homogeneity and “ground-glass” changes were often reported in primary herpes, whereas inclusions were predominant among cases with post primary infection. The studies of Vesterinen and associates 63 and other workers have not confirmed this observation, although the World Health Organization has adapted these observations differentiating primary and recurrent herpes. Morphologically, HSV 2 and HSV 1 cannot be differentiated. Either such diagnosis can be rendered only by appropriate serologic reactions or viral culture studies ( Fig. 7.45 ).

Fig. 7.45 Herpes genitalis. Note both intranuclear inclusions and “ground-glass” nuclei in the same group of cells. LBGS (Papanicolaou × MP).

Cytomegalovirus
These large (1800–2000 Å) DNA viruses belong to the herpes group that includes HSV 1, HSV 2, varicella-zoster, and EB virus. CMV is ubiquitous and circulates commonly in the general population. As with herpes viruses, CMV establishes itself in the host and causes persistent infection and recurrent disease. In the genital tract, reinfection may occasionally occur. Unlike other herpes viruses, however, clinically overt manifestations of viral replication are seen only rarely. The infection spreads by intimate contact through body secretions, including saliva, tears, urine, endocervical mucus, semen, and transplanted organs.
Nearly 50–60% of adult women have circulating antibodies to CMV. Serologic evidence of infection is more common in low socioeconomic groups. Cervical shedding of CMV occurs in nearly 10% of the female population. 21 These figures vary in different populations; e.g. the recovery rate for CMV in cervical specimens has been reported to be 28% in Japan. In a number of studies summarized by Kumar and co-workers, 68 the incidence of primary genital CMV infection has been reported as between 0.2 and 2.2%. Most cases of primary genital CMV infection are clinically asymptomatic. Stagno and colleagues reported symptomatic episodes in only 1 of 21 cases, 68a and Griffiths and associates 69 in only 1 of 14 pregnant women with primary CMV infection. Almost always these symptoms are infectious mononucleosis-like and include lethargy, malaise, and fever.
Characteristically, epithelial tissues, including salivary gland, alimentary tract, bronchial and alveolar lining cells, hepatocytes, renal tubule cells, hematopoietic cells, and endocervical and endothelial cells, are targeted by CMV infection. Cytologically, the infected cells, endocervical columnar, perhaps, occur more commonly in the cervical smears than is recognized. According to Naib, 70 the CMV-bearing cells occur within the endocervical glands, and not many cells may be observed in the epithelium of the endocervical canal. Proper cellular specimens, as can be obtainable with a Cytobrush or similar technique, may be more rewarding. In addition, because almost all women are asymptomatic, very little effort may be made to screen these smears critically for CMV-associated changes. In a certain number of patients, concomitant CMV and HSV infections may occur. It is not infrequent to overlook the not-so-obvious CMV-infected cells in such cases. At times, the morphologic identification of CMV and its differentiation from HSV may be extremely difficult.
The CMV-infected cells may be multinucleated and somewhat enlarged. The nuclear degenerative changes may be similar to those in HSV. We have observed that the internuclear molding tends to be less obvious in CMV-infected cells than in those infected with HSV. Infected endocervical cells are anisocytotic. They contain round intranuclear inclusions, generally acidophilic, which are disproportionately large when compared with the total size of the nuclei. These inclusions have a clear zone of halo around them, and frequently threads of chromatin material may stretch between the inclusion and the inner leaf of the nuclear membrane, which may be considerably thickened with the chromatin material in apposition against it, giving a wheel spoke appearance. The infected cells may have an intracytoplasmic, irregular inclusion ( Figs 7.46 and 7.47 ).

Fig. 7.46 (A) Cytomegalovirus cervix. Note the chromatin threats. (B) Intracytoplasmic inclusions, LBGS (Papanicolaou × OI).

Fig. 7.47 Cytomegalovirus cervicitis (H&E × MP).
Infected endocervical cells may be buried among inflammatory or other epithelial cells and may be hard to screen for in a routine fashion. In selected cases, CMV-specific monoclonal antibodies and in situ techniques can be used to establish the presence of CMV-infected cells.

Herpes zoster
Varicella-zoster is related to varicella (chickenpox) that is so common in childhood and infancy. Herpes zoster (shingles) infection may involve the vulva and vagina. Lesions often appear in patients who were exposed to the varicella virus in childhood. They occur in older individuals along the distribution of sensory nerves as extremely painful vesiculae or blisters, and tend to be unilateral.
A smear from the base of the lesion often reveals numerous multinucleated giant cells with little intercellular molding. Numerous infected single cells may occur. Intranuclear inclusions may be basophilic, large, and inseparable from the markedly thickened inner nuclear membrane. Infected parabasal-type cells may show some cytoplasmic degenerative changes. Intracytoplasmic vacuolation and hyalinization may be present ( Fig. 7.48 ).

Fig. 7.48 Herpes zoster. (A) Note the basophilic intranuclear inclusion. (B) Higher power showing the texture and details of the inclusions. Vaginopancervical smear (Papanicolaou (A) × MP, (B) × OI).

Human Papillomavirus (HPV)
HPVs belong to the family Papovaviridae , which includes double-stranded DNA members, papillomaviruses, and polyomaviruses. The papillomavirus genome is approximately 8000 base pairs in length. It has three functioning areas, including genes for early viral function, the late region containing genes for viral structural proteins, and a noncoding regulatory region. The viral capsid has two proteins and polypeptides. Papillomavirus has been isolated from more than 60 animal species, including mammals, reptiles, and amphibians. The vast majorities of these viruses infect epithelial surfaces of either the skin or mucosa and cause self-limiting warty growth. The papillomaviruses are species-specific and generally do not cross-infect.
In humans, using molecular hybridization restriction enzyme analyses and polymerase chain reaction (PCR), greater than 100 distinct HPV types have been identified. These different HPV types tend to be site-specific. Approximately 40 different HPV types affect the anogenital tract. HPV types are generally separated into two major groups, low and high risk, depending on their risk for the development of cervical cancer. Low-risk types of HPV (most commonly types 6 and 11) have essentially no risk association for invasive cervical carcinoma, but tend to cause condyloma acuminatum, flat condyloma, and some low-grade squamous dysplasia (cervical intraepithelial neoplasia [CIN] I). 71 The high-risk or “oncogenic” HPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73, and 82) are often found in high-grade cervical dysplasia (CIN II and III), cervical carcinoma in situ, and invasive squamous cell carcinoma and adenocarcinoma and its precursors. High-risk HPV types are also commonly seen in CIN I lesions 72a , 72b ( Table 7.7 ).
Table 7.7 Common human papillomavirus lesions (over 100 types) HPV type Lesion 1, 2, 3, 4, 10, 28 Common warts 6, 11, 31, 42 Anogenital condyloma, low-risk (CIN grade I) lesions 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73, 82 High-risk (CIN grade II and III) lesions—anogenital, laryngeal, esophageal, lung cancer (possibly) 26, 27 Warts, immune deficiency, renal transplant
HPV replicates in the nuclei of squamous epithelial cells. In the superficial squamous epithelium, the virus reaches full maturity as a permissive infection and undergoes koilocytic changes with HPV virions filling the nucleus and upon cell death, capable of reinfecting other cells. In the basal layers of squamous and glandular epithelium, the virus is latent and its DNA can only be detected by molecular techniques. 73 In many low-grade precancerous lesions, HPV DNA is usually retained intact in an episomal form. In contrast, in invasive cancers and some high-grade precancers, viral DNA is integrated into the host DNA, and retains only genes associated with oncogenesis. 73

Papillomavirus and Cancer
The virus is believed to enter the body through small, inconspicuous cuts or abrasions on the skin or mucous membrane. It stimulates the growth of the prickle cell layer. The growth is by clonal expansion and, as suggested by Broker and Butcher, 74 it pushes aside the normal epithelium to form benign warts. In these low-grade processes virus replication occurs episomally in the nucleus.
In addition to the numerous benign, self-limiting warty growths, papillomaviruses are associated with a number of neoplasms occurring in animals, such as rabbits and cattle. The most convincing evidence for the association of HPV with cervical cancer comes from the studies by Bosch et al. 75 and Walboomers et al., 76 who elegantly demonstrated the presence of HPV DNA in 99.7% of cervical cancers (932 cervical cancer cases from 32 hospitals in 22 countries). HPV DNA and RNA has also been demonstrated in at least 80% of all cervical, vulvar, and penile squamous cell carcinomas, in a similar proportion of premalignant CIN lesions, and in 95% of genital condylomas. 77, 78 HPV types 6 and 11 predominate in benign warty lesions, whereas types 16 and 18 occur in 60–70% of all cervical tumors investigated. Other oncogenic HPV types (31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68), in conjunction with HPV types 16 and 18, are responsible for more than 95% of cervical squamous carcinoma and its precursors and in excess of 90–95% cervical adenocarcinoma and its precursor adenocarcinoma in situ. HPV types 16 and 18 have also been found in several human cervical cancer cell lines, including the famous HeLa cell. Success has been achieved in transformation experiments conducted using HPV subgenomic particles and human epidermoid cell lines.
Additional molecular evidence linking HPV to cervical cancer has been demonstrated by experimental data showing that the viral genes E6 and E7 of high-risk HPVs can extend the life span of genital epithelial cells. E7 can disrupt the cell cycle by binding to pRB with up-regulation of cyclin E and p16 INK4 . E6 can interrupt cell death by binding to p53 and preventing replicative senescence by up-regulation of telomerase. Both genes have been shown to induce centrosome duplication and genomic instability. 73, 79 - 83 More recently, the strong association of HPV with cervical cancer has resulted in the development of vaccines directed against the most common oncogenic HPV types 16 and 18. 108 These vaccines have been shown to prevent infection and the development of precancerous lesions due to these subtypes, and one has been FDA approved for use in girls aged 9–26.

Historical Perspective
The typical cytomorphologic changes now associated with HPV infection were first documented by Ayre in 1949. 85 Papanicolaou, in his Atlas of Exfoliative Cytology published in 1954, presented magnificent illustrations depicting cellular changes currently recognized as HPV-associated. 86 Koss and Durfee in 1956 used the term koilocytic atypia to describe the surface epithelial changes of the cervix and its relationship with cancer. 87 Naib and Masukawa in 1961 published a paper entitled “Identification of Condyloma Acuminata Cells in Routine Vaginal Smears.” 88 Additional papers on the same topic were published by Ayre, 89 Sagiroglu, 90 and De Girolami. 91 Ayre for the first time suggested a possible viral etiology of cervical dysplastic lesions. 89 In the mid-1970s, Meisels and Fortin, and Purola and Savia independently defined a set of cellular changes that are associated with cervical condylomatous lesions and HPV infection. 72, 92 Also, flat condyloma, indistinguishable from cervical dysplasia, was recognized as a separate viral-related entity. Laverty and co-workers 93 demonstrated the presence of virus particles within the koilocytes in cervical tissues ultrastructurally. In 1980, Woodruff and colleagues demonstrated the presence of HPV capsid antigen in genital condylomas using polyclonal rabbit antihuman wart virus antibodies. 94 Gupta and associates employed Papanicolaou and immunoenzymatic stained cells in cervical smears and demonstrated viral particles in cervical dysplastic cells identified by the viral antigen detection system. 95, 96
Immunoenzymatic studies may reveal the presence of HPV antigens in 0.5–5% of the cells in cervical smears. Abundance of HPV infection, accompanying inflammatory reaction ( G. vaginalis ), and the degree of dysplastic changes determine the proportion of the antigen-positive cells. In up to 80% of these cases, corresponding cervical biopsy tissue may reveal identical results. Antigenically stained cells may include parabasal, intermediate, and metaplastic types ( Fig. 7.49 ). Antigen positivity has no relationship to age, but is instead dependent on the degree and depth of condylomatous change in the squamous epithelium. Nearly 90% of the cases with two-thirds thickness involvement may reveal the HPV antigen. Condylomatous and low-grade lesions (CIN grade I) reveal HPV antigens more commonly.

Fig. 7.49 Human papillomavirus (HPV) infection of the cervix. These smears have been stained with rabbit anti-HPV antibodies using immunoenzymatic reaction (immunoenzymatic stain with chromogen DAB × HP).
Nucleic acid hybridization was described in 1975 by Southern. 97 It has been used extensively for study of HPV DNA in vaginal secretions and infected cells, both in the smears and in tissues. Brahic and Haase developed in situ hybridization, which permits localization of labeled DNA or RNA probes within the cellular preparations and tissue sections. 98 In 1987, Gupta and co-workers used radiolabeled HPV DNA probes and established the presence of viral genomes in infected dysplastic cells in cervical smears 96 ( Fig. 7.50 ). RNA probes have been developed and used by Stoler and Broker among others. 99 Radiolabeled probes have more recently been replaced by immunoenzymatic techniques. Filter dot and PCR techniques have been used to investigate the presence of extremely small quantities of the viral genome in infected specimens and to identify latent infection. This has also allowed investigation of the relationship of specific viral types to human disease. More recently, a signal amplification hybrid capture technique was developed by Lorincz to detect the presence of HPV in liquid cytology samples. 100, 101

Fig. 7.50 Human papillomavirus (HPV) infection. This slide has been processed for in situ hybridization using an S35-labeled HPV-16 probe. The radiolabeled granules are distinctly seen on the surface of the infected nuclei. Vaginopancervical smear (autoradiograph with hematoxylin × HP).
The worldwide understanding of the role of HPV as a necessary, but not sufficient, cause of invasive cervical cancer, coupled with the development of liquid cytology providing residual cervical sample for testing and molecular techniques, has led to multiple changes in cervical cancer screening and triage algorithms over the past decade. The NCI-sponsored multicenter randomized ASC-US LSIL Triage Study (ALTS) demonstrated the clinical utility of “reflex” HPV high-risk (HPV HR) testing for the cytologic diagnosis of atypical squamous cells of undetermined significance (ASC-US). HPV HR testing for ASC-US was found to have greater sensitivity (96%) to detect CIN III or worse lesions and similar specificity as compared to a repeat cytology specimen. 104 The strong scientific evidence led to changes in the Bethesda System (TBS) classification and terminology and the American Society for Colposcopy and Cervical Pathology (ASCCP) and the American Cancer Society (ACS) management guidelines for abnormal cervical cytology. 103 - 105 ASC-US high-risk (HR) HPV positive cases are recommended to be referred to colposcopy and biopsy, while negative HR HPV ASC-US cases can return to annual screening. 107 It is estimated that 80% of ASC-US cases are now triaged using HPV HR testing. 105
HPV HR testing can also be used as an adjunct to cervical cancer screening with cytology in women aged 30 and older. The FDA approved this use in 2003 for assessment of overall cervical cancer risk due to the 99.2% negative predictive value for developing CIN III or invasive cervical cancer in the next three years if both tests are negative. The ACS and the American College of Obstetrics and Gynecology (ACOG) have defined this dual screening method as an option. Women who have a negative Pap cytology and negative HPV HR test are at a very low risk for developing cervical cancer in the next 3–5 years (0.8%) and can increase their screening interval safely to 3 years. 77, 84 The combined cytology and HPV testing is not appropriate for women under age 30, since they frequently test positive for HPV, which in this age group is most frequently a transient and/or clinically insignificant infection.
The more recent development of the HPV vaccine and its implementation will most likely impact upon cervical cancer screening cytology volume and protocols in the future. Importantly, despite the development of molecular techniques to detect HPV and changes to cervical cancer screening algorithms, the molecular tests do not substitute or replace cytologic morphology and diagnosis. Cytomorphology corroborates molecular changes as illustrated by the high incidence of HR HPV in low-grade squamous intraepithelial lesion (LSIL) and worse lesions, making HPV HR testing in these diagnoses unnecessary. 107, 107a In addition, HPV HR infection is very common, but only a minority of women will develop precancerous changes detectable by cytology, and even a smaller number will develop invasive carcinoma. It is estimated that the lifetime risk of becoming infected with one or more of the sexually transmitted HPV types is 50–79% of all women who have had sexual intercourse as compared to the 1.3% lifetime risk of developing invasive cancer. 71 Currently, cytomorphologic examination for diagnosis remains as important as ever.

Cytomorphology
Cytomorphologic identification of cellular changes is currently the most convenient, rapid, and economical, procedure for detection of HPV infection in the genital tract. As discussed previously, HPV infection generally manifests as verrucous or flat-surface epithelial lesions. Both lesions have the same basic pathognomonic features; additionally, the papilliferous lesions reveal surface hyper- and parakeratosis and papillomatosis.
HPV, being a DNA virus, affects both the nucleus and the cytoplasm of the infected cells. It is generally believed that the virus gains entrance into the susceptible cell through the plasma membrane. The cytoplasmic changes of the infected cells—dyskeratosis—are a prominent feature of HPV infection. HPV DNA may occur within the epithelial cell nucleus as either unintegrated/episomal or integrated forms. Consequently, nuclear changes are commonly seen with HPV infection. These, however, tend to be more pronounced in cases of HPV DNA integration with the epithelial cell nuclear DNA. Such lesions typically appear as dysplastic and atypical. Chromosome and ploidy alteration, a hallmark of CIN or dysplastic change, may occur in these cases.
The classic manifestation of HPV infection is the presence of the koilocyte, so named by Koss and Durfee, 87 also been called nearo-carcinoma by Ayre 89 and balloon cells by Meisels and Fortin. 72
Typical koilocytes are squamous epithelial cells ( Fig. 7.51 ). Most commonly intermediate cells and sometimes metaplastic-type koilocytic cells may occur. The latter may be indistinguishable from parabasal cells.

Fig. 7.51 Typical koilocytic cells. Vaginopancervical smear (Papanicolaou × MP).
The HPV-infected cells show blunting of the sharp angles of the squamous cells. The squamous angular forms tend to become rounded, and the cell assumes a softer, rounded or ovoid appearance. Typically, the cytoplasm shows a peripheral condensation and produces a “wire looping” effect, in which the cytoproteins gel at the margins, leaving an almost empty shell. This cytoplasm generally appears structureless and opaque, or waxy. It may be acidophilic and appear as a brighter reddish-orange, resembling a deep pumpkin red or a shade thereof. The typical koilocytic cell has a large cavity or halo intracytoplasmic space. This space has a sharp peripheral margin, and the nucleus contained within this most often is eccentrically located; that is, this is a large paranuclear, and not a perinuclear, halo. The latter is generally small with soft margins and is seen in the infectious process, such as in trichomoniasis. It may also have a variable pale or faint cytoplasm within it. Occasionally, phagocytosed material may be observed within the koilocytic space ( Fig. 7.52 ).

Fig. 7.52 Koilocytic cell. Vaginopancervical smear (Papanicolaou × HP).
Ultrastructurally, viral particles can be more easily detected in warty lesions occurring in the skin than in those in the genital tract and airways. Intranuclear icosahedral nonenveloped virions of 40 to 50 nm can be observed. At times, some particles may also be seen within the cytoplasm. Such findings are believed to be artifactual.
Immunocytologic investigations may reveal high-molecular-weight cytokeratin within the koilocytic space. No viral antigens can be demonstrated with the paranuclear halo of koilocytes by immunoenzymatic techniques.
Diagnostically, the koilocyte is an excellent indicator of HPV infection. It has a high degree of specificity. Properly selected and evaluated, nearly 90% of the cases of condyloma with koilocytic change were found by immunoenzymatic techniques to have demonstrable HPV antigen. In ALTS, greater than 84% of LSIL cases contained HR HPV. 104 Although highly specific, the koilocyte alone has low sensitivity for detection of HPV infection. Based on examination of routine vaginopancervical smears, in our experience nearly 60% of the cases of condyloma reveal obvious koilocytes; an additional 20% of the smears will reveal the diagnostic cells if carefully screened. About one-third of the cases of HPV infection may be missed cytologically if the diagnosis is made solely on the basis of koilocytic changes. Schneider and colleagues have reported that only 20% of cases of HPV infection may be detected cytologically. 109 We feel this is too low and probably not correct. It is obvious that the conflicting reports in the literature about the diagnostic value of vaginopancervical smears for HPV infection are based on excessive reliance on the koilocyte as the sole diagnostic feature.
It is well known that the most vulnerable area of the ectocervical–endocervical interface is the transformation zone, an area situated between the squamous-lined ectocervix and the columnar-lined endocervix. 73 HPV infection is believed also to originate in this region. Once established, the infection may move proximally or distally and affect the adjacent epithelial cells. The ectocervical squamous epithelial changes that manifest as koilocytes and other dyskeratotic forms are more obvious and more common, but the proper recognition of these nonclassic changes can improve the diagnostic value of routine vaginopancervical smears.
Quite often, a diagnosis of HPV infection or condylomatous changes can be suspected with the lower magnification examination of the vaginopancervical smear. The infected cells may occur with or without inflammation, concomitant infection, or endocervical component. It is not uncommon to be able to suggest a diagnosis of HPV infection in an otherwise less-than-satisfactory smear that is rich in vaginal components and has minimal endocervical representation.
The HPV-infected cells tend to stick together. The presence of groups or clumps of deeply acidophilic, opaque squamous cells in an otherwise well-fixed (no air-drying artifacts in conventional smears) and -stained smear may be telltale evidence of HPV infection ( Fig. 7.53 ). A careful evaluation of such groups may reveal a number of squamous cells that have lost their sharp, polygonal shapes; they have become rounded and blunt. The cells may be overlapping, as commonly seen in pregnancy and late postovulatory smears, but they do not remain transparent and thin. They become dense and may be opaque. The normal gradual centrifugal thinning of the intermediate and superficial cell cytoplasm is lost. Any degree of change between a normal-appearing squamous cell and the typical koilocyte may be present in these cells. Besides the HPV-infected cells occurring as aggregates, syncytial or pseudoepithelial formation may be observed.

Fig. 7.53 Human papillomavirus (HPV) infection. This picture represents the cellular clumping or plaque formation commonly observed in this infection. Vaginopancervical smear (Papanicolaou × LP).
Ectocervical involvement by HPV may result in infected cells with dyskeratosis. Although frequently observed in cases of squamous cell carcinoma, some of the abnormal cytoplasmic features, especially changes in tinctorial character, occur commonly in HPV-infected cells. Abnormal shape is another feature affecting the squamous epithelial cells. Frayed edges ( Fig. 7.54 ), fiber, and tadpole formations are some of the features observed in HPV-infected squamous cells. Nuclear changes must be evaluated in these cases for proper interpretation of cytomorphology.

Fig. 7.54 Human papillomavirus (HPV) infection. This photomicrograph reveals the dyskeratosis that frequently accompanies the viral infection. Some of the cells are elongated, fiber-shaped, and bizarre. These keratinized cells, if not carefully examined, can be mistaken for evidence of invasive cancer. Vaginopancervical smear (Papanicolaou stain × MP).
Parakeratosis ( Fig. 7.55 ) is another feature that sometimes is seen in association with HPV infections. This finding is important in cases that may not have an inflammatory background. The parakeratotic cells can reveal abnormal keratinization or dyskeratosis. Hyperkeratosis may be seen as anucleated squames. This per se is not considered suggestive of HPV infection, and a better cellular sample should be examined in such cases.

Fig. 7.55 Human papillomavirus (HPV) infection. This infection appears to involve the ectocervix and the transformation zone simultaneously. Both mature squamous cells and immature metaplastic-type cells are observed. Some of the cells show small tissue fragment formation. Multinucleation and keratinization are also observed. Vaginopancervical smear (Papanicolaou stain; ×172).
In cases in which the HPV infection may dominate the transformation zone or involve the endocervical canal, the infected cells as observed in vaginopancervical smears appear cyanophilic and may not be the keratinized squamous type. Atypical repair-type cells (discussed elsewhere) may dominate the smear. Very commonly, these cases have immature, metaplastic-type cells infected with HPV in the smear ( Figs. 7.56 , 7.57 ). They appear as single and basophilic with dense cytoplasm and round or oval shapes. They may be as small as parabasal or appear as big as intermediate cells. Some anisocytosis is common in these cells. The cytoplasm may show peripheral condensation similar to that seen in typical koilocytes. The nuclei may be eccentrically located, reveal size variation, and show features of increased activity such as enlargement, chromatin granularity, and bi- or multinucleation. The nucleoli are inconspicuous. When the infection is nested in the transformation zone, parakeratotic and metaplastic-type cells may predominate. Small tissue fragments can be seen sometimes. In cases with infection involving the ecto- and endocervical regions, a mixture of mature, parakeratotic squamous cells and metaplastic-type cells with evidence of HPV infection may all appear together ( Fig. 7.57 ). A number of these morphologic changes observed either singly or in various combinations may be grouped as ASC-US.

Fig. 7.56 A diagrammatic representation of cellular changes in the cervix following human papillomavirus (HPV) infection (see text). (1) Normal squamous cells. (2) Koilocytes. (3) Plaques of HPV-infected cells. (4) Parakeratotic cells. (5) Dyskaryosis with atypical forms. (6) Hyperkeratosis with anucleate squames. (7) Metaplastic-type HPV-infected cells. These cyanophilic cells have eccentric nuclei and peripheral cytoplasmic condensation. (8) HPV-infected cells from the transformation zone. These appear undifferentiated. (9) HPV-infected cells from the transformation zone. These cells are of unclassified type.

Fig. 7.57 Human papillomavirus (HPV) infection cervix. Left transformation zone, ecto and endocervix. (a) Hyperkeratotic papillae, (b) squamous plaques, (c) koilocytes transformation zone changes, (d) endocervical cells changes, (e, f) endocervix. (H&E × LP, [(a) × LP, (b-d) × HP, (e, f) × MP] Papanicolaou.
Using the diagnostic criteria mentioned previously and as summarized in Table 7.8 and Fig. 7.57 , we have excellent correlation between the cytodiagnosis of HPV, its histology, and detection of HPV by in situ hybridization (ISH) molecular techniques. The sampling procedures and the interval between the cytosmear and appropriate tissue studies are important. Changes observed in the endocervical cells in HPV infection, in our experience, have not been sufficiently specific to be useful diagnostically.
Table 7.8 Cytomorphologic features of human papillomavirus infection AFFECTED CELL Keratinized, mature squame Keratinized, immature squame Undifferentiated transformation zone CELLULAR CONFIGURATION Clumps Single Tissue fragments CELLULAR SHAPE Loss of polygonal form, blunt or rounded corners CELLULAR MARGIN Thickened, wire loop appearance CELL SIZE Iso- and anisocytosis CYTOPLASM Variable translucency and condensation Koilocytosis NUCLEUS Karyomegaly Bi- and multinucleation Minimal dyskaryosis ASSOCIATED CHANGES Parakeratosis Hyperkeratosis
Nuclear changes, although not specific, are commonly observed among HPV-infected cells. The nuclei may not show any changes; they may degenerate and appear hyperchromatic and pyknotic or reveal chromatin margination and abnormal clearing. Bi- and multinucleated forms in the squamous cells occur often. The nuclei may be enlarged, but changes most often are proplastic with pale, finely divided chromatin granules distributed uniformly within the enlarged nucleus. The nucleoli are absent or inconspicuous in these cells. The nuclear membrane may be loose and wavy. It may be degenerated and appear folded and somewhat wrinkled, giving the typical raisin like appearance so often observed in HPV infections ( Fig. 7.51 ).

HPV and Cervical Dysplasia
The established Bethesda System of gynecologic cytology reporting recognizes changes of HPV infection or condyloma and varying degrees of dysplasia. The same terminology is being used to describe the HPV-related lesions. Morphologically, it can be difficult to distinguish condyloma from mild dysplasia or a CIN I lesion, both cytologically and histologically. An awareness of this difficulty and no difference in clinical management has led to the two lesions being grouped together as low-grade squamous intraepithelial lesion. When distinguished or subclassified, CIN grade I lesions have enlarged nuclei and hyperchromasia in contrast to condyloma. Chromatin granules are uniformly distributed and they tend to be isodiametric in size, be moderately coarse, and appear identical; no parachromatin clearing and abnormal clumping are observed ( Fig. 7.58 ). At times the nuclear membrane may be undulating and appear folded and wrinkled. Most affected cells are intermediate squames. In the Bethesda System for reporting gynecologic cytodiagnoses, CIN grades II and III are grouped together as “high-grade” cervical intraepithelial lesions. 103 CIN grade II, or moderate dysplasia, has distinct nuclear chromatin granularity, which is uniform and coarse and may be unevenly distributed within the affected intermediate and parabasal-type cells. Parachromatin clearing is inconspicuous ( Fig. 7.59 ). This is essentially a state of further exaggeration of the changes seen in CIN grade I. Whereas all condylomas and three-quarters of CIN grade I may reveal HPV antigen-positive cells, only two-thirds of CIN grade II smears may manifest such findings. CIN grade III, or severe or marked dysplasia, may be indistinguishable from carcinoma in situ. The biologic behavior and management of these lesions are essentially identical.

Fig. 7.58 Human papillomavirus (HPV) infection, low-grade squamous intraepithelial neoplasm (LSIL, CIN I) changes. Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.59 Human papillomavirus (HPV) infection, high-grade squamous intraepithelial neoplasm (HSIL, CIN II/III) changes. Compared with Fig. 7.54 , the nuclear chromatin pattern is more irregular, coarser, and hyperchromatic. Vaginopancervical smear (Papanicolaou × HP).
The affected cells tend to appear mostly singly and are more immature. They have distinct hyperchromasia with coarse but uniform chromatin granularity, nuclear membrane undulation, and uniformity. No nucleoli or parachromatin clearing is generally seen. At times, the affected cells may be in small groups and fragments. Atypical, keratinized, and bizarre dyskeratotic forms may appear in some cases.
Rarely (less than 5%), CIN grade III cells may be antigen-positive. The integrated HPV genome, however, can be detected within the infected cells by molecular hybridization techniques and PCR in greater than 90% cases. Given the high frequency of oncogenic HPV DNA positivity in LSIL+ positive lesions, there is little clinical utility for HPV HR testing in LSIL or HSIL cases. Likewise, there is a limited role for HPV HR testing in ASC-H or AGC cases. At present, there is no alternative marker to predict which patients with these lesions are at greatest risk for progressing to cancer. All patients with these significant cytologic abnormalities are recommended to undergo colposcopy and biopsy, as outlined by the ASCCP management guidelines. 107 Some exceptions in management apply to younger compliant patients who are more likely to have a transient infection.
Using more sensitive techniques such as PCR, HPV genomes have been detected in a large number of tumors, including endocervical and ovarian adenocarcinomas; and vulvar, vaginal, cervical, and anal carcinomas. 110 The interrelationship of immunologic, cellular, and mutagenic events and the role of other genital infections such as HSV and CMV infections may all be important in understanding the biology of this common genital infection.
It is not uncommon for HPV to manifest clinically in the condition of altered immune response. Pregnancy and physical and emotional stress may precipitate clinical HPV disease. Similarly, prolonged immunosuppression, as after a kidney transplant, and chemotherapy may cause clinical HPV disease. In fact, the first case reported by Gupta and co-workers of cervical dysplasia after azathioprine (Imuran) therapy we believe now to represent a case with condylomatous changes in the cervical epithelium following immunosuppression. 111 Increased incidence of condyloma in patients with acquired immunodeficiency syndrome (AIDS) is also well documented.

Molecular testing and ancillary techniques:
As noted earlier, HPV HR testing is commonly used as an adjunctive tool in cervical cancer screening for ASC-US triage/reflex testing, and in women aged 30 and older with negative cytology to assess overall cervical cancer risk. As the 2002 ASCCP management guidelines noted: “HPV testing has matured, appears clinically validated and should become integral to both screening and clinical management.” 107
There are a variety of molecular techniques available for detecting and quantifying HPV in clinical specimens, including cytology and histology. Historically, tests for HPV using nuclei acid probes have been commercially available since the late 1980s, but were cumbersome and involved radioactive labeling. Nucleic acid ISH was also commercially available. Currently, available HPV tests include three major types of nucleic acid hybridization methodologies: direct probe methods, hybridization signal amplification, and target amplification. 112
Direct probe methods include Southern blot, the gold standard for HPV genomic analysis, and ISH. The latter provides the advantage of correlating the presence of HPV with morphology (cytologic or histological). Disadvantages of direct probe methods include low sensitivity, labor intensity, and potential need for highly purified DNA, which is a challenge in formalin-fixed tissue. Signal amplification methods are proprietary technologies that boost sensitivity of direct probe methods by detection innovations. Examples of signal amplification include multimeric layering of reporter molecules on DNA probes (hybrid capture), branched DNA (bDNA), and isothermal signal amplification (invader technology). Target amplification methods include PCR and can be home brew or commercially available. Target amplification is highly sensitive and can be used for detection, viral load quantitation, DNA sequencing, and mutation analysis, and can be performed in multiplex assays. Patented HPV sequences limits the applicability of PCR as an in vitro diagnostic test due to legal and/or proprietary restrictions. Home-brew assays also lack interlaboratory standardization. 112, 113
At the time of this writing, there is only one FDA-approved HPV test for in vitro diagnostic use/clinical testing, Digene Hybrid Capture 2 (HC2), a signal amplification methodology. HC2 utilizes a cocktail of specific RNA probes directed toward individual DNA sequences of 13 HR HPV genotypes to be detected. A proprietary antibody directed toward DNA–RNA hybrids in the capture and detection steps is developed by a chemiluminescence system. 100, 101 This test also has the most clinical experience to date, having been utilized in ALTS with well-documented positive and negative predictive values and has formed the basis of current cervical cancer screening guidelines incorporating HPV HR testing. Despite its success, this assay is time-consuming and labor intensive, contains no internal control to evaluate for adequate cellular material, and can cross-react between low-risk and probable high-risk HPV types. Although clinically sensitive, improvement in clinical specificity and positive predictive value is also desirable. 106 The ability to correlate HPV presence with cytomorphology and histology provided by ISH is preferred by some. Automated ISH is available for HPV detection in liquid cytology and histology specimens (Ventana), but variable reports of sensitivity and specificity which do not allow advocating for widespread clinical use at present have been documented in the literature. In some ISH platforms, episomal versus integrated HPV patterns can be distinguished, but whether this distinction is predictive of significant concurrent or future disease is still for investigative purposes only ( Figs. 7.60 , 7.61 ). In our experience using Ventana probes, the results for the detection, specificity, sensitivity, and predictive values are comparable to the HC2 detection. There is slightly increased detection among the normal and HSIL groups, and somewhat decreased detection in ASC-US cases (unpublished data).

Fig. 7.60 Human papillomavirus (HPV) infection cervix, LSIL. (A) LBGS (Papanicolaou 60 × MP). (B) In situ hybridization on the same specimen reveals intranuclear HPV high-risk localization (in situ hybridization with DAB changes × MP).

Fig. 7.61 Human papillomavirus (HPV) infection cervix, HSIL. (A) LBGS (Papanicolaou 60 × MP). (B) In situ hybridization on the same specimen reveals intranuclear HPV high-risk localization in an integrated form. (In situ hybridization with DAB changes × MP).
Despite the fact that emergence of other HPV assays is expected and desired, utilization of non-FDA approved tests for HPV HR testing has raised considerable concern in the literature. 106
The future of HPV testing is continuing to evolve and in the USA, additional FDA-approved assays are anticipated following well-controlled, statistically powered clinical validation trials. Time will tell whether some of these assays will include ISH, viral load, and detection of individual HPV genotypes.

Polyomavirus
In addition to HPV, another member of the papovavirus family—BK or polyomavirus—may rarely be observed in the vaginopancervical smear. BK virus is seen frequently as an urothelial infection following renal transplantation with immunotherapy and immunosuppression. Rarely, intranuclear large basophilic inclusions may be seen in the vaginal smear. Cross-contamination of the vaginal contents from the urinary tract, however, is a distinct possibility.

Molluscum Contagiosum
Thomas Bateman first described this in 1814. Like other poxviruses, the genome is a single, linear molecule of double-stranded DNA. The virus contains a virus-specified, DNA-dependent RNA polymerase and has not been cultured in in vitro systems.
A typical molluscum contagiosum lesion consists of a localized mass of hypertrophic, hyperplastic epidermal cells that push the basement membrane down and produce on the epidermal surface a pearl-white, somewhat umbilicated nodule. The germinal cells in the lesion multiply rapidly. Each cell enlarges and is filled with dense acidophilic intracytoplasmic inclusion called molluscum bodies ( Fig. 7.62 ). To be diagnostic, each molluscum body must be intracytoplasmic and should have a compressed hematoxylinophilic nucleus on the outside. The tinctorial character of the inclusion may change with the age of the lesion.

Fig. 7.62 Molluscum bodies. Note the dark, large intracytoplasmic inclusions (arrow). Vulvar smear (Papanicolaou × HP).

Adenovirus
Adenovirus infection may occur in the female genital tract and be asymptomatic ( Fig. 7.63 ). Virus often affects the columnar cell, producing intranuclear inclusions. These may be acidophilic or basophilic, depending on the duration of infection. Intracytoplasmic inclusions are not well recognized in the cervical smears. Newborn infants may be infected from the maternal adenovirus infection. 114

Fig. 7.63 (A) Adenovirus infection. Note the numerous intranuclear basophilic inclusions within the endocervical cells. Vaginopancervical smear (Papanicolaou stain × HP). (B) Adenovirus infection with eosinophilic inclusions. Vaginopancervical smear (Papanicolaou × HP).

Chlamydial Infection
Chlamydia trachomatis infection is one of the most common sexually transmitted diseases. 115 Although documented perhaps earlier, Halberstaedter and Von Prowazek first visualized C. trachomatis in 1907 in a conjunctival scraping. 116 The growth cycle of C. trachomatis was described in the early 1930s. 117, 118 The first isolation of C. trachomatis from a nonlymphogranuloma venereum case was done in 1959 by Jones and co-workers. The patient was the mother of a newborn infant with ophthalmia neonatorum. In addition to the well-known culture techniques, 117 immunodiagnostic techniques, including immunofluorescence and immunoenzymatic techniques, radioimmunoassays, and more recently molecular hybridization techniques, have been introduced in the diagnostic armamentarium for C. trachomatis .
Nearly 50% of women with C. trachomatis infection may be asymptomatic. Genitourinary disease, including urethritis, vaginitis, cervicitis, endometritis, salpingitis, and pelvic inflammatory disease (PID), may occur in women infected with C. trachomatis organisms. 119 The infection can also affect the neonate during passage through the birth canal and cause neonatal pneumonia or may be transmitted to the sexual partner, causing urethritis, prostatitis, and epididymitis. Systemic disease may be associated with C. trachomatis infection; perihepatitis, endocarditis, and gastroenteritis have all been associated. The cases are few, and precise data are lacking. The clinical chlamydial diseases are summarized in Table 7.9 . The precise role of C. trachomatis infection in all of the clinical diseases mentioned in Table 7.10 is variable. Taylor-Robinson and Thomas summarized the data presented in Table 7.10 . 53
Table 7.9 Chlamydial diseases Infections in men Infections in women Infections in infants Nongonococcal urethritis Cervicitis Conjunctivitis Postgonococcal urethritis Conjunctivitis Pneumonia Conjunctivitis Subclinical genital infection Gastroenteritis (possibly) Subclinical genital infection Salpingitis   Epididymitis Dysplasia (possibly)   Reiter's syndrome Infertility   Systemic disease Systemic disease  
Table 7.10 Relationship of Chlamydia trachomatis and various associated diseases Acute and chronic cervicitis Established Follicular cervicitis Established Salpingitis Established Lymphogranuloma venereum Established Nongonococcal urethritis Established Postgonococcal urethritis Established Reiter's syndrome Established Neonatal pneumonia Established Conjunctivitis Establishe Epididymitis Questionable Cervical atypia Questionable Abortion Questionable
Chlamydia trachomatis is an obligate intracellular organism. It shares certain bacterial and viral characteristics. The organism does not stain with a Gram stain. It contains both RNA and DNA, is susceptible to antibodies, divides by binary fusion, and has a rigid cell wall. A comparison of some of the common characteristics of Chlamydia , bacteria, and viruses is given in Table 7.11 .

Table 7.11 Common characteristics of chlamydia, bacteria, and viruses
Chlamydia organisms occur in two major forms: the infective, extracellular elementary body, and the intracellular initial and intermediate forms or bodies, also referred to as reticulate particles. Elementary bodies measure about 300 nm. These are liberated from the infected cells and are phagocytosed by the susceptible cells, which most frequently are squamous metaplastic or endocervical columnar cells. There is some evidence of Chlamydia organisms infecting the parabasal cells of the lower genital tract. Shurbaji and associates 44, 120 have documented the occurrence of Chlamydia organisms within the prostate epithelium and urothelium. As seen by life cycle studies, once within the infected cells, the elementary bodies reorganize and enlarge. They repeatedly replicate by binary fision and, after undergoing the intermediate stages of development, produce intracytoplasmic inclusions that are large pockets of numerous elementary bodies. The infected cell finally lyses, liberating numerous elementary bodies to restart the cycle. This life cycle in a cell culture, e.g. McCoy cell line, is completed within 48 to 60 hours.
The infected epithelial cell can undergo a number of proplastic and retroplastic changes. These include enlargement of the nuclei and cytomegaly, hyperchromasia, and nucleolar prominence. Multinucleation may occur. The retroplastic changes commonly include cytoplasmic vacuolation and protein precipitation that may appear as nonspecific inclusions. Some serologic evidence exists for the possible association of C. trachomatis infection and cervical dysplasia, but it is not well accepted.
In a symptomatic patient, the body reacts to C. trachomatis infection, and the occurrence of intense polymorphonuclear leukocytic response in the vaginopancervical smears is the most common single feature of symptomatic C. trachomatis infection. This was one of the major features observed among chlamydial cases by Kiviat and associates. 121 Quinn and co-workers found the presence of intense polymorphonuclear leukocytic exudation as the single most common feature among cases of genital C. trachomatis infection. 122
Although direct detection of Chlamydia in appropriately stained representative smears has been practiced for the longest time, the diagnostic value of direct smear examination in specimens other than conjunctival smears has been questioned repeatedly. A number of advances in Chlamydia diagnosis have been made in more recent years. Improved culture techniques, direct immunofluorescence antigen detection and enzyme-linked immunosorbent assay (ELISA)-based tests, have been introduced with a high degree of specificity and sensitivity. In a recent summary statement, the CDC, however, enumerated the state of Chlamydia diagnosis. It stated, “Despite the encouraging improvement in diagnostic capability, current tests are not ideal. They are relatively difficult to perform, require considerable experience and have limited application.” 115
Until recently, tissue cell culture has been used as a “gold standard” for the diagnosis. Kellogg critically reviewed the various diagnostic methods and their limitations. 123 Briefly stated, the number of swabs cultured, site of epithelial sampling, type of collection device, storage, evaluation of symptoms, serum immunoglobulin levels and antibodies, any treatment to stabilize the cellular membranes including antibodies and steroid preparations and other drugs, maturations, and various inhibiting substances in the vaginal secretions can have a bearing on the diagnostic value of Chlamydia cultures. Also, selection of container, cell line and pretreatment, inoculation technique and culture duration, concomitant infections and contamination, and type of stain used to detect growth of the organism in the cell line all have a bearing on the performance of the culture techniques. Cultures for Chlamydia are generally more valuable in younger patients with primary first-time infection. Antigen assays, on the other hand, depend on the population makeup. These results tend to be less positive in asymptomatic and older patients. The size of the population, clinical disease, quality of the specimen, and reading threshold also determine the results of direct Chlamydia diagnostic procedures.
Along with Gupta and co-workers 52 and Bibbo and Wied, 13 Naib has documented the cytologic detection of Chlamydia . 70, 124 Elementary bodies that are abundant in symptomatic individuals and may be easily cultured cannot be detected in Papanicolaou stained specimens. Romanowsky stained specimens can be helpful for such identification, especially when the sample is examined under dark-field illumination. A golden-yellow discoloration is observed within the Chlamydia organisms. In addition, the presence of heavy, acute inflammation may cause cellular obscuring and degeneration, rendering the fine cytoplasmic details incomprehensible. Cellular degenerative changes, on the other hand, can cause intracytoplasmic structures that may mimic chlamydial inclusions. It must be appreciated that cytoplasmic degenerative changes may occur in the presence of C. trachomatis infection also, but they per se are not sufficient for a diagnosis.
Intracytoplasmic aggregates of minute elementary bodies occurring within the metaplastic cells are the earliest discernible feature of C. trachomatis infection. These may undergo degeneration with small, pinhead structures surrounded by a halo that is thin-walled ( Fig. 7.64 ). These intermediate forms may be prominent and give the infected cell a “moth-eaten” appearance. A careful study of such cells often shows clumps of elementary bodies also. 52 At times the intracytoplasmic aggregates of elementary bodies can condense and produce distinctly identifiable intracytoplasmic nebular inclusions. 125 These are large, homogeneous or finely granular structures with ill defined and indistinct walls. Ultrastructurally, nebular inclusions contain the Chlamydia organisms. The Chlamydia organisms can be identified using appropriate antigen detection techniques in most of the infected cells ( Fig. 7.65 ). When properly done, the cytodiagnosis of Chlamydia has a high degree of specificity and sensitivity. The cytomorphologic features are summarized in Table 7.12 .

Fig. 7.64 Chlamydia trachomatis infection. Moth-eaten, rarified appearance and numerous fine-walled intracytoplasmic vacuolated structures. Vaginopancervical smear (Papanicolaou × OI).

Fig. 7.65 Chlamydia trachomatis infections. Intracytoplasmic organisms seen by monoclonal antibodies, elementary bodies. (A) Various intermediate forms and (B) 60, 50a, FITC, 1/p with AEC changes; (C) monoclonal antibodies with immunoperoxidase and AEC Chromogen (Cervicovaginal Smear 60 (A) × OI, (B) × HP, (C) × MP).
Table 7.12 Cytologic features of Chlamydia trachomatis infection Background Acute inflammation with numerous polymorphonuclear leukocytes and macrophages Infected cells Metaplastic columnar cells, and possibly parabasal cells either singly or in tissue fragments Morphology Intracytoplasmic elementary bodies; faint, acidophilic coccoid structures occurring diffusely or focally; moth-eaten appearance; reticulate and intermediate bodies occurring intracellularly as thin-walled target forms   Nebular forms occurring as dense intracytoplasmic structures; multinucleation and cellular reactive changes
Cellular degeneration with intracytoplasmic inclusion formation as commonly observed after radiation and chemotherapy, secretions, and coccoid organisms all can be mistaken for Chlamydia . The value of obtaining a proper medical history and a high-quality representative smear and staining cannot be overemphasized. Cytomorphological recognition of Chlamydia in genital smears is not recommended in routine practice. The detection is made using more specific and sensitive molecular techniques.

Fungal Infections
Candida albicans and Torulopsis glabrata are now grouped together, the latter being called Candida glabrata . Candida infection generally involves the vulva, the vagina, and sometimes the cervix. A large proportion (about 40%) of women with detectable Candida organisms may be asymptomatic. Clinically, the infection produces a white, cheesy, thick discharge with a burning sensation and intense itching. Pruritus is a common symptom of Candida infection involving the vulvar region. Sometimes there may be minimal vaginal discharge. In Papanicolaou stained smears, numerous filamentous organisms may occur, revealing pseudo- and true hyphal and yeast forms ( Fig. 7.66 A). Only yeast-budding forms may be observed, especially among asymptomatic women ( Fig. 7.66 B). Sometimes the organisms cause a peculiar fern-like (herringbone/shish kabob) arrangement of epithelial cells. Fungal organisms can be suspected in such cases also when the background contains numerous fragmented leukocytic nuclei (polydust) ( Fig. 7.67 ). Although the organisms can be conveniently seen in potassium hydroxide (KOH) preparations, a high degree of correlation between the Papanicolaou detection of Candida and fungal cultures exists. With Candida infections, squamous cells often show increased evidence of maturation, with parakeratosis and hyperkeratosis. Occasional groups of mature squamous cells can show prominent perinuclear degenerative vacuolar change which can be mistaken for koilocytosis. In addition, some intact squamous cells may show nuclear enlargement that can prompt an interpretation of atypical squamous cells of undetermined significance.

Fig. 7.66 Candida infection. (A) LBGS (Papanicolaou × HP). (B) Budding yeast forms in Candida infection. These occur commonly among asymptomatic women. Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.67 “Polydust” background in Candida infection. Vaginopancervical smear (Papanicolaou × MP).
Rare cases of blastomycosis have been observed in vaginal smears. Alternaria , Aspergillus , and other fungi seen in the vaginal specimens are most often contaminants.
Key features of Candida infection
• Infection may be asymptomatic,
• Polydust and herringbone patterns are useful diagnostically;
• Only yeast forms may occur; and
• Cellular changes may resemble epithelial atypia.

Parasitic Infections

Trichomonas vaginalis
Trichomonas vaginalis is a protozoan and is the most common parasitic organism in cervicovaginal specimens. It is one of the four species ( Trichomonas tenax , Trichomonas hominis , Trichomonas fecalis , and T. vaginalis ) and is the only pathogen for human beings. The most common infection occurs in the lower female genital tract, although nongenital infections including neonatal pneumonia, perinephric abscess, cutaneous lesion, and alimentary tract infection all have been documented, as has been human prostatic disease. Host factors such as endocervical glands and mucus, various immunoglobulins, the complement system and leukocytes, macrophages, and the polymicrobial vaginal environment are some of the important contributing factors for clinical symptoms.
Clinical disease is often described as occurring in acute, chronic, and latent phases. Nearly 50% of the women who have this infection harbor this parasite in the latent phase and are asymptomatic. During the symptomatic phase, the organisms occur in the vagina and occasionally in the secretions of the Skene and Bartholin glands. In approximately 10–20% of the women, lower urinary tract infection may occur and present as dysuria and urethral discharge. The organism may be recovered from clean-catch urine specimens. It has also been recovered from purulent tubal material. The precise role, if any, of T. vaginalis infection in the development of PID, although documented, is controversial.
The incubation period of T. vaginalis infection is between 4 and 28 days. A foamy vaginal discharge occurs in 10–25% of patients. It may be malodorous, copious, frothy, and greenish-yellow. Vulvar vaginitis and symptoms of PID including inguinal lymphadenopathy may occur. “Strawberry vagina” with reddening of the mucosa and small, punctate hemorrhagic spots is typical. The strawberry cervix, although classic, is seen in less than 5% of infected patients.
The disease may exacerbate in the latent phase during or immediately following the menstrual period. Clinical disease is more common during pregnancy. Frost observed Trichomonas in 19% of pregnant women. 10 The effect of Trichomonas infection on the newborn infant and postpartum endometrial infection is documented but not universally accepted.

Cytomorphologic Features
Although a number of techniques such as dark field, hanging drop and wet mount preparations, and PAS, Romanowsky, and immunoenzymatic staining methods can observe the organism, excellent morphologic details are seen using wet fixation and Papanicolaou stained vaginal smears. “Routine” vaginopancervical smears are most valuable diagnostically. Because the organisms occur in the vagina, vaginal pool material or secretions from the posterior fornix are more sensitive. Vaginal douching in the preceding 24 hours, dilution of the secretions by menstrual blood, inflammation with cellular obscuring, marked cytolytic changes as in pregnancy or late luteal phase, and atrophic cellular changes all may make the detection of T. vaginalis difficult. In cytologic preparations from asymptomatic women, the organisms are best detected by examining cervical material obtained by endocervical canal scraping or aspiration.
In well-stained and representative vaginopancervical smears, T. vaginalis can be suspected by the presence of a number of features. These include the occurrence of aggregates of leukocytes covering the surfaces of the isolated, mature squamous epithelial cells ( Fig. 7.68 ). These leukocytic agglomerations have been called “BB shots” and “cannonballs” and represent a number of T. vaginalis organisms feeding on the squamous epithelial cell that, in turn, is phagocytosed by the leukocytes and macrophages. The attachment of T. vaginalis to the margins of the squamous epithelial cells can be easily studied by appropriate immunoenzymatic techniques ( Fig. 7.69 ). Similar changes to a lesser degree can be observed in the LBGS also.

Fig. 7.68 (A) Trichomonas vaginalis infection revealing “BB shot” or “cannonball” appearance (LBGS × LP). (B) Trichomonas attached to the surface of squamous cells (LBGS × MP).

Fig. 7.69 Trichomonas vaginalis attached to the edges of the squamous epithelial cell in the center of the field (LBGS Papanicolaou × MP).
Sometimes, T. vaginalis may be suspected by observing Leptotrichia in the vaginal smears. These are large, generally nonbranching, curved bacillary structures distinct from lactobacilli and were described in detail earlier. Bibbo and Wied observed Leptotrichia and Trichomonas together in 95% of more than 1000 cases examined. 13
In women having a total hysterectomy, T. vaginalis occurs most often without the accompanying inflammatory reaction. In such smears, numerous mature squamous cells and T. vaginalis organisms occur with minimal changes, and only some bacteria may occur as accompaniments.
Cytomorphologically, T. vaginalis frequently appears as small, round or oval structures. Their size may vary from that of the nucleus of a leukocyte to that of the parabasal cell. Extremely large, giant forms of Trichomonas , 150–200 nm in size, have been observed. The organisms stain a cytoplasmic color that may vary according to the pH of the vagina, staining quality, and fixation. Generally, they are cyanophilic or delicate lavender in color. Bizarre, elongated, or tadpole forms may occur ( Fig. 7.70 ). Although they frequently occur singly, in cases of severe infection and immunosuppression, large microcolonies of the organisms may appear in the smear. The organisms always have a distinct, faint, vesicular nucleus. Most often it is eccentric. Sometimes a number of acidophilic, uniformly sized granules may be observed within the organisms. Trichomonas vaginalis multiplies by binary fission, and the organisms may be observed in mitosis. 125a , 125b
Key features of Trichomonas vaginalis
• May be suspected by “BB Shots” (tight clusters of neutrophils) and leptothrix;
• Inflammatory reaction absent after total hysterectomy;
• Variable size, shape, and color;
• Must have faint, vesicular nucleus for diagnosis; and
• Presence of flagella—better seen in LBGS.

Fig. 7.70 (A) Trichomonas vaginalis various morphological forms. LBGS (Papanicolaou × MP). (B) Trichomonas vaginalis organisms have been stained with mouse anti- Trichomonas vaginalis monoclonal antibody and immunoenzymatic technique. Note the varied forms of the organisms in this picture. Vaginopancervical smear (IP × MP).
Morphology is better recognized in LBGS-based preparations; flagella and pear-shaped forms as seen in the living organisms may be recognizable. A specificity of 93%, with a sensitivity of 50%, a positive predictive value of 77%, and a negative predictive value of 80% for the detection of T. vaginalis were obtained in an evaluation of LBGS-based specimens when compared to the wet-mount and culture studies. 125b
Some correlation has been observed between the size of the T. vaginalis and their pathogenicity. It is believed that the smaller organisms cause more fulminant and symptomatic infection.
Histologically, the organisms, although inconspicuous in routine hematoxylin and eosin-stained cervical tissues, may be detected by PAS, Masson's trichrome, and appropriate immunodiagnostic techniques. 125a They often occur on the surface of both squamous and endocervical epithelial cells but may be seen within the epithelial cells ( Fig. 7.71 ) especially in the LBGS. Pericellular (“chicken wire”) edema may occur among the squamous epithelial cells. Reserve cell hyperplasia, squamous metaplasia, and epithelial papillomatosis with capillary proliferation may occur. Some of these changes may be nonspecific in nature but are often seen in cases of trichomoniasis.

Fig. 7.71 Trichomonas vaginalis organisms within the metaplastic endocervical cells. Vaginopancervical smear (LBGS × MP).
Although reactive and atypical epithelial cellular changes may occur with Trichomonas , there is no convincing evidence of pure T. vaginalis infection causing nuclear chromosomal changes or aneuploidy, and thus dysplasia or preneoplastic epithelial changes. Concomitant infection with Chlamydia , herpes, and HPV, among others, may occur in cases with T. vaginalis infection. There is some evidence for suspecting the presence of Chlamydia within T. vaginalis organisms. Bare nuclei, degenerated neutrophils, and mucus fragments must be distinguished from trichomonads.

Enterobius vermicularis
Enterobius vermicularis is a nematode commonly found in the tropics. Although alimentary tract infection is common, occasional reports exist of its occurrence in the endometrium, fallopian tubes, and other sites. Most often the eggs of E. vermicularis occur as a contaminant in vaginal pool material, especially among women with poor personal hygiene. These eggs are 50–60 μm by 20–25 μm. They are flattened on one side ( Fig. 7.72 ). The shell is double-walled and smooth. Within the egg, an embryo can often be recognized. Only rarely may the larvae be seen in the vaginal smear ( Fig. 7.73 ).

Fig. 7.72 Enterobius vermicularis eggs . Vaginopancervical smear (Papanicolaou × MP).

Fig. 7.73 Enterobius vermicularis larva in the smear . Vaginopancervical smear (Papanicolaou × LP).

Trichuriasis
This is a nematode infection caused by Trichuris trichiura . The eggs of this alimentary tract parasite may occur in the vaginal smear as a contaminant. The eggs are barrel shaped and 50–55 nm by 20–25 nm. They have a thick shell with a brownish discoloration. Bipolar colorless prominence is typical of the eggs and is diagnostic. The infection is seen in tropical countries.

Entamoeba histolytica
Entamoeba histolytica is a protozoan that may infect the lower genital tract. Clinically, the affected area is ulcerated and fungating and can be misdiagnosed as neoplasm. Intermixed with necrotic material, numerous trophozoite forms can be seen ( Fig. 7.74 ). These histiocytic-type organisms have biphasic cytoplasm with vesicular nuclei containing a central karyosome. Ingested red blood cells are often seen within the cytoplasm. Only rarely may a cyst form be observed in the vaginal smear ( Fig. 7.75 ). This most likely represents a contaminant from the alimentary tract.

Fig. 7.74 Entamoeba histolytica . In contrast to the Entamoeba gingivalis seen in Fig. 7.32 , the cytoplasm of E. histolytica contains red blood cells and not leukocytic debris. Vaginopancervical smear (Papanicolaou × HP).

Fig. 7.75 Cyst of Entamoeba histolytica . Vaginopancervical smear (Papanicolaou × HP).

Entamoeba gingivalis
These protozoa may be seen in vaginopancervical smears in association with Actinomyces . Such a relationship was documented by de Moraes-Ruehsen and associates. 45 These histiocytic-type organisms have numerous ingested fragments of leukocytic nuclei and cellular debris in the cytoplasm. They do not contain the intracytoplasmic red blood cells commonly observed in E. histolytica infection.

Filariasis
The most common nematode causing vascular space infection is Wuchereria bancrofti . Microfilariae may sometimes be seen in vaginal smears in endemic areas. 126 The microfilariae are generally 175–260 nm in length. They have a pointed tail with an elongated terminal nucleus. A 5- to 15-nm caudal space is present beyond the nucleus terminally.

Ascariasis
This nematode infection is caused by Ascaris lumbricoides . Although alimentary in habitat, the organism may migrate to various parts of the body and be seen in pulmonary and vaginal specimens. In vaginal smears, both unfertilized and fertilized ova may be observed. 127

Cysticercosis
Hymenolepis nana is a cestode that causes alimentary tract symptoms. The eggs may occur in the vaginal smear as a contaminant from the feces. 128 Eggs of the various tapeworms can be distinguished by detailed morphologic examination.

Arthropods
Water fleas or mites, but more commonly human body louse Pediculus humanus or human pubic louse Phthirus pubis may be seen as contaminants in the vaginal smear.
Detached ciliary tufts ( Fig. 7.76 ) can masquerade as parasites and have been reported as “ciliated protozoa.” The appearance of DCT is a physiologic occurrence when ciliated tufts are shed in the vaginal smear. 128a

Fig. 7.76 Detached ciliary tuft (DCT). Vaginopancervical smear (Papanicolaou × OI).
Other organisms including Schistosoma , Toxoplasma , varicella, Balantidium coli , and Turbatrix aceti (vinegar eels) have been observed in vaginal smears.

Acquired Immunodeficiency Syndrome (AIDS)
Most of the parasitic infections, although uncommon in Western populations, are being observed with increased frequency among patients with AIDS. Single and multiple infections occur in women who are positive for the human immunodeficiency virus (HIV) and who have a history of intravenous drug abuse or heterosexual relationships. An association of HIV-induced immunosuppression with HPV infections and CIN has more recently been reported. 129 HPV-associated lesions occur more commonly among HIV-positive women. There appears to be correlation between the high-grade cervical intraepithelial changes and lower CD4 values.

Concluding Remarks
In this chapter we document the cytomorphologic features and identification of the various female genital tract infections. Genital tract infections are far more common than the neoplasm, and cause considerable morbidity and economic loss. Personal hygiene, sexual habits, self-medication, and contraceptive practices contribute to the occurrence of many of these diseases and associated cellular changes. While an accurate diagnosis of these infections ( Candida ) can reduce complications and morbidity, familiarity and recognition of cytomorphologic features (IUD, herpes) is critical in proper interpretation and diagnosis of many epithelial neoplastic processes. International Academy of Cytology recommended the identification of many of these infections as “an integral part of the general diagnostic workup and of good patient care.” 130
Some uncommon neoplastic lesions (Langerhans histiocytosis-X, pemphigus) have been included; they have morphologic features that mimic infectious and neoplastic processes and an awareness and accurate diagnosis is critical. A number of uncommon infections are included not only for a comprehensive coverage but to emphasis the changing environment and emerging geographic cytopathology.

References

1. Eckert L.O. Clinical practice. Acute vulvovaginitis. N Engl J Med . 2006;355:1244-1252.
2. Spiegel C.A., Eschenbach D.A., Amsel R., et al. Curved anaerobic bacteria in bacterial (nonspecific) vaginosis and their response to antimicrobial therapy. J Infect Dis . 1983;148:817-822.
3. Larsen B., Galask R.P. Vaginal microbial flora: composition and influences of host physiology. Ann Intern Med . 1982;96:926-930.
4. Takei H., Ruiz B., Hicks J. Cervicovaginal flora. Comparison of conventional pap smears and a liquid-based thin-layer preparation. Am J Clin Pathol . 2006;125:855-859.
5. Wied G.L., Bahr G.F. Vaginal, cervical and endocervical cytologic smears on a single slide. Obstet Gynecol . 1959;14:362-367.
6. Hollander D.H., Gupta P.K. Hematoidin cockleburrs in cervico-vaginal smears. Acta Cytol . 1974;18:268-269.
7. Frost J.K., editor. Concepts Basis to General Cytopathology, 4th edn, Baltimore: Johns Hopkins Press, 1972.
8. Geirsson G., Woodworth F.E., Patten S.F.Jr., et al. Epithelial repair and regeneration in the uterine cervix. I. An analysis of the cells. Acta Cytol . 1977;21:371-378.
9. Ramzy I., editor. Clinical Cytopathology and Aspiration Biopsy. Fundamental Principles and Practice. Norwalk, CT: Appleton and Lange, 1989.
10. Frost J.K. The cell in health and disease. An evaluation of cellular morphologic expression of biologic behavior. 2nd, revised edition. Monogr Clin Cytol . 1986;2:1-304.
11. Novotny D.B., Maygarden S.J., Johnson D.E., et al. Tubal metaplasia. A frequentpotential pitfall in the cytologic diagnosis of endocervical glandular dysplasia on cervical smears. Acta Cytol . 1992;36:1-10.
12. Saqi A., Pasha T.L., McGrath C.M., et al. Overexpression of p16INK4A in liquid-based specimens (SurePath) as marker of cervical dysplasia and neoplasia. Diagn Cytopathol . 2002;27:365-370.
13. Bibbo M., Wied G., editors. Cytology of Inflammatory Reactions, Tissue Repair, Effects of IUD, Contaminants and Microbiologic Classification Including Chlamydial Organisms, 3rd edn, Chicago: Tutorials of Cytology, 1982.
14. Wied G., Bibbo M. Microbiologic classification of the cellular sample. In Management of Patients with Vaginal Infections: An Invitational Symposium. J Reprod Med . 1972;9:1-16.
15. Gardner H.L., Dukes C.D. Hemophilus vaginalis vaginitis. Ann N Y Acad Sci . 1959;83:280-289.
16. Blackwell A., Barlow D. Clinic diagnosis of anaerobic vaginosis (non-specific vaginitis). A practical guide. Br J Vener Dis . 1982;58:387-393.
17. Hillier S.L. The complexity of microbial diversity in bacterial vaginosis. N Engl J Med . 2005;353:1886-1887.
18. Gardner H.L., Dukes C.D. Haemophilus vaginalis vaginitis: a newly defined specific infection previously classified non-specific vaginitis. Am J Obstet Gynecol . 1955;69:962-976.
19. Fredricks D., Marrazzo J. Bacteria associated with bacterial vaginosis. N Eng J Med . 2006;354:202-203.
20. Petersdorf R.G., Curtin J.A., Hoeprich P.D., et al. A study of antibiotic prophylaxis in unconscious patients. N Engl J Med . 1957;257:1001-1009.
21. Ledger WJ, Campbell C, Willson JR. Postoperative adnexal infections. Obstet Gynecol 968;31:83-89.
22. Fredricks D.N., Fiedler T.L., Marrazzo J.M. Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med . 2005;353:1899-1911.
23. Wynn R., editor. Vaginitis and Related Problems. Proceedings of a Symposium Sponsored by the Department of Obstetrics and Gynecology, University of Arkansas for Medical Science. Norwalk, CT: The Purdue Frederick Company, 1980.
24. Schnadig V., Davie K., Shafer S. The cytologist and bacterioses of the vaginal-ectocervical area: Clues, commas and confusion. Acta Cytol . 1988;33:287-297.
25. Shands K.N., Schmid G.P., Dan B.B., et al. Toxic-shock syndrome in menstruating women: association with tampon use and Staphylococcus aureus and clinical features in 52 cases. N Engl J Med . 1980;303:1436-1442.
26. Hjelm E., Hallen A., Forsum U. Motile anaerobic curved rods in nonspecific vaginitis. Eur J Sex Transmit Dis . 1982;1:9.
27. Spiegel C.A., Amsel R., Eschenbach D., et al. Anaerobic bacteria in nonspecific vaginitis. N Engl J Med . 1980;303:601-607.
28. Horowitz B.J., Mardh P.A., Nagy E., et al. Vaginal lactobacillosis. Am J Obstet Gynecol . 1994;170:857-861.
29. Korenek P., Britt R., Hawkins C. Differentiation of the vaginoses—bacterial vaginosis, lactobacillosis and cytolytic vaginosis. Internat J Adv Nurs Pract . 2003;6:1-20.
30. Shirley R.L. Acute vulvovaginitis. N Engl J Med . 2006;355:2791. author reply 2791
31. Murphy R. Desquamative inflammatory vaginitis. Dermatol Ther . 2004;17:47-49.
32. Wright C., Pipingas A., Grayson W., et al. Pemphigus vulgaris of the uterine cervix revisited: case report and review of the literature. Diagn Cytopathol . 2000;22:304-307.
33. Kapila K., Verma K. Intracellular bacilli in vaginal smears in a case of malacoplakia of the uterine cervix. Acta Cytol . 1989;33:410-411.
33a. Angrish K., Verma K. Cytologic detection of tuberculosis of the uterine cervix. Acta Cytol . 1981;25(2):160-162.
34. Qualman S.J., Gupta P.K., Mendelsohn G. Intracellular Escherichia coli in urinary malakoplakia: a reservoir of infection and its therapeutic implications. Am J Clin Pathol . 1984;81:35-42.
35. Case record of the Massachuttes General Hospital. N Eng J Med . 1970;282:862-868.
36. Thomas R., Barnhill D., Bibro M., et al. Histiocytosis-X in gynecology: a case presentation and review of the literature. Obstet Gynecol . 1986;67:46S-49S.
37. Kumari T.R., Kini U. Unraveling the mystique case of pseudotumor syndrome of pelvis. Diagn Cytopathol . 2007;35:67-70.
38. Gupta P.K. Intrauterine contraceptive devices: vaginal cytology, pathologic changes and clinical implications. Acta Cytol . 1982;26:571-613.
39. Burkman R., Schlesselman S., McCaffrey L., et al. The relationship of genital tract actinomycetes and the development of pelvic inflammatory disease. Am J Obstet Gynecol . 1982;143:585-589.
40. Luff R.D., Gupta P.K. Actinomycetes-like organisms in wearers of intrauterine contraceptive devices. Am J Obstet Gynecol . 1977;129:476-477.
41. de la Monte S.M., Gupta P.K., White C.L.3rd. Systemic Actinomyces infection. A potential complication of intrauterine contraceptive devices. JAMA . 1982;248:1876-1877.
42. Hager W.D., Majmudar B. Pelvic actinomycosis in women using intrauterine contraceptive devices. Am J Obstet Gynecol . 1979;133:60-63.
43. Gupta P.K., Hollander D.H., Frost J.K. Actinomycetes in cervico-vaginal smears: an association with IUD usage. Acta Cytol . 1976;20:295-297.
44. Shurbaji M.S., Gupta P.K., Myers J. Immunohistochemical demonstration of Chlamydial antigens in association with prostatitis. Mod Pathol . 1988;1:348-351.
45. de Moraes-Ruehsen M., McNeill R., Frots J. Amebae resembling Entamoeba gingivalis in the genital tracts of IUD users. Acta Cytol . 1980;24:413-420.
46. Bibbo M., Harris M., Wied G., editors. Microbiology and Inflammation of the Female Genital Tract. Tutorials of Cytology: Chicago, 1976.
47. Gupta P.K., Burroughs F., Luff R.D., et al. Epithelial atypias associated with intrauterine contraceptive devices (IUD). Acta Cytol . 1978;22:286-291.
48. Graham R., editor. The Cytologic Diagnosis of Cancer, 3rd edn, Philadelphia: WB Saunders, 1972.
48a. Lipshutz B. Untersuchungen über die aetiologic der krankeiten der herpesgruppe. Arch Derm Syph (Berlin) . 1921:136-428.
49. Mardh P.A., Stormby N., Westrom L. Mycoplasma and vaginal cytology. Acta Cytol . 1971;15:310-315.
50. Roberts T.H., Ng A.B. Chronic lymphocytic cervicitis: cytologic and histopathologic manifestations. Acta Cytol . 1975;19:235-243.
51. Halford J.A. Cytological features of chronic follicular cervicitis in liquid-based specimens: a potential diagnostic pitfall. Cytopathology . 2002;13:364-370.
52. Gupta P.K., Lee E.F., Erozan Y.S., et al. Cytologic investigations in Chlamydia infection. Acta Cytol . 1979;23:315-320.
53. Taylor-Robinson D., Thomas B.J. The role of Chlamydia trachomatis in genital-tract and associated diseases. J Clin Pathol . 1980;33:205-233.
54. Hollander D.H., Gupta P.K. DCT, CCP and pseudoprotozoa. Acta Cytol . 1979;23:258-259.
55. Schneweis K.E. [On the antigenic structure of herpes simplex virus.]. Z Immun exp ther . 1962;124:173-196.
56. McDougall J., editor. In Situ Cytological Hybridization to Detect Herpes Simplex Virus RNA in Human Tissue. New York: Plenum, 1979.
57. Anthony D.D., Wentz W.B., Reagan J.W., et al. Induction of cervical neoplasia in the mouse by herpes simplex virus type 2 DNA. Proc Natl Acad Sci USA . 1989;86:4520-4524.
58. Aurelian L., Gupta P.K., Frost J.K., et al. Fluorescence-activated separation of cervical abnormal cells using herpesvirus antigenic markers. Anal Quant Cytol . 1979;1:89-102.
59. Gilman S.C., Docherty J.J., Clarke A., et al. Reaction patterns of herpes simplex virus type 1 and type 2 proteins with sera of patients with uterine cervical carcinoma and matched controls. Cancer Res . 1980;40:4640-4647.
60. Gupta P.K., Aurelian L., Frost J.K., et al. Herpesvirus antigens as markers for cervical cancer. Gynecol Oncol . 1981;12:S232-258.
61. Jones K.W., Fenoglio C.M., Shevchuk-Chaban M., et al. Detection of herpes simplex virus type 2 mRNA in human cervical biopsies by in situ cytological hybridization. ARC Sci Publ . 1978:917-925.
62. Kessler II. Venereal factors in human cervical cancer: evidence from marital clusters. Cancer . 1977;39:1912-1919.
63. Vesterinen E., Purola E., Saksela E., et al. Clinical and virological findings in patients with cytologically diagnosed gynecologic herpes simplex infections. Acta Cytol . 1977;21:199-205.
64. Corey L., editor. Genital Herpes. New York: McGraw Hill, 1984.
65. Lipschutz B., editor. Untersuchunger uber die aetiologic der krankheiten der herpesgruppe. Berlin: Arch Derm Syph, 1921, 136.
66. US Department of Health and Human Services. Public Health Service. Sexually Transmitted Disease Fact Sheet , 39th edn. Atlanta: Centers for Disease Control; 1981.
67. Patten SJ (ed.). Diagnostic Cytopathology of the Uterine Cervix , 2nd edn. 1978
68. Kumar M.L., Gold E., Jacobs I.B., et al. Primary cytomegalovirus infection in adolescent pregnancy. Pediatrics . 1984;74:493-500.
69. Griffiths P.D., Campbell-Benzie A., Heath R.B. A prospective study of primary cytomegalovirus infection in pregnant women. Br J Obstet Gynaecol . 1980;87:308-314.
70. Naib Z., editor. Exfoliative Cytopathology, 3rd edn, Boston: Little Brown, 1985.
71. Koutsky L. Epidemiology of genital human papillomavirus infection. Am J Med . 1997;102:3-8.
72. Meisels A., Fortin R. Condylomatous lesions of the cervix and vagina. I. Cytologic patterns. Acta Cytol . 1976;20:505-509.
72a. Muñoz N., Bosch F.X., de Sanjosé S., et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med . 2003;348(6):518-527.
72b. Cogliano V., Baan R., Straif K., et al. WHO International Agency for Research on Cancer. Carcinogenicity of human papillomaviruses. Lancet Oncol . 2005;6(4):204.
73. Koss L., Melamed M., editors. Koss Diagnostic Cytology and Its Histopathologic Bases, 5th edn, Philadelphia: Lippincott Williams and Wilkins, 2006.
74. Broker T., Butcher M., editors. Papillomaviruses: Retrospectives and prospectives. Cold Spring Harbor, NY: Cold Spring Harbor Publications, 1986.
75. Bosch F.X., Manos M.M., Munoz N., et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group. J Natl Cancer Inst . 1995;87:796-802.
76. Walboomers J.M., Jacobs M.V., Manos M.M., et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol . 1999;189:12-19.
77. Sherman M.E., Lorincz A.T., Scott D.R., et al. Baseline cytology, human papillomavirus testing, and risk for cervical neoplasia: a 10-year cohort analysis. J Natl Cancer Inst . 2003;95:46-52.
78. zur Hausen H. Papillomaviruses in human cancer. Cancer . 1987;59:1692-1696.
79. Crum C.P., Berkowitz R.S. Human papillomaviruses. Applications, caveats and prevention. J Reprod Med . 2002;47:519-528. discussion 528-529
80. Duensing S., Lee L.Y., Duensing A., et al. The human papillomavirus type 16 E6 and E7 oncoproteins cooperate to induce mitotic defects and genomic instability by uncoupling centrosome duplication from the cell division cycle. Proc Natl Acad Sci USA . 2000;97:10,002-10,007.
81. Heselmeyer K., Schrock E., du Manoir S., et al. Gain of chromosome 3q defines the transition from severe dysplasia to invasive carcinoma of the uterine cervix. Proc Natl Acad Sci USA . 1996;93:479-484.
82. Klingelhutz A.J., Foster S.A., McDougall J.K. Telomerase activation by the E6 gene product of human papillomavirus type 16. Nature . 1996;380:79-82.
83. Munger K., Howley P.M. Human papillomavirus immortalization and transformation functions. Virus Res . 2002;89:213-228.
84. Khan M., Castle P., Lorincz A. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst . 2005;97:1072-1079.
85. Ayre J.E. The vaginal smear; precancer cell studies using a modified technique. Am J Obstet Gynecol . 1949;58:1205-1219.
86. Papanicolaou G., editor. Atlas of Exfoliative Cytology. Cambridge, MA: Harvard University Press, 1954. color plate A-IV
87. Koss L.G., Durfee G.R. Unusual patterns of squamous epithelium of the uterine cervix: cytologic and pathologic study of koilocytotic atypia. Ann NY Acad Sci . 1956;63:1245-1261.
88. Naib Z.M., Masukawa N. Identification of condyloma acuminata cells in routine vaginal smears. Obstet Gynecol . 1961;18:735-738.
89. Ayre J.E. Role of the halo cell in cervical cancerigenesis. A virus manifestation in premalignancy? Obstet Gynecol . 1960;15:481-491.
90. Sagiroglu N. Progression and regression studies of precancer (anaplastic or dysplastic) cells, and the halo test. Am J Obstet Gynecol . 1963;85:454-469.
91. De Girolami E. Perinuclear halo versus koilocytotic atypia. Obstet Gynecol . 1967;29:479-487.
92. Purola E., Savia E. Cytology of gynecologic condyloma acuminatum. Acta Cytol . 1977;21:26-31.
93. Laverty C.R., Russell P., Hills E., et al. The significance of noncondylomatous wart virus infection of the cervical transformation zone. A review with discussion of two illustrative cases. Acta Cytol . 1978;22:195-201.
94. Woodruff J.D., Braun L., Cavalieri R., et al. Immunologic identification of papillomavirus antigen in condyloma tissues from the female genital tract. Obstet Gynecol . 1980;56:727-732.
95. Gupta J.W., Gupta P.K., Rosenshein N., et al. Detection of human papillomavirus in cervical smears. A comparison of in situ hybridization, immunocytochemistry and cytopathology. Acta Cytol . 1987;31:387-396.
96. Gupta J.W., Gupta P.K., Shah K.V., et al. Distribution of human papillomavirus antigen in cervicovaginal smears and cervical tissues. Int J Gynecol Pathol . 1983;2:160-170.
97. Southern E.M. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol . 1975;98:503-517.
98. Brahic M., Haase A.T. Detection of viral sequences of low reiteration frequency by in situ hybridization. Proc Natl Acad Sci USA . 1978;75:6125-6129.
99. Stoler M., Broker T. In situ hybridization detection of human papillomavirus DNAs and messenger RNAs in genitali condylomas and a cervical carcinoma. Hum Pathol . 1986;17:1250-1258.
100. Lorincz A.T. Hybrid Capture method for detection of human papillomavirus DNA in clinical specimens: a tool for clinical management of equivocal Pap smears and for population screening. J Obstet Gynaecol Res . 1996;22:629-636.
101. Lorincz A.T. Molecular methods for the detection of human papillomavirus infection. Obstet Gynecol Clin North Am . 1996;23:707-730.
102. Lorincz A.T., Reid R., Jenson A.B., et al. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gynecol . 1992;79:328-337.
103. The 2001 Bethesda system for reporting cervical/vaginal cytological diagnoses. NCI Workshop . 2002.
104. Solomon D., Schiffman M., Tarone R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst . 2001;93:293-299.
105. Solomon D., Breen N., McNeel T. Cervical cancer screening rates in the United States and the potential impact of implementation of screening guidelines. CA Cancer J Clin . 2007;57:105-111.
106. Stoler M.H., Castle P.E., Solomon D., et al. The expanded use of HPV testing in gynecologic practice per ASCCP-guided management requires the use of well-validated assays. Am J Clin Pathol . 2007;127:335-337.
107. Wright T.C.Jr., Cox J.T., Massad L.S., et al. 2001 Consensus guidelines for the management of women with cervical cytological abnormalities. JAMA . 2002;287:2120-2129.
107a. Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: baseline data from a randomized trial. The Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. J Natl Cancer Inst . 2000;92(5):397-402.
108. Koutsky L.A., Ault K.A., Wheeler C.M., et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med . 2002;347:1645-1651.
109. Schneider A., Meinhardt G., De-Villiers E.M., et al. Sensitivity of the cytologic diagnosis of cervical condyloma in comparison with HPV-DNA hybridization studies. Diagn Cytopathol . 1987;3:250-255.
110. Katz R., Ferre F., Sneige N. Comparison of the sensitivity of cytomorphology, slot-blot hybridization and polymerase chain reaction for the detection of human papillomavirus in cervical brushing. Acta Cytol . 1989;33:696.
111. Gupta P.K., Pinn V.M., Taft P.D. Cervical dysplasia associated with azathioprine (Imuran) therapy. Acta Cytol . 1969;13:373-376.
112. Hubbard R.A. Human papillomavirus testing methods. Arch Pathol Lab Med . 2003;127:940-945.
113. Seabrook J.M., Hubbard R.A. Achieving quality reproducible results and maintaining compliance in molecular diagnostic testing of human papillomavirus. Arch Pathol Lab Med . 2003;127:978-983.
114. Montone K.T., Furth E.E., Pietra G.G., et al. Neonatal adenovirus infection: a case report with in situ hybridization confirmation of ascending intrauterine infection. Diagn Cytopathol . 1995;12:341-344.
115. Chlamydia trachomatis infection: Policy guidelines for prevention and control. In: Control CfD, ed. 34 vol: MMWR; 1985:53–74
116. Halberstaedter L., von Prowazek S. Uber zelleinschlusse parasitarer natur beim Trachom. Arb Gesundheitsa . 1907;26:44-47.
117. Schachter J., Dawson C., editors. Human Chlamydial Infections. Littleton, MA: PSG, 1978.
118. Thygeson P. The etiology of inclusion blennorrhea. Am J Ophthalmol . 1934;17:1019-1035.
119. Paavonen J., Brunham R., Kiviat N., editors. Cervicitis-etiologic, Clinical and Histopathologic Findings in Chlamydial Infections. New York: Elsevier, 1982.
120. Gupta P.K., Shurbaji M.S., Mintor L.J., et al. Cytopathologic detection of Chlamydia trachomatis in vaginopancervical (Fast) smears. Diagn Cytopathol . 1988;4:224-229.
121. Kiviat N.B., Paavonen J.A., Brockway J., et al. Cytologic manifestations of cervical and vaginal infections. I. Epithelial and inflammatory cellular changes. JAMA . 1985;253:989-996.
122. Quinn T.C., Gupta P.K., Burkman R.T., et al. Detection of Chlamydia trachomatis cervical infection: a comparison of Papanicolaou and immunofluorescent staining with cell culture. Am J Obstet Gynecol . 1987;157:394-399.
123. Kellogg J.A. Clinical and laboratory considerations of culture vs antigen assays for detection of Chlamydia trachomatis from genital specimens. Arch Pathol Lab Med . 1989;113:453-460.
124. Naib Z.M. Cytology of TRIC agent infection of the eye of newborn infants and their mothers genital tracts. Acta Cytol . 1970;14:390-395.
125. Shiina Y. Cytomorphologic and immunocytochemical studies of chlamydial infections in cervical smears. Acta Cytol . 1985;29:683-691.
125a. Gupta PK, Frost JK. Cytopathology and histopathology of the female genital tract in Trichomas vaginalis infection . Ch. 14 Trichanamonads Parasitic in Humans . 1989 Hingberg BM (ed) Springer-Verlag. New York.
125b. Lara-Torre E., Pinkerton J.S. Accuracy of detection of Trichomas vaginalis organisms on a liquid based Papanicolaou smear. Am J Obstet Gynecol . 2003;188:354-356.
126. Chandra K., Annousamy R. Letters to the editor: An unusual finding in the vaginal smear. Acta Cytol . 1975;19:403-404.
127. Bhambhani S. Eggs of ascaris lumbricoides in cervicovaginal smear. Acta Cytol . 1984;28:92.
128. Bhambhani S., Milner A., Pant J., et al. Ova of Taenia and Enterobius vermicularis in cervicovaginal smears. Acta Cytol . 1985;29:913-914.
128a. Hollander D.H., Gupta P.K. Detached ciliary tufts in cervico-vaginal smears. Acta Cytol . 1979;18:367-369.
129. Henry M.J., Stanley M.W., Cruikshank S., et al. Association of human immunodeficiency virus-induced immunosuppression with human papillomavirus infection and cervical intraepithelial neoplasia. Am J Obstet Gynecol . 1989;160:352-353.
130. International Academy of Cytology (IAC). Policy statement on the frequency of gynecologic screening. IAC Newsletter . 1981;8:29.
CHAPTER 8 Benign Proliferative Reactions, Intraepithelial Neoplasia, and Invasive Cancer of the Uterine Cervix

G. Peter Vooijs, Anniek J.M. van Aspert-van Erp, Johan Bulten

Contents
Introduction
The Normal Uterine Cervix
Histology and Cytology
Benign Proliferative Reactions
Hyperkeratosis
Parakeratosis
Squamous Metaplasia
Atypia
Reactive and Regenerative Changes
Squamous Intraepithelial Neoplasia
Terminology
Papanicolaou Classification
Dysplasia
Cervical Intraepithelial Neoplasia
Invasive Cancer of the Uterine Cervix
Epidemiology
Microinvasive Carcinoma
Invasive Cervical Cancer
Efficacy of Cervical Cytology in the Detection of Cervical Abnormalities
Liquid-Based Cervical Cytology
Screening Programs for Cervical Cancer
Observer Variability
Quality Control
Review
Concluding Remarks

Introduction
Diagnostic cytology of the uterine cervix was not the first application of cytology in clinical diagnosis of diseases, but it is definitely the most widespread and best known. In its early days, vaginal cytology was primarily directed at the diagnosis of invasive cancer of the uterine cervix and of the endometrium, but cytologists later began to realize that cervical lesions were best recognized in direct scrapes of the cervical mucosa. The concept that invasive carcinoma of the cervix is antedated by an intraepithelial neoplastic change—carcinoma in situ —was postulated at the beginning of the 20th century. 1 These intraepithelial changes were believed to be potentially progressive precursors of invasive cancer. It subsequently became known that the spectrum of abnormal changes of the epithelial lining of the uterine cervix was much wider than previously thought. Cytologists soon became aware that these noninvasive epithelial abnormalities of the uterine cervix could also be diagnosed in the direct scrapings. With meticulous comparison between the characteristics of the cells in cytologic smears and the histologic changes found in the same patients, the cytologic characteristic of intraepithelial lesions of the cervix became better defined and the accuracy of cervical cytology in predicting the histopathologic change improved.
All epithelial abnormalities of squamous character derive from ectocervical squamous basal cells and endocervical reserve cells. Depending on the strength of the negative stimulus on the differentiating and maturing basal cells and endocervical reserve cells and, in a later stage, immature metaplastic cells, the ultimately resulting cells have a more or less differentiated aspect. This explains not only common morphologic features in the different variants of dysplasia and ultimately in the most dedifferentiated intraepithelial variant, carcinoma in situ, but also the extremely common co-occurrence of carcinoma in situ and dysplastic changes of different severity.
Despite a relatively high proportion of incorrect diagnoses, in large-scale population-screening programs, cervical cytology has proved to be the most effective tool for the diagnosis of cervical cancer. After an initial increase in the number of severe epithelial lesions diagnosed in the population, the majority of lesions diagnosed are of slight-to-moderate severity. It is well known that the majority of these lesions regress to a less abnormal change after a variable time. In this respect, it is justified to monitor these changes cytologically, when the cytodiagnostic service is of high quality.
For correct interpretation of abnormalities, a thorough knowledge of the cytologic and histologic characteristics of cervical cancer and its potential precursor lesions is necessary.

The Normal Uterine Cervix

Histology and Cytology

Ectocervix
The vagina and the outer portion of the uterine cervix—the ectocervix—are lined with nonkeratinizing squamous epithelium. Embryologically, this epithelium is derived from a solid epithelial plate growing inward from the urogenital sinus up to the level of the later endocervical canal. This solid plate replaces, at the level of the vagina and the ectocervix, the primitive cuboidal epithelium from the fused Müllerian ducts, from which also originates the columnar epithelium, lining the fallopian ducts, the uterine corpus, and the endocervical part of the cervix. The fusion site of these two types of epithelium is called the squamocolumnar junction ( Fig. 8.1 ). The location of this fusion site varies considerably, depending on physiologic and pathologic conditions. During reproductive years, the squamocolumnar junction is located at the entrance of the endocervical canal, the external os , but owing to hormonal influences or as a result of stromal edema that has changed the endocervical canal or cervix, it may be located inside the endocervical canal or on the surface of the cervix. When located on the external surface of the cervix, part of the columnar cell lining of the endocervical canal is present on the ectocervical face. This causes a usually well-demarcated reddening of the surface that is easily recognizable at inspection. This outward bulging of the endocervical mucosa, referred to as ectropion , eversion , or false erosion , should be differentiated from a true erosion of the epithelium, which by definition is a loss of the lining mucosa, leaving the underlying cervical stroma barren. During childhood and after menopause, the squamocolumnar junction is located inside the canal.

Fig. 8.1 Squamocolumnar junction. Fusion site of the stratified nonkeratinizing squamous epithelium of the ectocervix and the mucus-producing columnar epithelium lining the endocervical canal (H&E × HP).
During reproductive years, the morphology of the stratified squamous epithelium, the function of which is primarily a protective one, is cyclically changing under the influence of ovarian hormones. In its fully matured stage, the epithelium of the vagina and ectocervix can be subdivided into several layers ( Fig. 8.2 ). For correlation with the different cell types present in cervical cytologic specimens, a subdivision into three layers is most practical. Beginning with the deepest layer, these are the (1) basal cell and parabasal cell layer, (2) intermediate cell layer, and (3) superficial cell layer. Under physiologic conditions, epithelial regeneration takes place in the basal layer, usually composed of a single layer of relatively primitive cells with scarce cytoplasm and large oval-to-round nuclei with prominent nucleoli. Under normal conditions, true basal cells are not present in cervical smears unless, when the smear is taken, the entire epithelial layer has been removed, leaving the underlying stroma denuded. Parabasal cells are unusual in smears from women in reproductive years, unless certain pathologic processes, particularly a reduced or absent estrogenic hormonal stimulation, occur. They are the predominant cell type in smears from women in postmenopause and from children.

Fig. 8.2 Normal mature stratified nonkeratinizing squamous epithelium, covering the ectocervical surface (H&E × HP).
Parabasal cells are round to oval and have a small cytoplasmic body and relatively large oval-to-round nuclei ( Fig. 8.3 ; see also Fig. 8.16 ). The cytoplasm is relatively dense, has distinct borders, may contain vacuoles, and commonly stains cyanophilic. After drying of the vaginal or cervical surface due to absence of estrogenic stimulation or after exposure to air, the cytoplasm stains eosinophilic.

Fig. 8.3 Parabasal cells. Round-to-oval cells with relatively large nuclei and scant dense, cyanophilic cytoplasm. Admixture of leukocytes is due to inflammation (Papanicolaou × HP).
The major part of the thickness of the epithelium is formed by the intermediate cell layer. With maturation of the cells toward the surface, the cells become better differentiated, reflected by an increase in the cytoplasmic volume and signs of specific functional qualities of the cytoplasm, such as storage of glycogen or secretory products. Cells are bound to each other by intracellular bridges called desmosomes. Cytoplasm of intermediate squamous cells is cyanophilic in staining. Nuclei become only slightly reduced in size during the passage of the cells through the intermediate cell layers. The round-to-oval nuclei have a diameter of about 8 to 10 μm, have a clearly defined nuclear membrane ( Fig. 8.4 ; see also Fig. 8.18 ), and contain an evenly distributed, finely granular chromatin.

Fig. 8.4 Normal superficial and intermediate squamous cells desquamated from normal nonkeratinizing squamous epithelium (Papanicolaou × OI).
The most superficial layers are composed of fairly large polygonal cells loosely attached to each other. These cells do not proliferate and represent an end stage in the maturation process of nonkeratinizing stratified squamous epithelium. Intercellular attachments—desmosomes—become loose, and cells constantly exfoliate from the surface. Cells are polygonal and have a clear, translucent, usually pink-staining, occasionally cyanophilic cytoplasm; sharply defined boundaries; and a small, often pyknotic central nucleus. Superficial cells have a diameter of approximately 40 μm. The nuclear diameter is 3 to 5 μm. The eosinophilic staining of the cytoplasm is caused by the presence of prekeratin proteins. Fibrillary strands of keratin can sometimes be recognized in the cytoplasm of the most superficial eosinophilic staining cells. Rarely, superficial cells may contain keratohyalin granules in the cytoplasm. These granules are supposedly derived from the granular cell layer of an altered—keratinized—stratified squamous epithelium.
The entire maturation cycle of a normal squamous cell takes approximately 4 days. The maturation process can be accelerated significantly under estrogenic stimulation. Although the influence of estrogenic hormones is predominantly evident on the vaginal epithelium, the ectocervical squamous and endocervical columnar epithelium, contrary to what is often stated, also show a cellular reaction to ovarian hormones. During the reproductive years, reduced maturation of the cervical squamous epithelium may be found as a result of the action of oral hormonal contraceptives. Before reproductive age, the start of which is marked by the first menstrual period, or menarche, the epithelial lining of the vagina and ectocervix is relatively thin and composed of less mature cells of a parabasal type ( Fig. 8.5 ). Cells do not contain glycogen. Also, during pregnancy, as a result of the relative predominance of progesterone, maturation of the epithelium becomes reduced. The epithelial lining is then composed of intermediate-type cells, with cyanophilic-staining cytoplasm, often with a pronounced outer zone. These rather characteristic cells are also called navicular cells . The underlying stroma sometimes shows a large-cell—decidual—reaction, as is physiologic in the endometrial stroma ( Fig. 8.6 ). These large stromal cells, when present in smears, may cause differential diagnostic problems with cells from invasive processes, particularly adenocarcinomas, because of large, prominent nucleoli ( Fig. 8.7 ). Their occurrence in sheets and as single cells, not in clusters, together with the evenly distributed, finely granular nuclear chromatin, against the background of the pregnancy, should provide the correct diagnosis.

Fig. 8.5 Atrophy of ectocervical squamous epithelium. Epithelial layer is reduced in thickness and composed almost entirely of immature parabasal-type cells (H&E × HP).

Fig. 8.6 Atrophy of the ectocervical squamous epithelium during pregnancy. Decidual change of cervical stromal cells (H&E × MP).

Fig. 8.7 Large stromal cells in a cervical smear during pregnancy show large nuclei with prominent nucleoli and an evenly distributed finely granular chromatin (Papanicolaou × MP).
Also, after the last menstrual period—menopause—the epithelium reduces in thickness and again becomes atrophic. The epithelial cells do not contain glycogen, and the epithelium in its atrophic state becomes highly vulnerable to even slight trauma and often shows signs of inflammation (see Figs 8.3 , 8.16 , and 8.17 ). An absolute or relative reduction in the level of circulating estrogenic hormones leads to lysis of the cytoplasm of vaginal and cervical squamous cells, as is evident during the second, or luteal, phase of the menstrual cycle, during pregnancy, during lactation, and after menopause. Cytolysis may be conspicuous, resulting in a large number of bare nuclei throughout the smear ( Fig. 8.8 ).

Fig. 8.8 Cytolysis. Bare nuclei are due to lysis of the cytoplasm. Döderlein bacilli are conspicuous (Papanicolaous × HP).
Continuous stimulation by estrogenic hormones causes hypertrophy of the superficial layers of the ectocervical squamous epithelium and induces hypersecretion of the endocervical columnar epithelium. The superficial squamous layers are composed of large polygonal cells with clear, slightly eosinophilic staining cytoplasm and very small, almost completely pyknotic nuclei. The presence of these cells in smears from women in postmenopause, particularly when endometrial cells are also found, should be a warning signal of an endometrial abnormality induced by the continuous estrogenic growth stimulus and requires additional diagnostic evaluation.

Endocervix
The epithelial lining of the endocervical canal is formed by a single layer of tall columnar cells ( Fig. 8.9 ). Owing to an arrangement of the nuclei on different levels and variability in size of the columnar cells, this epithelial lining often has a pseudostratified appearance. The surface of the endocervical canal is irregularly shaped, with invaginations extending up to 8 mm deep into the cervical stroma ( Fig. 8.10 ). In tissue sections, these invaginations are often tangentially cut and appear as round or oval epithelial stromal inclusions. They are then often referred to as endocervical glands , even though the uterine cervix does not contain glands in the strict sense. The lining columnar epithelial cells are the source of the mucus that covers the epithelium and composes the mucus plug that fills the endocervical canal. The configuration of the epithelium, the amount of mucus produced, and the consistency of the mucus vary with the hormonal status.

Fig. 8.9 Mucus-producing columnar cells lining the endocervical canal (H&E × HP).

Fig. 8.10 Low-power view of the uterine cervix. Deep stromal invaginations of endocervical mucosa greatly increase the surface of the mucus-producing columnar epithelium (H&E × LP).
In a normal cervical smear, endocervical columnar cells appear as tall columnar cells with a large body of clear, finely or coarsely vacuolated, faintly cyanophilic, clear cytoplasm ( Fig. 8.11 ). Ciliated cells may be present. Endocervical cells usually are arranged in strips of parallel-arranged cells or in tight sheets. When present in sheet-like arrangements, cell boundaries may create a honeycomb pattern ( Fig. 8.12 ). It is not unusual to find stripped, bare nuclei due to lysis of the fragile cytoplasm. Basally located nuclei are round to oval and often rather variable in size, with a finely granular, evenly distributed chromatin. In most nuclei, one or two small nucleoli can be observed. In cases of increased proliferative activity of the endocervical epithelium, the cells and their nuclei show a rather wide variation in size and shape and nucleoli may become prominent and variable in size. Multinucleation of cells is not uncommon. In well-preserved cytologic specimens, ciliated columnar cells may be recognized. The morphologic variants of ciliated and mucus-producing cells refer to the mullerian origin of the endocervical epithelium. The presence of ciliated cells has no special significance. It is therefore no sign of an epithelial abnormality.

Fig. 8.11 Superficial squamous cell and tall columnar cells with eccentrically located nuclei and abundant cytoplasm. Ciliated cells may be present (Papanicolaou × OI).

Fig. 8.12 Sheet of columnar cells. Sharply outlined cytoplasmic boundaries create a honeycomb pattern (Papanicolaou × OI).
In tissue sections, a single layer of primitive cells can often be observed beneath the columnar cell layer. These cells, also called reserve cells , are thought to be multipotential germinative cells, which under physiologic conditions produce normal endocervical columnar cells ( Fig. 8.13 ). In pathologic states, reserve cells may proliferate and, depending on the severity of the stimulus, produce abnormal, less well-differentiated columnar cells or, through the process of metaplasia, squamous metaplastic cells.

Fig. 8.13 Single-layer of reserve cells beneath endocervical columnar cells (H&E × MP).

Immunocytochemistry of Normal Cervical Epithelium
The structure and shape of a cell are maintained by an internal cytoplasmic structure, the cytoskeleton. In this cytoskeleton, on the basis of their ultrastructural appearance and biochemical composition, three different types of filaments can be distinguished. Next to microtubules and microfilaments, filaments measuring 8 to 11 nm in diameter are commonly seen in mammalian cells. These so-called intermediate-sized filaments , which are extremely insoluble and have a biochemical composition entirely different from that of microtubules and microfilaments, often constitute a considerable part of the intracellular matrix. There are five types of intermediate filaments, and they occur in tissue-specific combinations. 2, 3
Keratins have been recognized as epithelium-specific intermediate filament proteins and as comprising a family of at least 19 different polypeptides (not including the hair keratins). The tissue-specific intermediate filament proteins are retained during malignant transformation. Tumors of epithelial origin thus retain cytokeratin as the structural protein for the intermediate filaments. 3 Keratin immunocytochemistry has proved to be a valuable additional technique in the routine diagnosis of cancers that pose problems on morphologic examination. Broad-spectrum monoclonal antibodies can be used to separate epithelial tissues from nonepithelial tissues. Combinations of the 19 different keratin proteins are distributed in a more or less tissue-specific fashion as initially detected by two-dimensional gel electrophoresis. 2
Mainly based on gel electrophoretic studies, normal ectocervical epithelium was found to contain keratins 1, 4, 5, 6, 13–15, and 19, with some variability in the expression of keratins 2, 8, 10, 11, 16, and 17. Endocervical cells contain keratins 7, 8, 18, and 19, with variability in the expression of keratin 4. 4 Reserve cells show an unequivocal distribution of keratin 18 and contain keratins 5, 7, 8, and 14–19. Immature squamous metaplasia has a keratin expression pattern that on the one hand is characteristic of endocervical columnar cells and on the other hand characteristic of an epithelium that has undergone squamous differentiation. This change becomes emphasized when we compare immature squamous metaplasia with mature squamous metaplasia. The pattern of keratin expression in immature squamous metaplasia was shown to differ from that in normal squamous epithelium. Keratin 19 is present in the full thickness of immature squamous metaplastic epithelium, as opposed to normal squamous epithelium, in which only the basal cell component reacts positively. Keratins 8 and 18, indicative of a columnar differentiation of the cells, become absent. The expression of keratins 4, 10, 13, and 14 increases with squamous differentiation ( Figs 8.14 and 8.15 ). 5

Fig. 8.14 Expression of keratin 4 in intermediate and superficial cell layers of mature stratified nonkeratinizing squamous epithelium. Basal cell layers do not stain (immunoperoxidase method × LP).

Fig. 8.15 Expression of keratin 10 indicating keratin production in the intermediate cell layer of stratified nonkeratinizing squamous epithelium (immunoperoxidase method × MP).
Puts and co-workers studied the presence of vimentin-positive cells present in normal ectocervical and endocervical epithelium, subcolumnar reserve cell hyperplasia, and squamous metaplastic and dysplastic epithelium of the uterine cervix. 6 They demonstrated a relatively large number of vimentin-positive and Langerhans cells in normal ectocervical stratified squamous metaplastic epithelium, a small number in endocervical columnar epithelium, and a larger number in subcolumnar reserve cell hyperplasia and in immature squamous metaplasia. Mature squamous metaplastic epithelium showed a great resemblance to normal ectocervical stratified squamous epithelium, in both numbers and distribution of Langerhans cells.

Atrophy
A smear from an atrophic epithelium usually does not cause diagnostic problems. Cells are of the basal–parabasal cell type, with a high nucleocytoplasmic ratio. Cells are often arranged in syncytia with indistinct cell borders. Nucleoli are usually absent.
In cases of inflammation and atrophy (senile vaginitis), cell changes due to infection and degeneration may cause diagnostic problems ( Figs 8.16 and 8.17 ). Nuclear chromatin becomes coarsely granular and hyperchromatic. Owing to erosion or ulceration of the superficial stromal layers, regeneration of the epithelium is induced. From these parabasal-type cells with relatively large nuclei, prominent nucleoli may appear. 7 In these cases, differentiation from an epithelial abnormality may become difficult and at times virtually impossible. However, a short course of locally applied or oral estrogenic hormones induces maturation ( Fig. 8.18 ). Because epithelial abnormalities do not react to the estrogenic stimulus, or at least not to the same degree as normal epithelia do, abnormal cells stand out clearly and diagnosis can be readily made. In our material, after a short course of oral estrogenic hormones, the number of smears with significant drying artifacts was reduced from 66 to 32% and the percentage of smears with a marked-to-moderate inflammatory exudate from 73 to 55%.
Key features of atrophy
• Smear composed of parabasal type cells;
• Cells arranged in aggregates with indistinct cell borders (syncytia-like);
• Cytoplasm is cyanophilic to amphophilic;
• Round to oval nuclei;
• Relatively high nucleus-to-cytoplasmic ratio;
• Coarsely granular chromatin with hyperchromasia; and
• Nucleoli not present.

Fig. 8.16 Parabasal-type cells and leukocytes in a cervical smear taken after menopause (Papanicolaou × OI).

Fig. 8.17 Cervical smear taken after menopause with evidence of inflammation, degenerative nuclear changes, and drying artifacts (Papanicolaou × MP).

Fig. 8.18 Superficial and intermediate squamous cell in a cervical smear after a short course of estrogenic hormones because of epithelial atrophy postmenopause (same case as that in Figs 8.16 and 8.17 ) (Papanicolaou × HP).

Benign Proliferative Reactions

Hyperkeratosis
The covering epithelium of the vagina and ectocervix apparently still has the potential for further “differentiation,” as is demonstrated when this epithelium comes under the influence of chronic, rather severe stimulation. An example of such a chronic stimulation is descensus uteri (prolapse of the uterus), but it may also occur with inflammatory processes or as a reaction to hyperestrinism of long duration. The epithelium increases its protective role by increasing its overall thickness (acanthosis). In addition, a granular layer and the development of several layers of keratinized cells—hyperkeratosis—may occur ( Fig. 8.19 ). Hyperkeratosis implies excessive formation of keratin over the surface of the stratified squamous epithelium. It should be emphasized that keratinization of the stratified squamous epithelium of the vagina and cervix represents an abnormal differentiation. At clinical examination, this area may appear as a white patch, a sign of leukoplakia. In the cytologic smear, leukoplakia can be recognized by the presence of numerous anucleated squames found singly or in sheets 8 ( Fig. 8.20 ). These are often folded and are pale yellowish pink. Remnants of nuclei may be visible as a central clear zone, so-called nuclear ghosts. Cells from the granular cell layer may be encountered in the smear, resembling intermediate or superficial squamous cells, containing eosinophilic or cyanophilic keratohyalin cytoplasmic granules.

Fig. 8.19 Hyperkeratosis. Multiple layers of keratin composing the surface of the squamous epithelium. A granular cell layer has been formed beneath the keratin layers (H&E × MP).

Fig. 8.20 Anucleated squames desquamated from the superficial layers of keratinized—hyperkeratotic—squamous epithelium (Papanicolaou × HP).

Parakeratosis
Parakeratosis is another protective reaction of the nonkeratinizing squamous epithelium of the genital tract, characterized by the presence of various numbers of layers of small squamous cells, sharply demarcated from the underlying superficial zone. The nuclei are small, frequently pyknotic, and hyperchromatic.
In cytologic specimens, cells from parakeratosis appear as relatively small, superficial squamous cells, either isolated or in sheets ( Fig. 8.21 ). Shapes vary from round or oval to polygonal or spindle shaped. Cytoplasmic staining usually is dark or light eosinophilic, rarely cyanophilic.

Fig. 8.21 Parakeratosis. Small superficial squamous cells with sharply outlined cytoplasmic borders and small, often pyknotic nuclei (Papanicolaou × MP).
Nuclei are small and often hyperchromatic owing to pyknosis. Although hyperkeratosis and parakeratosis are usually associated with a relatively mature squamous epithelium, a counterpart may overlay an abnormal change such as dysplasia or squamous cell carcinoma. Patients with cellular evidence of hyperkeratosis or parakeratosis should be reexamined to preclude a more serious lesion camouflaged by the overlying hyperkeratotic or parakeratotic epithelial layer. Physicians should be advised to take two smears in succession. The first scrape is intended to remove the superficial, abnormally keratinized layers. In the material obtained with the second scrape, the true nature of the underlying epithelium becomes apparent.
Key features of parakeratosis and hyperkeratosis
• Isolated cells or large sheet-like aggregates with anucleate squames;
• Eosinophilic cytoplasm;
• Small round-to-oval nuclei, frequently pyknotic;
• Slight to moderate hyperchromasia; and
• Nucleoli not present.

Squamous Metaplasia
The most common protective mechanism of the endocervical epithelium of the uterine cervix is squamous metaplasia. The term metaplasia implies the transformation of one cell type into another type of cell, the latter being of a lower organizational order. As applied to the uterine cervix, the term refers to the process of replacement of simple columnar epithelium lining the endocervical canal and glands by a stratified squamous epithelium.
Squamous metaplasia may be arbitrarily subdivided into the following:
• Reserve cell hyperplasia;
• Immature squamous metaplasia; and
• Mature squamous metaplasia.
Reserve cell hyperplasia is transformed in immature squamous metaplasia, which with increasing differentiation gradually evolves into mature squamous metaplasia. The maturation of immature squamous metaplasia tends to be more pronounced in the distal part of the endocervical canal.
Factors in the initiation and promotion of squamous metaplasia are chronic irritation of a physical nature, such as that caused by an intrauterine contraceptive device (IUD), chemical irritants, inflammation with cell destruction, and endocrine changes at the beginning of, during, and after reproductive age. Some of the chemical stimuli that induce squamous metaplasia in subcolumnar reserve cells are also capable of inducing cancer in the uterine cervix of experimental animals.
Squamous metaplasia as such should not be regarded as a change that necessarily and inevitably precedes the development of cancer, but the concept of squamous metaplasia is of great importance in the understanding of carcinogenesis in the uterine cervix.

Basal Cell Hyperplasia—Reserve Cell Hyperplasia
Reserve cell hyperplasia is defined as the appearance of one or more layers of primitive, undifferentiated cells in a subcolumnar position between an overlying endocervical lining epithelium and an underlying basement membrane. The earliest form of reserve cell hyperplasia is a single layer of subcolumnar cells 9 ( Fig. 8.22 ). Proliferation of the subcolumnar reserve cells may involve only one or two layers of cells beneath columnar epithelium or may attain a considerable thickness.

Fig. 8.22 Reserve cell hyperplasia. The earliest form of reserve cell hyperplasia is the appearance of a single layer of primitive cells beneath the endocervical columnar lining epithelium (H&E × HP).
Much controversy surrounds the origin of these reserve cells. Hypotheses about their origin include 10 :
1. Ingrowth of basal cells from the stratum germinativum of adjacent normal stratified squamous epithelium;
2. Origin from fetal squamous basal cells in the preexisting stratified squamous epithelium lining of the urogenital sinus;
3. Origin from undifferentiated fetal rests;
4. Origin from endocervical columnar cells; and
5. Origin from cervical stromal cells.
Fluhmann arbitrarily states that the primitive subcolumnar cells are of epithelial origin and arise above the basement membrane directly from the columnar cells by a process termed prosoplasia . 11 Evidence to definitely exclude their origin from stromal cells is lacking.
The epithelium lining the endocervical canal is derived embryologically from the coelomic epithelium lateral to the urogenital ridge and the subsequently developed mullerian system. The stroma of the uterine cervix, derived from the primitive mesoderm at the site of the urogenital ridge, may regain certain of its embryologic potentialities to supply replacement cells through a poorly defined basement membrane. The most logical derivation, however, is from the same coelomic epithelium as that from which the columnar cells are derived.
Contrary to what is often stated, these reserve cells are not comparable to the basal cells of the original stratified squamous epithelium, because these cells are already dedicated to the formation of squamous cells, whatever the degree of final differentiation (maturation) may be. These two types of reserve cells also differ in keratin phenotype.
Reserve cell hyperplasia per se is not a significant reaction biologically, but it is a frequently occurring nonspecific reaction of the endocervical mucosa. 12, 13

Cytology
Cells are usually arranged in the form of a sheet. Cell borders are usually poorly defined, and thus the cell aggregates often have the appearance of a syncytium, lacking the loss of polarity and the disorganization usually observed in carcinoma in situ.
In cervical smears, it is not unusual to find a single layer of columnar endocervical cells tightly attached to the margin of a sheet of reserve cells. Pure reserve cells are infrequently identifiable in cervical smears. The presence of reserve cells in cervical smears probably implies that the overlying columnar layer has been dislodged. When present, these reserve cells are usually arranged in larger syncytial aggregates called microbiopsies. Cells are relatively small, irregular, and polygonal. The small amount of cytoplasm, which is ill defined, is cyanophilic and may be finely vacuolated. Nuclei are small, relatively uniform in size and shape, and bean shaped, round, or oval and may show longitudinal grooves. Nuclear chromatin is finely granular and is comparable to the nuclear chromatin of the normal interphase nucleus of the columnar cell. Hyperchromasia is uncommon, but in marked proliferation, nuclear chromatin may be arranged in coarser chromatin masses. 14 Nucleoli are not identifiable. 12, 13
Key features of reserve cell hyperplasia
• Cells typically arranged in a 2-dimensional sheet;
• Cytoplasmic borders often indistinct;
• Group polarity and organization maintained;
• Columnar endocervical cells may be attached to the group;
• Cells are small, irregular to polygonal;
• Small amount of ill-defined cytoplasm which is cyanophilic and vacuolated;
• Nuclei are small, bean-shaped, round to oval, with grooves; and
• Nuclear chromatin is finely granular typically without nucleoli.
The cells arising in reserve cell hyperplasia are noteworthy, because in some instances they are reminiscent of those seen in carcinoma in situ. Proliferation of the subcolumnar reserve cells may involve only one or two layers of cells beneath the columnar epithelium or may attain a considerable thickness. The proliferation not only simulates carcinoma in situ but actually may represent a developmental stage of this process. 14 In the cytologic specimen, cells are usually arranged in the form of a sheet. Cell borders are often poorly defined, giving the cell aggregates the appearance of a syncytium but lacking the loss of polarity and the disorganization usually observed in carcinoma in situ.
Reserve cell hyperplasia represents theoretically the earliest stage in immature squamous metaplasia in the uterine cervix. The concept of reserve cell hyperplasia as a stage in squamous metaplasia based on the embryonic rest hypothesis was initially proposed by Eichholz. 15
Reserve cell hyperplasia is in all instances related to the endocervical canal. 12 The predominant site of pure reserve cell hyperplasia is in the proximal part of the endocervical canal, somewhat more proximal than the site of maximal involvement of immature squamous metaplasia.
A surface reaction with any degree of differentiation toward a squamous cell type is more logically placed within the category of immature squamous metaplasia despite the persistence of an overlying endocervical columnar epithelium.

Immature Squamous Metaplasia
Immature squamous metaplasia represents the morphologic spectrum of epithelial changes from a single or multiple layers of reserve cells to an epithelium composed of three or more layers of cells with features of mature nonkeratinizing squamous epithelial cells ( Figs 8.23 to 8.25 ; see also Fig. 8.30 ). Unlike reserve cells, cells derived from areas of immature squamous metaplasia are more often isolated ( Figs 8.26 to 8.28 ). Their tendency to occur as single cells is correlated with the degree of maturation of the parent epithelium. The majority of cells are round to oval, with the number of polygonal cells increasing with maturation. The cytoplasm of immature squamous metaplasia cells is dense, homogeneous but sometimes vacuolated, and cyanophilic in staining reaction ( Fig. 8.29 ). Cytoplasmic vacuolation is frequently observed in the presence of inflammation or as a consequence of degeneration. Nuclei, particularly in the more immature cells, are large, creating a high nucleocytoplasmic ratio (see Fig. 8.28 ). In cervical smears, cells from immature squamous metaplasia changes often demonstrate some degree of atypia. Cells and nuclei show a slight irregularity in size and shape, which is understandable because most metaplastic changes occur under the influence of some irritating factor ( Fig. 8.30 ; see also Fig. 8.29 ). Differential diagnosis with dysplastic changes should be made on the basis of the evenly distributed, finely granular, nonhyperchromatic chromatin.
Key features of immature squamous metaplasia
• Cells in sheet-like aggregates with usually distinct cell borders;
• Densely cyanophilic cytoplasm;
• Relatively large round-to-oval nuclei;
• Increased nucleus-to-cytoplasmic ratio;
• Finely granular, evenly distributed normochromatic chromatin;
• Conspicuous micronucleoli; and
• Macronucleoli sometimes present.

Fig. 8.23 Immature squamous metaplasia. Under the single layer of endocervical columnar cells, a layer of immature squamous metaplastic cells showing slight irregularity in size and shape has been formed. In the vesicular nuclei are finely granular nonhyperchromatic chromatin and prominent nucleoli (H&E × HP).

Fig. 8.24 Immature squamous metaplasia. Layers of relatively immature squamous cells beneath endocervical columnar cells (H&E × HP).

Fig. 8.25 Immature squamous metaplasia. Multiple layers of relatively mature squamous metaplastic cells have replaced the endocervical columnar lining (H&E × MP).

Fig. 8.26 Immature squamous metaplasia. Immature squamous metaplastic cells lying singly and in a sheet (Papanicolaou × OI).

Fig. 8.27 Immature squamous metaplasia. Cells lying singly and in a sheet with relatively large nuclei, indistinct cell borders, and homogeneous or finely vacuolated cytoplasm (Papanicolaou × OI).

Fig. 8.28 Immature squamous metaplasia. Cells with a high nucleocytoplasmic ratio. Large nuclei with prominent nucleoli and a finely granular chromatin (Papanicolaou × HP).

Fig. 8.29 Slightly atypical immature squamous metaplastic cells. Finely granular nuclear chromatin and small multiple nucleoli (Papanicolaou × OI).

Fig. 8.30 Immature squamous metaplasia. Epithelial lining composed of relatively mature squamous metaplastic cells. In the superficial layers, maturation is still incomplete, resulting in cells with relatively high nucleocytoplasmic ratios. Cytoplasmic borders are distinct. A small island of columnar cells is recognizable in the superficial layer (H&E × MP).

Mature Squamous Metaplasia
Squamous metaplasia is represented by a spectrum of epithelial changes resulting in an admixture of cells of varying maturity in the cellular sample.
The squamocolumnar junction bears no constant relationship to the anatomic external os, and the external os has no histologic landmarks to delineate it. The increase in linear extent of squamous metaplasia in the endocervical canal and the region of the transformation zone with increasing age is inversely related to the reduction of the linear extent of reserve cell hyperplasia.
Mature squamous epithelium encompasses the classic three layers of nonkeratinizing squamous epithelium, making mature squamous metaplastic epithelium virtually indistinguishable from the original ectocervical squamous epithelium. Foci of mature squamous metaplasia may be indistinguishable from the normal ectocervical mucosa. The only clue to its metaplastic origin is underlying endocervical glands ( Fig. 8.31 ).

Fig. 8.31 Mature squamous metaplasia. An island of mature squamous metaplastic epithelium bordered by endocervical columnar epithelium. Mature squamous metaplastic epithelium has great resemblance to stratified squamous epithelium of the ectocervix (H&E × MP).
The presence of squamous metaplasia in cervical smears was reported by von Haam and Old to be 41.5%. 16 Howard and co-workers reported an 83% incidence, 17 and Carmichael and Jeaffreson found it to be present in 41% of cervices examined histologically. 18
The relationship of age to prevalence of squamous metaplasia was found by von Haam and Old to be 86.2% in the third decade of life and 69.2% in women older than 60 years. 16 In 101,000 first cervical smears from women 35 to 54 years of age, 62.5% contained squamous metaplastic cells. The prevalence of squamous metaplasia rises significantly from the third to the fifth decade. 19

Cytology
Cells originating from squamous metaplasia tend to be isolated, less frequently occurring in loose sheets. The number of cells varies with the extent of the epithelial change, the localization of the lesion, and the method of sampling. 12, 13 Cells from squamous metaplasia characteristically have distinct borders and are predominantly round, oval, or polyhedral. In immature squamous metaplasia, the cytoplasm is homogeneous and cyanophilic, whereas in a more mature type of metaplasia, it is characterized by a more densely staining outer zone or ectoplasm and a clear central perinuclear zone or endoplasm ( Fig. 8.32 ). At the periphery of the cells, remnants of the fibrillar apparatus, observed in normal squamous cells, may be demonstrated. The nuclei are relatively small, round or oval, usually centrally located, and uniform in size and have a basically finely granular chromatin in which there are small aggregates or chromocenters. 14 Mature squamous metaplastic cells can sometimes be differentiated from cells derived from the original ectocervical squamous epithelium by their slightly denser staining cytoplasm.
Key features of mature squamous metaplasia
• Cells tend to be isolated or less frequently in loosely aggregated sheets;
• Cytoplasmic borders are distinct;
• Cells are round to oval, or polyhedral;
• Cytoplasm displays a densely staining outer zone (ectoplasm) and a central clear perinuclear zone (endoplasm)—cyanophilic staining; and
• Nuclei are relatively small, centrally located, and uniform in size and shape, and show finely granular, evenly distributed chromatin with occasional micronucleoli.

Fig. 8.32 Mature squamous metaplasia. Singly lying, relatively mature round-to-oval cells with relatively large nuclei. Cytoplasm showing dense outer zone, “ectoplasm,” and lighter inner zone “endoplasm” (Papanicolaou × OI).
Cells from a slightly to moderately atypical squamous metaplasia change may pose major problems to the inexperienced cytologists ( Figs 8.33 to 8.35 ). Within this group of diagnostically difficult lesions is a subset associated with abnormal cytologic smears. These lesions are colposcopically abnormal and have the histologic and cytologic features of atypical immature squamous metaplasia. 20

Fig. 8.33 Slightly atypical relatively mature squamous metaplastic cells. Cells resembling ectocervical squamous cells with slightly irregular somewhat enlarged nuclei. (Papanicolaou × OI).

Fig. 8.34 Atypical squamous metaplasia. Round, oval, and polygonal cells with slightly enlarged, slightly irregular nuclei. Nuclear chromatin is finely granular and evenly distributed (Papanicolaou × HP).

Fig. 8.35 Cells from atypical squamous metaplasia simulating a dysplasia because of abnormal size and shape of the cells. Nuclei are enlarged but relatively small in comparison with the size of the cytoplasmic body (Papanicolaou × HP).
The cytoplasm gives information related to the maturity and specific tasks of the cell, whereas the nucleus reflects not only maturity but also the degree of dedifferentiation of a cell (malignant potential). Nuclei in atypical squamous metaplasia appear to be large but only relative to the size of the cytoplasm. Absolutely, the nuclei in atypical squamous metaplasia are much smaller than nuclei in dysplastic changes. In squamous metaplasia, hyperchromasia is generally absent. Hyperchromasia in nuclei of dysplastic cells reflects an abnormality in DNA synthesis, for example, an abnormal number of chromosomes in that particular nucleus.
Mature squamous metaplasia is also referred to as:
• Epidermization ( Fig. 8.36 );
• Complete squamous metaplasia; and
• Squamous prosoplasia, stage V (Fluhmann). 11

Fig. 8.36 Epidermization. Relatively mature squamous metaplastic epithelium bridging an invagination of the endocervical mucosa into the stroma. Complete blocking of the invagination may lead to large mucus-filled cysts, Naboth's ovula in the cervix (H&E × LP).

Atypia
The word atypia means not normal, or not typical for a normal cell of this particular tissue. The term atypia is a minimally defined descriptor in diagnostic cytopathology. It should not be used as a noun but only as a descriptor of an observation with additional specification of the severity of the atypia. When the word atypia is used in a descriptive diagnosis, the degree of aberration from the normal should be further specified, as should what specific type of cells the observer has referred to. The use of the word atypia without further specification should be avoided, because it then can be used too often as a substitute for a careful description and definition. The word atypia is commonly used as a descriptive diagnosis when indicating minimal-to-slight aberrations from normal. Features most frequently causing such a diagnosis of slight-to-minimal atypia are nuclear enlargement and aberrations from the normal configuration of the cells ( Fig. 8.37 ). Causative processes most often are inflammation, regenerative reactions, certain deficiency states such as folic acid deficiency, and the earliest changes in an epithelium that is in neoplastic transformation.

Fig. 8.37 Minimally atypical squamous epithelium. Irregular arrangement of cells and slightly abnormal nuclei in the basal and parabasal layers. Reduced maturation of cells and slight nuclear enlargement in more superficial layers (H&E × HP).
In 70,625 first smears, minimally atypical squamous cells and squamous metaplasia cells were identified in 16.8% of smears. When related to the mode of contraception, these findings of minimal atypia were noted in 14.1% of smears from women using oral contraceptives and in 24.3% of smears from women using lUDs. Diagnoses of minimal-to-slight atypical changes such as “some abnormal squamous cells present” and “atypical squamous metaplastic cells present” should be followed by a repeat smear after 12 months. 21 Patten advises maintaining patients with evidence of cytologic atypia under surveillance at yearly intervals. 8 Moderate atypia, frequently observed in reparative reactions, should be followed by a repeat smear after 3 months. 21

Reactive and Regenerative Changes
A reactive change is an epithelial reaction to injury characterized by the presence of sometimes highly atypical cells of endocervical and squamous metaplastic origin.
The epithelia covering the uterine ectocervix and endocervical canal are under the constant influence of physiologic stimuli but also of ever-changing external stimuli. 8
Regenerative epithelial reactions are commonly found in patients after:
• Radiotherapy;
• Recent hysterectomy;
• Cautery or biopsy;
• Cryocoagulation diathermy;
• Past history of severe cervicitis; and
• Partial or complete destruction of the epithelium by infection and inflammation.
These environmental changes result in various morphologic responses, which may be classified as destructive, protective, or reparative. All of these reactive changes may result in some form of epithelial atypia in that certain morphologic features that are present represent a departure from the normal.
Repair epithelium in experimental animals has been found to be more susceptible to the action of carcinogenic agents than nontraumatized tissue. 22
Regeneration of cells as a manifestation of a reparative change can occur in squamous epithelium, in squamous metaplasia epithelium, and in columnar epithelium. 23 Geirsson and colleagues found 51.3% of the cells in tissue repair to be of glandular origin, 38.7% of squamous metaplasia origin, and 10% of squamous origin. 24
Reparative reactions are frequent in patients who have had severe recurrent cervicitis and in patients who have had recent treatment such as punch biopsies, conization, cryosurgery, laser therapy, and endocervical curettage. This type of reaction is also found after hysterectomy, together with evidence of granulation tissue, in the postirradiation stage and in cases of true erosion or ulceration of the cervical stroma, which may be caused by a prolapsed uterus, by pressure necrosis from a ring or a shield pessary, or by an IUD.


Cytology
Reparative changes are characterized morphologically by significant nuclear enlargement and usually the presence of large, prominent nucleoli as a sign of active protein synthesis in the fast-growing cells, which try to replace the damaged epithelial cells ( Fig. 8.38 ).

Fig. 8.38 Reparative reaction. Syncytial arrangement of immature cells with relatively large, round-to-oval nuclei, prominent nucleoli, and a finely granular evenly distributed nuclear chromatin. Variation in nuclear size (Papanicolaou × OI).
Cells from reparative epithelium usually desquamate as large, sheet-like aggregates with indistinct cytoplasmic boundaries. In these aggregates, mitoses may be present ( Fig. 8.39 ). It is not unusual to find that leukocytes have infiltrated the larger aggregates of epithelial cells. Rarely, abnormal singly lying cells are found. The cells have a wide variation in size and shape. The cytoplasm is usually cyanophilic and sometimes is finely vacuolated or may contain large vacuoles. Nuclei are mostly round to oval, with some nuclear enlargement and variation in nuclear size. Nucleoli are prominent, and multiple macronucleoli are sometimes present. As a rule, the nuclear chromatin is finely granular, almost always evenly distributed, and not hyperchromatic. Cells have essentially the characteristics as immature columnar cells ( Fig. 8.40 ), immature squamous cells, or immature squamous metaplastic cells. 25 Depending on the severity of the stimulus causing the epithelial damage, the replacing epithelium—regeneration and repair—will be blocked in its maturation and show a degree of abnormal configuration or hyperchromasia ( Figs 8.41 and 8.42 ).
Key features of repair
• Nuclear enlargement with wide variability of size and shape;
• Prominent nucleoli;
• Chromatin typically finely granular, evenly distributed, and not hyperchromatic;
• Mitoses may be present;
• Sheet-like 2-dimensional arrangements;
• Leukocyte infiltration of groups;
• Rarely abnormal isolated cells are found; and
• Abundant cytoplasm with “taffy pull” appendages.

Fig. 8.39 Reparative reaction. Syncytium of immature cells with irregular arrangement of nuclei, finely granular chromatin, and prominent nucleoli (Papanicolaou × OI).

Fig. 8.40 Reparative reaction. Aggregate of immature cells with round-to-oval, eccentrically located nuclei, finely granular nuclear chromatin, and one to two prominent nucleoli (Papanicolaou × OI).

Fig. 8.41 Atypical reparative reaction. Sheet-like aggregate of immature cells with variation in cellular size and shape, together with a somewhat coarse, slightly hyperchromatic nuclear chromatin and prominent nucleoli (Papanicolaou × OI).

Fig. 8.42 Atypical reparative reaction. Sheet-like aggregate of immature cells with indistinct cytoplasmic borders, round-to-oval and irregular nuclei, coarse, somewhat hyperchromatic nuclear chromatin, and very prominent, often pleomorphic macronucleoli. In the right upper and lower corner slightly atypical squamous metaplastic cells (Papanicolaou × OI).
Differentiating between cells from reparative changes and cells from invasive neoplastic processes may be difficult. The predominant arrangement of cells in sheet-like aggregates, even though the cytoplasmic boundaries may be indistinct, together with the normochromatic, finely granular, evenly distributed chromatin, and the presence of macronucleoli usually can provide the correct diagnosis. In no other epithelial abnormality and particularly in no invasive process do these three characteristics occur simultaneously.

Inflammation-Associated Cellular Changes
Inflammation alone causes minor cytologic abnormalities, such as a dual staining reaction, lysis or vacuolation of the cytoplasm, slightly disproportionate nuclear enlargement, and an increase of the nucleocytoplasmic ratio ( Fig. 8.43 ). Nuclear chromatin is more often hypochromatic than hyperchromatic (see Chapter 7 , Microbiology, Inflammation, and Viral Infections, for additional information).

Fig. 8.43 Slight atypia of squamous cells in a cervical smear with evidence of a bacterial infection (Papanicolaou × MP).

Degenerative Changes
Degenerative changes of nuclei such as folding of the nuclear membrane, karyorrhexis, karyolysis, and pyknosis in cases of inflammation must be differentiated from abnormal nuclear changes in premalignant or malignant epithelial lesions.
Cytomorphologic alterations caused by inflammation or physical or chemical trauma usually are nonspecific. Changes are cell destruction, cytolysis, karyorrhexis ( Fig. 8.44 ), and karyolysis. In cases of nuclear alterations, such as nuclear enlargement, binucleation, and multinucleation, as well as coarse clumping and irregular distribution or a complete loss of structure of the chromatinic material, the differential diagnosis with true atypical changes, such as dysplastic lesions or even invasive carcinoma, becomes relevant ( Fig. 8.45 ). The correct diagnosis usually can be made on the basis of the cytoplasmic vacuolation due to hydropic degeneration and the fading of nuclear contours due to autolysis.

Fig. 8.44 Karyorrhexis. Loss of nuclear structure, dissolved nuclear membrane, irregular clumping of chromatinic material (Papanicolaou × OI).

Fig. 8.45 Degenerative changes in endocervical columnar cells. Hyperchromasia due to condensation of the chromatinic material. Loss of nuclear structure. Differential diagnosis with a severe epithelial lesion on the basis of clearly outlined cytoplasmic borders (Papanicolaou × HP).

Unsatisfactory Quality
Extreme admixture of inflammatory cells can occasionally obscure epithelial cells or dilute the number of diagnostic cells in a specimen, thus reducing the chance of detection of abnormal cells. In women participating in a large population-screening program, the number of smears unsatisfactory for cytologic diagnosis was 6.8%. 26 Poor quality of the smears was caused by too few epithelial cells present or a strong admixture of erythrocytes or inflammatory cells. Of the smears considered evaluable, endocervical columnar cells and squamous metaplastic cells were found significantly more frequently in smears showing signs of bacterial inflammatory reaction or the presence of Trichomonas vaginalis . A significant proportion of smears without signs of inflammation were less reliable for cytologic diagnosis.
An inflammatory exudate has been reported to be associated with approximately 32% of dysplastic lesions. 27 In a large population study, a significantly higher percentage of smears without signs of epithelial abnormalities were found in the absence of signs of inflammation. 26 In these smears, the occurrence of minimal epithelial atypia was even significantly lower than expected. The prevalence of mild and moderate dysplasia was also lower than expected but not significantly so. The percentage of smears without any sign of inflammation rose significantly with age. The distribution of epithelial changes in the groups of smears with signs of bacterial infection did not differ significantly from the expected numbers. In contrast, epithelial abnormalities were significantly more common in smears showing evidence of T. vaginalis ( Fig. 8.46 ), as were changes consistent with severe dysplasia and carcinoma in situ. In smears with evidence of Candida albicans (moniliasis), significantly more minimally atypical epithelial changes were found, which seemed more a reaction of the epithelium to the inflammatory stimulus than a dysplastic change. This was not paralleled by a higher proportion of smears with squamous metaplastic cells present, nor was there evidence of a higher proportion of mild-to-moderate dysplastic changes.

Fig. 8.46 Evidence of Trichomonas vaginalis infection. Trichomonads can be recognized next to a slightly atypical squamous metaplastic cell with vacuolated cytoplasm (Papanicolaou × OI).

Epithelial Abnormalities
The finding that the presence of inflammatory signs or microorganisms is more common in smears consistent with dysplasia or carcinoma, and the lack thereof to be suggestive of atypia or dysplasia, 28, 29 was not confirmed. The exception to that finding was T. vaginalis ; this parasite was found four times more frequently than expected in smears consistent with severe dysplasia and carcinoma in situ.
La Vecchia and co-workers also found significant associations between a history of T. vaginalis and C. albicans and cervical intraepithelial neoplasia (CIN) lesions but not invasive cancer. 30 Frisch found 4% of smears originally diagnosed as inflammatory atypia to be underreported, because in subsequent smears these atypias had “progressed” to CIN. 28
Schachter and colleagues reported a significantly increased risk for cervical neoplasia in women with antibodies to Chlamydia trachomatis , 31 and Hanekar and associates found in patients with Chlamydia- associated epithelial abnormalities that the progression rate to CIN grade III after 2 years of follow-up was significantly higher than in a control group with comparable epithelial changes but without evidence of Chlamydia infections. 32 They suggest that Chlamydia may be a cocarcinogen or a potentiating agent in the progression of cervical intraepithelial lesions. In performing colposcopically directed biopsies on patients with persistent inflammatory atypia, Noumoff found underlying CIN in about one-third of the patients; again, about one third were of greater severity than CIN grade I. On the basis of these data, he advises that all patients with persistent inflammatory atypia undergo colposcopic evaluation. 33
It may be difficult to decide whether the epithelial abnormality is a nonspecific response to the inflammatory stimulus or a true preneoplastic intraepithelial lesion. The cytologic report should clearly state the diagnostic dilemma, and recommendation for a repeat examination after a follow-up interval should be part of the report.
This minimal-to-mild atypia may eventually be followed by the appearance of cellular changes consistent with mild dysplasia. Smears showing signs of atypia related to inflammation should therefore be repeated after 1 year. 21

Immunosuppression
Immunosuppression either due to immunodeficiencies or caused by medication conveys a significant risk for infections with herpes simplex virus type 2 and human papillomavirus (HPV) and for developing neoplastic conditions ( Fig. 8.47 ). 34, 35 Patients at risk are those receiving immunosuppressive drugs because of transplants or for various other conditions, patients with Hodgkin's disease, and patients with cancer following treatment with cytostatics. 27, 36, 37 The reported significantly increased risk for cervical cancer with the number of pregnancies is an interesting association in view of the fact that pregnancy is considered a transient state of immunodepression. 30, 38

Fig. 8.47 Changes consistent with a herpes simplex type 2 virus. Multinucleated giant cell with nuclear enlargement and dense opaque nuclear chromatin (Papanicolaou × OI).

Intrauterine Contraceptive Devices
The most severe reactive changes, those that most closely mimic intraepithelial neoplasia, both of squamous and of glandular type, are associated with the presence of an IUD. 39, 40 Composing cells may show severe cellular and nuclear polymorphism, an increased nucleocytoplasmic ratio, prominent nucleoli, and cytoplasmic vacuolization. Severely atypical cells arranged in clusters are often present. There is usually a marked inflammatory reaction. The cells may be extremely difficult to differentiate from those derived from an adenocarcinoma, but the cytopathologist can avoid this erroneous diagnosis because these IUD-related changes usually are found in relatively young women who are using an IUD. This, however, emphasizes the importance of correct and complete clinical information accompanying the request for a cytologic examination. It may sometimes be necessary to remove the device and to repeat the cytologic examination after an interval of 4 to 6 weeks. The epithelial abnormalities usually will have disappeared by then.

Human Papillomavirus and Cervical Cancer
In the late 1970s epidemiological studies suggested that transfer of a sexually transmitted factor could explain the occurrence of cervical cancer in women. 41, 42 Already in 1973, zur Hausen pointed out the possible etiological role of HPV in cervical carcinoma. 43 Two decades earlier, Koss introduced the term “koilocytotic atypia” for abnormal squamous epithelial cells in smears from patients with CIN or cervical cancer. 44 Koilocytotic cells exhibit vacuolization of the cytoplasm with condensation of the chromatin and slight atypia of the nuclei ( Fig. 8.48 ). Not until the introduction of high-resolution imaging by electron microscopy could viral particles be demonstrated in koilocytotic cells of CIN lesions. 45 After the introduction of molecular techniques as southern blot hybridization and polymerase chain reaction (PCR), many epidemiological and clinicopathological studies demonstrated that HPV is associated with the development of CIN and cervical cancer. 46 - 51

Fig. 8.48 Koilocytotic atypia. Cervical smears with LSIL, showing cells with vacuolization of the cytoplasm and slight atypia of the nuclei (Papanicolaou × MP).
To date, more than 100 different genotypes of HPV have been identified thus far. Two large groups of HPVs can be distinguished based on the site of infection: cutaneous and genital (mucosal/cervical). Most cutaneous HPV types (HPVs 1, 2, 3, 4, 10, 28, 29, and 38) cause benign lesions of the skin such as verruca plantaris, verruca vulgaris, and verruca plana. The group of genital HPVs is further subdivided based on the behavior of the resulting cervical lesions. Low-risk cervical HPV types (HPVs 6, 11, 40, and 42–44) are predominantly present in condyloma acuminatum, ASCUS, and CIN1 lesions. Eighteen out of more than 35 genotypes (HPVs 16, 18, 26, 31, 33, 35, 39, 45, 51–53, 56, 58, 59, 66, 68, 73, and 82) that infect the mucosal types of epithelium are considered high-risk HPV since they have been detected in the cervix of patients with cervical carcinoma. Infections with HPV 16, 18, and 31 together account for almost 80% of all squamous cell carcinomas of the cervix. 52
The effects of HPV on the infected cervical cell are described in more detail in chapter 7 . In short, in low-grade cervical lesions the double-stranded circular HPV DNA genome of approximately 8000 base pairs is in an episomal state. With progression of the cervical lesion, the HPV DNA becomes disrupted and integrates in the host genome. In cervical carcinoma, HPV is always in an integrated state. Not until integration of HPV in the genome of the host cell takes place does its disadvantageous effects pop up being the result of blocking of p53 (the “guardian” of the genome) as well as blocking of pRb (the “brake” of the cell cycle). Both effects of HPV, DNA instability and hyperproliferation, may finally lead to invasive carcinoma of the cervix.
The presence of high-risk (hr)-HPV genotypes rises with increasing severity of the CIN lesion and depends on the method of detection. Two popular and widely used HPV detection techniques are PCR, which uses very sensitive primers (MY09/11 and GP5/6) by which the HPV target is amplified, and hybrid capture (HC) assays. The latter system has recently been approved by the US FDA and has now a second-generation (HC2) that allows for the detection of 13 hr-HPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68). It works with a reaction tube in which cervical cells become destroyed in order to release HPV DNA that is hybridized with complementary chemical RNA probes, after which the hybridized product is visualized with a staining procedure on a micro titration plate. HC2 is more automated but somewhat less sensitive than PCR and its advantage over PCR is that it is less prone to cross-specimen contamination as PCR is able to produce millions of copies from a single set of DNA base pairs.
With sensitive primers as MY09/11 and GP5/6 in the PCR it appeared that more than 99% of cervical cancers are HPV DNA positive. 50 CIN lesions that are hr-HPV positive are more likely to progress than those that are negative. 53 Further, it appeared that hr-HPV infections lead to a 100- to 300-fold increased risk for patients with persistent hr-HPV infections to getting a CIN 3 lesion. 54, 55 All these studies assume that (persistent) hr-HPV is associated with progression of CIN. However, it was also shown that young women often have transient infections and more than 80% of hr-HPVs do become cleared. 56
Recently, many studies have investigated the value of HPV detection in addition to conventional cytology in order to increase sensitivity and specificity of cervical cancer screening. The summary of meta-analysis of Arbyn et al. gives a good overview of these studies. 57 Their study concludes that HPV triage with the HC2 assay is slightly more accurate than repeat cytology in women with “equivocal” (ASCUS) Pap smear results. Triaging women with low-grade squamous intraepithelial lesions (LSIL) (or more) does not show a significant higher sensitivity, but even a significant lower specificity if it is compared to repeat Pap smear testing. Partly based on this knowledge, the United States recently started to combine cytology and HC2 primary screening in women older than 30 years. In Europe, primary screening with the Pap smear alone still remains the standard. Several large randomized controlled trials that test the role of HPV in primary cervical screening are now underway. These trials are near to completion (Sweden 2006; The Netherlands 2007; UK 2007, Italy 2007, and Finland 2009). 58
One of the disadvantages of HPV testing with PCR or HC2 is that the cost of a single test is rather high and its technology is not as widely available as is the Pap smear. Another unresolved problem with HPV testing is how to communicate the test results when a tested woman is “HPV-positive.” This woman must be told that she harbors a sexually transmitted viral infection that eventually may cause cervical carcinoma. Must she change her sexual habits? What about the legal complications? Thirdly, suppliers of an HPV test can guarantee high-quality standards for the test. What will happen when more tests become available? An HPV test harbors many variables that a simple Pap test does not have. 59 Moreover, it is remarkable that despite the tremendous amount of research on HPV (with PCR and/or HC2), there are just a few in situ hybridization studies dealing with the physical state (episomal or integrated) of HPV in the infected cell. Studies with PCR or HC2 only give information regarding the mere presence of HPV in cervical swabs or brushes and do not inform us about the physical state of the HPV, where it is located and what it is doing in the cervix.. 60, 61

Condylomatous Lesions and Epithelial Abnormalities
Intraepithelial neoplastic lesions of the human cervix, first defined as koilocytotic warty atypia by Koss and Durfee, 44 would now be diagnosed as condylomatous lesions, caused by an infection with HPV. In particular the subspecies 6 and 11 were known as causative agents in warty condyloma acuminatum of the vulva and vagina and were thought to be important sexually transmitted factors in the genesis of cervical cancer. 62 Condyloma acuminatum is found in younger age groups than is dysplasia and is common in sexually active teenagers. 63
Types 6 and 11 are most commonly associated with condyloma acuminatum of the cervix and with CIN lesions of mild severity.
In the differential diagnosis with CIN, it is relevant that condylomata occur at a younger mean age, are found in the transformation zone and the cervical portio, are polyploid, often regress, and contain HPV antigens in the majority of cases. Koss, contrary to what most authors favor, prefers to use the term CIN grade I with features of condyloma. 64
For a more detailed overview of the relation between HPV and epithelial abnormalities of the cervix, see Chapter 7 , Microbiology, Inflammation, and Viral Infections (as well as the previous section on HPV).

Histology
Cervical lesions are of three types. The most frequent form is a flat, acanthotic epithelial change with well-preserved basal layers and marked nuclear degeneration with perinuclear halos toward the surface. It also shows dyskeratotic changes ( Fig. 8.49 ). The classic proliferative, papillomatous condyloma is found much less frequently, and the third type, even rarer, is an endophytic “inverted” condyloma. The flat and endophytic condylomata represent new lesions previously not noted on the cervix. Many of the condylomatous lesions were described as dysplasias. Meisels and co-workers were thus led to believe that this sexually transmitted viral lesion represented a precursor of cervical neoplasia in view of the fact that condyloma acuminatum has been proved to undergo malignant transformation in a few cases and that it behaves epidemiologically similar to carcinoma of the cervix. 65

Fig. 8.49 Condylomatous change with slight nuclear atypia and irregular arrangement of the cells in the squamous cell layer (H&E × HP).

Cytology
Although certain features are characteristic of condyloma and CIN, these two lesions cannot always be clearly distinguished by morphologic means. Some features may help in differentiating between them.
Koilocytes are the predominant cellular features of infection with HPV. 66 Koilocytes must be differentiated from cells with perinuclear halos found in other types of infection, such as trichomoniasis. In these cells, however, nuclear abnormalities are not as apparent and halos tend to be smaller and less well demarcated. Once the cellular pattern is recognized, condyloma acuminatum becomes the most frequent epithelial lesion diagnosed in a mass screening program. 63
The pathognomonic change is the koilocytotic cell. This is a superficial or intermediate squamous cell that displays a large perinuclear halo with irregular, clear-cut edges and a dense, often amphophilic, sometimes almost hyaline cytoplasm in the area surrounding the perinuclear cavity ( Fig. 8.50 ). In condyloma, cells are generally mature and often display an ample amphophilic cytoplasm. The chromatin is usually poorly defined, and the nuclei may show various stages of degeneration. The nuclear membrane is not distinct, and nucleoli usually are absent. In intraepithelial abnormalities, cells are mostly immature, with scant cyanophilic cytoplasm. The chromatin details are distinct, the nuclear membrane is clearly visible and somewhat irregular, and small nucleoli can be recognized. 66, 67 Abnormal differentiation becomes evident by an underdeveloped cytoplasmic body and a relatively enlarged nucleus, leading to an increased nucleocytoplasmic ratio ( Fig. 8.51 ). Nuclear shapes are round to oval and often irregular. Particularly in the more severe abnormalities, the nuclear chromatin is increased and often irregularly distributed. Cellular changes originate in preexistent squamous epithelium or more often in squamous metaplastic epithelium.
Key features of condylomatous lesions (CIN 1, LSIL)
• Koilocyte is the classic cell—superficial or intermediate type cell with large clear-cut perinuclear halo;
• Cytoplasm is densely cyanophilic and often amphophilic;
• Nuclei may be degenerated with poorly defined, clumped chromatin and indistinct nuclear membrane;
• Chromatin is normochromic to slightly hyperchromatic unevenly distributed;
• Nucleoli are absent;
• Multinucleation is common; and
• Non-koilocytes may show markedly enlarged nuclei with increased nucleus to cytoplasmic ratio.

Fig. 8.50 Koilocytotic cells. Large perinuclear halo. Dense outer zone of the cytoplasm and an enlarged nucleus with finely granular nuclear chromatin (Papanicolaou × HP).

Fig. 8.51 Condylomatous change in mature squamous epithelium. Irregularity in nuclear size and shape and conspicuous nuclear halos (H&E × HP).

Sensitivity
Cytology may detect the majority of lesions with koilocytotic features but may not detect focal CIN grades I, II, and III lesions associated with these condylomatous changes. 68
Cellular changes due to low-risk HPV infection may mimic the changes found in cases of well-differentiated squamous cell carcinoma. The right diagnosis may usually be made on the basis of the less apparent nuclear abnormalities in low-risk HPV-induced epithelial lesions.
Neither cytology nor histology can detect an asymptomatic, colposcopically negative, latent HPV infection, which can be detected by in situ hybridization tests only. The great intraobserver and interobserver variability in the cytologic and histologic diagnosis of condyloma and CIN reflects the great amount of subjectivity in evaluating the morphologic characteristics of a lesion. The correlation between morphologic signs of HPV infection and the results of HPV-DNA hybridization indicates that the sensitivity of the cytologic and histologic features of HPV infection are in the range of 15 to 36%. 69 Future diagnosis of HPV by PCR should include also in situ hybridization. 67

Squamous Intraepithelial Neoplasia

Terminology

The primary function of a diagnostic terminology is to communicate to the referring physician the interpretation of a specimen in descriptive terms that will have clear implications for appropriate patient management. Many practicing physicians do not have detailed knowledge of cytopathology. Thus the cytologic report should not only be scientifically accurate but also easily understood. 40
The terminology currently used in cytopathology is variable, inconsistent, and sometimes ambiguous. Many cytopathologists and clinicians are thus confused and uncertain about the meaning of some terms. Failure to understand the clinical meaning of a cytopathologic evaluation may lead to inappropriate treatment of patients.
The cytology report should consist of a concise description of abnormal cellular findings in well-defined and generally accepted terms, followed, if appropriate, by a prediction of the histologic condition, and should also include a recommendation for the further treatment of the patient. 21, 26, 70 - 72
In pathology, the basis for any system of classification is morphologic. However, where possible, the classification should relate to the biologic significance or potential of the process. 73

Papanicolaou Classification
The Papanicolaou classification of cytologic findings in epithelial lesions has led to different interpretations. Also, the Papanicolaou system does not provide for the diagnosis of noncancerous lesions.
“The groups of the Papanicolaou system do not reflect the current understanding of cervical neoplasia and the Papanicolaou classes do not have an equivalent in tissue diagnostic terminology.” 74, 75 Epithelial abnormalities of the uterine cervix form a morphologic and most likely also a biologic continuum.
The rationale for distinguishing between dysplasia and carcinoma in situ on a cytomorphologic basis is to permit the best possible correlation between the cellular and tissue sample and the final clinical outcome of the epithelial change.
Years of experience have demonstrated the difficulty of categorizing dysplastic cervical lesions into the diagnostic triad of mild, moderate, and severe dysplasia. In our own experience, cytologic diagnoses of mild or severe dysplasia and carcinoma in situ proved to be fairly accurate. The cytologic diagnoses of moderate dysplasia were less accurate. A large proportion (57%) of moderate dysplasias were absent or had regressed to a less severe lesion at follow-up 6 months after primary diagnosis. On the other hand, in about 20% of cases, during follow-up a more severe lesion was found. 76, 77
In the USA a National Cancer Institute (NCI) working group (aka Bethesda 1988 and Bethesda 1991) recommended discontinuing the use of the Papanicolaou classification as a means for reporting. The Papanicolaou classification is deficient in advising about the true nature of abnormalities unless it is accompanied by a verbal description of the cytologic findings.
In order to establish some order, all lesions known to be reactions to infections, inflammation, or reparative reactions (regeneration) should be excluded, as well as benign proliferative reactions in the endocervical canal such as reserve cell hyperplasia and squamous metaplasia, because these in themselves are not considered to be stages in the process of carcinogenesis. 74, 75 In the Bethesda System the inflammatory diseases, the infectious conditions, and other reactive cellular changes such as repair (due to trauma, infection, IUCD), radiation, and atrophy are classified as “negative for intraepithelial lesion or malignancy.”
The NCI working group adopted a new classification based on only two main categories for intraepithelial cell abnormalities:
Low-grade squamous intraepithelial lesions (LSIL)
• Changes consistent with HPV infection: multinucleation, perinuclear halos, and slight nuclear atypia; and
• Mild dysplasia (CIN grade 1).
High-grade squamous intraepithelial lesion (HSIL)
• Moderate dysplasia (CIN grade 2);
• Severe dysplasia (CIN grade 3); and
• Carcinoma in situ (CIN grade 3).
Very slight epithelial abnormalities—conditions between “negative” and true intraepithelial lesions (LSIL and HSIL)—are placed into a fourth category as “atypical squamous cells (ASC).
This fourth category was chosen because everybody agreed that the presence of (slightly) atypical squamous cells of “undetermined significance” implied a significant risk for an underlying high-grade cervical intraepithelial lesion (SIL).
Abnormalities of glandular cells, as well as ASC, LSIL, and HSIL, are placed in the group of (intra)epithelial cell abnormalities and subdivided as “atypical glandular cells,” “cervical adenocarcinoma in situ,” and “adenocarcinoma.”
Finally, in 2001 Bethesda introduced a new category, “other,” in which disorders such as the presence of endometrial cells in a postmenopausal woman can be placed. 78 For further details on the Bethesda System see Chapter 6 .
It was thought that using these four categories would reduce the present inconsistencies in terminology. The proposed terminology is related to the expected clinical behavior of the epithelial abnormalities and is management oriented in that it includes a recommendation for the preferred follow-up procedure in an individual case. The new classification is thought to provide a basis for communicating the diagnostic interpretation of a specimen in unambiguous descriptive or diagnostic terms that have clear implications for proper treatment of patients. 74, 75
Not all countries have adopted the Bethesda System in their screening programs. Several countries continue using the World Health Organization (WHO) terminology (mild, moderate, and severe dysplasia; carcinoma in situ; and carcinoma) or the CIN categories (CIN 1, 2 and 3; invasive carcinoma).
The Bethesda System terminology can be easily translated into these two other systems of categorization.
For an overview of all currently used classification systems see Table 8.1 (Papanicolaou, WHO, CIN, Bethesda).

Table 8.1 Main different reporting systems for cervical cytological squamous epithelial abnormalities

Dysplasia
The term dysplasia means “disordered form” (disordered differentiation). Reagan and associates introduced the term dysplasia to describe these disordered growth patterns and reported that the majority of these intraepithelial changes would regress spontaneously or persist unchanged if left untreated. 79
In the uterine cervix, the term dysplasia is applied to a spectrum of heteroplastic reactions involving stratified squamous or squamous-like (metaplastic) epithelium. As the term implies, this group of reactions is characterized by malformation or disordered development, manifested morphologically by variations in cytoplasmic maturation in association with certain nuclear abnormalities. 9, 12
The definition of dysplasia is derived from the definition of carcinoma in situ, which is defined as “an intraepithelial abnormality in which throughout its whole thickness, no differentiation takes place.”
All other disturbances of differentiation in the squamous epithelial lining, the glands, or covering of the surface are to be classified as dysplasia. They may be characterized as a high or a low degree, terms that are preferable to suspicious and nonsuspicious, as the proposed terms describe the histologic appearance and do not express an opinion. 80
In the definition of the World Health Organization, dysplasia is “a lesion in which part of the thickness of the epithelium is replaced by cells showing varying degrees of atypia.” The lesions were further graded as mild, moderate, and severe. 81 Most pathologists determine these grades of dysplasia on the basis of the proportion of the epithelial thickness occupied by the abnormal cells, taking into account the degree of atypia of these cells. The latter grading has stimulated much discussion, because a number of investigators do not believe that this grading is sufficiently correlated to biologic behavior. The dual terminology has led gynecologists to assume that dysplasia and carcinoma in situ are two biologically distinct entities with different progressive potential and that dysplastic lesions may not require treatment. 82 Some studies have shown that the behavior of these intraepithelial lesions cannot be predicted with accuracy and that even mildly abnormal changes, when left untreated, may progress to carcinoma in situ and invasive cancer. 83, 84 Burghardt reported on the direct development of invasive cancer from dysplasia with an interphase of carcinoma in situ. 85
The marked variation in morphology, even from one site to another within the same cervix, is undoubtedly the cause for the confusion in terminology and the difficulties in evaluating the biologic significance of these reactions.
Fundamentally, dysplasia represents a reaction to injury in the sense that a stimulus acting on a normal epithelium results in some morphologic alteration of that epithelium.
It is unpredictable where (or when) the stimulus will initiate an abnormal reaction (i.e. in mature stratified squamous epithelium, mature squamous metaplasia, or immature squamous metaplasia). The form of dysplasia that results may then depend on the maturity of the epithelium that reacts. The morphologic features of a dysplastic lesion involving an immature squamous metaplastic epithelium are relatively uniform and differ from a similar reaction involving mature stratified squamous epithelium.
The stimulus exerts its effect on the cells of the basal layer of mature ectocervical stratified epithelium or on the reserve cells or immature squamous metaplastic cells in the endocervical canal. This would initially be demonstrated by a defect in the mitotic mechanism of the cell, resulting in scattered cells with an abnormal DNA content, demonstrated by the appearance of enlarged nuclei with an abnormally distributed hyperchromatic chromatinic material. Differentiation of the cytoplasm is disturbed owing to an abnormal stimulus from the nucleus.
With a continuing stimulus, persistence of some of the abnormal cells capable of complete division and survival might lead to an increase in the number of abnormal cells. Continued selection and the persistence of the abnormal cells increasingly involve multiple layers of the lining epithelium, thus increasing the severity of the lesion.

Cervical Intraepithelial Neoplasia
Because of a great deal of confusion about the clinical significance of the terms dysplasia and carcinoma in situ, the term cervical intraepithelial neoplasia was introduced in an effort to bring a more active approach to the evaluation and treatment of noninvasive epithelial abnormalities of the cervix. 86
The concept of CIN reflects the basic unity of precancerous changes, regardless of the phenotype of such epithelial abnormalities. 86 - 89 It was meant to replace the prior system of nomenclature, such as dysplasia and carcinoma in situ.
Data on the progression rate of dysplasias of different grades of severity and carcinoma in situ emphasize the continuous nature of the epithelial alterations ultimately leading to invasive squamous cell cancer. On this basis, it seemed advisable to abandon the artificial distinction between dysplasia and carcinoma in situ. 89
The terminology of CIN subclassified into grades of dedifferentiation essentially has not provided an advantage over the former subclassification of dysplasias. It has always been recognized that within the group of dysplasias, a proportion of the lesions may progress to a more severe abnormality, eventually even to an invasive process. Many cases of CIN are not truly neoplastic in character but represent a nonspecific response to injury due to chronic irritation or inflammation. Although the presence of an inflammatory process does not preclude the possibility that the epithelial abnormality is not truly neoplastic in nature, it remains true that in a residue of cases, nonspecific reactive changes very closely mimic those of true CIN. 7 Richart and Barron apply strict criteria to lesions before they can be included in the group of CIN. 89 However, these are applicable only in retrospect, when epithelial changes have proved to persist for a certain period. As stated by these investigators, the progressive or indolent nature of dysplastic lesions cannot be judged on the basis of mere light microscopic examination. This, however, is equally impossible in grading intraepithelial neoplasias. A relatively large percentage of the lesions do not progress during long-term follow-up. Within the morphologic spectrum of CIN, one must include lesions known as koilocytotic atypia, recognized as due to papillomavirus infection. CIN can be graded from 1 to 3 to reflect the degree of epithelial abnormality, provided that no prognostic significance is attached to this classification. The implication of the CIN concept is that all patients with abnormalities, whatever their grade, must be referred for colposcopic examination to be evaluated further. 64 Koss advises following CIN grade 1 (mild dysplasia) with cytology or destroying the lesion. 64 In daily practice, the subclassification into CIN grades 1 through 3 has not reduced the number of women referred for colposcopic evaluation or biopsies, and the diagnoses CIN grades 1 through 3 still include a large number of lesions that spontaneously regress because they were of a reparative or a reactive nature.

Electron Microscopy
Although it is possible to subclassify CIN into mild, moderate, and severe dysplasia and carcinoma in situ by light microscopy, such a subclassification is difficult at the subcellular level.
The nuclear enlargement may reflect the increase in DNA synthesis that accompanies the decreased generation time in dysplasia and carcinoma in situ. Mitotic figures are rarely observed in normal epithelium but are frequently seen in dysplasia and carcinoma in situ, in keeping with the more rapid growth rate. The mitotic figures in dysplasias do not generally differ from those in normal epithelium, but in rare mitotic figures that are observed, the number of chromosomes in nuclei appear to be increased. 90
In a comparative study between normal and dysplastic cells using scanning electron microscopy, Kenemans and co-workers demonstrated differences in surface architecture between the basal and luminal side of normal intermediate squamous cells and between normal and abnormal cells, both in exfoliated cells and in cells from tissue specimens. 91, 92 The luminal side of epithelial cells bears microridges, and the basal side contains microvilli. All cases of histologically established (moderate or severe) dysplasia showed a remarkable increase in the number of microvilli and a decrease in the number of desmosomes. Abnormal configuration of microvilli was also observed in epithelial abnormalities. These changes have also been documented in cells from invasive carcinoma. The decrease in number of desmosomes is in accord with the finding that neoplastic cells are more loosely attached to one another than are normal cells.
The lack of adhesiveness between the epithelial cells in dysplasia and carcinoma in situ could also account for the increased number of leukocytes in the altered epithelium, because it would facilitate the penetration of the epithelium by inflammatory cells.
Surface cells in dysplasia, although they may appear flattened under the light microscope, have cytoplasm containing large numbers of ribosomes, numerous mitochondria, and decreased or absent storage of glycogen, which is in keeping with their less-differentiated state. Increased numbers of ribosomes and polyribosomes have also been reported in cervical squamous cell carcinoma. 90, 93

Cervical Intraepithelial Abnormalities

Origin and Localization
Dysplastic lesions occur significantly more frequently on the anterior than on the posterior lip of the uterine cervix 22 and are usually localized in the endocervical mucosa at the transformation zone in the region of the cervical os or in the epithelium covering the ectocervix or portio vaginalis.
The majority of abnormal surface reactions begin within the area of the endocervical-lining epithelium and mimic in an abnormal fashion various stages in the process of squamous metaplasia. 73 There is some confusion about the exact site and cells of origin of cervical intraepithelial abnormalities. Koss identifies the site of origin of more than 90% of the intraepithelial abnormalities to be the area of squamous epithelium bordering the columnar epithelium, known as the squamocolumnar junction or transformation zone 94 ( Fig. 8.52 ). The remaining fewer than 10% of lesions are then thought to originate in the area of the columnar epithelium. This would mean that the cell of origin of most of the intraepithelial lesions would be the basal cell of the original ectocervical squamous epithelium overlying an area without gland-bearing stroma. This is in conflict with the observation by Reagan and Patten that only 11.1% of a series of dysplasias were located in an area of the cervix without underlying glands. 95 In 48.9% of specimens, the site of involvement was partially related to gland-bearing stroma, and the remaining 40% of the lesions were confined to a site with underlying glands. 8 From these data, it appears that in 88.9% of their specimens there had been at least partial involvement of the arena in which squamous metaplasia occurs as the basic reaction to injury of the columnar epithelium. The latter distribution is more in line with the experience from routine clinical practice that the majority of intraepithelial changes and virtually all of the severe intraepithelial abnormalities extend into the invaginations of the columnar epithelium. It is a common observation that the severity of the change in the invaginations is less than that of the surface change.

Fig. 8.52 Transformation zone. The most distal part of the endocervical columnar epithelial lining bordering the original stratified squamous nonkeratinizing epithelium has been replaced by mildly atypical squamous metaplastic epithelium (H&E × MP).

Morphology Related to Origin
From the spectrum of morphologic patterns of intraepithelial abnormalities occurring on the ectocervix and in the epithelial lining of the endocervical canal, it is evident that the potential of reserve cells to differentiate into squamous epithelium through an intermediate stage of immature squamous metaplasia gradually diminishes in the proximal direction along the canal. This is reflected in the epithelial abnormalities originating at different sites. The dysplastic lesions occurring close to the squamocolumnar junction usually still demonstrate some maturation into squamous-like epithelium in the superficial layers, whereas abnormalities occurring proximally in the canal completely lack this squamous differentiation and seem to be composed entirely of undifferentiated primitive cells (atypical reserve cells). The resemblance to squamous epithelium of the more distal lesions does not necessarily mean that these, although often erroneously named “better differentiated” when compared with classic carcinoma in situ, would have a less malignant potential. In this respect, the relatively large-cell severe dysplasia should be considered of comparable severity as the more classic carcinoma in situ and treated accordingly.

Grading
The current concept of dysplasia and carcinoma in situ is that these form a continuous spectrum of a developing intraepithelial abnormality. Many have believed such a continuum to be in conflict with a subdivision into grades. However, on the basis of histologic criteria, it is not only possible to grade the severity of the change but also clinically relevant, because grading provides a basis for the mode of treatment. Grading lesions also makes these lesions accessible for retrospective and prospective studies, which are the only available means to achieve more insight into the true nature of these changes. The fact that it has been proved to be very difficult to diagnose these lesions accurately and reproducibly can never be the reason to refrain from categorizing these lesions.
Grading of dysplastic lesions may give more insight into the morphologically different types of dysplasia as well as into the biologic potential of these types.
The more closely the lesion resembles normal epithelium, the less severe that lesion is thought to be. The more that primitive cell types that lack signs of differentiation predominate, the more severe a lesion is considered to be. Thus, it is possible to apply the terms minimal, slight, moderate, and severe to differentiate on a morphologic basis—phenotypically—one lesion from another.
Unfortunately, the degree of morphologic abnormality does not necessarily correlate with the biologic potential of the epithelial change. 12 Because of this frequent lack of correlation, Friedell recommended a morphologic classification incorporating both qualitative and quantitative information. 96
Fundamentally, the intraepithelial neoplastic—dysplastic—reaction is characterized by premature keratinization of component cells and abnormal differentiation of a variable number of cell layers that, in the most severe forms, encompasses all cell layers throughout the whole thickness of the mucosal lining.
In all dysplastic changes, by definition, in the uppermost layers of the epithelium, cells that remain have features of normal epithelial cells. This is usually more apparent in dysplasia originating in the original squamous epithelium than in changes involving the squamous metaplastic epithelium.
Cells in the upper layers of dysplasia have features reflecting differentiation. However, the nuclei are usually enlarged and more hyperchromatic than in normal squamous epithelial cells, and the nuclei of the cells show remarkable variation, in contrast to the rather uniform size and shape of nuclei in normal epithelial cells at a comparable level.
In lesions involving mature squamous epithelium, individual cell keratinization and keratohyalin pearl formation are not uncommon.
Mitoses are usually found. In less severe lesions, mitoses, both normal and abnormal forms, are confined to the lower half or lower two-thirds of the epithelium. In more severe changes, mitoses can also be found in the upper third of the mucosal lining.

Occurrence
The mean age at which cervical intraepithelial lesions are diagnosed seems to be gradually decreasing. This might be related to changes in sexual behavior within Western countries and a younger age of first sexual activity. Although several observers have reported that invasive cancers are also diagnosed with greater frequency in younger women, 69 this has so far not resulted in a greater mortality from cervical cancer in women in younger age groups. The reported increased frequency of intraepithelial lesions in younger women may in part be due to an increased frequency of cytologic testing and may not be a true increase in the number of intraepithelial abnormalities.
On the basis of cellular evidence of dysplasia, Patten reported a prevalence of 0.98% in 57,469 nongravid women. 13 Reagan and co-workers reported a prevalence of 0.77% in a study of 10,533 women. 97 On the basis of histologic evidence, the prevalence has been reported to vary between 1.2 98 and 3.2%. 97 In our own experience with women ages 35 through 55 years, mild and moderate dysplasia was found in 1.6% and severe dysplasia and carcinoma in situ in 0.3% of women at first screening. 76

Age at Detection
The stages of evolution of the dysplastic process might be reflected in the age at detection. Based on cellular evidence of dysplasia, the mean age at detection was 34.7 years. 9 Patients with slight dysplasia were found to have a mean age of 32.0 years. With moderate dysplasia, the mean age was 35.7 years, and in patients with severe dysplasia, the mean age was 38.4 years. 9
Reagan and associates reported a mean age of 34.2 ± 1.6 years in women having only minimal-to-slight dysplastic lesions and of 41.4 ± 3.0 years in reactions classified as severe dysplasia. 97 In our own series, the average age of detection of moderate dysplasia was 36.8 years and of severe dysplasia 35.7 years ( Figs 8.53 and 8.54 ).

Fig. 8.53 Moderate dysplasia. Histologic diagnoses. Age distribution–average age is 36.8 years (SD 9.1) and n = 420.
(Data from the Nijmegen Registry of Cervical Cytology, 1978–1987.)

Fig. 8.54 Severe dysplasia. Histologic diagnoses. Age distribution–average age is 35.7 years (SD 8.7) and n = 609.
(Data from the Nijmegen Registry of Cervical Cytology, 1970–1987.)

Histology
In the dysplastic lesion, abnormal cells are present throughout all layers of the epithelium. Signs of differentiation occur in inverse relation to the severity of the lesion, from the basal layers to the surface. The morphology of cells composing the superficial layers of a dysplastic lesion is related to the severity of the dysplastic lesion. The dysplastic lesion is characterized by premature keratinization of component cells, abnormal differentiation of the cells composing the epithelial lining, and abnormally large nuclei in association with various degrees of cytoplasmic maturation (usually various degrees of abnormal maturation of the cytoplasm).
Cells present in the upper layers of the mucosa reflect in their morphology the entire cascade of maturation steps throughout the epithelium. The more dedifferentiated the cell in the most basal layers, the less influence maturation stimuli have during this cell's passage through the layers of the epithelium, and the greater the remaining abnormality of the cell that finally reaches the superficial layers. From the morphology of these superficial cells, which are mechanically removed from the superficial layers of the mucosa, an experienced cytopathologist can rebuild an image of the histopathologic appearance of the mucosal lining at the site of the scrape and thus give an impression of the severity of the abnormality at that site.

Cytology
Shortly after Ayre introduced the spatula for making cervical smears, 99, 100 it was reported that precancerous and cancerous changes still confined to the mucosa of the cervix could be detected in these cytologic samples. 100 - 103 The number of abnormal cells in a cell preparation taken from epithelial lesions of comparable severity may differ from one case to the other, depending on the method of collection and the skill of the person taking the sample. 26 In general, the number of abnormal cells is related to the severity (and the extent) of the lesion. The lesser the abnormality, the fewer abnormal cells are found in the specimen. The more severe the lesion, the higher the number of abnormal cells. On the basis of cell population evaluation, one can obtain rather specific information on differentiation characteristics of the parent lesion. 73 With experience, a more definitive interpretation can be made on the basis of cellular specimens, because both cancerous and noncancerous changes in the surface mucosa of the uterine cervix are reflected in the desquamated cells. 14
Abnormal cells originating from the surface of epithelial abnormalities may be subdivided morphologically into two groups. Those showing signs of differentiation of squamous type have, depending on the degree of maturation, features reminiscent of superficial, intermediate, or parabasal squamous cells ( Figs 8.55 and 8.56 ). When signs of differentiation are almost completely absent, cells bear a resemblance to reserve cells.

Fig. 8.55 Mild dysplasia. Cells of squamous type with slightly enlarged nuclei and some hyperchromasia (Papanicolaou × OI).

Fig. 8.56 Mild dysplasia. Cells of squamous type with slightly enlarged nuclei. The nuclear chromatin is finely granular and very slightly hyperchromatic (Papanicolaou × OI).
Patten, in an attempt to provide a morphologic terminology for dysplasia that might be applicable to both histologic and cellular material and might provide evidence for biologic potential, introduced a subclassification of dysplasia for routine use. 12 The major subdivisions were (1) keratinizing (ectocervical) dysplasia, (2) nonkeratinizing dysplasia, and (3) metaplastic dysplasia.
In a series of 2453 cases, the nonkeratinizing variant was observed approximately 7 times more frequently than the metaplastic type and 25 times more frequently than the keratinizing variant. Admixtures of these different types are most often represented by the simultaneous occurrence of cells consistent with the nonkeratinizing and metaplastic variants. Of this series, about 85% of dysplasias in a 5-year period progressed to carcinoma in situ. The reactions are further classified according to severity by adding the terms minimal, slight, moderate, and marked (severe). 73

Arrangement of Cells
Abnormal cells from dysplastic lesions usually appear singly and have well-defined cell borders. In the majority of specimens, atypical cells are also found in sheets. Cell borders usually are still recognizable. The presence of cell aggregates with indistinct cell borders in a cellular specimen indicates a reduced tendency to maturation in the mucosal lining. This reduced maturation is a reflection of the dedifferentiation of the component epithelial cells that lack the stimulus to mature. In such an aggregate, the component cells are regularly arranged with relation to one another. Less frequently, cells may be arranged in syncytial masses. Here the component cells are irregularly arranged with relation to one another and have indistinct cell borders. A syncytial arrangement is more commonly associated with carcinoma in situ and invasive cancer.
In almost all cases of dysplasia, sheet-like arrangements are found. Syncytial masses are found in only 10% of specimens and are always associated with a severe epithelial abnormality. 8, 12

Cell Size
The volume and the condition of the cytoplasm are a reflection of the state of maturation and differentiation of the cells. Usually found is an admixture of cells of various stages of maturity. In view of the site of origin of dysplasias and the preceding metaplastic process in the endocervical canal, the size of cells involved in dysplastic changes may vary from almost the size of a normal superficial squamous cell in minimal abnormalities to the size of an immature basal cell or a very immature squamous metaplastic cell in more severely abnormal changes.
Relating cell size to the severity of the histopathologic change in dysplasias, those samples containing dysplastic cells with cell areas predominantly in the range of normal squamous cells tend to originate from a less atypical (less severe) dysplastic lesion. These reactions are more frequently located on the portio vaginalis or in relation to the external cervical os. Those samples with dysplastic cells possessing cell characteristics more reminiscent of squamous metaplastic cells are more likely to have arisen in the area of the distal portion of the endocervical canal (transformation zone).
Most of the abnormal cells observed in the presence of dysplasia, carcinoma in situ, and invasive cancer possess cell sizes as observed in cells from immature squamous metaplasia and reserve cell hyperplasia. The relative nuclear area, nuclear shape, and particularly intranuclear chromatin architecture should then provide the basis for the right diagnosis. 8

Cell Configuration
The shape of an abnormal cell in a sample may also reflect the maturity of the parent tissue reaction. Dysplasia is characterized by a predominance of polygonal cells accounting for about 55% of the abnormal cell population (52.7 27 and 56% 104 ). Round or oval forms, indicative of a less mature reaction, represented about 40% of the abnormal cells in samples studied (41.6 27 and 40% 104 ).
A predominance of polygonal forms, often found together with eosinophilic staining of the cytoplasm, suggest an origin from a dysplastic lesion, originating in original squamous stratified epithelium. A predominance of oval forms is suggestive of a dysplastic lesion in an area of squamous metaplasia, most likely the transformation zone of the endocervical canal. The presence of spindle-shaped or elongated cells may indicate the presence of (abnormal) keratinization at the surface of the epithelium. The elongated cells often show fibrillary structures in the cytoplasm. The presence of these fibrils is indicative of keratinization in the dysplastic process. Keratinization overlying a dysplastic lesion most often occurs in the ectocervical epithelium, but a keratinizing dysplasia may occasionally be found in the endocervical canal in a metaplastic epithelium.
Anucleated squames, sometimes with pale yellow cytoplasm, may also be found. When contamination from the vulvar mucosa can be precluded, the presence of these squames should always lead to an extra awareness on the part of the cytopathologist. Severe dysplastic lesions or even keratinizing squamous cancer may sometimes be covered by a thick layer of hyperkeratosis.
When making a scrape from this area, the cellular material obtained may be restricted to keratinized squames. Owing to the resistant cover, deeper layers have not been sampled and the true lesion remains obscure.
When anucleated squames are diagnosed in a cell sample, even without atypical cells being observed, a repeat smear should be advised and the physician should be specifically instructed to sample any area of leukoplakia (literally white patch ) very carefully, preferably by taking successive smears from the same area, thus gradually uncovering the nonkeratinized part of the lesion.

Nuclear Morphology
Nuclear characteristics are the main determinants for the grading of an epithelial abnormality. Although cytoplasmic features may provide additional information about the origin and degree of maturation of a cell, the main important denominator of the severity of an epithelial abnormality remains nuclear changes.
Nuclear atypia should be classified as mild, moderate, or severe. Cytoplasmic changes should be classified according to quantity, density, staining quality, and shape.
The morphology of the cell nuclei in cases of epithelial changes comprises a combination of any number of the following: 71
• Disproportionate nuclear enlargement;
• Irregularity in form and outline;
• Hyperchromasia;
• Irregular chromatin condensation;
• Abnormalities of the number, size, and form of the nucleoli; and
• Multinucleation.
Papanicolaou introduced the term dyskaryosis to designate certain cytologic patterns observed in vaginal and cervical smears from cases of early carcinoma and some other pathologic lesions of the uterine cervix. 105, 106 In these lesions, the exfoliated cells are characterized by marked nuclear abnormalities consistent with the generally accepted criteria of malignancy, although the cells as a whole may show no significant deviation from their standard normal type. He described the morphology of dyskaryosis cells as follows: “The nuclei show distinct abnormal features such as enlargement, hyperchromasia, anisokaryosis, bi- or multinucleation et cetera.” Patten strongly advocates avoiding the use of subjective terminology to describe morphologic changes and thus rejects the use of the term dyskaryosis “which although useful during the developmental stages of applied cytology, presently has no place in the vocabulary of the diagnostic cytologist except for historical reflection.” 12 The working party of the British Society for Clinical Cytology, 71 however, endorsed the recommendation made by Spriggs and associates 107 to use the terms dyskaryosis and dyskaryotic in the description of nuclear abnormalities in both squamous and endocervical cells in intraepithelial lesions as well as in invasive carcinoma.

Relative Nuclear Area
The relative nuclear area is an expression of nuclear area in relation to cytoplasmic area. The relative nuclear area increases with the severity of the lesion from minimal dysplasia to carcinoma in situ.
Nuclei of dysplastic cells are relatively large when compared with their normal counterparts. The greater size of the nuclei is a reflection of the reduced maturation of the cell, indicated by a relatively large nucleus and a relatively small amount of cytoplasm.
Actual nuclear area is of less practical importance in routine diagnostic cytology because of the lack of a possibility for an accurate comparison with an object of known size.

Mild Dysplasia (CIN Grade 1, Low-Grade SIL)

Histology
In mild dysplasia, a slight disturbance of the regular arrangement of cells is seen. The upper two-thirds of the epithelium usually shows a relatively regular arrangement of cells with preserved stratification. These layers are composed of cells, recognizable as intermediate-type and superficial squamous-type cells, with slightly reduced cytoplasmic volume and slightly increased nuclear size. Nuclei are usually of normal round-to-oval shape and have a minimally hyperchromatic nuclear chromatin. Aberrations of the nuclear morphology are predominantly limited to the most basal layers of the epithelium.

Cytology
Cells from mildly atypical lesions such as mild dysplasia (CIN grade 1) usually have plentiful clear, translucent cytoplasm with well-defined angular borders. Cells resemble intermediate- and superficial-type squamous cells with a somewhat reduced cytoplasmic body and a slightly enlarged nucleus, occupying less than one-third of the total area of the cell ( Fig. 8.57 ; see also Figs 8.55 and 8.56 ). Nuclear chromatin is finely granular, evenly distributed, and only slightly hyperchromatic ( Fig. 8.58 ).
Key features of mild dysplasia (CIN 1, LSIL)
• Singly lying cells of superficial and intermediate type—slightly reduced cell size;
• Abundant cyanophilic and eosinophilic translucent cytoplasm with well-defined angulated borders;
• Cells resemble intermediate or superficial squamous cells;
• Enlarged nucleus (3–5× intermediate cell nucleus) occupies one-third of cytoplasmic area;
• Chromatin finely granular, evenly distributed, with slight hyperchromasia; and
• Nucleoli not present.

Fig. 8.57 Mild dysplasia of metaplastic type. Enlarged nuclei and increased nucleocytoplasmic ratio. Compare cell size with the superficial squamous cell and the leukocytes (Papanicolaou × OI).

Fig. 8.58 Mild dysplasia. Irregular arrangement of cells in the basal and parabasal layers. Reduced maturation and slightly enlarged, somewhat hyperchromatic nuclei in the most superficial layers (H&E × HP).

Moderate Dysplasia (CIN Grade 2, High-Grade SIL)

Histology
In moderate dysplasia, there is a moderate disturbance of stratification. Usually only the upper third of the epithelium still shows evidence of stratification of cells of superficial and intermediate cell size. The uppermost layers are still composed of flat squamous cells, although nuclei may be enlarged and slightly hyperchromatic ( Fig. 8.59 ). The surface layers occasionally are keratinized, with loss of nuclei and the formation of a granular layer. Nuclear abnormalities may be seen throughout the epithelium, particularly in the more basal layers. Cell arrangement is disturbed in as much as two-thirds of the thickness of the epithelium ( Figs. 8.60 and 8.61 ). In these disturbed layers, mitoses, sometimes abnormal, are present in increased numbers.

Fig. 8.59 Moderate dysplasia. Irregular arrangement of cells in basal and intermediate cell layers. Reduced maturation, nuclear enlargement, and irregularities in nuclear size and shape still appear in the more superficial layers (H&E × HP).

Fig. 8.60 Moderate dysplasia. Disturbed maturation in all cell layers. In the uppermost layers, stratification is still recognizable. Moderate increase of nuclear size. Irregular nuclear shapes and some hyperchromasia (H&E × HP).

Fig. 8.61 Moderate dysplasia of metaplastic type bordering endocervical mucus-producing columnar cells. Disturbed maturation of superficial layers. Nuclei show irregular sizes and shapes. Hyperchromasia of nuclear chromatin (H&E × HP).

Cytology
The size of abnormal cells is more variable. Next to some abnormal cells of the superficial squamous cell type, smaller cells of the intermediate and parabasal cell type are usually found ( Fig. 8.62 ). Most cells are round to oval, but spindle cells and elongated and bizarre shapes may occasionally be found. Cytoplasmic staining is cyanophilic, but a relatively high number of cells may show eosinophilia of the cytoplasm. Nuclei are enlarged and round to oval, sometimes elongated or irregularly shaped ( Fig. 8.63 ). Nuclear chromatin is evenly distributed and slightly to moderately hyperchromatic ( Fig. 8.64 ). Nucleoli are usually absent. The nucleocytoplasmic ratio is increased, both by nuclear enlargement and by reduction of the cytoplasmic volume ( Figs 8.65 and 8.66 ). The nucleus generally occupies less than half of the total area of the cell.
Key features of moderate dysplasia (CIN 2, HSIL)
• Singly lying cells and cells in sheet-like arrangement of superficial, intermediate, and parabasal type;
• Moderately reduced cell size;
• Cells predominantly round to oval, but occasional spindled, elongate, or bizarre cells may be seen;
• Cytoplasm is cyanophilic, but more mature cells may be eosinophilic with distinct cell borders;
• Nuclei enlarged, round to oval, with some irregular or elongate;
• Chromatin irregularly distributed with slight to moderate hyperchromasia;
• Nucleoli usually not present; and
• Elevated nucleus-to-cytoplasmic ratio usually one-half of cytoplasmic area.

Fig. 8.62 Moderate dysplasia. Cells of squamous and squamous metaplastic type. Nuclear enlargement, moderate hyperchromasia of the finely granular chromatin. Increased nucleocytoplasmic ratio (Papanicolaou × OI).

Fig. 8.63 Moderate dysplasia. Nuclei are enlarged and show irregularities in shape. Nuclear chromatin is evenly distributed and finely granular (Papanicolaou × OI).

Fig. 8.64 Moderate dysplasia. Irregularly shaped cells and enlarged nuclei. Moderate hyperchromasia of finely granular nuclear chromatin. Compare size of cells and nuclei with superficial squamous cell (Papanicolaou × OI).

Fig. 8.65 Moderate dysplasia. Round, oval, and elongated cells with enlarged, somewhat irregularly shaped nuclei. Increased nucleocytoplasmic ratio. Finely granular, slightly and moderately hyperchromatic nuclear chromatin. Compare with intermediate squamous cells (Papanicolaou × OI).

Fig. 8.66 Moderate-to-severe dysplasia. Cells vary in size and shape. Nuclei are round to oval and irregular in shape. Nuclear chromatin is in part finely, in part coarsely, granular and moderately hyperchromatic. Increased nucleocytoplasmic ratio (Papanicolaou × OI).
According to Koss, virtually all lesions classified as moderate dysplasia (CIN grade 2) must be considered as neoplastic events in the squamous epithelium. 64 Some of the abnormalities may resemble flat condylomata, and in a proportion of these moderately abnormal lesions HPV may be documented, particularly with the application of recent highly sensitive in situ hybridization techniques. Meisels and associates proposed the term atypical condylomata for these lesions. 108 Because of their progressive potential, Koss advises eradicating all moderately abnormal lesions under colposcopic control. 64 This also applies to the atypical condylomata associated with other forms of CIN.

Severe Dysplasia and Carcinoma In Situ (CIN Grade 3, High-Grade SIL)

Histology
In severe dysplasia, cells show a greatly disturbed arrangement in all three layers of the epithelium ( Fig. 8.67 ). Stratification is present in only the most superficial layers ( Fig. 8.68 ). Throughout the entire epithelium, cells show a reduced maturation with loss of cytoplasmic volume and an increased nuclear size. Cells and nuclei vary in size and shape and often have irregular forms. Differentiation in intermediate-type and superficial squamous-type cells may be lost ( Fig. 8.69 ). Nuclei have a hyperchromatic, irregularly distributed, coarsely granular chromatin. Mitoses may be found throughout all epithelial layers. The abnormal changes often extend into the stromal invaginations of the endocervical epithelium ( Fig. 8.70 ).

Fig. 8.67 Severe dysplasia. Irregular arrangement and disturbed maturation of cells involving almost the entire thickness of the epithelium. Only in the most superficial layers do cells show an increased cytoplasmic volume resulting in a lower nucleocytoplasmic ratio (H&E × HP).

Fig. 8.68 Severe dysplasia. Palisade arrangement of cells in the lower half of the epithelium. The upper half of the epithelium is composed of abnormal cells with irregular nuclei and hyperchromatic nuclear chromatin lying parallel to the surface. The sudden change in the polarity of the cells in this abnormal epithelium may reflect a sudden change in the environment during the development of the lesion (H&E × HP).

Fig. 8.69 Severe dysplasia. Only in the most superficial layers are signs of maturation identifiable by a change in polarity and a somewhat increased cytoplasmic volume of the cells. Vacuolation of the cytoplasm of the superficial cells (koilocytosis) may be caused by infection with human papillomavirus but is more often a sign of degeneration (H&E × HP).

Fig. 8.70 Severe dysplasia of metaplastic type. (A) Extension of the epithelial abnormality into the stromal invagination, with (B) replacement of the lining columnar epithelium (H&E (A) × HP and (B) × HP).

Cytology
The size of the cells in severely abnormal intraepithelial changes is comparable with the parabasal cell type. Cytoplasm is usually sparse, typically forming a small rim around the nucleus ( Fig. 8.71 ). Cells are round to oval and often irregular or elongated ( Figs 8.72 and 8.73 ). The keratinizing squamous lesions sometimes contain large cells with plentiful, often eosinophilic, cytoplasm. Cells are seen singly as well as in aggregates. In the most severe intraepithelial lesions, aggregates have a syncytial composition, with indistinct cell borders and irregularly arranged nuclei ( Fig. 8.74 ). The morphology of these aggregates is consistent with the histologic evidence of the lack of cell maturation and the irregular arrangement of nuclei in the most superficial layers of severely abnormal epithelium ( Figs 8.75 and 8.76 ). The nucleus usually occupies at least two-thirds of the total area of the cell. Nuclei have a hyperchromatic, irregularly distributed, coarsely granular chromatin ( Fig. 8.77 ). In actively proliferating lesions, eosinophilic-staining nucleoli may be observed, but these more often are obscured by the dense hyperchromatic chromatin. Some severe dysplasias show an extreme irregularity in shape and size of the composing cells. In cervical smears, these lesions may present with large, bizarre-shaped cells with highly abnormal hyperchromatic nuclei. Differentiation from invasive squamous cell cancer may at times be extremely difficult ( Figs 8.78 and 8.79 ).
Key features of severe dysplasia (CIN 3, HSIL)
• Cells similar in appearance to parabasal cells;
• Sparse cytoplasm in rim around nucleus;
• Cells isolated or in syncytial groups with indistinct cytoplasmic borders and loss of polarity;
• Enlarged nucleus occupies three-quarters of cytoplasmic area;
• Dense hyperchromatic chromatin often with irregular coarseness;
• Nucleoli usually absent; and
• Spindled, elongate, or bizarre forms may be present.

Fig. 8.71 Severe dysplasia. Cells of parabasal cell type with a small rim of cytoplasm around a large, often irregularly shaped, hyperchromatic nucleus. High nucleocytoplasmic ratio (Papanicolaou × OI).

Fig. 8.72 Severe dysplasia. Cells and nuclei are irregularly shaped. Finely granular and coarse, moderately hyperchromatic nuclear chromatin (Papanicolaou × OI).

Fig. 8.73 Severe dysplasia. Large, elongated cell with relatively large nucleus and dense, hyperchromatic nuclear chromatin (Papanicolaou × OI).

Fig. 8.74 Severe dysplasia. Singly lying cells and cells in a syncytial aggregate with indistinct cell borders and irregularly shaped hyperchromatic nuclei (Papanicolaou × OI).

Fig. 8.75 Severe dysplasia. Irregular arrangement of abnormal squamous cells in the most superficial layers (H&E × HP).

Fig. 8.76 Dysplasia. Cells desquamating from the superficial layers reflect the lack of maturation and the grade of abnormality of the epithelial lesion (H&E × HP).

Fig. 8.77 Severe dysplasia. Cells in a syncytial aggregate and singly lying with relatively large irregular nuclei and a dense, coarsely granular, irregularly distributed hyperchromatic nuclear chromatin (Papanicolaou × HP).

Fig. 8.78 Dysplasia of metaplastic type. Irregular arrangement of sometimes bizarre-shaped abnormal cells, bordered by regular endocervical columnar epithelium (H&E × HP).

Fig. 8.79 Large abnormal squamous cell. with multiple irregular, hyperchromatic nuclei. Compare the size of this cell with the other abnormal squamous cells (Papanicolaou × OI).

Biologic Significance Follow-up
The evolution of invasive squamous cell cancer involves a number of stages with increasing intraepithelial abnormality designated as dysplasia, carcinoma in situ, and microinvasive carcinoma. Although it is not usually possible to predict the malignant potential of an epithelial abnormality (premalignant lesion), evidence suggests that mild dysplasias are more prone to regress spontaneously, and conversely, severe dysplasia and carcinoma in situ are more likely to persist or progress.
Dysplasia may follow a variable course.

Regression
Dysplastic lesions may undergo spontaneous regression. A lesion may disappear in a matter of weeks and become undetectable at a short-term follow-up screening, or it may take a much longer time, becoming gradually less abnormal in a matter or months or even years. In some cases, the disappearance may be caused by desquamation of the abnormal epithelium due to minimal trauma.
A dysplastic epithelium tends to separate more easily from the supporting stroma than the normal epithelium does. This may account in part for the disappearance of dysplastic lesions after initial cytologic diagnosis in women during cytologic follow-up. In these instances, the diagnostic test essentially becomes a cure for the lesion.
In studies reporting on regression or progression after confirmation of the cytologic diagnosis by histology, part of the observed regression during follow-up must be ascribed to the surgical procedure, as may be the case in studies reporting on follow-up of dysplasias diagnosed during pregnancy, due to the traumatic effects on the cervix during childbirth. 109 Spontaneous regression may therefore be lower than cited in these studies, whereas progression may be higher in cases that have not had bioptic intervention or that are diagnosed during pregnancy.
Even if regression does not follow the biopsy, however, it is entirely possible that the biology of the lesion is altered in various ways. If the smear is positive after a biopsy and later becomes negative, the assumption that the lesion regressed spontaneously may be invalid. 86
Koss and colleagues, in a long-term follow-up study of individual patients with carcinoma in situ and related lesions, noted that a single biopsy could eradicate an area of intraepithelial neoplasia. 110 There can be no doubt that punch biopsies can eradicate areas of CIN completely, either directly by complete removal or indirectly by altering the balance between the host and the neoplasm so that areas of residual CIN regress, thus producing an immediate cure, a delayed cure, or a change in the distribution of an area of CIN. 86 In follow-up studies after biopsies, no valid figure for the regression rate can be derived, because the proportion of cures inadvertently produced by the diagnostic procedures rather than occurring spontaneously cannot be determined.

Persistence
Dysplasia may also persist during a variable period of time before regressing or progressing to a more severe lesion, such as marked dysplasia, carcinoma in situ, or invasive cancer. Documented cases of marked dysplasia have persisted for as long as 20 years without showing malignant progression. There is unanimity among investigators that dysplasia in certain circumstances can progress to carcinoma in situ and finally to invasive cancer.

Progression
Slight dysplasias may also antedate the appearance of a severe intraepithelial lesion or an invasive cancer by many years. Most prospective studies, when carried out without intercurrent intervention through biopsies or frequent repeat smears, indicate that fewer than 15% of dysplasias progress. 73

Follow-up Studies: Literature Review
Christopherson monitored more than 200 patients with dysplasia from 1 to 13 years. 111 Patients were included on the basis of a biopsy diagnosis of dysplasia that later was cytologically confirmed. During the observation period, 30% of the lesions regressed to normality, 49% of lesions persisted, 20% showed progression to carcinoma in situ, and 1.3% evolved to invasive cervical cancer.
In a study of 120 women with cytologic evidence of dysplasia, during long-term follow-up for up to 9 years, Scott and Ballard found regression or complete involution in 60% of women; progression to carcinoma in situ in this series was 4.1%. 112 In 223 women monitored for an average of 3.8 years, progression to carcinoma in situ was 5.8%. 113 In a retrospective study of 364 women with dysplastic lesions of varying severity that were initially detected and closely monitored by cytology alone for a minimum of 9 months and a maximum of 24 months, Patten found the reaction to regress or disappear in 71.9% with slight dysplasia, 44.2% with moderate dysplasia, and 16.3% with marked dysplasia. 9 Furthermore, during this period of observation, only those cases initially classified as marked dysplasia progressed to carcinoma in situ. Of cases initially diagnosed as marked dysplasia, 6.8% (3) developed carcinoma in situ from 10 to 17 months later. None of these cases were seen to evolve to invasive cancer. In a retrospective study of 102 women with moderate dysplasia, Patten noted regression in 44.2%, persistence in 42.2%, and progression in 13.6% during follow-up from 9 to 23 months. 13
Progression occurred on the average 3 years after initial diagnosis of dysplasia. Patten stated that on the basis of data available, 5–10% of dysplastic lesions could be expected to antedate the subsequent appearance of cancer in situ. 12 From reported series, it seems likely that in fewer than 1.5% of patients with evidence of dysplasia will an invasive cancer subsequently develop. 98, 111, 112
Nasiell and associates monitored 894 women with cytologically diagnosed moderate dysplasia by cytology with minimal treatment. 114 During a follow-up period of 78 months, they observed regression of the moderate dysplasia in 54%. Average follow-up time for patients with continuous normal cytology after disappearance of dysplasia was 53 months. During follow-up for 51 months, progression was observed in 30% and persistence in 16%. In 54% of patients, biopsies were performed. In patients without biopsies, regression was 50%, progression 35%, and persistence 15%, which implies a statistically significant difference between patients with and without biopsy. In patients over 51 years of age, fewer lesions progressed than in younger patients, and the progression time was also significantly longer.
In 3.8% of patients with persistent moderate dysplasia observed for an average of 50 months, cytology periodically gave no evidence of an abnormality. The risk of progression of moderate dysplasia was 5 to 9 cases per 100 women per year. When related to an incidence of carcinoma in situ of 4 in 100,000 women per year, the yearly progression risk of moderate dysplasia can be calculated to be 2000 times greater than for a woman without cervical dysplasia. Regression showed only slight variation between different age groups.
Tanaka and co-workers monitored 230 cases with an initial diagnosis of mild dysplasia cytologically and colposcopically for 2 to 10 years. 115 Regression of the epithelial abnormality was observed in 73.5% of cases, persistence in 20%, and progression in 6.5% of cases. Of 15 cases that showed progression, 10 cases of carcinoma in situ or microinvasive carcinoma developed. The average period after which progression was diagnosed was 54.8 months.
An important factor in explaining the wide variation between the results obtained in different follow-up series is the highly subjective character of the histologic diagnosis of intraepithelial neoplasia. Various pathologists studying the same lesion return a diagnosis varying from mild dysplasia to carcinoma in situ. 116
Clinical management of intraepithelial lesions of the cervix depends greatly on the cytologic and histologic definitions of dysplasia. The diagnosis of dysplasia is often considered inconsequential by uninformed physicians, who fail to take appropriate action. 117

Rate of Progression: Transition Time
On the basis of all available evidence, it may be concluded that the rate of progression of dysplastic lesions is strongly related to the severity of the lesion as evaluable by cytologic analysis.
Data from long-term population studies in a population that has been screened intensively allow true incidence figures to be measured and can provide evidence from the ratios between mild and severe intraepithelial lesions that the majority of mild and moderate intraepithelial changes eventually regress or persist for a prolonged period.
In the Nijmegen screening program, the ratio between mild and moderate dysplasias on the one hand and severe dysplasias and carcinomas in situ on the other hand was 4:1 at second screening. It is often not understood that the ratio of precancerous lesions to invasive cancer is probably on the order of 10:1, possibly even higher. 118
A suggestion of a fast transition time between normal epithelium and severe dysplasia should be made only after the more likely possibility of false-negative cytology because of poor sampling has been precluded. In cases of severe epithelial abnormalities, the number of unsatisfactory or less reliable smears is much increased, owing to admixture of inflammatory cells, cell debris, or blood.
The results of these studies strongly support a conservative approach in the clinical treatment of patients diagnosed with mild to moderately atypical epithelial abnormalities.

Rate of Progression: Review of the Literature
Richart and Barron monitored 557 patients with cervical dysplasia detected by cytologic and colposcopic examinations without interference of punch biopsies or other treatment. 89 They estimated the time spent in each stage of dysplasia before its progression to carcinoma in situ.
At the end of the follow-up period, transition probabilities were calculated and transition times of mild, moderate, and severe dysplasia to carcinoma in situ were computed. The transition times ranged from a median of 86 months for patients with very mild dysplasia, 58 months for mild dysplasia, and 38 months for moderate dysplasia to 12 months for severe dysplasia. The median transition time to carcinoma in situ for all dysplasias was 44 months. In a 10-year period, moderate dysplasia was calculated to progress to a more severe lesion in about 90% of the cases. They reported the progression rates to be relatively stable, and they concluded that almost all cases of dysplasia would in time develop into carcinoma in situ if left untreated. Although evolution from one smear class to another might be an age-related phenomenon, no obvious gradient in the ages was found. Similarly, there was no evidence that progression from one class to another was solely time dependent. During long-term follow-up, a substantial proportion of cases remained in the stage in which they were detected and did not progress to a higher stage disease during a definite period. These data cannot be considered as representative of true regression and persistence rates because patients entered in the study were highly selected: One of the criteria for acceptance was persistence of the lesion for at least a three-smear interval. This eliminated many lesions owing to repair and other processes that cytomorphologically cannot be discriminated from dysplasia and that had regressed during the interval.
Bamford and co-workers reviewed the smears preceding the histologic diagnosis in 100 cases of CIN grade 3 diagnosed in an intensively screened population. 119 These suggest that the transition time from normality to CIN grade 3 may be shorter than has generally been assumed. However, they provide no information about the quality and cellular composition of the smears that were reviewed.

Cytologic Follow-up
Given a correct cytologic diagnosis, we do not consider it necessary to perform immediate biopsy on lesions of minimal or moderate severity, because only a small proportion of these lesions will progress to a more marked abnormality, whereas the time required for a lesion to evolve provides ample time to detect a lesion at successive cytologic examinations. In view of the relatively large proportion of lesions that regress spontaneously and the median duration of progression time to a more severe lesion, patients with initial diagnoses of mild and moderate dysplasias should be monitored by regular cytologic smears for various periods of time before further therapy is recommended. Only after persistence of the lesion has been confirmed are further follow-up procedures, including colposcopy, warranted.
A cytologic diagnosis of severe dysplasia or carcinoma in situ is usually followed by colposcopy and biopsy, which at confirmation of the process are followed by deep excision, cryocautery, laser treatment, conization, or hysterectomy.

Follow-up Interval
There is no unanimity about when and to what extent follow-up examinations of cytologically diagnosed cervical abnormalities should be carried out. 21, 120 - 124 In cases of minimal epithelial abnormalities, a repeat smear should be advised after 12 months. In cases consistent with mild-to-moderate dysplasia, a repeat smear after 3 months should be recommended. When the abnormality is also diagnosed in the repeat smear, the patient should be referred for further evaluation including colposcopy. This should also be the procedure when a more severe lesion is diagnosed in the repeat smear. When the lesion is not confirmed in the repeat smear, a second repeat smear should be made again after 3 months. If the lesion remains absent or appears to be of less severity, the follow-up interval can be doubled to 6 months and later to 12 months.
It is well known that initial findings of mild-to-moderate atypia at follow-up sometimes prove to be of a more serious kind. 77, 125 - 128 With a follow-up procedure as described, it is therefore mandatory that the execution of repeat examinations be well supervised by the laboratory. Only when this condition is fulfilled is an initial cytologic follow-up of mild-to-moderate abnormalities warranted. 21
The procedure described has been adopted for the nationwide screening program in the Netherlands. 21, 121, 129 In cases of cytologic diagnoses consistent with mild dysplasia or a more severe lesion detected outside the screening program, an immediate referral for colposcopic evaluation is agreed on by all parties involved.

Colposcopy
The colposcope is well suited for follow-up studies of CIN because of its lack of influence on the natural history of the disease. 22 An optimal cancer detection system for preclinical asymptomatic cervical lesions should combine a cytologic examination with a colposcopic follow-up examination. The principal goal of cervical cytology is to detect precancerous lesions and asymptomatic, preclinical cancer. The purpose of cytologic screening is to signal these abnormalities and to induce further evaluation of lesions that have proved to be persistent or that may progress. Colposcopy could have an important place in the evaluation of these patients. Colposcopically guided biopsies may clarify inconclusive cytologic findings and give an assessment of the location, size, and extent of a lesion. There is a good correlation between the presence of abnormalities detected by colposcopy and histologically but little correlation between colposcopic categories and histologic grades of intraepithelial lesions. 69

Cellular Reactions Simulating Dysplasia
Certain stimuli may cause reactions in the cervical mucosa that may stimulate dysplasia. Chronic inflammatory processes and instrumental treatment for erosion or other surface abnormalities of the cervix (electrocautery, cryotherapy, laser treatment) may induce changes that may be confused with dysplastic changes when relevant clinical information is lacking. Certain medications, particularly those used in the systemic treatment of malignancies or in the course of immunosuppression in organ transplant recipients, may also induce cellular changes of a dysplastic nature. In rare instances, folic acid deficiency may cause an increase of both cytoplasm and nucleus. Even though extremely large nuclei may be found, the nucleocytoplasmic ratio remains within the limits of a minimal abnormality because of a correlated increase of the cytoplasmic body, often creating rather characteristic giant cells. In contrast to dysplastic changes, nuclei are hypochromatic. 130 Similar changes may be found as an effect of radiotherapy or treatment with alkylating drugs. 64

Prognosis of Dysplasia and DNA Cytophotometry
Light microscopically observed severely abnormal epithelial lesions cannot be subdivided into regressive, persistent, or progressive subtypes. An alternative to visual light microscopic evaluation of these lesions is the study of DNA ploidy. 131 In one study, the DNA content in dysplasia proved to be higher than normal, but the mode of the DNA index distribution fell entirely within the diploid range. Some cells had DNA values in the tetraploid and the hyperdiploid range, but a significant relationship between aberrant DNA values in dysplasia and DNA profiles of carcinoma in situ and invasive carcinoma could not be established. 132
The wide range of DNA values found in dysplasia and carcinoma in situ indicates that dysplasia and carcinoma in situ contain a highly variable population of cells whose range of DNA content is similar to that found in invasive cancer. A high proportion of cells in these lesions have abnormal chromosome numbers, but no chromosomal feature is distinctive of dysplasia, carcinoma in situ, or microcarcinoma, nor is any marker chromosome characteristic for early cervical neoplasia. 107 In some studies, DNA ploidy analyses indicated a high percentage of aneuploidy in CIN grade 3 lesions. On this basis, DNA aneuploidy was considered to be a marker for progression. 133, 134 In dysplasia and carcinoma in situ, clonal proliferation may already be occurring. However, there is also much evidence of the opposite kind: wide scatter of chromosome counts without a recognizable similarity from cell to cell. DNA estimations usually do not indicate a dominant aneuploid stem line. Evidence suggests that by the time microinvasion is identifiable, a new clone has overgrown the rest of the epithelium. Most cases have one distinct stem line, but some can have several. 135, 136
According to some investigators, quantitative DNA determinations in cytomorphologically equivalent dysplastic cervical cells do not offer additional means of predicting the outcome of the epithelial change. In patients with moderate cervical dysplasia, no significant differences between cell populations from moderate dysplasias that subsequently progressed to carcinoma in situ and those from lesions that regressed to normality were observed. The DNA distribution pattern of both groups was different from that of normal cells. 137
A striking association was found between DNA ploidy and age. In women who were younger than 35 years and who had CIN grade 3 lesions, aneuploidy was present in 27%, and the majority of the lesions showed a polyploid pattern. In women older than 50 years, aneuploidy was found in 88%. These findings suggest that processes finally progressing to invasive cancer may have different biologic characteristics in these two age groups. A diploid-like DNA pattern does not necessarily imply a regressive or persistent behavior of intraepithelial abnormalities. 135, 136, 138, 139

Postirradiation Dysplasia
After successful radiotherapy for a malignancy of the cervix, a small percentage of patients, following a latent period varying from 6 months to more than 20 years, develop an abnormality of the cervical or vaginal mucosa that has the characteristics of dysplasia. The characteristics of a postirradiation dysplasiaare essentially those of classic dysplasia. Accompanying the cellular evidence of dysplasia is the evidence of increased maturation of squamous cells, comparable to estrogen-induced maturation and the characteristics of irradiation changes such as multinucleation, dual staining reaction, and vacuolation of the cytoplasm. The appearance of superficial squamous cells may antedate the appearance of a cellular abnormality. 140 The biologic significance of a postirradiation dysplasia remains uncertain.

Accuracy
With the exception of invasion, the basic changes of importance in the recognition of primary cancer are apparent in both tissues and cells. 14 Because there is no single distinguishing feature that is in itself invariably pathognomonic of cancer in the cell or the epithelial lesion of origin, more than one fundamental change must be present to warrant an interpretation of cancer. Cell samples usually show dysplastic cells with a variable degree of atypia.
A prerequisite for the study of cellular pathology is an intimate knowledge of the component cells in the parent tissues in order to learn the origin of various cells identified in cellular preparations. Because the cellular sample represents a very comprehensive sample of the surface changes, in many instances the cellular evidence of a change is not apparent on so-called punch biopsy samples. For this reason, the presence of a lesion can be precluded only when the pathologic study is comprehensive. This is an important consideration when dealing with evidence gained by cell studies. 14
Clinicians increasingly show an inclination to reduce the amount of tissue removed for diagnostic or even therapeutic purposes. For diagnostic purposes, multiple biopsy specimens preferably taken under colposcopic control or a very shallow conus are most often used. Thus, a severe lesion of limited extent bordered by less severe changes may not be present in the histologic material. When comparing cytologic and histologic findings, this may lead to the conclusion that a severe lesion diagnosed cytologically apparently was overestimated: a false-positive diagnosis .
With the use of collection techniques that provide comprehensive samples in the uterine cervix, most reactions occurring in the cervix can be recognized on the basis of the cellular changes alone. 12 The number of cells diagnostic for a specific, benign proliferative reaction depends on the extent of the epithelial change and the method used for collecting the cells. In a well-sampled specimen, cells from these multiple atypical mucosal changes may be found next to one another. In a less adequate sample in which only part of the circumference of the ectocervix and the endocervical canal is represented by the cells in the sample, only part of the abnormal changes may be recognizable. This may lead to an underestimation of the severity of the lesion or, when the sample is highly inadequate, to a false-negative diagnosis .
Moreover, in cases in which sampling of the cervix has been done less expertly, the number of abnormal cells may be relatively low. In routine screening situations, this low number of cells may remain unobserved or, because of the paucity of abnormal cells, may erroneously lead to an underestimation of the severity of the lesion.
The initial cellular sample collected from the cervix usually contains the greatest number of abnormal cells. Subsequent samples may contain few abnormal cells even after an interval of 2 to 3 weeks. Because after only a short interval samples are taken from less-matured lesions, diagnoses on repeat studies made after too short an interval are likely to overestimate the severity of the lesion. Therefore, when follow-up of the lesion is done cytologically or when confirmation of the lesion is required before biopsy, the repeat study should not be performed within 4 weeks.
In all cases of discrepancies between cytologic and histologic diagnoses, both the cytologic and the histologic specimens must be reviewed. When the original cytologic diagnosis is confirmed at review, a repeat histologic examination should be requested.

Carcinoma In situ
Broders introduced the term carcinoma in situ (CIS) to describe epithelial lesions composed entirely of cells that have all the features of malignant cells but that do not exhibit invasive growth. 141 This term has become widely used since. Other terms that have been in use are incipient cancer, surface cancer, Bowen's disease of the cervix, intraepithelial cancer, carcinomatoid change, and preinvasive or noninvasive cancer. 12, 79, 142
Carcinoma in situ is now generally accepted as a precursor of invasive squamous cell carcinoma.

Origin, Localization, and Extent
The morphologic variations that can be observed when studying the spectrum of lesions directly associated with carcinoma in situ suggest a common pathway in the development of dysplasia and carcinoma in situ. The basic factor in the mechanism of the genesis of carcinoma in situ is reserve cell hyperplasia. The cells arising in reserve cell hyperplasia are noteworthy, because in some instances they are reminiscent of those seen in carcinoma in situ. A proliferation of the subcolumnar reserve cells may involve only one or two layers of cells beneath columnar epithelium or may attain considerable thickness. The latter not only simulates carcinoma in situ but actually may represent a developmental stage of this process. 14
The stimulus that has induced a proliferation of reserve cells as such also blocks the differentiation of these reserve cells into immature and then mature squamous metaplastic cells. Thus, by proliferation, primitive undifferentiated cells finally compose the entire lesion. Each pathway in the development of carcinoma in situ, localized in the endocervical canal, begins with or mimics reserve cell hyperplasia. If some differentiation occurs in the epithelial substrate before conversion to carcinoma in situ, such as occurs in squamous metaplasia or dysplasia of metaplastic type, the ensuing lesion is of large cell type. A lack of differentiation of the epithelial substrate would result in an in situ lesion composed entirely of small primitive cells. 12
The majority of cases of carcinoma in situ originate in the area of the transformation zone, bordering the original anatomic separation between stratified squamous epithelium and columnar epithelium. The farthest limit of CIN is determined by the farthest limit of reserve cell hyperplasia (squamous metaplasia, repair epithelium). 22 In the most distal part of this area, the stimulus for a metaplastic change of the columnar epithelium is apparently the strongest. This may also be the reason why large-cell carcinomas in situ are much more frequent distally in the canal than proximally, where small-cell variants are more often seen. It may well be that the potential for metaplastic change in the more proximal parts of the canal is lower, resulting in a lower frequency of mature squamous metaplastic changes and a higher frequency of poorly differentiated epithelial abnormalities.
Epithelial changes preceding a small-cell carcinoma in situ likely do not develop through precursor lesions showing some squamous differentiation. It is probably very difficult to differentiate between early small-cell intraepithelial changes and the finally resulting small-cell carcinoma in situ. In a small number of women participating in the Nijmegen population-screening program, a small-cell lesion was identified as composed of rather immature small cells, which histologically showed a pseudo-stratified, somewhat columnar arrangement. In a few women who had no intercurrent smears or biopsy specimens taken for various reasons, classic small-cell carcinomas in situ were found at repeat examination 3 to 4 years later.
The biologic significance of these pseudostratified small-cell precursor lesions, which biologically may be comparable to the more distally located severe dysplasias, needs further clarification. On a cytomorphologic as well as on a histologic basis, it is possible to subclassify abnormal intraepithelial lesions specifically into categories that can be arbitrarily labeled dysplasia and carcinoma in situ.

Age at Detection
Most researchers report an average age at detection of about 40 years (41.6 years 12, 79, 142 - 144 and 42.5 years, with a range of 22 to 91 years). 12 In our own material, average age at first histologic diagnosis of carcinoma in situ was 39.4 years, with a range of 17 to 82 years ( Fig. 8.80 ). More recently the mean age at detection has been reported to be earlier. This reduction in average age has been related to an earlier onset of sexual activity but may at least in part be attributed to more intense cytologic screening.

Fig. 8.80 Carcinoma in situ (CIS) and invasive cancer (INV CA). Histologic diagnoses. Age distribution–average age of patients with carcinoma in situ was 39.2 years (SD 9.8) and n = 879. Average age of patients with invasive cancer was 53.0 years (SD 13.9) and n = 435.
(Data from the Nijmegen Registry of Cervical Cytology, 1970–1987.)
In our own registry the age-distribution curve showed a bimodal pattern, with peak incidences at 35 and 50 years. 76, 139, 145 These peaks were correlated with peak incidences of invasive squamous cancer at about 48 years and after 60 years of age.

Histology
The subjective nature of morphologic classification criteria initially caused a wide variation in the spectrum of histologic substrates diagnosed as carcinoma in situ. However, experience in the classification of carcinoma in situ combined with clinically correlated studies has increasingly narrowed the histologic spectrum to a point of relatively uniform agreement on certain morphologic patterns. 12
The best description of the characteristic cytomorphologic and architectural changes of carcinoma in situ is given in the definition adopted by the International Committee on Histological Terminology for Lesions of the Uterine Cervix in Vienna in 1961 80 : “Only those cases should be classified as carcinoma in situ ” that, in the absence of invasion, show a surface epithelium in which, throughout its whole thickness, no differentiation takes place ( Fig. 8.81 ). The process may involve the cervical glands without creating a new group.
It is recognized that the cells of the uppermost layers may show some flattening [ Fig. 8.82 ]. The very rare case of an otherwise characteristic carcinoma in situ that shows a greater degree of differentiation belongs to the exception for which no classification can provide. 80

Fig. 8.81 Carcinoma in situ. Almost complete lack of maturation and parallel arrangement of cells throughout the entire epithelium. Disturbed palisade arrangement of cells in the most basal layers (H&E × MP).
The WHO definition of a carcinoma in situ is “a lesion in which all or most of the epithelium shows the cellular features of carcinoma.” 81 This definition also includes an epithelial abnormality with some evidence of maturation in the most superficial layers and thus also encompasses a lesion that many pathologists would regard as severe dysplasia.
In general, the term carcinoma in situ is used to describe a reaction replacing the normal surface epithelium or the epithelium of the invaginations of the surface epithelium or both, in which all the layers of the epithelium are composed of abnormal poorly differentiated or largely undifferentiated cells. 12
The lesion may be composed of large or small cells, but essential in the classification of a lesion as carcinoma in situ is that the entire thickness of the epithelium is composed of poorly differentiated cells virtually without signs of maturation (differentiation) toward the surface (see Fig. 8.82 ). Continuous with the surface lesion, it is not unusual to find in the endocervical invaginations an epithelial change that shows better maturation, with the more superficial layers composed of cells recognizable as squamous (metaplastic).

Fig. 8.82 Carcinoma in situ. Irregular arrangement of cells and almost complete lack of maturation. Most superficial layers show parakeratosis. Mitoses can be recognized at all levels (H&E × HP).
Thus, together with a characteristic carcinoma in situ on the surface of the endocervical canal, it is not unusual to find moderate-to-severe dysplasia (of squamous metaplastic type) in the invaginations of the endocervical canal. 12 Both normal and atypical mitoses are present at all levels of the epithelial reaction (see Fig. 8.82 ). This contrasts with moderate and severe dysplastic lesions, in which mitoses usually are absent in the most superficial layers. On a histologic and cytologic basis, Reagan and Hamonic 14 proposed a subclassification of carcinoma in situ into small- and large-cell carcinoma in situ . Although useful in recognizing morphologic variations and studying the morphogenesis of these lesions, this classification has not provided significant information on the biologic potential of these lesions. 12 Patten subdivides carcinomas in situ into large-cell type, intermediate type, and small-cell type. 73 The intermediate type seemed to be the dominant morphologic variant in his material. The large-cell type, although relatively rare, appeared to be increasing in frequency. The once-dominant cell type of small-cell carcinoma in situ was observed with decreasing frequency.
At the time of his reporting, it was not yet evident whether this change in the distribution of morphologic variants of carcinoma in situ was paralleled by a change in the distribution of morphologic variants of invasive cancer. The lesion is characteristically located in the area of the transformation zone. The overall extent and distribution of carcinoma in situ are comparable to that observed for reserve cell hyperplasia and immature squamous metaplasia. Extension of the surface change into the invaginations occurs in more than 90% of cases. 146 Richart observed that extension into the portio vaginalis occurred in about 55% of cases and that 3% of cancers in situ had extension into the vaginal fornix. 22

Cytology
The number of abnormal cells in cellular samples is usually greater in cases of carcinoma in situ than in dysplasias. The cellular changes of importance in the recognition of cancer in situ are those exhibited by cells desquamated or forcibly scraped from the epithelial surface.

Arrangement of Cells
Carcinoma in situ, because of the undifferentiated nature of the component cells, lacks the characteristic cytoplasmic changes of differentiation (maturation). Either the cells are isolated or, because of a disturbance in cytoplasmic division during cell division, are adherent together and arranged as syncytial aggregates ( Figs. 8.83 and 8.84 ). Most samples do contain isolated cells, but aggregates of abnormal cells predominate. A scraping of the cervix forcibly removes cells from the surface. In view of the high number of cells arranged in syncytial masses, in these situations, taking a smear is essentially performing multiple microbiopsies. In a syncytial group, cells are arranged irregularly and have indistinct cell borders and overlapping nuclei ( Fig. 8.85 ). The latter two features differentiate these syncytial aggregates from sheets found in the presence of dysplasias and in which the more distinct cell borders and the more regular arrangement of the cells are a reflection of the relatively higher differentiation (maturation) of the dysplastic epithelium, in comparison with the epithelium composing the carcinoma in situ.

Fig. 8.83 Carcinoma in situ. Cells occur singly but predominantly in syncytial aggregates with indistinct cell borders. Cells have only a minimal amount of cytoplasm. Nuclei, although relatively small, vary greatly in size (Papanicolaou × MP).

Fig. 8.84 Carcinoma in situ. Syncytial aggregate of cells with indistinct cytoplasmic borders. Nuclei vary in size and shape and are frequently overlapping. Nuclear chromatin is hyperchromatid, irregularly distributed, and coarsely granular (Papanicolaou × OI).

Fig. 8.85 Carcinoma in situ. Syncytial aggregate of cells. Nuclei vary in size and shape and frequently overlap. Nuclear chromatin is irregularly distributed and coarsely granular (Papanicolaou × OI).

Cell Size and Shape
Cells from a carcinoma in situ lesion are relatively small compared with cells from normal stratified squamous epithelium or dysplastic cells. Cells from the histologically large-cell variant of carcinoma in situ are predominantly in the range of small immature squamous metaplastic cells. Similarly, cells from a histologically small-type carcinoma in situ are in the range of reserve cells. From a series of cytologic specimens in cases of histologically proven carcinomas in situ, Patten computed a mean cell area of 238μm 2 . 12 The cells are predominantly round to oval, reflecting their immature character. Irregular or elongated cell forms are related to a specific superficial change of the lesion. The sparse cytoplasm stains predominantly basophilic, a cytoplasmic staining reaction that reflects the lack of keratinization in these undifferentiated cells. The few cells showing an eosinophilic staining reaction are most likely derived from a coexisting dysplasia. Owing to the highly vulnerable cytoplasm in cases of carcinoma in situ, the finding of cells with damaged cytoplasm or with bare nuclei is frequent ( Fig. 8.86 ).

Fig. 8.86 Carcinoma in situ. Syncytial aggregate with indistinct cell borders and multiple isolated bare nuclei. Compare the size with the intermediate squamous cells and columnar cells in the upper half of this field. Nuclei have sharply outlined membranes, an irregular distribution of finely granular chromatin, and nuclear grooves (Papanicolaou × HP).

Nuclear Morphology
The lack of a differentiation stimulus, usually associated with relatively rapid cell growth, becomes apparent from the round-to-oval shape of the nuclei. On the average, nuclei of cells derived from carcinoma in situ are usually somewhat smaller than nuclei derived from dysplastic cells. Patten computed a mean nuclear area of 125μm 2 . 12 The nuclear membrane in isolated cells is usually well defined; at higher magnification, nuclear grooves may be seen and are rather characteristic of these undifferentiated cells, even though they are most likely artifacts due to fixation ( Fig. 8.87 ; see also Fig. 8.86 ). The chromatin varies from finely granular and unevenly distributed to coarsely granular and hyperchromatic (see Fig. 8.87 ). The number of small nuclei with coarse hyperchromatic chromatin is correlated with the degree of dedifferentiation of the cells from the carcinoma in situ. Another characteristic of carcinoma in situ cells, recognizable at high magnification, is the interrupted nuclear membrane due to irregular sedimentation on the membrane of the coarsely granular chromatin. It is rare to find eosinophilic-staining nucleoli, but chromocenters or “false” nucleoli are rather common (see Fig. 8.87 ). “True” nucleoli may be obscured by the coarse hyperchromatic chromatin but are definitely present, because they are related to the high proliferation rate of the lesion. The absence of nucleoli would be in contradiction to the relatively high proliferative activity, which is evident from the presence of mitoses even in the uppermost layers of the epithelium. Macronucleoli, which are usually present in nuclei from invasive cancer cells, are only rarely seen in carcinomas in situ.

Fig. 8.87 Carcinoma in situ. Coarsely granular irregularly distributed nuclear chromatin. Large irregular condensations of chromatin—chromocenters—should be differentiated from nucleoli. (Papanicolaou × OI).

Relative Nuclear Area
The relative nuclear area (nuclear area in relation to cytoplasmic area) with an average value of 50% better reflects the primitive character of these cells from carcinomas in situ than just the size of the cell or the nucleus.
Key features of carcinoma in situ
• Larger number of abnormal cells than other grades of dysplasia;
• Cells arranged predominantly in syncytial aggregates with indistinct cell borders;
• Some isolated cells;
• Small cells with little cyanophilic cytoplasm and without signs of maturation;
• Round-to-oval nuclei occupy half of cellular area;
• Bare nuclei are commonly noted;
• Even to irregular chromatin distribution with fine to coarsely granular texture;
• Nuclear membrane is commonly disrupted and irregular (cerebriform); and
• Micronucleoli rarely noted—obscured by dense chromatin.
Reagan and Hamonic 14 reported the presence of inflammatory exudate in specimens from cancers in situ in 68% of cases, compared with 32% in specimens from dysplasias. In our own material from a population-screening program, the percentage of smears showing an inflammatory exudate in cases of carcinoma in situ and severe dysplasia (85.1%) was not significantly different from the percentages of smears found with less severe changes or without signs of epithelial abnormalities. 72

Biologic Significance
It is generally agreed that the majority of lesions classified as carcinoma in situ are actively proliferating lesions that, if left alone, will finally evolve into invasive cancer. However, with currently available knowledge and techniques, it is virtually impossible from cytologic, histologic, or clinical data to predict which reaction will regress and which will progress to a more severe lesion.
It is generally accepted that all invasive cancers of the uterine cervix develop from a carcinoma in situ or from severe dysplasia. However, massive evidence shows that not all carcinomas in situ or severe dysplasias progress into an invasive process; many regress into a lesion of less severity or disappear completely. This means that these lesions diagnosed as intraepithelial cancer are essentially not malignant in nature.
Many studies have tried to uncover the characteristics of those carcinomas in situ that progress to an invasive process, to discriminate these from the lesions that show regression but that appear identical on light microscopy. Data on the follow-up of severe dysplasias and carcinomas in situ are not abundant, because most lesions are immediately treated when diagnosed. Available data indicate that only a relatively small number of lesions eventually evolve to an invasive squamous cell carcinoma. 110, 116, 147 - 149
Prospective studies designed to monitor the course of carcinoma in situ suggest a relatively slow evolution to invasive carcinoma. Analyses of incidence rates and mortality indicate that the length of the preinvasive stage is on the order of 12 to 15 years. Reports of so-called fast-growing cancers, which are claimed to have passed very quickly through a premalignant preinvasive phase, as should be evidenced by a nondiagnostic cytologic examination shortly before the diagnosis of the invasive process, appear with some periodicity in the literature. However, on careful analysis of available evidence, the accuracy of the preceding cytologic diagnosis usually does not stand, because false-negative diagnoses due to sampling errors cannot be precluded. In our own registry of premalignant and malignant lesions of the uterine cervix, encompassing 17 years before and during a population-screening program, difference in mean age between cancer in situ and invasive cancer was 13.7 years (carcinoma in situ: mean age 39.2 years; invasive cancer: mean age 52.9 years).

Developmental Carcinoma in Situ
Reagan and colleagues 79 described the morphologic characteristics of a group of epithelial lesions that they considered to be developmental carcinoma in situ. Comparable cytologic features in women who ultimately were proved to have carcinoma in situ were described by Koss and Durfee. 44 Compared with classic carcinoma in situ, cells more frequently lie singly than in arranged syncytial masses. When present in aggregates, cells more often had a sheet-like rather than a syncytial arrangement. In almost all instances, a significant admixture of dysplastic cells was found in the smears. Average cell size was between the mean sizes for dysplastic cells and carcinoma in situ cells. Also, for the distribution of cell shapes, values proved to be between the values found for dysplasias and those for carcinomas in situ. Nuclear chromatin was predominantly finely granular and unevenly distributed. The lesions usually showed a more orderly growth pattern and a relatively low mitotic activity.

Cytology
The cellular characteristics of this so-called developmental carcinoma in situ were reminiscent of cells found in proliferating reserve cells. The mean age at which this developmental cancer in situ was found antedated the mean age of classic carcinoma in situ by 5 years.
Fawdry found “early recurrences” of carcinoma in situ within a year after hysterectomy in 0.9% of women who had a hysterectomy after a diagnosis of carcinoma in situ. 150 During subsequent long-term follow-up, invasive squamous cell carcinomas of the vaginal vault were found in 0.3% of patients. These figures emphasize the importance of cytologic follow-up by vaginal vault smears in women who have had a hysterectomy for a severe epithelial abnormality of the cervix.

Co-occurrence of Dysplasia and Carcinoma in Situ
An area in which discrepancies in the grading of lesions still occurs is the discrimination between severe dysplasia and carcinoma in situ. Although for the sake of a better understanding of the biologic behavior of intraepithelial lesions it is still advocated to differentiate as much as possible between severe dysplasia and carcinoma in situ, in routine diagnosis such a strict separation is not very relevant. Both changes are considered to be of a high potential for progression to an invasive lesion, and clinical management is usually identical. In view of the influence of local environmental factors and because of the origin from different stem cells, a markedly dysplastic change, which usually still shows some maturation of the most superficial layers into squamous cells, originating in the original ectocervical squamous epithelium, is likely to be fully comparable to the characteristic carcinoma in situ, originating in the endocervical canal through a squamous metaplastic interphase and characteristically showing no sign of maturation, even in the most superficial layers.
Approximately 80% of carcinomas in situ coexist with dysplasia. Graphic distribution of these lesions suggests that when coexistence occurs, dysplasia lies distal from carcinoma in situ. 12 Both dysplasia and carcinoma in situ occur more frequently on the anterior lip of the cervix.
A possible explanation is that the anterior lip is more frequently traumatized. The radial extent of CIN on the portio epithelium is greatest in the population with carcinoma in situ. Also, the size and distribution of carcinoma in situ on the exposed portion of the cervix are more constant than in dysplasia.
Of patients with carcinoma in situ, 45% had only dysplasia on the exposed portion of the cervix. The difference in the distribution of carcinoma in situ and dysplasia may be accounted for by the increased size of the lesion in patients with carcinoma in situ 22 and by the age-related size differences between dysplasia and carcinoma in situ. 95
The hypothesis that dysplasia represents a younger (smaller and less extensive) precursor of carcinoma in situ is supported by the colposcopic and cytologic observation in the individual case. The lesions are not biologically different, but a patient with a carcinoma in situ has a lesion that has been present for a longer time and involves a larger area of the cervix. Our knowledge of the biologic behavior of carcinoma in situ is still incomplete; an untreated lesion may develop into an invasive cancer, regress to a less severe lesion, completely disappear, or persist for an indeterminate time. 79 Both dysplasia and carcinoma in situ occur in the area of the transformation zone.

Co-occurrence of Squamous and Columnar Intraepithelial Abnormalities
Carcinogenic changes in the uterine cervix can be considered to be part of a field carcinogenic process. This means that the carcinogenic stimulus does not exert its action on an isolated cell that becomes the stem cell of a malignant proliferation but on a larger area of the epithelium. This may be the explanation for the common observation that next to a severe lesion, changes of different, often lesser severity are found. It also explains the common multifocal occurrence of epithelial abnormalities.
This field stimulus apparently influences not only squamous and squamous metaplastic epithelium but also the columnar epithelium of the endocervical canal. The co-occurrence of abnormal columnar cell changes together with abnormalities of the squamous and squamous metaplastic epithelium is becoming increasingly frequent.
The reduction in the incidence of squamous cell carcinoma of the uterine cervix and mortality from cervical cancer is a result of cytological screening and the subsequent detection and removal of precursor lesions. The diagnosis of cervical glandular cell lesions or combined squamoglandular cell lesions is based on the same cytological principles.
In a series of 42,863 first cervical smears, minimally to severely atypical columnar cell changes were observed in 69% of cases of severe dysplasia, carcinoma in situ, and microinvasive and invasive cancer ( Figs 8.88 and 8.89 ). In 13% of cases, this atypia was diagnosed as an adenocarcinoma in situ ( Figs 8.90 and 8.91 ).

Fig. 8.88 Mild-to-moderate atypia of endocervical columnar epithelium. Pseudostratified columnar epithelium with irregular arrangement of nuclei, variation in nuclear size, and scattered nuclear hyperchromasia (H&E × HP).

Fig. 8.89 Mildly atypical columnar cells . Irregular arrangement of nuclei and variation in nuclear size and shape. Slight hyperchromasia of regularly distributed nuclear chromatin. One to two nucleoli are usually present (Papanicolaou × OI).

Fig. 8.90 Co-occurrence of squamous cell carcinoma in situ and adenocarcinoma in situ. (Papanicolaou × HP).

Fig. 8.91 Adenocarcinoma in situ. Aggregate of cells with a gland-like structure with partly overlapping nuclei. Variation in size of predominantly round-to-oval nuclei. Moderate hyperchromasia of predominantly finely granular and regularly distributed nuclear chromatin. One to multiple nucleoli are present (Papanicolaou × OI).
The increased prevalence of these atypical columnar cell changes may be due to a new factor that also causes columnar cell abnormalities but is more likely caused by screeners' increased sensitivity to columnar cell changes, correlated with a growing awareness of the significance of endocervical columnar cells as a quality parameter for cervical smears. 72, 151
Nevertheless published results indicate that glandular cell abnormalities were frequently overlooked or underestimated in the cytological specimen. 152 - 161 This is at least in part due to glandular abnormalities being less common than cervical squamous cell lesions. Furthermore many cytotechnologists and cytopathologists are focused on the cytological characteristics of squamous cell lesions, whereas characteristics of glandular cell lesions (GCLs) are less familiar and therefore are likely to be overlooked in a relatively high number of cases.
Data from the literature indicate that in approximately 50% of severe GCLs (range, 25–75%), a coexisting squamous cell lesion was present. 152, 153, 162 - 170 Our most recent study 170 addressed the predictive value of the PAP smear. A data search of the Dutch National Pathology Archive (PALGA) revealed that 721 of 1141 registered histological cervical glandular cell abnormalities (63.2%) were in fact cases of a severe combined squamoglandular cell lesion.
In addition the data search of the PALGA registry showed that in 547 cases (51.9%) a cytological diagnosis of a severe glandular cell lesion (with or without a squamous cell component) was made. Prediction of a severe glandular cell lesion in the PAP smear was found to be more accurate in cases of histologically confirmed pure glandular cell abnormalities than in cases with a histological diagnosis of a combined lesion. The cytological prediction was found to be correct in 75.2% of cases of pure adenocarcinoma in situ and in 47.3% of cases of adenocarcinoma in situ with coexistent high-grade squamous intraepithelial lesion.
These figures illustrate a relatively poor “performance” in the diagnosis of GCLs, suggesting that more attention must be given to specific morphologic characteristics of minor and severely atypical cervical GCLs.
Studies addressing the prediction of severe cervical GCLs on cervical Pap smears indicated great differences with regard to predictive accuracy. 152 - 161 A prerequisite for the recognition of GCLs or combined lesions is knowledge of and agreement on the significance of all cytomorphologic features characteristic of the different grades and types of cervical glandular cell neoplasia. Furthermore cytotechnologists and cytopathologists must always consider the possibility that two different cervical neoplasms are present in a given patient.
In our previous work on endocervical columnar cell intraepithelial neoplasia, various cytologic, architectural, cellular, and nuclear characteristics proved useful in diagnosing intraepithelial cervical GCLs and adenocarcinoma and in assessing the severity of these lesions 171 - 175 ( Fig. 8.92 ).

Fig. 8.92 Carcinoma in situ (HGSIL) and adenocarcinoma in situ (AIS) in histology and cytology. (A) Left: portion of the adenocarcinoma in situ lesion: Severely atypical glandular epithelium, exhibiting pseudostratification. Right: HGSIL (H&E × HP). (B) Left: acinic structure composed of severely atypical cervical glandular cells derived from adenocarcinoma in situ. Right: highly abnormal squamous cells that contain polymorphic nuclei consistent with carcinoma in situ (Papanicolaou × HP). (C) Left: polypoid structure composed of severely atypical cervical glandular cells with elongated nuclei, derived from adenocarcinoma in situ. Right: highly abnormal squamous cells with polymorphic nuclei consistent with carcinoma in situ (Papanicolaou × HP)
The findings of one of our studies showed that application of cytomorphologic features directed to cervical glandular cell lesions such as pseudostratification, cellular crowding, glandlike structures, feathering, vacuolated cytoplasm, and nuclear and chromatin features increased the accuracy of diagnosing combined severe lesions of the cervix. 176 Most lesions as well as different subtypes from glandular lesions were recognized in the cytological specimens by means of special architectural and cellular features.
Reevaluation of the cytological specimens revealed that in some cases, failure to detect the GCL was attributable to the abundance of abnormal squamous cells in the smear sample, whereas in other cases, abnormal glandular cells were overlooked or underestimated in severity at initial diagnosis. It also is possible for an abundance of abnormal glandular cells to mask the presence of a squamous cell abnormality.
Cytological interpretation of GCLs as with squamous lesions is subjective in nature and thus demonstrates considerable interobserver variability.
We therefore advocate a thorough evaluation of smears having been reported with glandular cell abnormalities, by correlating cytological and histological findings and intra- and interlaboratory exchange of cases with glandular abnormalities in different grades of severity in order to achieve consensus upon the interpretation of cytomorphological characteristics. This will improve diagnostic accuracy.
Cytological identification of cervical glandular cell abnormalities is important for patients with coexisting squamous cell lesions, because the treatment of these combined lesions differs from the treatment of pure squamous cell lesions ( Figs 8.93 to 8.97 ). Diagnosis of only a squamous cell abnormality may result in insufficient surgical treatment. Even circular biopsy (shallow conization) may not reveal glandular abnormalities, which typically are located more proximally in the endocervical canal.

Fig. 8.93 Carcinoma in situ (HGSIL) and adenocarcinoma in situ/ adenocarcinoma in histology and cytology. (A) Carcinoma in situ in collision with adenocarcinoma in situ/adenocarcinoma (H&E × MP). (B) Increased magnification of the boxed area in (A). Left: endocervical lining with severely atypical glandular cells from adenocarcinoma in situ (left) in collision with severely atypical squamous cells from carcinoma in situ (top) (H&E × HP). (C) Cells derived from carcinoma in situ (top left) and small strips composed of pseudostratified atypical glandular cells derived from adenocarcinoma in situ (top right and bottom middle) (Papanicolaou × MP). (D) Severely atypical squamous cells exhibiting nuclear overlap, polymorphism, hyperchromasia, and a coarse granular chromatin pattern (Papanicolaou × HP). (E) Strip-like arrangement composed of pseudostratified glandular cells that contain elongated nuclei and vary in size (Papanicolaou × MP). (F) Acinic structure derived from adenocarcinoma in situ. The cytoplasm is directed toward the center of the structure: the elongated peripherally located nuclei vary in size and shape. The chromatin is finely granular (Papanicolaou × HP).

Fig. 8.94 Intestinal adenocarcinoma in situ. (A) “Back-to-back” cellular arrangement and prominent cytoplasmic vacuolation are evident (H&E × HP). (B) Cervical glandular cells with abundant cytoplasmic vacuolation can be seen (Papanicolaou × HP).

Fig. 8.95 Endocervical/endometrioid adenocarcinoma in situ. (A) Endometrioid type (right) and endocervical components can be seen (left) (H&E × MP). (B) On the right-hand side, relatively small cells corresponding to the endometrioid component can be seen; on the left-hand side, a rosette-like structure derived from the endocervical component of adenocarcinoma in situ can be seen (Papanicolaou × HP).

Fig. 8.96 Villoglandular (papillary) adenocarcinoma in situ/adenocarcinoma . (A) Papillary growth pattern, with severe pseudostratification visible in the upper right-hand corner (H&E × MP). (B) Papillary structure composed of severely atypical glandular cells (Papanicolaou × MP).

Fig. 8.97 Carcinoma in situ (HGSIL) and adenocarcinoma in situ/adenocarcinoma in liquid-based cytology specimen (LBCS). (A) Red arrow: highly abnormal squamous cells with scanty cytoplasm and polymorphic nuclei corresponding with carcinoma in situ. Blue arrow: strip-like arrangement composed of glandular cells exhibiting elongated nuclei, varying in size (LBCS, Pap × LP). (B) Glandular cells showing elongated nuclei, varying in size. The chromatin is coarsely granular and irregularly distributed. Feathering is demonstrated (LBCS, Papanicolaou × HP). (C) Highly atypical squamous cells exhibiting polymorphism, hyperchromasia, and a coarse chromatin pattern (LBCS, Papanicolaou × HP).
Because the prognosis for patients with adenosquamous carcinoma or carcinoma with glandular differentiation is said to be poorer than the prognosis for patients with squamous cell carcinoma, 177 - 179 every effort must be made to increase the reliability of cytological diagnosis and the quality of cervical specimens, as cytological screening is likely to remain for the foreseeable future the most widely used method for detecting cervical abnormalities (both squamous and glandular).
More accurate identification of intraepithelial GCLs or combined squamoglandular lesions of the cervix may eventually lead to a decrease in the incidence of cervical adenocarcinoma, just as increases in diagnostic accuracy have led to decreases in the incidence of squamous intraepithelial lesions and invasive squamous carcinoma of the cervix.
For a detailed description of the cytomorphologic characteristics of adenocarcinoma in situ, see our studies on columnar cell abnormalities 171 - 174 180 and Chapter 9 , Glandular Neoplasms of the Uterine Cervix.

Dysplasia and Carcinoma in Situ during Pregnancy
In general, the cellular features of dysplastic changes in pregnant women are identical to the changes observed in nongravid women. Dysplastic changes often remain undetected in view of the significantly larger number of smears of unsatisfactory quality. Particularly during the first trimester of pregnancy, smears may contain a large number of relatively small, immature, atypical cells. In general, dysplastic lesions during the first months of pregnancy tend to be composed of relatively small cells, thus suggesting a severe abnormality. These relatively immature dysplastic changes may lead to a dilemma: whether the patient should undergo histologic evaluation. In most instances, with progression of pregnancy, a reduction of the severity of a dysplastic lesion occurs. In a study of 87 women with dysplasia during pregnancy, Slate and Merritt found 45% of lesions to regress, 29% to persist, and 25% to progress to a more severe abnormality after a variable period of time. 109
As a general rule, unless clinical symptoms suggest a severe lesion, routine cytologic diagnosis should be avoided during the first trimester. Contrary to widespread belief, there is no evidence that dysplastic lesions and carcinoma in situ behave differently during pregnancy. The morphologic features of carcinoma in situ do not differ from those presented for nongravid patients. The growth rate of preinvasive epithelial changes is not greater during pregnancy.
Surgical procedures may induce overstimulation or may damage the cervix and thus may interfere with the pregnancy. In general, it is fully warranted to monitor the abnormality carefully during pregnancy, even in cases of carcinoma in situ, and to postpone treatment of the epithelial change until after childbirth. It is advocated that surgical treatment be instituted only after cytologic reconfirmation of the lesion, when normal menstrual cycling has resumed and epithelial atrophy has been reversed. In some of the cases, the epithelial abnormality disappears after childbirth, possibly owing to the passage of the newborn through the cervical canal or because of a change in hormonal influence on the cervical epithelium, potentially altering the balance (or imbalance) between the host tissue and the abnormal epithelium.

Dysplasia and Carcinoma in Situ in Postmenopausal Patients
Dysplasia and carcinoma in situ occurring during postmenopause are often difficult to diagnose because of the lack of mature squamous and columnar cells in the smears. Owing to the lack of estrogenic stimulation, cells remain rather small, often the size of parabasal cells or relatively immature squamous metaplastic cells. Aggregates of small cells with densely staining nuclei are a common finding in smears from women in postmenopause. Relative nuclear enlargement due to a reduced development of the cytoplasm, which is in turn caused by a low estrogenic stimulus, may mimic the nucleocytoplasmic ratio found in epithelial lesions ( Figs 8.98 and 8.99 ). These cells are often erroneously diagnosed as dysplastic cells or cells consistent with carcinoma in situ.

Fig. 8.98