Practical Surgical Neuropathology: A Diagnostic Approach E-Book
987 pages
English

Vous pourrez modifier la taille du texte de cet ouvrage

Practical Surgical Neuropathology: A Diagnostic Approach E-Book

-

Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus
987 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

Description

Practical Surgical Neuropathy—a volume in the new Pattern Recognition series— offers you a practical guide to solving the problems you encounter in the surgical reporting room. Drs. Arie Perry and Daniel J. Brat present diagnoses according to a pattern-based organization that guides you from a histological pattern, through the appropriate work-up, around the pitfalls, and to the best diagnosis. Lavish illustrations capture key neuropathological patterns for a full range of common and rare conditions, and a "visual index" at the beginning of the book directs you to the exact location of in-depth diagnostic guidance. No other single source delivers the practical, hands-on information you need to solve even the toughest diagnostic challenges in neuropathology.

  • Provides all the information essential for completing a sign-out report: clinical findings, pathologic findings, diagnosis, treatment, and prognosis.
  • Illustrates key pathologic and clinical aspects of disease entities through over 1430 superb, high-quality full-color images that help you evaluate and interpret biopsy samples.
  • Presents a team of internationally recognized experts for authoritative and up-to-date information from leading diagnosticians in neuropathology.
  • Features a user-friendly design with patterns color-coded to specific entities in the table of context and text and key points summarized in tables, charts, and graphs so you can quickly and easily find what you are looking for.
  • Directs you to the chapter and specific page number of the in-depth diagnostic guidance you need through a unique, pattern-based visual index at the beginning of the book.
  • Details key diagnostic features associated with rare and esoteric conditions in a visual encyclopedia with distinctive findings and artifacts for unusual patterns at the end of the book.

Sujets

Ebooks
Savoirs
Medecine
Derecho de autor
Realimentación
Lesión
Organización Mundial de la Salud
Hodgkin's lymphoma
Parkinson's disease
Neurocytoma
Oncology
Electroencephalography
Amnesia
Meningitis
Pineocytoma
Photocopier
Alzheimer's disease
Pulmonary pathology
Germinoma
Schwannoma
Optic nerve glioma
Leukoencephalopathy
Primary central nervous system lymphoma
AIDS
Neurodegeneration
Ganglioglioma
Ganglioneuroma
Medulloblastoma
Neuroimaging
Temporal lobe epilepsy
Embryonal carcinoma
Neurofibroma
Cerebral hemorrhage
Neuroma
Neurofibromatosis type II
In Debt
Malignant peripheral nerve sheath tumor
Antihemorrhagic
Neuropathology
Neuroradiology
Krabbe disease
Dermatopathology
Progressive supranuclear palsy
Neurofilament
Neoplasm
Chapter (books)
Traumatic brain injury
Astrocytoma
Glioblastoma multiforme
Meningioma
Ependymoma
Oligodendroglioma
Prolactinoma
Choriocarcinoma
Vestibular schwannoma
Pituitary adenoma
Differential diagnosis
Acute lymphoblastic leukemia
Enthusiasm
Demyelinating disease
Hypopituitarism
Subarachnoid hemorrhage
Melanoma
Histopathology
Vasculitis
Glioma
Stroke
Infarction
Tuberous sclerosis
Pulmonology
Renal cell carcinoma
Biopsy
Lesion
Aneurysm
Archive
Cauterization
Osteosarcoma
Teratoma
Medical imaging
Lissencephaly
Mentorship
Prolactin
Hydrocephalus
Cytopathology
Artifact
Spasticity
Atherosclerosis
Atkins diet
Hypertension
Non-Hodgkin lymphoma
Pituitary gland
Gynaecology
Adrenoleukodystrophy
X-ray computed tomography
Multiple sclerosis
Dementia
Encephalitis
Brain tumor
Infection
World Health Organization
White matter
Unconsciousness
Transient ischemic attack
Tuberculosis
Data storage device
Epileptic seizure
Radiation therapy
Peripheral nervous system
Optic neuritis
Neurologist
Neurology
Mechanics
Magnetic resonance imaging
Myelin
Histology
General surgery
Feedback
Epilepsy
Chemotherapy
Chemical element
Brain abscess
Anxiety
Algorithm
Hypertension artérielle
Wick
Business
Necrosis
Pathology
Brain
Gene
Lésion
Archive (homonymie)
Consultant
Talent
Mentor
Electronic
Rétroaction
Maladie infectieuse
Organisation mondiale de la santé
Copyright
Hormone

Informations

Publié par
Date de parution 16 avril 2010
Nombre de lectures 0
EAN13 9781455706006
Langue English
Poids de l'ouvrage 65 Mo

Informations légales : prix de location à la page 0,0796€. Cette information est donnée uniquement à titre indicatif conformément à la législation en vigueur.

Exrait

  • Provides all the information essential for completing a sign-out report: clinical findings, pathologic findings, diagnosis, treatment, and prognosis.
  • Illustrates key pathologic and clinical aspects of disease entities through over 1430 superb, high-quality full-color images that help you evaluate and interpret biopsy samples.
  • Presents a team of internationally recognized experts for authoritative and up-to-date information from leading diagnosticians in neuropathology.
  • Features a user-friendly design with patterns color-coded to specific entities in the table of context and text and key points summarized in tables, charts, and graphs so you can quickly and easily find what you are looking for.
  • Directs you to the chapter and specific page number of the in-depth diagnostic guidance you need through a unique, pattern-based visual index at the beginning of the book.
  • Details key diagnostic features associated with rare and esoteric conditions in a visual encyclopedia with distinctive findings and artifacts for unusual patterns at the end of the book.

" />

Practical Surgical Neuropathology
A Diagnostic Approach

Arie Perry, MD
Professor of Pathology and Neurological Surgery, Vice Chair, Department of Pathology, Director of Neuropathology Division and the Neuropathology, Fellowship Program, University of California at San Francisco (UCSF), San Francisco, California
Daniel J. Brat, MD, PhD
Professor and Vice Chair, Translational Programs, Director, Division of Neuropathology, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, Georgia
Table of Contents
Cover image
Title page
Copyright
Contributors
Series Preface
Preface
Acknowledgments
Practical Surgical Neuropathology Major Patterns
Chapter 1: Neuropathology Patterns and Introduction
Central Nervous System Tumor Classification Schemes and Additional “Neuropathology Patterns”
Electron Microscopy
Immunohistochemistry
Molecular Diagnostics
Chapter 2: Normal Brain Histopathology
Cell Types
Tissue Organization
Features of Infancy and Childhood
Features of the Aging Nervous System
Chapter 3: Intraoperative Consultation and Optimal Processing
Types of Neurosurgical Specimens
Intraoperative Cytologic Preparations as a Complement to Frozen Tissue Sections
Fixation and Staining Options for Intraoperative Cytologic Preparations
Frozen Sectioning of Central Nervous System Tissue
Artifacts
Iatrogenically Introduced Hemostatic and Embolic Agents
The Bottom Line: What Does the Surgeon Need to Know?
Chapter 4: Neuroradiology: The Surrogate of Gross Neuropathology
Basic Noninvasive Diagnostic Imaging Techniques
Advanced Noninvasive Diagnostic Imaging Techniques
Basic Invasive Diagnostic Imaging Techniques
Advanced Invasive Therapeutic Techniques
Imaging Patterns in Neuroradiology
Advanced Strategies of Lesion Analysis
Conclusion
Chapter 5: Astrocytic and Oligodendroglial Tumors
Introduction and Brief Historical Overview
Diffuse Astrocytomas
Pilocytic Astrocytoma
Subependymal Giant-Cell Astrocytoma
Pleomorphic Xanthoastrocytoma
Oligodendroglioma
Oligoastrocytoma
Chapter 6: Ependymomas and Choroid Plexus Tumors
Definitions and Synonyms
Brief Historical Overview
Ependymal Tumors
Choroid Plexus Tumors
Chapter 7: Neuronal and Glioneuronal Neoplasms
Gangliogliomas and Gangliocytomas (“Ganglion Cell Tumors”)
Desmoplastic Infantile Astrocytoma and Ganglioglioma
Dysplastic Gangliocytoma of the Cerebellum (Lhermitte-Duclos Disease)
Central Neurocytoma and Other Neurocytic Neoplasms
Cerebellar Liponeurocytoma
Dysembryoplastic Neuroepithelial Tumor
Papillary Glioneuronal Tumor
Rosette-Forming Glioneuronal Tumor of the Fourth Ventricle
Hypothalamic Hamartoma
Chapter 8: Pineal Parenchymal Tumors
Introduction, Definitions, and Synonyms
Brief Historical Overview
Incidence and Demographics
Localization and Clinical Manifestations
Radiologic Features and Gross Pathology
Histopathology
Histologic Variants and Grading
Principles of Diagnosis and Grading on Small or Artifact-Ridden Samples
Differential Diagnosis
Ancillary Diagnostic Studies
Treatment and Prognosis
Chapter 9: Embryonal (Primitive) Neoplasms of the Central Nervous System
Definition and Synonyms
Brief Historical Overview
Medulloblastoma (Cerebellar Primitive Neuroectodermal Tumor)
Central Nervous System Supratentorial Primitive Neuroectodermal Tumors and Related Embryonal Neoplasms
Atypical Teratoid Rhabdoid Tumor
Chapter 10: Meningiomas
Benign Meningioma (WHO Grade I)
Atypical Meningioma (WHO Grade II) (either of two major criteria)
Brain-Invasive Meningioma (WHO Grade II)
Anaplastic (Malignant) Meningioma (WHO Grade III) (either of two criteria)
Introduction and Proposed Etiologies
Definition and Analogies with Meningothelial Cells
Brief Historical Overview of Nomenclature and Histogenetic Notions
Meningiomas That Are Mostly Benign (WHO Grade I)
WHO Grade II Meningiomas
WHO Grade III (Malignant) Meningiomas
Chapter 11: Mesenchymal Tumors of the Central Nervous System
Brief Historical Overview
Incidence and Demographics
Lipoma and Liposarcoma
Fibroblastic–Myofibroblastic Tumors
Smooth Muscle Tumors
Skeletal Muscle Tumors
Vascular Tumors
Bone Tumors
Chondromatous Tumors
Chordoma
Undifferentiated Sarcoma
Miscellaneous Tumors and Tumor-like Lesions
Chapter 12: Peripheral Nerve Sheath Tumors
Traumatic Neuroma
Schwannoma
Neurofibroma
Perineurioma
Granular Cell Tumor
Miscellaneous Benign Neurogenic Tumors
Primary Malignant Tumors of Nerve
Malignant Granular Cell Tumor
Primitive Neuroectodermal Tumor of the Nerve
Chapter 13: Epithelial, Neuroendocrine, and Metastatic Lesions
Metastatic Epithelial and Epithelioid Neoplasms
Paraganglioma
Papillary Tumor of the Pineal Region
Cysts of the Central Nervous System
Chapter 14: Lymphomas and Histiocytic Tumors
Definition and Synonyms
Lymphomas
Histiocytic Disorders
Chapter 15: Germ Cell Tumors
Definition and Synonyms
Brief Historical Overview
Incidence and Demographics
Clinical Manifestations and Localization
Radiologic Features and Gross Pathology
Histopathology
Histologic Variants
Ancillary Diagnostic Studies
Differential Diagnosis
Genetics
Treatment and Prognosis
Germ Cell Tumors Metastatic to the Central Nervous System
Chapter 16: Melanocytic Neoplasms of the Central Nervous System
Definitions and Synonyms
Brief Historical Overview
Melanocytoma and Melanoma
Neurocutaneous Melanosis/Melanomatosis
Chapter 17: Other Glial Neoplasms
Angiocentric Glioma
Astroblastoma
Chordoid Glioma
Chapter 18: Pathology of the Pituitary and Sellar Region
The Sellar Region and Normal Pituitary
Anterior Pituitary Tumors
Invasion and Malignancy in Anterior Pituitary Tumors
Pituitary Hyperplasia
Other Primary Pituitary Tumors
Miscellaneous Sellar Region Tumors
Sellar Region Cysts
Inflammatory Lesions
Infectious Lesions
Acknowledgment
Chapter 19: Therapy-Associated Neuropathology
Radiation Necrosis and Other Forms of Radiation Injury
Therapy-Induced Leukoencephalopathies and Vasculopathies
Therapy-Induced Secondary Neoplasms
Chapter 20: Familial Tumor Syndromes
Neurofibromatosis Type 1
Neurofibromatosis Type 2
Schwannomatosis
Tuberous Sclerosis Complex
Von Hippel-Lindau Disease
Turcot Syndrome
Nevoid Basal Cell Carcinoma Syndrome (Gorlin Syndrome)
Cowden Disease
Li-Fraumeni Syndrome
Carney Complex
Rhabdoid Predisposition Syndrome
Other Syndromes
Chapter 21: Infections and Inflammatory Disorders
Meningitis
Encephalitis
Focal Discrete Central Nervous System Infection
Inflammatory Conditions of the Nervous System Mimicking Infections
Central Nervous System Manifestations of Rheumatoid Arthritis
Central Nervous System Manifestations of Wegener Granulomatosis
Acknowledgment
Chapter 22: White Matter and Myelin Disorders
Multiple Sclerosis
Neuromyelitis Optica
Acute Disseminated Encephalomyelitis
Acute Hemorrhagic Leukoencephalitis
Progressive Multifocal Leukoencephalopathy
Subacute Sclerosing Panencephalitis
HIV Leukoencepahlopathy and Vacuolar Myelopathy
Binswanger Encephalopathy
Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy
Fat Embolism
Leukodystrophies
Osmotic Demyelination Syndrome (Central Pontine and Extrapontine Myelinolysis)
Chapter 23: Pathology of Epilepsy
Malformations of Cortical Development (Cortical Dysplasia)
Hippocampal Sclerosis
Infarct Including Porencephalic Cyst
Rasmussen Encephalitis
Chapter 24: Vascular and Ischemic Disorders
Ischemic Cerebral Infarct
Hypertensive Cerebrovascular Disease
Cerebral Amyloid Angiopathy
Vasculitis Involving the Nervous System
Giant-Cell Arteritis
Primary Angiitis of the Central Nervous System
Polyarteritis Nodosa
Cerebral Aneurysms
Fusiform and Infective (“Mycotic”) Aneurysms
Vascular Malformations
Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy
Moyamoya Syndrome
Chapter 25: Biopsy Pathology of Neurodegenerative Disorders in Adults
Alzheimer Disease
Dementia with Lewy Bodies and Idiopathic Parkinson Disease
Frontotemporal Lobar Degeneration and Pick Disease
Human Prion Diseases (Transmissible Spongiform Encephalopathies) Including Creutzfeldt-Jakob Disease
Index
Copyright

PRACTICAL SURGICAL NEUROPATHOLOGY: A DIAGNOSTIC APPROACH
ISBN: 978-0-443-06982-6
© 2010 by Churchill Livingstone an affiliate of Elsevier Inc. All rights reserved .
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions .
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
With respect to any drug or pharmaceutical products identified, readers are advised to check the most current information provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to verify the recommended dose or formula, the method and duration of administration, and contraindications. It is the responsibility of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses, to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication Data
Practical surgical neuropathology : a diagnostic approach / [edited by] Arie Perry, Daniel J. Brat.
p. ; cm.
Includes index.
ISBN 978-0-443-06982-6
1. Brain–Tumors–Diagnosis. 2. Nervous system–Diseases–Diagnosis. 3. Pathology, Surgical. I. Perry, Arie. II. Brat, Daniel J. III. Title.
[DNLM: 1. Nervous System Diseases–surgery. 2. Nervous System Diseases–pathology. 3. Pathology, Surgical–methods. WL 368 P8947 2010]
RC280.B7P723 2010
616.99′481–dc22
2009053489
Acquisitions Editor : William Schmitt
Publishing Services Manager : Joan Sinclair
Printed in China
Last digit is the print number: 9 8 7 6 5 4 3 2 1
Contributors

Daniel J. Brat, MD, PhD , Professor and Vice Chair, Translational Programs, Director, Division of Neuropathology, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, Georgia
David A. Decker, MD , Department of Neurology, University of Florida College of Medicine, McKnight Brain Institute, Gainesville, Florida
Michelle Fèvre-Montange, PhD , Université de Lyon, Lyon, France
Christine E. Fuller, MD , Professor of Pathology, Director of Neuropathology and Autopsy Pathology, Medical College of Virginia/Virginia Commonwealth University, Richmond, Virginia
Gregory N. Fuller, MD, PhD , Professor of Pathology (Neuropathology), The University of Texas Graduate School of Biomedical Sciences, Professor and Chief Section of Neuropathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas
Eyas M. Hattab, MD , Associate Professor, Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Medical Director, Immunohistochemistry Laboratory, Clarian Pathology Library, Indianapolis, Indiana
Eva Horvath, PhD , Pathology Consultant, Endocrine Pathology Research, St. Michael’s Hospital, Toronto, Ontario, Canada
Anne Jouvet, MD, PhD , Associate Professor of Pathology, Centre de Pathologie et Neuropathologie EST, Groupement Hospitalier EST, Hospices Civils de Lyon, Lyon, France
Scott E. Kilpatrick, MD , Pathologists Diagnostic Services, Forsyth Medical Center, Winston-Salem, North Carolina
B.K. Kleinschmidt-DeMasters, MD , Professor and Head, Division of Neuropathology, Department of Pathology, University of Colorado Health Science Center, Denver, Colorado
Kalman Kovacs, MD, PhD , Professor of Pathology, Department of Laboratory Medicine, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada
M. Joe Ma, MD, PhD , Assistant Professor of Pathology, Department of Medical Education, University of Central Florida College of Medicine, Pathologist, Department of Pathology, Florida Hospital—Orlando, Orlando, Florida
Ricardo V. Lloyd, MD , Professor of Pathology, Department of Anatomic Pathology, Mayo College of Medicine, Rochester, Minnesota
Sonia Narendra, MD , Pathology Resident, SUNY Upstate Medical University, Syracuse, New York
Werner Paulus, MD , Professor of Neuropathology, University of Muenster, Director Institute of Neuropathology, University Hospital Muenster, Muenster, Germany
Arie Perry, MD , Professor of Pathology and Neurological Surgery, Vice Chair, Department of Pathology, Director of Neuropathology Division and the Neuropathology Fellowship Program, University of California at San Francisco (UCSF), San Francisco, California
Richard A. Prayson, MD , Head, Section of Neuropathology, Department of Anatomic Pathology, Cleveland Clinic Foundation, Cleveland, Ohio
Bernd W. Scheithauer, MD , Professor of Pathology, Department of Anatomic Pathology, Mayo College of Medicine, Consultant, Mayo Clinic, Rochester, Minnesota
Robert E. Schmidt, MD, PhD , Professor and Chief, Division of Neuropathology, Washington University School of Medicine, St. Louis, Missouri
Ana I. Silva, MD , Hospital Assistant, Department of Anatomic Pathology, Hospital São Marcos, Graga, Protugal
Robert J. Spinner, MD , Professor, Neurologic Surgery, Orthopedics and Anatomy, Mayo Clinic, Rochester, Minnesota
Tarik Tihan, MD, PhD , Professor, Department of Pathology, University of California San Francisco, San Francisco, California
Kenneth L. Tyler, MD , Reuler-Lewin Family Professor of Neurology and Professor of Medicine and Microbiology, Departments of Neurology, Medicine, Microbiology, University of Colorado Denver Health Sciences Center, Neurology Service, Denver VA Medical Center, Denver, Colorado
Alexandre Vasiljevic, MD , Centre de Pathologie et Neuropathologie EST, Groupement Hospitalier EST, Hospices Civils de Lyon, Lyon, France
Franz J. Wippold, II, MD
Professor and Chief, Neuroradiology Section, Mallinckrodt Institute of Radiology, Washington University School of Medicine, Attending Neuroradiologist, Barnes-Jewish Hospital, Attending Neuroradiologist, St. Louis Children’s Hospital, St. Louis, Missouri
Adjunct Professor of Radiology/Radiological Sciences, F. Edward Hérbert School of Medicine, Uniformed Services University of the Health Sciences, Bethesda, Maryland
James M. Woodruff, MD , Emeritus Attending Physician, Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York
Anthony T. Yachnis, MD , Professor and Director of Anatomic Pathology, Department of Pathology, Immunology, and Laboratory Medicine, University of Florida College of Medicine, Gainesville, Florida
Series Preface
Kevin O. Leslie, MD and Mark R. Wick, MD
It is often stated that anatomic pathologists come in two forms: “Gestalt”-based individuals, who recognize visual scenes as a whole, matching them unconsciously with memorialized archives; and criterion-oriented people, who work through images systematically in segments, tabulating the results—internally, mentally, and quickly—as they go along in examining a visual target. These approaches can be equally effective, and they are probably not as dissimilar as their descriptions would suggest. In reality, even “Gestaltists” subliminally examine details of an image, and, if asked specifically about particular features of it, they are able to say whether one characteristic or another is important diagnostically.


In accordance with these concepts, in 2004 we published a textbook entitled Practical Pulmonary Pathology: A Diagnostic Approach (PPPDA). That monograph was designed around a pattern-based method, wherein diseases of the lung were divided into six categories on the basis of their general image profiles. Using that technique, one can successfully segregate pathologic conditions into diagnostically and clinically useful groupings.
The merits of such a procedure have been validated empirically by the enthusiastic feedback we have received from users of our book. In addition, following the old adage that “imitation is the sincerest form of flattery,” since our book came out other publications and presentations have appeared in our specialty with the same approach.
After publication of the PPPDA text, representatives at Elsevier, most notably William Schmitt, were enthusiastic about building a series of texts around pattern-based diagnosis in pathology. To this end we have recruited a distinguished group of authors and editors to accomplish that task. Because a panoply of patterns is difficult to approach mentally from a practical perspective, we have asked our contributors to be complete and yet to discuss only principal interpretative images. Our goal is eventually to provide a series of monographs which, in combination with one another, will allow trainees and practitioners in pathology to use salient morphological patterns to reach with confidence final diagnoses in all organ systems.
As stated in the introduction to the PPPDA text, the evaluation of dominant patterns is aided secondarily by the analysis of cellular composition and other distinctive findings. Therefore, within the context of each pattern, editors have been asked to use such data to refer the reader to appropriate specific chapters in their respective texts.
We have also stated previously that some overlap is expected between pathologic patterns in any given anatomic site; in addition, specific disease states may potentially manifest themselves with more than one pattern. At first, those facts may seem to militate against the value of pattern-based interpretation. However, pragmatically, they do not. One often can narrow diagnostic possibilities to a very few entities using the pattern method, and sometimes a single interpretation will be obvious. Both of those outcomes are useful to clinical physicians caring for a given patient.
It is hoped that the expertise of our authors and editors, together with the high quality of morphologic images they present in this Elsevier series, will be beneficial to our reader-colleagues.
Preface
Arie Perry, MD and Daniel J. Brat, MD, PhD
When Kevin Leslie and Mark Wick approached us a few years ago to write a new neuropathology textbook for a patterns-oriented organ-based series, it was with some trepidation that we ultimately accepted. After all, there are already some excellent texts available on this topic, and we have both contributed chapters to some of these in the past. However, the patterns approach used in the Leslie and Wick Practical Pulmonary Pathology book is somewhat novel, and we were not aware of others placing a major emphasis on this tactic toward neuropathology diagnosis. As our work progressed, we found additional ways of enhancing the reader’s experience and we are quite excited about the final product! Our primary target audience is the general surgical pathologist and pathology trainees. However, while we focused most on common issues of surgical neuropathology, rarer entities and clinicopathologic correlations are also well covered and illustrated. Therefore, we believe that this book will also be useful to neuropathologists and clinical colleagues from related medical specialties such as neurosurgery, neurology, neuroradiology, neuro-oncology, and pediatrics. In order to readdress the important question of why one should buy yet another neuropathology textbook, we provide the following list of strengths.

•  Patterns-based diagnostic approaches: In addition to offering the traditional disease-based approach to nervous system pathology ( Chapters 5 through 25 ), this book provides instructive algorithms based on 8 major (scanning magnification) patterns (immediately following the introduction) and 20 minor (higher magnification) patterns ( Chapter 1 ). This material can be particularly helpful to less experienced morphologists who may feel lost or overwhelmed by the myriad diagnostic possibilities. After the reader obtains an appropriately focused differential, he or she can quickly turn to more detailed discussions of specific entities in later chapters of the book. Alternatively, one can start with basic clinicoradiologic patterns combining patient age, location, and neuroimaging features to create a differential diagnosis ( Table 1-1 ). In fact, these two approaches are easily combined to further narrow the differential. To further enhance this strategy, the key clinicopathologic features for 21 common differentials and the immunoprofiles for 26 common tumors are summarized in Tables 1-3 and 1-4 , respectively. Major neuroimaging patterns are listed in Box 4-1 .
•  Background data: The nervous system is particularly challenging because of its remarkable anatomic and cellular complexity. For instance, the histology changes completely from one area to another, engendering diverse diagnostic differentials depending on the site of involvement. Therefore, a review of basic neuroanatomy and histopathology may help ( Chapter 2 ). In addition, the use of ancillary techniques is rapidly evolving, and therefore an overview of immunohistochemistry, electron microscopy, and molecular diagnostics is provided in Chapter 1 .
•  Intraoperative consultation and optimal processing: Nothing seems to provoke a panic attack more reliably than the “neuro frozen,” yet there is often little practical guidance available for this common setting. Furthermore, artifacts induced by frozen sections and many other procedures implemented by either the neurosurgeon or the pathologist can present serious pitfalls and may preclude an accurate diagnosis. These important topics are discussed in Chapter 3 .
•  Neuroradiology: As will be mentioned several times in this book, neuroradiology increasingly provides the most relevant gross pathology for nervous system biopsy interpretation, particularly when the tissue sample is small. In this context the pathologist must become at least an amateur neuroradiologist so that important radiologic-pathologic correlations are not missed. This critical topic is summarized and illustrated in Chapter 4 .
•  The authors: In addition to being international authorities on their topics, the authors were carefully selected for their clarity and enthusiasm for teaching . They are highly sought conference speakers, writers, and recipients of teaching awards. One is also known for a somewhat unconventional but highly popular teaching method. Dr. Perry’s innovative use of “neuropathology songs” to help medical students remember key features of neurological disorders has been the topic of several newspaper and radio reports. By the time this book is published, a CD recording should be complete and readers interested in a fun approach to musically reinforcing their knowledge base should visit www.neuropathsongs.com .
•  The images: One can scarcely find a more visually oriented medical specialty than pathology. Therefore, if the average picture is worth 1000 words, then the average pathology picture must be worth at least 10,000. With this in mind, we took great care to find the best images possible, making sure that the text is amply illustrated with generously sized high-quality figures . Given the focus of this book on surgical neuropathology, most of the “gross photos” are naturally magnetic resonance images. Nonetheless, we did not hesitate to utilize some postmortem photos and discussions when these clearly enhanced the reader’s understanding. This was particularly true for the infectious/inflammatory, vascular, and neurodegenerative disorders covered in Chapters 21 24 , and 25 , respectively.
•  The text: In order to highlight the most salient features of each disorder, italics are used throughout the text for quick reference, as are helpful summary tables and boxes .
We have endeavored to create a practical guide for those who work with biopsies of the nervous system and the patients from whom they were derived. We sincerely hope that you find it useful and enjoyable.
Acknowledgments
Arie Perry, MD and Daniel J. Brat, MD, PhD
As with any project of this magnitude, it simply can’t be done alone. I am extremely grateful to my talented coeditor, Dan Brat, and to all my wonderful coauthors for injecting countless hours of additional time and effort into their already busy schedules in order to create an exceptional product. For any diagnostic prowess I may possess, I owe an incredible debt to my surgical neuropathology mentor, Bernd Scheithauer of the Mayo Clinic, as most of my “pearls of wisdom” are easily traced back to him. I was particularly thrilled that he agreed to contribute two chapters on topics for which he is clearly one of the world’s authorities: pituitary pathology and peripheral nerve sheath tumors. For autopsy neuropathology, Joe Parisi was an equally outstanding mentor. In addition, I would especially like to thank Robert Schmidt for being such a remarkably supportive “boss” and close friend over my 12 years at Washington University in St. Louis. I particularly enjoyed our cordial competitions over who could shoot (and improve with Adobe Photoshop) the best photomicrographs (Bob: I think I won!). Particularly useful for this book was our practice of sharing with one another images from interesting cases as they came through our clinical service. A number of Bob’s masterpieces are sprinkled throughout several chapters and perhaps a few of mine have snuck into his chapter. Special thanks also go to Franz (“Jay”) Wippold of the Mallinckrodt Institute for Radiology with whom I’ve coauthored several review articles for a series entitled “Neuropathology for the Neuroradiologist.” It seems only fitting that he now offers his remarkable expertise to teach us some basic “neuroradiology for the neuropathologist.” In broader terms, I’d like to thank my parents, Gabriel and Bathsheba, for their incredible support and for giving me an innate desire to excel. My brother Ron similarly supported me through some challenging times. Lastly, I’m eternally grateful to my wife, Andrea, and my kids, Ryan and Jaclyn, for putting up with me and my long hours at work over the last few years.
The writing and editing of a comprehensive and authoritative textbook should not be entertained by the impatient or the faint of heart. Because of his wealth of knowledge, high standards, persistence, and overall good nature, I can think of no better collaborator on such an effort than Arie Perry. I look forward to updated editions as well as new neuropathology songs in the years to come. The collection of authors that we were able to gently persuade to contribute to this text is truly impressive. They deserve our deepest appreciation for allowing us to tap into their hard-earned expertise for this project. For their efforts, I hope this text will be widely acknowledged for the excellence it brings to the field of neuropathology. My own abilities to assist in this effort are directly attributed to those who drew me into neuropathology and to those who trained me both in person and at a distance. Joe Parisi and Bernd Scheithauer were larger than life figures that attracted a young medical student at the Mayo Clinic into the field of neuropathology and have continued to be role models. Peter Burger provided mentorship and enormous opportunity during residency and fellowship at Johns Hopkins Hospital and is most responsible for any academic successes I have had or will have. Finally, the family of Brats has always been a source of stability, inspiration, and thorough entertainment. Thanks to Paul, Dave, Jim, and Nancy Elaine.
Practical Surgical Neuropathology Major Patterns

Pattern Diseases to Be Considered Parenchymal infiltrate with hypercellularity

Diffuse glioma
CNS lymphoma
Infections
Active demyelinating disease
Cerebral infarct
Reactive gliosis Solid mass (pure)

Metastasis
Ependymoma
Subependymoma
Subependymal giant-cell astrocytoma (SEGA)
Central or extraventricular neurocytoma
Pineocytoma
Embryonal tumor (e.g., AT/RT)
Choroid plexus papilloma
Hemangioblastoma
Paraganglioma
Pituitary adenoma Solid and infiltrative process

Pilocytic astrocytoma
Pleomorphic xanthoastrocytoma
Glioblastoma/gliosarcoma (and other high grade gliomas)
Ganglioglioma
Dysembryoplastic neuroepithelial tumor (DNT)
Embryonal tumor (e.g., medulloblastoma/CNS PNET)
Choroid plexus carcinoma
Germ cell tumors
Craniopharyngioma
CNS lymphoma
Sarcoma
Histiocytic disorders
Abscess and other forms of infection Vasculocentric process

CNS lymphoma
Intravascular lymphoma
Angiocentric glioma
Ependymoma
Vasculitis
Meningioangiomatosis
Active demyelinating disease
Amyloid angiopathy
Arteriolosclerosis
Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL)
Vascular malformations
Infections (e.g., aspergillosis)
Neurosarcoidosis
Thromboembolic disease Extra-axial mass

Meningioma
Hemangiopericytoma
Solitary fibrous tumor
Hemangioblastoma
Sarcomas
Schwannoma and other nerve sheath tumors
Metastasis
Melanoma or melanocytoma
Secondary lymphoma or leukemia
Paraganglioma
Pituitary adenoma
Neurosarcoidosis
Granulomatous infections
Inflammatory pseudotumors
Calcifying pseudotumor of the neuraxis
Primary bone tumors (e.g., chordoma)
Histiocytic disorders (e.g., Rosai-Dorfman disease) Meningeal infiltrate

Meningeal carcinomatosis, gliomatosis, melanosis, melanomatosis, sarcomatosis, or hemangioblastomatosis
Metastatic medulloblastoma/CNS PNET
Secondary lymphoma or leukemia
Histiocytic disorders
Meningitis
Neurosarcoidosis
Infectious granulomatous diseases
Collagen vascular disorders
Sturge-Weber syndrome Destructive/necrotic process

Cerebral infarct
Radiation necrosis or treatment effects
Infections
Vasculitis
CNS lymphoma in an immunosuppressed patient
Intravascular lymphoma
CADASIL
Severe demyelinating disease
Metabolic/toxic disease Subtle pathology or near-normal biopsy

Nonrepresentative biopsy specimen
Subtle diffuse glioma (WHO grade II)
Hypothalamic hamartoma
Cortical dysplasia or tuber
Mesial temporal sclerosis
Intravascular lymphoma
Meningioangiomatosis
Mild encephalitis
Cerebral malaria
Ischemic disease
Neurodegenerative diseases
Benign cysts
Metabolic or toxic disorder
Reactive gliosis or “glial scar”





Pattern 1 Parenchymal infiltrate with hypercellularity
Elements of the pattern : The brain biopsy specimen shows intact cortical architecture, but a hypercellular infiltrate is evident at scanning magnification. In this particular example, an additional finding is secondary structure formation, with subpial condensation, perivascular aggregates, and perineuronal satellitosis. This growth pattern is most common in diffuse gliomas.

Additional Findings Diagnostic Considerations Chapter:page Secondary structures of Scherer Diffuse gliomas Ch. 5:63 Extensive bilateral cerebral involvement

Gliomatosis cerebri
Lymphomatosis cerebri

Ch. 5:71, 80
Ch. 14:316 Angiocentric pattern

CNS lymphoma
Angiocentric glioma
Meningoencephalitis/Infections
Active demyelinating disease

Ch. 14:315
Ch. 17:361
Ch. 21:468
Ch. 22:485 Microcystic pattern Diffuse gliomas Ch. 5:63 Pleomorphism

Astrocytoma/glioblastoma
Infections, especially PML

Ch. 5:63
Ch. 21:470 Monomorphism

Oligodendroglioma
Some lymphomas

Ch. 5:93
Ch. 14:315 Lymphocytic infiltrate

Gemistocytic astrocytoma
CNS lymphoma
Meningoencephalitis/Infections
Active demyelinating disease

Ch. 5:70
Ch. 14:315
Ch. 21:468
Ch. 22:485 Foamy histiocytes

CNS lymphoma
Active demyelinating disease
Cerebral infarct

Ch. 14:315
Ch. 22:485
Ch. 24:528 Cytologic atypia or anaplasia

Diffuse gliomas
CNS lymphoma

Ch. 5:63
Ch. 14:315 Viral inclusions or organisms Meningoencephalitis/Infections Ch. 21:468 None Reactive gliosis Ch. 1:8 Ch. 5:74




Pattern 2 Solid mass (pure)
Elements of the pattern : The biopsy specimen shows a very sharply demarcated intracerebral mass. The increased cellularity imparts a blue color to the tumor, whereas foci of central necrosis appear pink. An additional finding was gland formation, consistent with metastatic adenocarcinoma.

Additional Findings Diagnostic Considerations Chapter:page Mucin-filled glands Metastatic adenocarcinoma Ch. 13:287 Perivascular pseudorosettes

Subependymal giant-cell astrocytoma
Ependymoma
Central or extraventricular neurocytoma
Pineocytoma
Metastasis (neuroendocrine)
Paraganglioma
Pituitary adenoma

Ch. 5:88
Ch. 6:103
Ch. 7:135
Ch. 8:152
Ch. 13:287
Ch. 13:296
Ch. 18:372 Nodularity

Subependymoma
Metastasis (neuroendocrine)
Paraganglioma
Pituitary adenoma

Ch. 6:104
Ch. 13:287
Ch. 13:296
Ch. 18:372 Gliofibrillary processes

Subependymal giant-cell astrocytoma
Ependymoma
Subependymoma

Ch. 5:88
Ch. 6:103
Ch. 6:104 Papillary pattern

Choroid plexus papilloma
Papillary ependymoma
Metastatic carcinoma
Pituitary adenoma

Ch. 6:113
Ch. 6:106, 109
Ch. 13:287
Ch. 18:372 Hypervascularity

Choroid plexus papilloma
Hemangioblastoma

Ch. 6:113
Ch. 20:440 Neuropil/neuronal rosettes

Central or extraventricular neurocytoma
Pineocytoma

Ch. 7:135
Ch. 8:152 Adjacent piloid gliosis

Craniopharyngioma
Hemangioblastoma

Ch. 18:402
Ch. 20:440 Epithelioid cytology

Choroid plexus papilloma
Metastatic carcinoma

Ch. 6:113
Ch. 13:287 Small primitive cells

Embryonal tumor (AT/RT)
Metastatic carcinoma (small cell)

Ch. 9:165, 179
Ch. 13:287 Melanin pigment Melanoma (usually metastatic) Ch. 13:291 Ch. 6:353 Clear cells

Clear cell ependymoma
Central or extraventricular neurocytoma
Pineocytoma
Hemangioblastoma
Metastatic carcinoma

Ch. 6:107
Ch. 7:135
Ch. 8:152
Ch. 20:440
Ch. 13:287 Cytologic anaplasia

Embryonal tumor (AT/RT)
Metastatic carcinoma

Ch. 9:165, 179
Ch. 13:287




Pattern 3 Solid and infiltrative process
Elements of the pattern : The biopsy specimen shows a mostly solid-appearing neoplasm ( left half ), but has fuzzy or ill-defined margins with the adjacent brain parenchyma, consistent with at least a partially infiltrative component as well ( right half , especially in white matter). Additional findings in this case were reticulin-rich spindled elements, GFAP-positive glial elements, and pseudopalisading necrosis, consistent with gliosarcoma.

Additional Findings Diagnostic Considerations Chapter:page Biphasic growth (compact and microcystic), EGBs, Rosenthal fibers

Pilocytic astrocytoma
Pleomorphic xanthoastrocytoma
Ganglioglioma
Dysembryoplastic neuroepithelial tumor

Ch. 5:82
Ch. 5:91
Ch. 7:125
Ch. 7:140 Pseudopalisading necrosis Glioblastoma or gliosarcoma Ch. 5:63, 66 Nodularity

Anaplastic oligodendroglioma
Dysembryoplastic neuroepithelial tumor
Ganglioglioma
Desmoplastic or nodular medulloblastoma
Germinoma or germ cell tumors

Ch. 5
Ch. 7:140
Ch. 7:125
Ch. 9:169
Ch. 15:336 Angiocentric pattern

CNS lymphoma
Infections

Ch. 14:315
Ch. 21:477 Fascicles of spindled cells

Gliosarcoma
Pleomorphic xanthoastrocytoma
Primary CNS sarcoma (rare)

Ch. 5:70
Ch. 5:91
Ch. 11:219 Inflammation-rich

Pleomorphic xanthoastrocytoma
Ganglioglioma
CNS lymphoma
Germinoma or germ cell tumors
Abscess and other infections

Ch. 5:91
Ch. 7:125
Ch. 14:315
Ch. 15:336
Ch. 21:477 Adjacent piloid gliosis

Pleomorphic xanthoastrocytoma
Craniopharyngioma

Ch. 5:91
Ch. 18:402 Glial cytology

Pilocytic astrocytoma
Pleomorphic xanthoastrocytoma
Glioblastoma or gliosarcoma
Ganglioglioma
Dysembryoplastic neuroepithelial tumor

Ch. 5:82
Ch. 5:91
Ch. 5:63, 66
Ch. 7:125
Ch. 7:140 Large ganglioid cells with vesicular nuclei and large nucleoli

Pleomorphic xanthoastrocytoma
Ganglioglioma
Dysembryoplastic neuroepithelial tumor
CNS lymphoma (anaplastic large cell)
Germinoma

Ch. 5:91
Ch. 7:125
Ch. 7:140
Ch. 14:330
Ch. 15:336 Epithelioid cytology

Choroid plexus carcinoma
Germ cell tumors
Craniopharyngioma

Ch. 6:113
Ch. 15:333
Ch. 18:402 Small primitive cells

Choroid plexus carcinoma
Medulloblastoma/CNS PNET
CNS lymphoma

Ch. 6:113
Ch. 9:165, 175
Ch. 14:315 Foamy cells

Pleomorphic xanthoastrocytoma
Glioblastoma (occasionally)
Histiocytic disorders
Infections

Ch. 5:91
Ch. 5:63, 66
Ch. 14:326
Ch. 21:455




Pattern 4 Vasculocentric process
Elements of the pattern: The biopsy specimen shows a disease process that is clearly centered on blood vessels. Additional findings in this case were foci of angionecrosis and vascular or perivascular inflammation, consistent with vasculitis.

Additional Findings Diagnostic Considerations Chapter:page Perivascular or intravascular infiltrate

CNS lymphoma
Meningoencephalitis/infection
Neurosarcoidosis
Active demyelinating disease
Vasculitis
Amyloid angiopathy with vasculitis

Ch. 14:315
Ch. 21:468
Ch. 21:481
Ch. 22:485
Ch. 24:537
Ch. 24:535 Intraluminal atypical cells Intravascular lymphoma Ch. 14:319 Perivascular glial or spindled cells

Ependymoma
Angiocentric glioma
Meningioangiomatosis

Ch. 6:103
Ch. 17:361
Ch. 20:433 Angionecrosis

Infections (aspergillosis)
Vasculitis
Thromboembolic disease

Ch. 21:464
Ch. 24:537
Ch. 24:528 Vascular hyalinization

Meningioangiomatosis
Amyloid angiopathy
CADASIL
Arteriolosclerosis
Vasculitis
Vascular malformations

Ch. 20:433
Ch. 24:535
Ch. 24:546
Ch. 24:533
Ch. 24:537
Ch. 24:542 Granular vascular deposits CADASIL Ch. 24:546 Granulomas or giant cells

Infections
Neurosarcoidosis
Vasculitis
Amyloid angiopathy with vasculitis

Ch. 21:455
Ch. 21:481
Ch. 24:537
Ch. 24:535 Cerebral hemorrhage

Infections (aspergillosis)
Amyloid angiopathy
Vascular malformations

Ch. 21:464
Ch. 24:535
Ch. 24:542 Cerebral infarcts or microinfarcts

Intravascular lymphoma
Infections
Neurosarcoidosis
Vasculitis
Amyloid angiopahty
CADASIL
Arteriolosclerosis
Thromboembolic disease

Ch. 14:319
Ch. 21:455
Ch. 21:481
Ch. 24:537
Ch. 24:535
Ch. 24:546
Ch. 24:533
Ch. 24:528 Disorganized, irregular blood vessels

Meningioangiomatosis
Vascular malformations

Ch. 20:433
Ch. 24:542




Pattern 5 Extra-axial mass
Elements of the pattern : The biopsy specimen shows a solid mass attached to a strip of dura in the upper portion of the image. Additional findings in this case were whorls of epithelioid cells and scattered psammoma bodies, consistent with meningioma.

Additional Findings Diagnostic Considerations Chapter:page Whorls or nests

Meningioma
Chordoma
Schwannoma (occasionally)
Metastatic carcinoma
Paraganglioma
Melanocytoma

Ch. 10:185
Ch. 11:231
Ch. 12:240
Ch. 13:287
Ch. 13:296
Ch. 16:353 Psammoma bodies

Meningioma
Psammomatous melanotic schwannoma
Metastatic carcinoma (rare)

Ch. 10:185
Ch. 12:251
Ch. 13:287 Peripheral or cranial nerve involvement

Pilocytic astrocytoma (optic pathway)
Orbital meningioma
Orbital sarcoma (rhabdomyosarcoma)
Schwannoma
Neurofibroma
Perineurioma
MPNST
Neurolymphomatosis

Ch. 5:82
Ch. 10:185
Ch. 11:226
Ch. 12:240
Ch. 12:251
Ch. 12:260
Ch. 12:272
Ch. 14:315 Biphasic (compact and loose) pattern with Verocay bodies

Meningioma (rare)
Schwannoma

Ch. 10:185
Ch. 12:240 Hypervascular

Angiomatous meningioma
Hemangiopericytoma
Hemangioblastoma

Ch. 10:194
Ch. 11:220
Ch. 20:440 Gaping “staghorn” blood vessels

Meningioma (rare)
Hemangiopericytoma
Solitary fibrous tumor

Ch. 10:185
Ch. 11:220
Ch. 11:220 Alternating “dark and light” regions

Hemangiopericytoma
Solitary fibrous tumor

Ch. 11:220
Ch. 11:220 Dense bundles of eosinophilic collagen

Clear cell meningioma
Solitary fibrous tumor

Ch. 10:200
Ch. 11:220 Inflammatory infiltrate

Lymphoplasmacyte-rich meningioma
Inflammatory myofibroblastic tumor
Secondary lymphoma/leukemia
Infections
Neurosarcoidosis
Collagen vascular disorders

Ch. 10:198
Ch. 11:221
Ch. 14:321
Ch. 21:455
Ch. 21:481
Ch. 21:481 Fibrillar to amorphous basophilic material Calcifying pseudoneoplasm of the neuraxis Ch. 10:211 Small primitive cells

Hemangiopericytoma
Other sarcomas (EWS/pPNET)
Metastatic carcinoma (small cell)
Secondary lymphomas/leukemias

Ch. 11:220
Ch. 11:233
Ch. 13:287
Ch. 14:321 Large anaplastic cells

Anaplastic meningioma
Metastatic carcinoma
Anaplastic large cell lymphoma
Myeloid sarcoma
Melanoma

Ch. 10:192, 203
Ch. 13:287
Ch. 14:320
Ch. 14:322
Ch. 16:353 Epithelioid cells

Meningioma
Metastatic carcinoma
Paraganglioma
Melanoma
Pituitary adenoma

Ch. 10:185
Ch. 13:287
Ch. 13:296
Ch. 16:353
Ch. 18:372 Clear cells

Clear cell meningioma
Hemangiopericytoma
Other sarcomas (leiomyosarcoma)
Metastatic carcinoma
Paraganglioma
Histiocytic disorders
Hemangioblastoma

Ch. 10:200
Ch. 11:220
Ch. 11:225
Ch. 13:287
Ch. 13:296
Ch. 14:326
Ch. 20:440 Foamy cells

Angiomatous meningioma
Schwannoma (histiocytes)
Histiocytic disorders
Hemangioblastoma

Ch. 10:194
Ch. 12:240
Ch. 14:326
Ch. 20:440 Granulomas or giant cells

Infections (TB, fungal meningitis)
Neurosarcoidosis
Collagen vascular disorders

Ch. 21:455
Ch. 21:481
Ch. 21:481



Pattern 6 Meningeal infiltrate
Elements of the pattern : The whole-mount brain section shows a markedly expanded subarachnoid space filled with blue cells. At higher magnification, the infiltrate consisted predominantly of neutrophils, consistent with acute meningitis.

Additional Findings Diagnostic Considerations Chapter:page Neoplastic cells

Metastatic medulloblastoma/CNS PNET
Meningeal sarcomatosis
Meningeal carcinomatosis
Secondary lymphoma/leukemia
Meningeal melanosis/melanomatosis
Meningeal gliomatosis

Ch. 9:165, 175
Ch. 11:220
Ch. 13:287
Ch. 14:321
Ch. 16:353, 357
Ch. 21: 465 Venous malformation Sturge-Weber syndrome Ch. 20:451 Neutrophil-rich infiltrate Acute bacterial meningitis Ch. 21:456 Lymphoplasmacytic infiltrate

Infectious meningitis
Chemical meningitis
Neurosarcoidosis
Collagen vascular disorder

Ch. 21:456
Ch. 21:465
Ch. 21:481
Ch. 21:481 Granulomas/giant cells

Infectious meningitis (TB, fungal)
Neurosarcoidosis
Collagen vascular disorder

Ch. 21:456
Ch. 21:481
Ch. 21:481 Clear to foamy cells

Meningeal carcinomatosis
Histiocytic disorders

Ch. 13:287
Ch. 14:326




Pattern 7 Destructive or necrotic process
Elements of the pattern : The brain biopsy specimen from a patient with known glioma shows extensive fibrinoid parenchymal and vascular necrosis, consistent with radiation necrosis.

Additional Findings Diagnostic Considerations Chapter:page Fibrinoid brain necrosis, vascular hyalinization, telangiectasias Radiation necrosis or treatment effects Ch. 19:417 Angionecrosis

Radiation necrosis or treatment effects
Infection (toxoplasmosis)
Vasculitis

Ch. 19:417
Ch. 21:476
Ch. 24:537 Vascular or perivascular inflammation

Lymphoma (immunosuppressed host)
Severe demyelinating disease (rare)
Vasculitis

Ch. 14:316
Ch. 22:485
Ch. 24:537 Intraluminal infiltrate Intravascular lymphoma Ch. 14:319 Granular vascular deposits CADASIL Ch. 24:546 Eosinophilic necrotic neurons Acute cerebral infarct Ch. 24:528 Neutrophil-rich infiltrate

Infection (abscess)
Acute cerebral infract (rare)

Ch. 21:478
Ch. 24:528 Macrophage-rich infiltrate

Severe demyelinating disease (rare)
Metabolic or toxic disorders
Cerebral infarct

Ch. 22:485
Ch. 22:506, 510
Ch. 24:528 Granulomas or giant cells

Infections (TB, fungal)
Vasculitis

Ch. 21:456
Ch. 24:537 Viral inclusions Encephalitis (HSV) Ch. 21:468




Pattern 8 Subtle pathology or near-normal biopsy specimen
Elements of the pattern : The brain biopsy specimen from a patient with chronic seizure disorder shows a nearly normal cortex. However, there is a subtle disarray of the laminar architecture and clustering of large superficial neurons in the center. Leptomeningeal gray matter heterotopia was also seen in other regions of the biopsy. This constellation of findings is consistent with a malformation of cortical development (i.e., cortical dysplasia).

Additional Findings Diagnostic Considerations Chapter:page Reactive gliosis or cerebral edema

Nonrepresentative biopsy
Subtle diffuse glioma
Hypothalamic hamartoma
Pineal cyst
Arachnoid cyst
Other developmental cyst
Metabolic or toxic disorders
Nonanatomic cause of epilepsy
Subtle form of cortical dysplasia

Chs. 3 – 5
Ch. 5:63
Ch. 7:146
Ch. 8:157
Ch. 13:309
Ch. 13:303
Ch. 22:506, 510
Ch. 23:515
Ch. 23:515 Glial atypia, clustering, or secondary structuring Diffuse glioma Ch. 5:63 Intraluminal atypical cells Intravascular lymphoma Ch. 14:319 Neuronal clustering and mild dysmorphism

Hypothalamic hamartoma
Subtle form of cortical dysplasia

Ch. 7:146
Ch. 23:515 Balloon cells

Focal cortical dysplasia, type IIb
Tuber

Ch. 23:518
Ch. 23:516, 518 Neuronal loss in hippocampus (CAI, CA4) Mesial temporal sclerosis/hippocampal sclerosis Ch. 23:520 Microglial nodules/scant perivascular inflammation

Encephalitis (infectious, paraneoplastic)
Rasmussen encephalitis

Ch. 21:469
Ch. 23:523 Intravascular pigment Cerebral malaria Ch. 21:474 Red necrotic neurons Acute cerebral infarct Ch. 24:528 Vascular hyalinization

Radiation effects
Meningioangiomatosis
Amyloid angiopathy
CADASIL
Arteriolosclerosis

Ch. 19:417
Ch. 20:433
Ch. 24:535
Ch. 24:546
Ch. 24:533 Granular vascular deposits CADASIL Ch. 24:546 Hemorrhage/hemosiderin

Epileptogenic “glial scar”
Amyloid angiopathy
Small cavernous angioma

Ch. 23:515
Ch. 24:535
Ch. 24:545 Neurofibrillary tangles or neuritic plaques Alzheimer disease Ch. 25:553 Spongiform changes in gray matter

Cerebral infarct
Creutzfeldt-Jakob disease (CJD)
Other neurodegenerative disorders (usually superficial spongiosis)

Ch. 24:528
Ch. 25:566
Ch. 25:559


For additional histopathology algorithms, see “Minor Histopathologic Patterns of Nervous System Tumors” in the next chapter.
1
Neuropathology Patterns and Introduction
Arie Perry,
Daniel J. Brat

Central Nervous System Tumor Classification Schemes and Additional “Neuropathology Patterns”   1
Electron Microscopy   1
Immunohistochemistry   11
Glial Markers  11
Neuronal Markers  11
Epithelial Markers  13
Proliferation Markers  13
Molecular Diagnostics   13

Central Nervous System Tumor Classification Schemes and Additional “Neuropathology Patterns”
The first comprehensive classification of nervous system tumors, formulated by Percival Bailey and Harvey Cushing in 1926, was founded on presumed parallels between embryologic and neoplastic cells. 1 In large part, this histogenetic “cell of origin” model still forms the basis for today’s nomenclature, although much of the terminology has changed considerably. Renewed interest in the role of developmental pathways in tumorigenesis has led to more recent studies focusing on cancer stem cells and progenitor cells. 2, 3 In 1949, however, as a means of enhancing the clinical utility of tumor classification, Kernohan contributed a tumor-grading system focusing on correlations with patient prognosis. 4 As progress was made over time, Russell and Rubinstein continued to modify and update the Bailey and Cushing system from the 1960s through the 1980s. Nonetheless, considerable variability in diagnostic practice existed on both sides of the Atlantic. In order to enhance consistency, an experts’ consensus scheme known as the World Health Organization (WHO) classification scheme was first completed in 1979 and then revised in 1993, 2000, and 2007. 5 This scheme is currently the most widely utilized by neuropathologists for typing and grading tumors.
The 2007 WHO “blue book” currently lists over 100 types of nervous system tumors and their variants. 5 This level of complexity can be daunting; therefore, an organized approach or algorithm is required. As a first step, consideration of clinical and radiologic characteristics is a critical way to narrow the differential diagnosis, often to a few fairly common entities. In fact, the combination of patient age and neuroimaging features (including tumor location) provides some of the most powerful diagnostic clues before any tissue is even sampled or examined under the microscope. For example, the differential varies considerably for supratentorial versus infratentorial, pediatric versus adult, and enhancing versus nonenhancing tumors. The most common diagnostic considerations are summarized by age, location, and imaging features in Table 1-1 , with each specific entity discussed in greater detail in subsequent chapters. Also, for a much more detailed background on the use of neuroimaging, the reader is referred to Chapter 4 . This is a particularly critical topic in surgical neuropathology, since brain and spinal cord biopsy specimens are often small and the neuroimaging essentially provides the “gross pathology.”

Table 1-1
Common Central Nervous System Tumor Diagnoses by Location, Age, and Imaging Characteristics Location Child/Young Adult Older Adult Cerebral/supratentorial
Ganglioglioma (TL, cyst-MEN, E)
DNT (TL, intracortical nodules)
PNET (solid, E)
AT/RT (infant, E)
Grade II-III diffuse glioma (NE, focal E)
GBM (E or rim E, “butterfly” mass)
Metastases (grey-white junctions, E or rim E)
Lymphoma (periventricular, E) Cerebellar/infratentorial/4th v.
PA (cyst-MEN)
Medulloblastoma (vermis, E)
Ependymoma (4th v., E)
Choroid plexus papilloma (4th v., E)
AT/RT (infant, E)
Metastases (multiple, E or rim E)
Hemangioblastoma (cyst-MEN)
Choroid plexus papilloma (4th v., E) Brainstem
“Brainstem glioma” (pons, ± E)
PA (dorsal, exophytic, cyst-MEN) Gliomatosis cerebri (multifocal, ± E) Spinal cord (intra-medullary)
Ependymoma (E, ± syrinx)
PA (cystic, E)
Drop metastases (cauda equina, E)
MPE (filum terminale, E)
Ependymoma (E, ± syrinx)
Diffuse astrocytoma (ill-defined, ± E)
MPE (filum terminale, E)
Paraganglioma (filum terminale, E) Spinal cord (intradural, extramedullary)
Clear cell meningioma (± dural tail, E)
Schwannoma (NF2, nerve origin, dumbbell shape, E)
Drop metastases (leptomeningeal, E)
Schwannoma (nerve origin, dumbbell shape, E)
Meningioma (± dural tail, E) Spinal cord (extradural)
Bone tumor spread (EWS/PNET, usually E)
Meningioma (± dural tail, E)
Abscess (E)
Vascular malformations (dilated vessels on imaging, ± E)
Herniated disc (T1-spin echo MRI, NE)
Postoperative scar (E)
Secondary lymphoma (E)
Metastases (E)
Abscess (E) Extra-axial/dural Secondary lymphoma/leukemia (E)
Meningioma (E with dural tail)
Metastases (E)
Secondary lymphoma/leukemia (E) Intrasellar
Pituitary adenoma (solid, E)
Craniopharyngioma (cystic, E)
Rathke’s cleft cyst (cystic, ± E)
Pituitary adenoma (solid, E)
Craniopharyngioma (cystic, E
Rathke’s cleft cyst (cystic, ± E) Suprasellar/hypothalamic/optic pathway/3rd v.
Germinoma (solid, E)
Craniopharyngioma (cystic, E)
PA (cyst-MEN)
Pilomyxoid astrocytoma (infant, solid, E)
Colloid cyst (3rd v., ± E)
Craniopharyngioma (cystic, E) Pineal Germinoma (solid, E)Pineocytoma (solid, E)Pineoblastoma (solid, E)Pineal cyst (cystic, NE) Pineocytoma (solid, E)Pineal cyst (cystic, NE) Thalamus
PA (cyst-MEN)
AA/GBM (E or rim E)
AA/GBM (E or rim E)
Lymphoma (E, ± multifocality) Cerebellopontine angle
Vestibular schwannoma (NF2, E, involves internal auditory meatus)
Choroid plexus tumor (E, component in 4th v.)
Vestibular schwannoma (E, involves internal auditory meatus)
Meningioma (E with dural tail) Lateral ventricle Central neurocytoma (E)SEGA (tuberous sclerosis, E)Choroid plexus papilloma (E)Choroid plexus carcinoma (infant, E, large, invasive)
Central neurocytoma (E)
SEGA (tuberous sclerosis, E)
Choroid plexus papilloma (E)
Subependymoma (± E)
Meningioma (E with dural tail) Nerve root/paraspinal
Neurofibroma (NF1, E)
MPNST (NF1, E, necrotic)
Schwannoma (E, dumbbell shape)
Meningioma (E with dural tail)
Secondary lymphoma (E)
Neurofibroma (NF1, E))
MPNST (E, necrotic)
AA, anaplastic astrocytoma; AT/RT, atypical teratoid/rhabdoid tumor; DNT, dysembryoplastic neuroepithelial tumor; E, enhancing; EWS, Ewing’s sarcoma; 4th v., fourth ventricle; GBM, glioblastoma; MEN, mural enhancing nodule; MPE, myxopapillary ependymoma; MPNST, malignant peripheral nerve sheath tumor; MRI, magnetic resonance imaging; NE, nonenhancing; NF, neurofibroma; NF1, neurofibromatosis type 1; PA, pilocytic astrocytoma; PNET, primitive neuroectodermal tumor; SEGA, subependymal giant cell astrocytoma; 3rd v., third ventricle; TL, temporal lobe.
The next set of clues is naturally provided by histopathology. The eight major patterns provided at the beginning of this textbook narrow the differential diagnosis considerably based purely on the overall low-magnification appearance, and the subheadings of additional findings provides a useful diagnostic algorithm. When presented with a challenging biopsy specimen, the pathologist can start with either the clinical or morphology-based approaches but is encouraged to incorporate all available data before making a final diagnosis. In the vast majority of cases, the clinical, radiologic, and pathologic features are all consistent with one another; if not, the pathologist should carefully reexamine the specimen to be sure that all appropriate possibilities have been considered and if necessary, excluded with ancillary studies. The use of common ancillary diagnostic techniques is briefly summarized in this chapter, with many more examples provided in the subsequent topic-specific chapters. As useful secondary algorithms, the major differential diagnosis based on an additional 20 minor histologic patterns is presented in Table 1-2 , with helpful clinicopathologic features summarized for 21 common differential diagnoses in Table 1-3 .

Table 1-2
Minor Histopathologic Patterns of Nervous System Tumors
“Fried Egg” or Clear Cells Lobulated, Nested, or Nodular

• Oligodendroglioma
• Glioblastoma, small cell variant
• Dysembryoplastic neuroepithelial tumor
• Clear cell ependymoma
• Central/extraventricular neurocytoma
• Pineocytoma
• Pilocytic astrocytoma
• Poorly preserved diffuse astrocytomas
• Autolyzed non-neoplastic brain
• Demyelinating disease (macrophages)
• Cerebral infarct (macrophages)
• Rosette-forming glioneuronal tumor
• Paraganglioma
• Clear cell meningioma
• Germinoma
• Pituitary adenoma

•  Desmoplastic/nodular medulloblastoma
•  Extensively nodular medulloblastoma
•  Dysembryoplastic neuroepithelial tumor
•  Oligodendrogliomas, mostly high-grade
•  Subependymoma
•  Ganglioglioma
•  Paraganglioma
•  Metastatic carcinoma
•  Pineal parenchymal tumors, mostly of intermediate differentiation
•  Plexiform neurofibroma or schwannoma
•  Epithelioid MPNST
•  Meningiomas: mostly meningothelial, chordoid, and atypical
•  Melanocytoma and melanoma
•  Germinoma
•  Pituitary adenoma Fascicles/Storiform Bundles of Spindled Cells Sheets of Epithelioid Cells

•  Fibrous, transitional, or anaplastic meningiomas
•  Schwannoma
•  Solitary fibrous tumor
•  Hemangiopericytoma
•  MPNST or other spindle cell sarcomas
•  Spindled glioblastomas/astrocytomas
•  Gliosarcoma
•  Pleomorphic xanthoastrocytoma
•  Desmoplastic gangliolioma/astrocytoma
•  Tanycytic ependymoma
•  Atypical teratoid/rhabdoid tumor
•  Melanocytoma or melanoma
•  Pituicytoma
•  Spindle cell oncocytoma of pituitary
•  Histiocytic disorders
•  Fibrous or granulomatous reactions

•  Metastatic carcinoma
•  Meningioma
•  Melanoma
•  Atypical teratoid/rhabdoid tumor
•  Pituitary adenoma
•  Anaplastic oligodendroglioma
•  Ependymoma
•  Choroid plexus tumors
•  Astroblastoma
•  Chordoid glioma
•  Germ cell tumors
•  Paraganglioma
•  Epithelioid glioblastoma
•  Epithelioid nerve sheath tumors
•  Craniopharyngioma
•  Anaplastic large cell lymphoma
•  Plasmacytoma Monomorphic Cytology Biphasic Pattern (Loose and Compact Areas)

•  Central/extraventricular neurocytomas
•  Pineocytoma
•  Oligodendroglioma
•  Pituitary adenoma
•  Pilomyxoid astrocytoma
•  Angiocentric glioma

•  Pilocytic astrocytoma
•  Pleomorphic xanthoastrocytoma
•  Ganglioglioma
•  Schwannoma
•  Hemangiopericytoma
•  Malignant peripheral nerve sheath tumor Microcystic Myxoid/Mucin-rich

•  Diffuse gliomas, mostly low-grade
•  Pilocytic astrocytomas
•  Pleomorphic xanthoastrocytoma
•  Subependymoma
•  Ganglioglioma
•  Schwannoma
•  Microcystic meningioma
•  Hemangioblastoma
•  Yolk sac tumor
•  Teratoma
•  Craniopharyngioma

•  Dysembryoplastic neuroepithelial tumor
•  Myxopapillary ependymoma
•  Chordoid glioma
•  Chordoid meningioma
•  Metaplastic chondromyxoid meningioma
•  Pilocytic/pilomyxoid astrocytomas
•  Diffuse gliomas, mostly low-grade
•  Rosette-forming glioneuronal tumor
•  Atypical teratoid/rhabdoid tumor
•  Nerve sheath tumors
•  Yolk sac tumor
•  Teratoma Rosette Forming Perivascular Pseudorosettes

•  Ependymoma (true ependymal)
•  Medulloblastoma (Homer Wright)
•  CNS PNET (Homer Wright, ependymoblastic)
•  Neurocytomas (neurocytic)
•  Pineocytoma (pineocytic)
•  Pineoblastoma (pineoblastic)
•  Embryonal tumor with abundant neuropil and true rosettes (ependymoblastic)
•  Pituitary adenoma (rosette-like pattern)

•  Ependymoma
•  Astroblastoma
•  Angiocentric glioma
•  Papillary glioneuronal tumor
•  Central/extraventricular neurocytomas
•  Medulloblastomas/PNETs (occasionally)
•  Glioblastoma (occasionally)
•  Papillary meningioma
•  Pituitary adenoma Palisading/Pseudopalisading Cells Papillary/Pseudopapillary

•  Glioblastoma (pseudopalisading necrosis)
•  Schwannoma (Verocay bodies)
•  Pilocytic astrocytoma (spongioblastic)
•  Oligodendroglioma (spongioblastic)
•  Ependymoma (spongioblastic)
•  Medulloblastoma/PNET (spongioblastic)
•  Angiocentric glioma (subpial palisades)

•  Choroid plexus tumors
•  Papillary ependymoma
•  Astroblastoma
•  Papillary meningioma
•  Hemangiopericytoma
•  Papillary glioneuronal tumor
•  Atypical teratoid/rhabdoid tumor
•  Papillary craniopharyngioma
•  Germ cell tumors (e.g., yolk sac tumor)
•  Papillary tumor of the pineal region “Small Blue Cells” (i.e., Primitive) Multinucleated Giant Cells

•  Medulloblastoma
•  CNS PNET
•  Pineoblastoma
•  Atypical teratoid/rhabdoid tumor
•  Lymphoma/leukemia
•  Glioblastoma, PNET-like variant
•  Choroid plexus carcinoma
•  Ewing’s sarcoma/PNET
•  Hemangiopericytoma
•  Malignant peripheral nerve sheath tumor (occasionally)
•  Melanoma (occasionally)

•  Giant cell glioblastoma
•  Pleomorphic xanthoastrocytoma
•  Subependymal giant cell astrocytoma
•  Melanoma
•  Choriocarcinoma
•  Giant cell ependymoma
•  “Ancient changes” in schwannoma, neurofibroma, or meningioma
•  Pilocytic astrocytoma (occasionally)
•  Ganglioglioma (occasionally)
•  Any poorly differentiated malignancy Extensive Calcification Desmoplasia or Sclerosis

•  Ganglioglioma
•  Central/extraventricular neurocytomas
•  Oligodendroglioma
•  Psammomatous meningioma
•  Meningioangiomatosis
•  Calcifying pseudoneoplasm of the neuroaxis
•  Craniopharyngioma
•  Tuber/focal cortical dysplasia
•  Vascular malformation
•  Subependymoma (occasionally)
•  Ependymoma (occasionally)
•  Astroblastoma (occasionally)
•  Choroid plexus papilloma (occasionally)

•  Astroblastoma
•  Desmoplastic/nodular medulloblastoma
•  Desmoplastic infantile ganglioglioma/astrocytoma
•  Meningioma, especially clear cell variant
•  Solitary fibrous tumor
•  Hemangiopericytoma or other sarcomas
•  Neurofibroma
•  Pleomorphic xanthoastrocytoma
•  Gliosarcoma
•  Ganglioglioma
•  Ependymoma (occasionally)
•  Abscess
•  Granulomas
•  Meningeal inflammation or neoplasm Hypervascular Inflammation-rich

•  Hemangioblastoma
•  Hemangiomas/vascular malformations
•  Vascular neoplasms (e.g., angiosarcoma)
•  Hemangiopericytoma
•  Angiomatous meningioma
•  Glioblastomas/high-grade gliomas
•  Pilocytic astrocytoma (occasionally)

•  Ganglioglioma
•  Pleomorphic xanthoastrocytoma
•  Gemistocytic astrocytoma
•  Germinoma
•  Chordoid glioma
•  Lymphoplasmacyte-rich meningioma
•  Inflammatory myofibroblastic tumor
•  Lymphomas and histiocytic disorders
•  Demyelinating diseases
•  Infections, granulomas, collagen vascular disorders Rosenthal Fibers/Eosinophilic Granular Bodies Discohesive

•  Pilocytic astrocytoma
•  Ganglioglioma
•  Pleomorphic xanthoastrocytoma
•  Dysembryoplastic neuroepithelial tumor (occasionally)
•  Piloid gliosis next to craniopharyngioma, hemangioblastoma, ependymoma, pineal cyst, or any slowly progressive process

•  Atypical teratoid/rhabdoid tumor
•  Papillary and/or rhabdoid meningiomas
•  Lymphomas/leukemias
•  Histiocytic disorders
•  Poorly differentiated malignancies






CNS, central nervous system; PNET, primitive neuroectodermal tumor.

Table 1-3
Helpful Features in Common Differential Diagnoses of Surgical Neuropathology
Atypical Gliosis vs. Diffuse Glioma (WHO Grade II)

• Evenly spaced astrocytes
• Abundant eosinophilic cytoplasm
• Radially oriented GFAP+ processes
• Other reactive changes, such as inflammatory infiltrates, macrophages, hemosiderin deposits, etc.

•  Clustered cells
•  “Naked nuclei”
•  Large, hyperchromatic, irregular nuclei
•  Ki-67 labels suspicious nuclei
•  Nuclei are strongly p53+ (not helpful if negative)
•  WT1+ processes
•  Demonstrable chromosomal alternations Diffuse Astrocytoma vs. Pilocytic Astrocytoma

•  MRI: Ill-defined, nonenhancing
•  Predominantly infiltrative
•  Clustered cells
•  “Naked nuclei”
•  Large, hyperchromatic, irregular nuclei
•  Nuclei are strongly p53+ (not helpful if negative)
•  Numerous intratumoral NFP+ axons

•  MRI: Demarcated, cystic, enhancing
•  Predominantly solid with focal invasion
•  Biphasic loose and compact areas
•  Long, thin, “hairlike” processes
•  Rosenthal fibers and EGBs
•  Multinucleate “pennies on a plate” cells
•  Hyalinized blood vessels
•  Strong, diffusely GFAP+
•  Few intratumoral NFP+ axons
•  Low Ki-67 labeling index, except in blood vessels Pleomorphic Xanthoastrocytoma vs. Giant Cell Glioblastoma

•  MRI: Demarcated, cystic, enhancing, often temporal lobe, minimal edema
•  Rare mitoses, despite pleomorphism
•  Spindled mesenchymal-like element
•  Foamy tumor cells (only in ~30%)
•  Eosinophilic granular bodies
•  CD34 + cells

•  MRI: Ring-enhancing, marked edema and mass effects
•  “Frankly anaplastic” cytology
•  Numerous mitoses
•  Atypical mitoses
•  Pseudopalisading necrosis
•  Extensively p53+ Diffuse Glioma/Glioblastoma vs. CNS Lymphoma

•  Secondary structures, such as perineuronal satellitosis
•  “Naked nuclei” or fibrillary processes
•  Eosinophilic cytoplasm
•  Nuclear hyperchromasia/pleomorphism
•  Microvascular proliferation
•  Pseudopalisading necrosis
•  GFAP or S-100 + or both

•  Angiocentric growth pattern
•  Discohesive on intraoperative smear
•  Open chromatin, large nucleoli
•  Rounded cells with scant blue cytoplasm
•  Prominent apoptosis
•  LCA or CD20 +
•  Intermixed reactive T lymphocytes High-grade Glioma or Lymphoma vs. Demyelinating Disease

•  Frankly anaplastic cells
•  Ill-defined lesional borders
•  Perineuronal satellitosis (gliomas)
•  Necrosis (gliomas or immunodeficiency-associated lymphomas)
•  Microvascular proliferation (gliomas)
•  GFAP+ or CD20 + atypical cells
•  EBV+ cells (immunodeficiency-associated lymphomas)

•  Relatively sharp demarcation
•  Fairly restricted to white matter
•  Myelin pallor with hypercellular infiltrate
•  Sheets of CD68 + histiocytes (better recognized on smear than frozen section)
•  Creutzfeldt cells
•  LFB-PAS shows marked myelin loss
•  NFP shows relative axonal preservation
•  CD20 + cells are all small and mature
•  GFAP+ cells are evenly spaced Glioblastoma vs. Abscess

•  Pseudopalisading necrosis
•  Microvascular proliferation
•  Infiltrative component
•  Secondary structure formation
•  GFAP+ atypical cells
•  Strongly p53+
•  WT1+ processes

•  Abundant neutrophils in necrotic foci
•  “Tissue culture” fibroblasts with variable nuclear atypia
•  Inflammatory rim
•  Brisk gliosis at edge of lesion
•  Trichrome reveals collagen deposition Glioblastoma vs. Metastasis

•  Infiltrative growth pattern
•  Secondary structure formation
•  Pseudopalisading necrosis
•  Microvascular proliferation
•  GFAP+
•  NFP+ intratumoral axons
•  Cytokeratin CAM 5.2- (Do not use cocktails such as AE1/AE3, which often cross-react with GFAP)

•  Sharp demarcation from brain
•  Glands or cytoplasmic mucin (adenoca)
•  Azzopardi effect (small cell ca)
•  Pigment (melanoma)
•  Hemorrhage (lung ca, melanoma, renal cell ca, choriocarcinoma)
•  CK7+, TTF1+ (lung ca)
•  CK20 + , CDX2+ (colon ca)
•  CK7+, mammaglobin+ (breast ca)
•  CD10 + (renal cell ca)
•  HMB45+, Melan-A+ (melanoma) Recurrence/Progression of Glioma vs. Radiation Necrosis/Radiation Effects

•  Pseudopalisading necrosis
•  Microvascular proliferation
•  Viable tumor with mitotic activity
•  High Ki-67 labeling index

•  Coagulative and fibrinoid parenchymal and vascular necrosis
•  Vascular hyalinization
•  Vascular telangiectasias
•  Rarefied hypocellular parenchyma
•  Dystrophic calcification
•  Radiation-induced atypia (bizarre bubbly nuclei and abundant pink cytoplasm) Anaplastic Oligodendroglioma vs. Small Cell Glioblastoma

•  Round uniform nuclei
•  Enlarged epithelioid cells with open chromatin and large nucleoli
•  Mucin-rich microcystic spaces
•  GFAP+ minigemistocytes and gliofibrillary oligodendrocytes
•  Chromosome 1p/19q codeletions

•  Oval uniform nuclei
•  Frequent mitoses despite “low-grade” cytology (delicate chromatin)
•  Pseudopalisading necrosis
•  GFAP+ thin cytoplasmic processes
•  EGFRvIII expression (~50%)
•  EGFR amplification (~70%)
•  Chromosome 10q deletions (>95%) Oligodendroglioma vs. Diffuse Astrocytoma

•  Round uniform nuclei with crisp nuclear membranes and small nucleoli
•  Clear haloes, no cytoplasm, or small eccentric belly of pink cytoplasm (mini- or microgemistocytes)
•  “Chicken wire” capillaries
•  Hypercellular nodules
•  Epithelioid/plasmacytoid cells with large nucleoli (anaplastic)
•  GFAP- or GFAP+ minigemistocytes and gliofibrillary oligodendrocytes
•  Mostly p53-
•  Chromosome 1p/19q codeletion

•  Variably elongate, irregular, hyperchromatic nuclei
•  “Naked nuclei,” elongate processes, or large eccentric belly of pink cytoplasm (gemistocytes)
•  Variably GFAP+ cytoplasm in most, although fibrillary and small cell variants may be negative due to minimal cytoplasm; high GFAP background makes interpretation difficult in others
•  Strongly p53+ (50%-60%) Oligodendroglioma vs. Dysembryoplastic Neuroepithelial Tumor

•  MRI: Cerebral nonenhancing tumor with significant mass effect
•  Extensive white matter component
•  Perineuronal satellitosis prominent
•  Chromosome 1p/19q codeletion

•  MRI: Gyriform, intracortical lesion, often mesial temporal lobe, with minimal mass effect; focal enhancement in a subset
•  Mucin-rich, patterned intracortical nodules
•  “Floating neurons”
•  Component resembling pilocytic astrocytoma in complex form
•  Adjacent cortical dysplasia
•  Rosenthal fibers/EGBs (occasionally) Oligodendroglioma vs. Central/Extraventricular Neurocytoma

•  Ill-defined margins
•  Perineuronal satellitosis
•  GFAP+ minigemistocytes and gliofibrillary oligodendrocytes
•  Entrapped NFP+ axons
•  Chromosome 1p/19q codeletion

•  Solid tumor with discrete borders
•  Neurocytic rosettes/neuropil formation
•  Diffusely synaptophysin+, including center of neurocytic rosettes
•  Neuronal features on EM Oligodendroglioma vs. Clear Cell Ependymoma

•  Ill-defined margins
•  Perineuronal satellitosis
•  GFAP+ minigemistocytes and gliofibrillary oligodendrocytes
•  Entrapped NFP+ axons
•  Chromosome 1p/19q codeletion

•  Sharp demarcation
•  Vague perivascular pseudorosettes, highlighted on GFAP stain
•  Nuclear grooves/folds
•  Dot-like cytoplasmic EMA+
•  NFP+ axons pushed to periphery of tumor
•  Ependymal features on EM Ependymoma vs. Diffuse Astrocytoma

•  Sharp demarcation
•  Perivascular pseudorosettes, highlighted on GFAP stain
•  Dot-like cytoplasmically EMA+
•  NFP+ axons pushed to periphery of tumor
•  Ependymal features on EM

•  Infiltrative growth pattern
•  Secondary structures
•  “Naked nuclei”
•  Numerous intratumoral NFP+ axons Cellular Ependymoma vs. Medulloblastoma/PNET

•  Solid growth pattern
•  Low mitotic/proliferative
•  Perivascular pseudorosettes with fibrillar processes, highlighted with GFAP
•  Dot-like cytoplasmically EMA+
•  NFP+ axons pushed to the periphery

•  Solid and infiltrative growth patterns
•  High mitotic/proliferative index
•  Homer Wright rosettes and occasional pseudorosettes with delicate neuropil
•  Synaptophysin positive
•  Ki-67 high Medulloblastoma/PNET vs. Atypical Teratoid/Rhabdoid Tumor

•  Mostly children/young adults
•  Solid and infiltrative growth patterns
•  Homer Wright rosettes
•  Retained INI1 expression
•  Synaptophysin+, GFAP focal+, most other markers negative
•  Genetics often shows i17q and/or MYC amplifications in anaplastic/large cell subtypes

•  Mostly infants (<3 years)
•  Mostly solid growth pattern
•  Variable PNET-like, carcinoma-like, and sarcoma-like foci
•  Rhabdoid cells present, but may be rare
•  Loss of INI1 expression
•  Polyphenotypic profile (vimentin, EMA, SMA, cytokeratin, etc.)
•  Genetics: INI1 gene mutations (may be germline), chromosome 22q losses Medulloblastoma/PNET vs. Glioblastoma

•  Mostly children/young adults
•  Solid and infiltrative growth patterns
•  Small blue cells with high mitotic and pyknotic indices
•  Homer Wright rosettes
•  Synaptophysin+
•  Genetics often shows i17q and/or MYC amplifications in anaplastic/large cell subtypes

•  Mostly adults
•  Infiltrative growth pattern
•  Secondary structure formation
•  Pseudopalisading necrosis
•  Microvascular proliferation
•  GFAP+
•  Genetics often shows EGFR amplification and/or chromosome 10q loss Meningioma vs. Schwannoma

•  Dural attachment
•  Whorls
•  Epithelioid cells
•  Psammoma bodies
•  Nuclear holes and pseudoinclusions
•  EMA+, S-100 patchy+

•  Encapsulated with parent nerve at periphery
•  Biphasic (Antoni A and B)
•  Verocay bodies
•  Wavy cells, wavy nuclei
•  Diffusely S-100+, collagen IV+ Meningioma vs. Hemangiopericytoma or Solitary Fibrous Tumor

•  Whorls
•  Storiform pattern
•  Epithelioid cells
•  Psammoma bodies
•  Nuclear holes and pseudoinclusions
•  EMA+ (~80%)
•  PR+ (more so in benign forms)

•  “Staghorn” vasculature
•  Marked hypercellularity with scattered pale islands (HPC)
•  Lacelike intercellular collagen (SFT)
•  Reticulin-rich
•  CD99 + , BCL2+
•  Diffuse CD34 + (SFT) Hemangioblastoma vs. Metastatic Renal Cell Carcinoma

•  Foamy clear cells
•  Delicate chromatin or degenerative nuclear atypia
•  Adjacent piloid gliosis
•  Low mitotic/proliferative indices
•  Inhibin+, D2-40+, S-100+, NSE+
•  PR+
•  Focally GFAP+ (occasionally)

•  Solid sheets of uniformly clear cells
•  Epithelioid cells with pink cytoplasm
•  Vesicular nuclei with prominent nucleoli
•  High mitotic/proliferative indices
•  Cytokeratin+, EMA+, CD10 +
•  RCC+ in a subset Hemangioblastoma vs. Angiomatous Meningioma

•  Parenchymal ± leptomeningeal (almost exclusively infratentorial)
•  Inhibin+, NSE+
•  Extramedullary hematopoiesis (~10%)

•  Dural-based mass
•  Foci of other meningioma subtypes
•  Psammoma bodies
•  EMA+




AE1/AE3, cytokeratins; BCL2, B-cell lymphoma 2 (gene); CAM, cell adhesion molecule; CD, cluster of differentiation; CK7, cytokeratin 7; EBV, Epstein–Barr virus; EGB, eosinophilic granular body; EGFR, epidermal growth factor receptor; EM, electron microscopy; EMA, epithelial membrane antigen; GFAP, glial fibrillary acidic protein; HPC, hemangiopericytoma; LCA, leukocyte common antigen; LFB-PAS, Luxol fast blue/periodic acid–Schiff; MRI, magnetic resonance imaging; NFP, neurofilament protein; NSE, neuron-specific enolase; PNET, primitive neuroectodermal tumor; PR, progesterone receptor; RCC, renal cell carcinoma; SFT, solitary fibrous tumor; SMA, smooth muscle antibody; TTF, thyroid transcription factor; WT1, Wilms tumor 1.

Electron Microscopy
Although electron microscopy (EM) has historically been vital in defining a number of diagnostic entities, its everyday use in surgical neuropathology is generally labor-intensive, time-consuming, and expensive, with interpretation typically being delayed by one to several weeks. For these reasons, EM has largely been supplanted by immunohistochemistry (IHC) in most medical centers. Nevertheless, EM remains extremely valuable in specific scenarios; for instance, it is still the gold standard for proving ependymal differentiation in diagnostically challenging examples. It can also be extremely useful in terms of classifying clinicopathologically relevant pituitary tumors (see Chapter 18 ). Ultrastructural pathology primarily provides insight into various forms of cellular differentiation by visualizing organelles, other cytoplasmic constituents, and cell membrane structures (intermediate filaments, neurosecretory granules, synapses, pinocytotic vesicles, intercellular junctions, cilia, microvilli, basement membrane, etc.). Further examples of EM use in surgical neuropathology are provided in subsequent disease-specific chapters.

Immunohistochemistry
IHC is a well established and frequently utilized ancillary diagnostic aid for complex surgical neuropathology cases, especially in the area of tumor neuropathology. Some of the most frequently utilized antibodies and expected immunoprofiles for common tumor types are summarized in Table 1-4 , although this list is by no means exhaustive and many exceptions to the rules may be seen in individual cases.

Table 1-4
Typical Tumor Immunoprofiles
Tumor + ± – Astrocytoma S-100, GFAP CK 1 , LCA, SYN, HMB-45 Oligodendroglioma S-100, GFAP 2 CK 1 , LCA, SYN Ependymoma S-100, GFAP, VIM, EMA (dot-like), CD99 CK LCA, SYN Choroid plexus tumors S-100, CK, VIM, transthyretin 3 GFAP EMA, CEA Metastatic carcinoma 12 EMA and CK, CK7 and TTF1 (lung), CK20 and CDX2 (colon), CK7 and mammaglobin (breast), CD10 (renal) CEA, S-100, SYN GFAP, LCA, HMB-45 Melanoma and melanocytoma S-100, HMB-45, Melan-A (MART-1) GFAP, CK, LCA Lymphoma LCA, CD20 (L26), CD79a EMA 4 , CD30 and ALK (+ in ALCL) CK, GFAP, HMB-45, SYN, CD3 Meningioma EMA, VIM, PR S-100, CD34, CK 5 GFAP, HMB-45 Hemangiopericytoma VIM, CD99, BCL-2, factor XIIIa 6 CD34 EMA, CK, GFAP, S-100 Solitary fibrous tumor CD34 (diffuse), VIM, CD99, BCL-2 EMA, CK, GFAP, S-100 Medulloblastoma SYN S-100, GFAP, NFP, NeuN CK, LCA, EMA Atypical teratoid rhabdoid tumor VIM, EMA, CK, actin SYN, GFAP, desmin, AFP INI1 11 , PLAP, β-hCG, LCA Ganglioglioma SYN, NFP, CG, GFAP, CD34 Neu-N CK, EMA, PLAP Central neurocytoma SYN, NeuN GFAP, S-100 NFP, CG, CK, LCA Schwannoma S-100 7 , Col IV 7 GFAP, calretinin, HMB-45 8 , CD34 EMA, NFP, CK Neurofibroma S-100/Col IV (patchy), CD34, NFP (entrapped axons) GFAP EMA, CK, calretinin Perineurioma EMA, Col IV NFP (axons in intraneural type) S-100, CK MPNST VIM, CD99, Col IV S-100, Leu-7, CK Actin, desmin, myogenin Paraganglioma SYN, CG, S-100 9 NFP, CK GFAP, HMB-45 Hemangioblastoma S-100, NSE, inhibin, PR GFAP CK, EMA Pituitary adenoma SYN, CG (~60%), hormones (PRL, GH, ACTH, FSH, LH, TSH) CK, Ki-67 >3% in atypical adenomas, p53+ in atypical adenomas S-100, GFAP, CK7, CK20, TTF1, CD99, germ cell tumor markers Germinoma PLAP, c-kit (CD117), OCT 3/4 β-hCG 10 , CK AFP, EMA, HMB-45, LCA Yolk sac tumor AFP, CK PLAP, EMA β-hCG, GFAP Choriocarcinoma β-hCG, CK, EMA PLAP AFP, GFAP, HMB-45, LCA Embryonal carcinoma CK, PLAP, CD30, OCT 3/4 β-hCG, AFP, EMA, LCA, HMB-45 Teratoma CK, PLAP, EMA AFP β-hCG


ACTH, adrenocorticotrophic hormone; AFP; α-fetoprotein; CEA, carcinoembryonic antigen; CG, chromogranin; CK, cytokeratin; Col IV, collagen type IV; EMA, epithelial membrane antigen; FSH, follicle-stimulating hormone; GFAP, glial fibrillary acidic protein; GH, growth hormone; β-hCG, β-human chorionic gonadotrophin; LCA, leukocyte common antigen; LH, luteinizing hormone; MART, melanoma antigen recognized by T cells; NeuN, neuronal nuclei (neuronal-specific protein); NFP, neurofilament protein; NSE, neuron-specific enolase; OCT 3/4, octomers 3 and 4; PLAP, placental alkaline phosphatase; PR, progesterone receptor; PRL, prolactin; SYN, synaptophysin; TSH, thyroid-stimulating hormone; TTF1, thyroid transcription factor 1; VIM, vimentin.
1 CAM 5.2 recommended, since AE1/AE3 frequently stains gliomas.
2 Strongly positive in microgemistocytes and gliofibrillary oligodendrocytes.
3 Not specific for choroid plexus.
4 Positive in myeloma and anaplastic large cell lymphomas (ALCL).
5 Positive in secretory variant.
6 Characteristic pattern of scattered, individual immunoreactive cells.
7 Diffuse, strong expression.
8 Positive in melanocytic variant.
9 Positive in sustentacular cells.
10 Positive in syncitiotrophoblasts.
11 INI1 is normally ubiquitously expressed, such that loss of nuclear expression is relatively specific for the diagnosis of AT/RT.
12 Metastatic prostate carcinomas are not included due to the rarity of CNS metastases, although they can involve the axial skeleton.

Glial Markers
As the name implies, expression of the intermediate filament protein, glial fibrillary acidic protein (GFAP) is fairly specific for tumors of glial lineage. Strong expression in astrocytic neoplasms is best recognized, but is also characteristic of ependymomas, pituicytomas, and oligodendroglial neoplasms containing minigemistocytes and gliofibrillary oligodendrocytes. Focal glial differentiation with GFAP expression may also be encountered in choroid plexus tumors, medulloblastomas/ central nervous system (CNS) primitive neuroectodermal tumors (PNETs), atypical teratoid/rhabdoid tumors, gangliogliomas, neurocytic tumors, and teratomas. Additionally, GFAP immunoreactivity may be encountered in a few nonglial neoplasms, such as Schwann cell (peripheral nerve sheath), myoepithelial, and cartilaginous tumors. On the opposite end, some astrocytomas have scant cytoplasm and intermediate filament synthesis, surprising the pathologist with minimal to no demonstrable GFAP expression. S-100 protein positivity is a useful and highly sensitive all-around “glial marker” in such cases, but is limited by the fact that it is considerably less specific. It is common to neuroectodermal cells in general, including melanocytes, glia, Schwann cells, chondrocytes, and the sustentacular cells of paraganglioma, pheochromocytoma, and olfactory neuroblastoma. To a lesser extent, it often stains neuronal tumors and fibrous meningiomas as well. Unfortunately, specific markers that reliably distinguish one glioma subtype from another are still lacking, although epithelial membrane antigen (EMA) has recently emerged as a useful marker of ependymal tumors, in that it displays a characteristic cytoplasmic dotlike pattern of reactivity. 6, 7

Neuronal Markers
Synaptophysin represents a component of presynaptic vesicle membranes and is one of the most commonly used neuronal cell markers in surgical neuropathology. It is a relatively sensitive marker of neuronal (and neuroendocrine) differentiation and is typically found even in the most primitive neuronal tumors, such as medulloblastoma and PNET. However, the characteristic staining of neuropil often makes it difficult to interpret, since it may be unclear whether this neuropil is part of the tumor or merely entrapped non-neoplastic tissue. Additionally, the specificity is not as high as one would like, increasingly being reported in a wide variety of glial neoplasms and other tumor types. Therefore, the use of multiple neuronal markers is advocated, and one should be cautious in diagnosing a tumor as “glioneuronal” based on the presence of synaptophysin alone. Chromogranin is similarly useful for highlighting neoplastic ganglion cells, as well as neuroendocrine tumors, such as metastastic small cell carcinomas, pituitary adenomas, carcinoids, and paragangliomas. This marker is considerably more specific than synaptophysin, but suffers from relatively low sensitivity.
NeuN is generally considered a marker of mature neuronal differentiation (e.g., cortical ganglion cells and neurocytes) and has the advantage of clearly staining tumor nuclei (and sometimes cytoplasm) rather than the neuropil. Surprisingly however, most neoplastic ganglion cells within gangliogliomas are negative for this marker. Nevertheless, it can still be useful in distinguishing immunonegative tumoral ganglion cells from entrapped cortical neurons, which are virtually always strongly reactive. Purkinje cells are also typically negative. Overall, NeuN is a relatively reliable marker of neuronal differentiation and despite being considered a marker of mature cells, is often found at least focally in primitive neuronal tumors as well, including medulloblastomas and CNS PNETs. Neurofilaments are intermediate filaments composed of three subunits that are relatively specific to neuronal and neuroendocrine cells, normally being expressed mostly in their axonal processes. There are multiple neurofilament protein (NFP) subtypes and variable states of phosphorylation; as such, individual antibodies differ in terms of their expression profiles. In general, however, axons stain readily for NFP in the normal brain, whereas neuronal cell bodies are negative. In gangliogliomas, NFP variably stains neuronal cell bodies, whereas primitive neuronal tumors such as medulloblastoma range from patchy cytoplasmic expression to being completely negative. Additionally, the staining of axons can be extremely useful for highlighting a tumor’s overall growth pattern. For example, discrete tumors such as metastases and ependymomas push axon-bearing parenchyma to the side, whereas diffuse gliomas will entrap NFP-positive axons even within more central portions of the tumor.

Epithelial Markers
Cytokeratins are a class of intermediate filaments used to demonstrate epithelial differentiation. Antibodies against cytokeratin are most commonly used in the diagnosis of metastatic carcinomas, but can also be used to identify primary epithelial or epithelium-containing CNS neoplasms, such as craniopharyngiomas, chordomas, teratomas, epithelial cysts, and choroid plexus tumors. Various cytokeratins can be used for subtyping various forms of epithelial differentiation or for narrowing possible sites of origin for metastatic carcinomas. For example, widespread CK7 positivity is most often seen in respiratory (including Rathke’s cleft, colloid, and neuroenteric cysts) or breast type epithelium, but is also common in choroid plexus tumors. Extensive CK20 expression is most often encountered in gastrointestinal-type epithelium, especially colorectal carcinomas. Other useful markers in metastatic carcinomas of unknown origin include TTF-1 (lung), mammaglobin (breast), GCDFP-15 (breast), CD10 (renal cell), and CDX2 (colorectal, gastric) (see Chapter 13 for detailed information). 8
Epithelial membrane antigen (EMA) is also a common constituent of normal and neoplastic epithelial cells, but is considerably less specific. In the CNS, EMA is a useful marker for meningiomas, which unlike true epithelial tumors, usually display minimal to no cytokeratin expression. The exception is secretory meningiomas. Unfortunately, most meningiomas only show patchy weak expression, so some laboratories have employed higher antibody concentrations or enhanced antigen retrieval for meningiomas as compared with carcinomas, which typically express EMA more robustly. As mentioned under glial markers, EMA is also useful for the diagnosis of ependymomas when producing a dot-like intracytoplasmic pattern.

Proliferation Markers
Most proliferation antigens are nuclear proteins that are actively expressed during one or more non-G 0 phases of the cell cycle. These markers are commonly used as ancillary aids to simple mitotic counts. For instance, the murine monoclonal antibody Ki-67 binds to a human nuclear protein expressed during the G 1 , S, G 2 , and M phases of the cell cycle. The MIB-1 antibody against Ki-67 works in paraffin sections and is a popular ancillary marker for quantifying proliferation in brain tumors. 9 Nonetheless, the WHO has resisted assigning any specific labeling index cutoffs in the grading of individual tumor types because there is too much interlaboratory variability. In other words, wide-ranging differences in staining results and counting methods make it difficult to extrapolate results from one medical center to another. Keeping in mind that each tumor type is different, a useful, though grossly oversimplified approach is to consider low, moderate, and high proliferative indices as less than 5%, 5% to 10%, and greater than 10%, respectively.
Another useful approach recently applied to both gliomas and meningiomas is phosphohistone-H3 (PPH3) immunohistochemistry, wherein the antibody specifically labels only mitotic figures. 9 This is an attractive alternative, since many grading schemes utilize mitotic index cutoffs. As all practicing pathologists will quickly concede, counting mitotic figures in hematoxylin and eosin (H&E) sections can be extremely cumbersome and time-consuming, particularly if the specimen is large or tissue preservation is poor, such that it is difficult to distinguish degenerating cells from mitoses. With PPH3 immunohistochemistry, mitotic hotspots are quickly and reliably identified. Unfortunately though, traditional mitotic cutoffs would have to be reestablished for each tumor type, since the sensitivity for picking up mitoses and therefore the mitotic indices themselves is considerably higher using the immunohistochemical approach. An additional pitfall is that archived specimens may lose some of their antigenicity over time using either the MIB-1 or PPH3 antibodies.

Molecular Diagnostics
Tumorigenesis is thought to typically occur through a series of multiple genetic mutations or “hits,” with some tumor-associated alterations already influencing our diagnostic approach. 10 Molecular diagnostics involves the detection of clinically useful genetic changes at the DNA, messenger RNA (mRNA), or protein levels. For instance, in situ hybridization (ISH) can be used to detect tumor-related mRNA expression. In neuropathology, diagnostic elevations of hormone-specific mRNAs are detectable in pituitary adenomas, whereas EBER ISH is used to identify Ebstein-Barr virus involvement in primary lymphomas of immunosuppressed patients. Many other tumor-associated transcripts are potentially testable, although ISH is not currently a frequently used technique in routine diagnostics, since larger fragments of RNA are often rapidly degraded in paraffin-embedded tissue, making this technique challenging.
Conventional cytogenetics (karyotyping) can be a very useful screening tool for detecting specific tumor-associated alterations, but they often need to be very large chromosomal events, since the resolution of this assay is relatively low. Also, low-grade tumors often do not proliferate sufficiently to yield diagnostic metaphases. Finally, the vast majority of tumors, especially in nonacademic settings, are submitted entirely for formalin fixation and paraffin embedding, limiting the types of genetic studies that can be applied.
As an alternative to karyotyping, DNA probes are relatively well suited to the study of chromosomal aberrations in paraffin-embedded tissue. Some tumors are associated with “signature” cytogenetic abnormalities, such as translocations, gene amplifications, and deletions. The most common and practical approaches for detecting DNA alterations include loss of heterozygosity ( LOH ), fluorescence in situ hybridization ( FISH ), and quantitative polymerase chain reaction ( PCR ) techniques. For detection of oncogene amplification, FISH and quantitative PCR are most practical. Thus far, no specific fusion transcripts have been associated with primary CNS neoplasms, with the exception of rare hematopoietic and soft tissue tumors that are first detected around the spine or cranium. However, at least one example of a nonbalanced translocation was recently described in oligodendrogliomas. 11, 12
Few molecular diagnostic tests have become routinely employed or “standard of care” in neuropathology. The most notable is the use of chromosome 1p and 19q testing as a prognostic/management tool for adult patients with oligodendroglial tumors. 10 As a technique, FISH has the advantage of simplicity, morphologic preservation, minimal tissue and purity requirements, and no need for microdissection or matching blood/non-neoplastic tissue. However, accurate interpretation requires considerable experience, especially in cases with aneuploid populations of tumor cells. It also uses large probes (100–300Kb) and is therefore incapable of detecting very small deletions. LOH has the advantage of simplicity and the probes are smaller, which makes them capable of identifying losses even in the presence of mitotic recombination (e.g., loss of wild type allele and duplication of mutant allele). However, normal patient DNA is required for comparison and false negatives are a problem if the tumor sample is not relatively pure. Other clinical applications of FISH utilized in selected cases include EGFR amplification or 10q deletions to distinguish small cell glioblastoma multiforme (GBM) from anaplastic oligodendroglioma, 22q deletion to distinguish atypical teratoid rhabdoid tumor (AT/RT) from variants of medulloblastoma, i17q in medulloblastomas, N- myc or c- myc amplifications in large cell/anaplastic medulloblastomas, and meningioma-associated deletions (22q, 1p, 14q) to distinguish anaplastic meningiomas from other malignancies or benign meningiomas from foci of meningothelial hyperplasia. These are each discussed in greater detail in subsequent chapters. It is also likely that many more applications of molecular diagnostics will become incorporated into diagnostic neuropathology labs in the future.

References

1. Bailey, P., Cushing, H.A Classification of Tumors of the Glioma Group on a Histogenetic Basis with a Correlation Study of Prognosis. Philadelphia: Lippincott, 1926.
2. Canoll, P., Goldman, J. E. The interface between glial progenitors and gliomas. Acta Neuropathol. . 2008; 116:465–477.
3. Eberhart, C. G. In search of the medulloblast: neural stem cells and embryonal brain tumors. Neurosurg Clin North Am. . 2007; 18:59–69. [viii–ix].
4. Kernohan, J. W., Mabon, R. F., . A simplified classification of the gliomas. Mayo Clin Proc. . 1949; 24:71–75.
5. Louis, D. N., Ohgaki, H., Wiestler, O. D., Cavenee, W. K.WHO Classification of Tumours of the Central Nervous System. Lyon: IARC, 2007.
6. Hasselblatt, M., Paulus, W. Sensitivity and specificity of epithelial membrane antigen staining patterns in ependymomas. Acta Neuropathol (Berlin) . 2003; 106:385–388.
7. Kawano, N., Yasui, Y., Utsuki, S., . Light microscopic demonstration of the microlumen of ependymoma: a study of the usefulness of antigen retrieval for epithelial membrane antigen (EMA) immunostaining. Brain Tumor Pathol. . 2004; 21:17–21.
8. Park, S. Y., Kim, B. H., Kim, J. H., . Panels of immunohistochemical markers help determine primary sites of metastatic adenocarcinoma. Arch Pathol Lab Med. . 2007; 131:1561–1567.
9. Takei, H., Bhattacharjee, M. B., Rivera, A., . New immunohistochemical markers in the evaluation of central nervous system tumors: a review of 7 selected adult and pediatric brain tumors. Arch Pathol Lab Med. . 2007; 131:234–241.
10. Fuller, C. E., Perry, A. Molecular diagnostics in central nervous system tumors. Adv Anat Pathol. . 2005; 12:180–194.
11. Griffin, C. A., Burger, P., Morsberger, L., . Identification of der(1;19)(q10;p10) in five oligodendrogliomas suggests mechanism of concurrent 1p and 19q loss. J Neuropathol Exp Neurol. . 2006; 65:988–994.
12. Jenkins, R. B., Blair, H., Ballman, K. V., . A t(1;19)(q10;p10) mediates the combined deletions of 1p and 19q and predicts a better prognosis of patients with oligodendroglioma. Cancer Res. . 2006; 66:9852–9861.
2
Normal Brain Histopathology
Daniel J. Brat

Cell Types   15
Neurons  15
Glia  17
Astrocytes  17
Oligodendrocytes  20
Ependyma  20
Choroid Plexus  20
Microglia  20
Blood Vessels  21
Meningothelial Cells  21
Melanocytes  22
Tissue Organization   22
Cerebral Cortex  22
White Matter  23
Basal Ganglia  23
Thalamus  24
Hippocampus  24
Pineal Gland  25
Pituitary Gland  26
Cerebellum  26
Brainstem  28
Midbrain  28
Pons  28
Medulla  29
Spinal Cord  29
Meninges  29
Peripheral Nerve, Schwann Cells, and Dorsal Root Ganglia  29
Features of Infancy and Childhood   31
Features of the Aging Nervous System   31
The practice of surgical neuropathology can be challenging for the generalist and specialist alike. Much of this difficulty results from the intrinsic complexity of the human central nervous system (CNS), an organ that is unrivaled in regional variation and specialized organization. Nevertheless, a basic understanding of the normal cellular and tissue organization of the brain is absolutely necessary for the practice of surgical neuropathology, since recognition of the abnormal rests on a firm knowledge of the normal. Even in the current age of advanced neuroimaging and image-guided biopsies, many neurosurgically sampled specimens contain normal or only slightly abnormal tissue that needs to be recognized as nondiagnostic. Indeed, much of the anxiety that arises at the time of frozen section seems to be due to a discomfort with distinguishing normal from abnormal rather than from correctly categorizing an abnormal biopsy specimen. The “sea of pink” noted under the microscope from a brain biopsy of normal brain tissue has been known to cause even the most experienced surgical pathologist a great degree of uncertainty. The changes of “reactive gliosis” only exacerbate the problem. Much like the pattern recognition approach used in this textbook for classifying diseases of the CNS, so too can the normal histology of the brain be approached based on recognition of tissue patterns . Like other forms of pattern recognition, this takes practice. For the neuropathologist, repeated exposure to normal CNS structure occurs from the regular review of autopsied brains. For generalists that practice surgical neuropathology, review of autopsied brain sections can add confidence in the recognition of normal CNS tissue patterns. This chapter introduces the cell types and normal histology of the human CNS at a depth necessary for routine diagnostic practice. Most of its content describes the normal adult brain, but certain aspects of age-related phenomena and developmental features that are routinely encountered in diagnostic neuropathology are also considered.

Cell Types
Given the high degree of functional complexity, it may be surprising that the brain parenchyma consists predominantly of only two cell types, neurons and glia. Both are large families with many members that have highly specialized functions, yet the underlying structure and cell biology of each retain some central features. Most challenging for the practicing surgical pathologist is the great degree of morphologic and geographic diversity of normal, reactive, and degenerative states of these two cell families.

Neurons
Neurons are the integrating and transmitting cells of the nervous system, communicating by chemical and electrical means. The spectrum of their morphology, connectivity, and function is enormous. As a rule, neurons have a cell body, branching processes called dendrites for integrating incoming signals, and a longer cell process—the axon—with a terminal synapse for chemically transmitting an electric signal over a short distance from one neuron to the next (or to a muscle cell through a neuromuscular junction). Cell body shape and size, as well as the number and arrangement of branching processes, vary considerably. For practicing pathologists, recognizing the major forms of neurons within their anatomic setting is crucial, since individual populations show differential vulnerability to injury and variable pathologic reactions in specific disease processes.
The pyramidal neurons of the cerebral cortex and subfields of the hippocampus represent a morphologic prototype ( Fig. 2-1A and B ). They are characterized by large, triangular cell bodies, a prominent apical dendrite extending toward the brain’s surface, and numerous finer branching basal dendrites. Measuring approximately 10 to 50 μm in greatest dimension, their cell bodies contain abundant cytoplasm, variable hematoxiphilic Nissl substance (rough endoplasmic reticulum) near the entry zone to processes, and a large nucleus with open chromatin and a prominent nucleolus (open chromatin and prominent nucleoli are typical of neurons and distinguish them from resting glia).
Figure 2-1 Neurons. A, Pyramidal neurons ( arrow ) of the cerebral cortex. B, Pyramidal neurons of the hippocampus. C, Betz cells (upper motor neurons) of the motor cortex ( arrow ). D, Granular neurons of the dentate fascia of the hippocampal formation. E, Purkinje cells (arrow) and granular cells (arrowhead) of the cerebellar cortex. F, Anterior horn cells (lower motor neurons) of the spinal cord. G, Dopaminergic neurons of the substantia nigra are deeply pigmented due to accumulation of neuromelanin.
Cortical granular (stellate) neurons are the smaller counterparts of pyramidal neurons in the cortex, typically measuring 15 μm or less in diameter. Being interneurons, they have numerous shorter processes that remain within the confines of the cortex.
Betz cells are the largest neurons of the cerebral cortex (70 to 100 μm) and are found in the primary motor cortex where they dwarf their neighboring cortical pyramidal cells ( Fig. 2-1C ). The amounts of cytoplasm and Nissl substance and the number of visible processes far exceed normal pyramidal cells. Betz cells are upper motor neurons.
Small, tightly packed granular neurons form the stratum granulosum of the dentate gyrus in the medial temporal lobe, intimately connected to the hippocampus proper ( Fig. 2-1D ). These neurons are nearly as small as cerebellar granular cells and have an extensive dendritic arbor that forms the adjacent molecular layer of the dentate gyrus.
Purkinje cells are large (50–80 μm), histologically distinctive neurons of the cerebellum with cell bodies that sit at the interface of the molecular and internal granular cell layers ( Fig. 2-1E ). Each neuron has a prominent pink cell body and an expansive dendritic tree with thick processes that extend into the molecular layer, as well as a large axon that travels centrally out of the cerebellar cortex.
Granular neurons of the cerebellar granular cell layer are tiny and densely packed, often displaying a linear arrangement or loose rosettes around delicate neuropil (see Fig. 2-1E ). Perinuclear cytoplasm is sparse, giving the appearance of only nuclei on hematoxylin and eosin (H&E) stains. This population can cause confusion on frozen section or cytologic preparations because they resemble “small round blue-cell tumors.”
Anterior horn cells are large, lower motor neurons (alpha motor neurons) that populate all levels of the spinal cord in the anterior horns and send long axonal processes via the anterior roots for their eventual termination on peripheral skeletal muscle endplates ( Fig. 2-1F ).
The CNS also contains a small number of highly specialized nuclei that contain neurons that produce specific bioaminergic neurotransmitters and project diffusely throughout the brain to affect global or regional tone. Rarely seen in biopsied material, these include the substantia nigra, locus ceruleus, raphe nuclei, and nucleus basalis of Meynert. The dopaminergic cells of the substantia nigra pars compacta (and the ventral tegmental area) are large, heavily pigmented neurons with “neuromelanin” (not to be confused with melanin of melanocytes), which accumulates in the cytoplasm as coarse brown granules and represents a combination of oxidized and polymerized dopamine within lysosomal granules ( Fig. 2-1G ; selectively vulnerable in Parkinson disease). Similarly, the locus ceruleus, located near the fourth ventricle in the rostral pontine tegmentum, contains a population of large, pigmented neurons that serve as a major source of norepinephrine in the brain (selectively vulnerable in Parkinson disease). Neurons located in the raphe nuclei, located along the midline of the brainstem, are similar in size and shape to the noradrenergic neurons of the locus ceruleus, but lack the pigmentation. These cells produce serotonin and have diffuse projections throughout the nervous system, but most heavily innervate limbic and sensory regions. Within the basal forebrain, inferior to the anterior commisure in a region called the substantia innominata, is the nucleus basalis of Meynert, a collection of large cholinergic neurons that project throughout the cerebral cortex (selectively vulnerable in Alzheimer disease).

Glia
Glia account for approximately 90% of all CNS cells and have been generally regarded as “glue,” providing structural and functional support for neuronal elements. In fact, they are functionally much more diverse and biologically important than originally suspected, such that neurobiologists have shifted away from this overly “neuronocentric” perspective. Glia are divided into the macroglia or true glia—astrocytes, oligodendrocytes, and ependyma—and the microglia, which are actually of hematopoietic rather than true glial derivation.

Astrocytes
Astrocytes are the multipolar, “starlike” glial cells of the CNS ( Figs. 2-2 and 2-3 ). They can be subdivided into protoplasmic and fibrillary families based on their location and morphology. Protoplasmic astrocytes reside in the cortex, whereas fibrillary astrocytes populate the white matter. In addition to similar cell shapes and numerous processes, all astrocytes contain abundant cytoplasmic intermediate filaments, largely composed of glial fibrillary acidic protein (GFAP). In the resting state, astrocyte nuclei are recognized on H&E-stained sections, but the scant, delicate cytoplasm and processes are not readily seen since they blend with surrounding neuropil. Nuclei are oblong with a chromatin pattern that is lighter and looser than either oligodendrocytes or neoplastic astrocytes ( Fig. 2-2A ). Nucleoli are not present in most resting astrocytes, in contrast to neurons. Many astrocytes have processes that terminate as end-feet on blood vessel walls, where they contribute to the blood–brain barrier . Others have processes that extend end-feet to the pial surface of the brain, contributing to the glia limitans of the brain–cerebrospinal fluid (CSF) barrier .
Figure 2-2 Normal glia. A, Normal white matter shows oligodendrocytes (arrow) , which have round dark nuclei often with a slight perinuclear halo, and astrocytes, with oblong nuclei (arrowheads) . Glial cytoplasm blends with the neuropil and cannot typically be noted in the resting state. B, Cytologic preparation of normal cortex demonstrates normal oligodendrocytes (short arrow) , astrocytes (arrowhead) , neuron (long arrow) , and capillaries (asterisk) .
Figure 2-3 Reactive glia. A, B, Gemistocytic astrocytes (arrows) are one form of reactive change in which the astrocytic cytoplasm is distended and esosinophilic processes are readily identified. C, Immunohistochemistry for GFAP highlights reactive astrocytes and emphasizes their “starlike” quality. D, E, Piloid gliosis is a highly fibrillar form of reactive gliosis that is composed of dense, elongate astrocytic processes that are tightly packed together and are often associated with numerous Rosenthal fibers. In this case, piloid gliosis forms the wall of a pineal cyst. F, Alzheimer type II astrocytes (arrow) have enlarged, clear nuclei and are seen in states of hyperammonemia. G, Bergmann gliosis occurs at the interface of the molecular and granular layers of the cerebellum, generally in response to Purkinje cell injury. In this case, a nearly normal complement of Purkinje cells (arrow) is seen on the right, whereas Purkinje cells have been replaced by one to three layers of Bergmann glia containing oval nuclei with long coarse cytoplasmic processes radiating toward the pial surface (left side) . The patient was an infant as evidenced by the thin remnant of the external granular layer. H, Creutzfeldt cells (granular mitoses) are reactive cells with fragmented nuclear material (arrow) that can be mistaken for mitotic figures.
Astrocytes are activated in response to a variety of pathologic conditions ( Fig. 2-3 ). The morphologic spectrum of reactive astrocytosis is critical to recognize because (1) it focuses attention on pathologically affected regions for further evaluation; (2) it validates the assumption that a disease process is present in the CNS (i.e., rather than artifact); and (3) reactive astrocytosis often causes diagnostic dilemmas due to its morphologic similarity to neoplastic conditions. Reactive astrocytosis involves both proliferation and hypertrophy of astrocytes, and its appearance varies with the chronicity and severity of the insult. The initial response is enlargement of cell body, processes, nuclei, and nucleoli. In H&E-stained sections, the presence of visible astrocytic cytoplasm and processes is almost always a pathologic finding ( Fig. 2-3B ). Immunohistochemistry for GFAP highlights the reactive nature of these astrocytes, demonstrating the extensive arborizing of their processes and the orientation of the reactive cells to the underlying injury ( Fig. 2-3C ). Reactive astrocytes of longer duration often take on a gemistocytic (from the Greek word gemistos , meaning “stuffed”) appearance, with large amounts of brightly eosinophilic cytoplasm in their eccentrically placed cell bodies ( Fig. 2-3A ). The relatively even spacing of these astrocytes and radially arranged processes help to distinguish them as reactive, rather than neoplastic.
Chronic reactive astrocytosis that occurs around a slowly growing lesion is often more fibrillar in nature, with numerous long astrocytic processes forming a layer of dense gliosis adjacent to injury. Rosenthal fibers are large, flame-shaped or globular proteinaceous deposits that may be seen in this type of long-standing process; when present, this form of reaction is often termed piloid gliosis , due to its morphologic overlap with the compact regions of pilocytic astrocytoma ( Fig. 2-3D and E ). It is most often encountered adjacent to slow-growing neoplasms (e.g., craniopharyngioma, ependymoma, hemangioblastoma) and benign cystic lesions (e.g., pineal cyst, spinal syrinx).
Alzheimer type II astrocytes are a reactive form seen in states of elevated blood ammonia, usually related to renal or hepatic disease ( Fig. 2-3F ). They are present in highest concentration in the basal ganglia where cells show nuclear swelling, marked chromatin clearing, and micronucleoli. Cytoplasmic hypertrophy is not prominent in this form of astrocytosis, and usually, the nuclei have no appreciable cytoplasm.
Bergmann glia are specialized astrocytes located between the molecular and granular layers of the cerebellum. Cells are only one to two layers thick and can go unnoticed in resting states. In response to cerebellar injury, especially to individual Purkinje cell loss from ischemia or hypoxia, the reactive proliferation of this cell layer is referred to as Bergmann gliosis. On H&E sections, the Purkinje cells are replaced with one to three layers of oval nuclei associated with coarse GFAP-positive fibrillary processes radiating toward the pia ( Fig. 2-3G ).
Creutzfeldt cells are another form of reactive astrocytes that have abundant cytoplasm and “granular mitoses,” the fragmenting of nuclear material that gives the impression of multiple micronuclei ( Fig. 2-3H ). They are not specific, but are seen most often in active inflammatory diseases (classic in demyelinating disease). It is important not to mistake them for the mitoses of an infiltrating astrocytoma.
Markedly enlarged and cytologically atypical astrocytes can be seen in many non-neoplastic conditions, although the nuclear atypia is often most pronounced in reactions to radiation ( radiation atypia ; see Chapter 20 ) and progressive multifocal leukoencephalopathy (PML; see Chapter 22 ). In both conditions, nuclei can be bizarre with marked hyperchromasia, multilobation, and irregular outlines. The context of additional microscopic changes and the clinical history are often required to avoid a misdiagnosis of neoplasm.

Oligodendrocytes
Oligodendrocytes are the myelinating cells of the CNS and are therefore more numerous in white than in gray matter (see Fig. 2-2 ). With their thin, short, cellular processes extending in all directions, oligodendrocytes provide internodes of myelination to multiple axonal processes in their environment. In H&E-stained sections, only the nucleus of oligodendrocytes is usually visible due to the blending of cellular processes with the neuropil. A clear zone surrounding the nucleus, the so-called perinuclear halo , often highlights oligodendrocytes as well as tumors with similar cytologic features (i.e., oligodendrogliomas); in either case, this appears due to a retraction artifact of formalin fixation. Nuclei are generally round and regular , but vary from small and darkly basophilic (accounting for a majority) to slightly larger with pale vesicular nuclei. Nucleoli are not usually noted by standard light microscopy. In the white matter, oligodendrocytes are disposed along the length of axonal processes, whereas in the cerebral cortex, they are scattered within the neuropil and concentrated immediately surrounding neuronal cell bodies (satellite cells). In the latter location, they may serve as a progenitor population.

Ependyma
Ependyma are cuboidal to columnar epithelioid glial cells that form a single-layered covering of the ventricular system ( Fig. 2-4 ). On their ventricular (apical) surface, ependyma have microscopically visible cilia and microvilli, whereas their lateral surfaces are tethered to one another by desmosomes, forming a functional CSF–brain barrier. Ependymal cytoplasm is pale to eosinophilic, and nuclei are oval and hyperchromatic. Within the supra- and infratentorial compartments, ependyma are fairly homogeneous, varying slightly by anatomic location in their cell height and degree of ciliation ( Fig. 2-4C and D ). Within the spine, the central canal is lined by ependyma and serves as a conduit for CSF during childhood. In adulthood, the central canal is collapsed and vestigial, remaining only as a central collection of clustered ependyma throughout the spinal cord length ( Fig. 2-4E ). Along the lateral ventricles, especially posteriorly, it is fairly common to encounter either entrapped outpouchings of ependyma or small clusters forming canals ( Fig. 2-4F ). These do not represent hamartomas or malformations, but only clinically inconsequential remnants of imperfect development.
Figure 2-4 Choroid plexus and ependyma. A, B, Choroid plexus (arrowhead) is a tufted aggregate of vascular channels lined by a single layer of choroid epithelium, which contains large pink cells with a cobblestone-like surface. Choroid plexus extends from its entry zone in the lateral ventricle, where it transitions from the ependymal lining (arrow) . C, D, Ependyma are a single layer of cuboidal to columnar epithelioid glial cells that line the ventricular system and form the brain–CSF barrier. Cilia can usually be noted on the ventricular surface of ependyma. The distinction between choroid plexus and ependymal cells can be seen at high magnification. E, In the spinal cord, a central collection of poorly organized ependymal cells forms the vestigial central canal, which can sometimes be mistaken for a neoplasm in small biopsy specimens. F, Extensions of the ependymal lining (arrow) can sometimes be noted within the white matter at a distance from the ventricular system, especially in the occipital lobe.

Choroid Plexus
The choroid plexus is a functionally differentiated region of ependyma that extends into the ventricular space as frondlike tufts of epithelium that secrete the ultrafiltrate of CSF (see Fig. 2-4 ). Individual cells are found as a single layer on a fibrovascular core . Compared with ependyma, they have larger, cobblestone-shaped cell bodies and contain small bland, basally located nuclei. Microvilli extend from the apical surface. Tight junctions and desmosomes are present between choroid plexus cells to ensure a viable blood–CSF barrier.

Microglia
Microglia are small, elongated cells located throughout the CNS gray and white matter ( Fig. 2-5 ). In the resting state, microglia are easily overlooked because of their small size and bland appearance, yet they account for nearly 20% of the cellular population. In standard H&E sections, the nuclei of activated microglia are long, thin, and dark—leading to their designation as “rod-cells” —but their cytoplasm is difficult to visualize. Special stains based on silver carbonate or lectins provide contrast to small processes and delicate branches that extend from their tips. Microglia are not neuroepithelial in origin, but rather are derived from a monocyte–macrophage lineage that incorporates into the CNS early in development. Once established, they serve as antigen-presenting cells for immune surveillance and participate in inflammatory responses, particularly against viral pathogens.
Figure 2-5 Microglia. A, Microglia have thin, elongated, and hyperchromatic nuclei that stand out from the neuropil (arrow) . Microglia are most readily identified in their reactive state, when they are referred to as “rod cells.” B, The rodlike quality of activated microglia is also appreciated on cytologic preparations (arrows) .
On activation, microglia proliferate and migrate to sites of damage, and in this state (“microgliosis”), cells are more readily identified. Activation also causes increased expression of proteins such as major histocompatibility complex (MHC) I and II, which can be detected immunohistochemically. When microglia and astrocytes aggregate around a central focus of injury, such as a virally infected neuron, they form a microglial nodule . Another population of monocyte–macrophage-derived cells resides in the perivascular compartment, between the outer basement membrane of the vessel and the glia limitans. As distinguished from parenchymal microglia, these perivascular macrophages are in continuity with the circulating monocyte population. Both perivascular and circulating populations of monocytes are recruited into the CNS parenchyma in response to severe injury, where they differentiate into tissue macrophages , often with foamy clear cytoplasm (a.k.a. gitter cells ) in order to perform phagocytic and immunologic functions. They are sometimes referred to as the “garbage collectors” of the CNS, since they clean up all necrotic debris, metabolic byproducts, and foreign material.

Blood Vessels
Similar to other organs, the brain has a population of vascular and perivascular cells essential for its oxygen and nutrient supply. Compared with their extracranial counterparts, the large arteries that run within the subarachnoid space have thinner muscular walls, less adventitia, and lack external elastic lamina ( Fig. 2-6 ). As they penetrate into the brain parenchyma, larger arteries retain both a thin covering by the pia-arachnoid and a perivascular space, the Virchow-Robin space , which represents a continuation of the subarachnoid space, although its function and content remain controversial. No such space exists once the vessels become small capillaries and the endothelium is intimately associated with neuropil. Capillaries are formed by individual endothelial cells forming delicate tubular structures with widths suitable only for passage of individual circulating blood cells. The physiologically critical blood–brain barrier is formed predominantly by the specialized nature of CNS endothelial cell junctions and cannot be identified histologically. In particular, these endothelial cells lack fenestrations between them and are joined by specialized tight junctions that functionally preclude the free movement of substances between vascular and CNS spaces. Astrocytic end-feet and basal lamina elements contribute to the integrity of this blood–brain barrier.
Figure 2-6 Blood vessels. A, Large arteries within the subarachnoid space supply the brain by penetrating into the parenchyma, where initially they maintain a perivascular space (Virchow-Robin space) that separates the vessel from the neuropil (arrow) . B, Smaller capillaries (arrows) within the brain consist of a thin, delicate tube lined by a single layer of endothelial cells that either directly abut the brain parenchyma or have a smaller perivascular space. The blood–brain barrier is due in large part to the specialized tight junctions between endothelial cells.

Meningothelial Cells
Meningothelial cells are scattered within the arachnoid membranes throughout the neuroaxis, but they are most concentrated at the tips of arachnoid granulations and the outermost layers of the arachnoid just under the adjacent dura, where they are called arachnoid cap cells ( Fig. 2-7A and B ). Meningothelial cells are epithelioid to slightly spindled and are typically seen in small clusters (10–20 cells), where they have a tendency to form whorls and psammoma bodies , similar to their neoplastic counterparts in meningiomas. Cells have moderate amounts of eosinophilic cytoplasm and oval nuclei with dispersed chromatin, often giving the appearance of central clearing.
Figure 2-7 Meningothelial cells and melanocytes. A, Meningothelial cells are scattered within the arachnoid membranes and are most frequent within the outermost layers as arachnoid cap cells. They are typically spindled to polygonal, have moderate amounts of eosinophilic cytoplasm and bland oval nuclei, and usually occur in small clusters. B, Melanocytes (arrowheads) are infrequent, flattened, highly pigmented cells of the pia and arachnoid membranes that are generally dispersed individually and are in highest density over the ventral brainstem. Meningothelial cells often form small whorls at the outer surface of the arachnoid membranes (arrow) . C, High magnification of melanocytes.

Melanocytes
Melanocytes are normal, neural crest-derived constituents of the human leptomeninges that are intimately associated with pia and subarach- noid membranes ( Fig. 2-7B and C ). They are widely scattered in most supratentorial regions and are noted histologically only following intense searching or fortuitous tissue sectioning. Their highest density is over the ventral surface of the superior spinal cord, brainstem, and base of the brain . Almost always seen as individual dendrite-shaped cells rather than clusters, leptomeningeal melanocytes are thin, elongated, and show slight branching and pigmentation in proportion to cutaneous pigmentation. As such, these cells are often most conspicuous in African American patients. Melanin pigment is made within cytoplasmic melanosomes and premelanosomes and is therefore similar to dermal melanocytes rather than the neuromelanin of the substantia nigra.

Tissue Organization
Cerebral Cortex
The vast majority (>90%) of cerebral cortex in humans is neocortex , an evolutionarily late form of cortical development that is distinguished from paleocortex (mostly limbic and olfactory cortices) and archicortex (hippocampal structures), which are more primitive. Neocortex differs from primitive cortex in its anatomic location and architecture. All neocortical areas—also called isocortex —go through developmental periods in which their elements are laid down in six layers . Many regions retain this layered appearance throughout life. Paleo- and archicortex do not share this developmental pattern or six-layered structuring into adulthood.
Cerebral cortex contains two dominant neuronal types: the granular (stellate) cell and the pyramidal cell (see section on Neurons). Pyramidal cells account for two thirds of cerebral cortical neurons and are the primary output. They have prominent apical dendrites that extend toward the cortical surface. Their axons extend long distances to terminate within the ipsilateral or contralateral cortex or travel to subcortical regions. Granular cells are smaller and are considered to be the primary interneurons of the neocortex. Other less common neurons are the horizontal cells (of Cajal), common in the superficial cortex in development; fusiform cells, most frequent in the deepest cortical layers; and cells of Martinotti, present in lesser numbers in all cortical layers.
The practice of neuropathology requires basic familiarity with neocortical structure ( Fig. 2-8A ), since subtle abnormalities underlie diseases such as developmental migration disorders, cortical dysplasia, epilepsy, neurodegenerative diseases, and hypoxic–ischemic injury (see Chapters 23 through 25 ). The six layers of the cortex, from the surface to the white matter, are (I) the molecular layer, which has very few neurons in adulthood; (II) outer granular cell layer; (III) outer pyramidal cell layer; (IV) inner granular cell layer; (V) inner pyramidal cell layer; and (VI) multiforme layer, which is populated primarily by fusiform neurons. These layers are more histologically apparent in some regions than others, often best appreciated in considerably thicker sections than are normally cut for routine surgical neuropathology. Regions with primary output function, such as primary motor cortex, have mostly pyramidal cells, whereas regions with primarily integrating or sensory function contain mainly granular cells. In either instance, dominance by a single cell type results in less apparent layering due to a loss of architectural contrast. Regions with nearly equal compliments of granular and pyramidal cells demonstrate the most apparent horizontal layering.
Figure 2-8 Cerebral cortex. A, Cerebral neocortex (isocortex) contains six layers, numbered sequentially from superficial to deep: I, molecular layer; II, external granular cell layer; III, external pyramidal cell layer; IV, internal granular cell layer; V, internal pyramidal cell layer; VI, multiforme layer. WM, white matter. B, The primary visual cortex of the occipital lobe has a distinctive histologic arrangement. Cortical layer IV is greatly expanded due to the high number of visual inputs and is divided into layers IVa, IVb, and IVc. Prominent bands of Baillarger are present in layers IV and V in primary visual cortex. The greatly expanded band of Baillarger in layer IV can be seen grossly as the line of Gennari.
Myelin staining of the cortex reveals parallel, horizontal bands of myelinated fibers that are not as readily apparent on H&E-stained sections. The two most prominent bands are in layers IV and V and are referred to as the external and internal bands of Baillarger, respectively. Primary visual cortex (Brodmann’s area 17), located on either side of the calcarine fissure in the occipital lobe, is characterized by a greatly widened band of Baillarger in layer IV due to the large input of visual afferent fibers from the lateral geniculate nucleus ( Fig. 2-8B ). This enlarged zone divides layer IV into three distinct layers and can be seen grossly as the “line of Gennari.”

White Matter
The white matter of the CNS is relatively uniform ( Fig. 2-9A ). It is generally more deeply eosinophilic than the overlying cortex, and its matrix is coarser. Its architecture is dictated by the arrays of axonal processes that extend to and from gray matter structures. Individual axons themselves are difficult to appreciate on H&E sections of normal brain since they are thin and blend with the background neuropil (though they can be noted in disease states in which neuropil is disrupted). However, neurofilament immunohistochemistry or silver stains can highlight axons ( Fig. 2-9B ). Oligodendrocytes, fibrillary astrocytes, and microglia are all oriented along the length of axons with a fairly rigid periodicity. When viewed in the plane of white matter tracts, units of approximately 5 to 10 oligodendrocytes are disposed in linear, parallel arrays along axonal processes and interrupted by single interspersed fibrillary astrocytes. Microglia are also located at regular intervals, albeit with much less frequency than oligodendrocytes, with cell bodies oriented parallel to axons.
Figure 2-9 White matter. A, Sweeping linear arrays of axons are the backbone of the white matter but cannot be readily identified on hematoxylin and eosin stains. Oligodendrocytes, astrocytes, and microglia are dispersed linearly along the length of axons with a fairly rigid periodicity. B, Axons in the white matter are highlighted in black by silver staining.

Basal Ganglia
The caudate, putamen, and the nucleus accumbens (a.k.a. the neostriatum ) are developmentally related and histologically similar ( Fig. 2-10A ). They contain a variety of small- and large-sized neuronal populations that have relatively uniform density. About 95% are small- and medium-sized (10–18 μm) γ-aminobutyric acid (GABA)-ergic spiny neurons that provide projections to the globus pallidus (a.k.a. the paleostriatum ). These have extensive dendritic trees packed with spines for connection with the large array of input fibers from the cerebral cortex, thalamus, and brainstem. Other populations consist of large cholinergic neurons (approximately 2% of neurons) and smaller cells containing neuropeptide Y, somatostatin, or nitric oxide synthetase. Interspersed among the neurons and neuropil of the striatum are small white matter bundles of the internal capsule that can only be seen microscopically. These “pencil fibers of Wilson” are specific for this region and serve as a guide to location when included in small biopsy specimens.
Figure 2-10 Basal ganglia and thalamus. A, The caudate, putamen, and globus pallidus contain a variety of small- and medium-sized neurons interspersed in a rich neuropil. Pencil fibers of Wilson are small white matter bundles embedded within the gray matter neuropil that are unique to these deep nuclei of the cerebrum (arrow) . B, The thalamus has large projection neurons as well as a less frequent population of smaller, inhibitory interneurons.

Thalamus
The thalamus is the main integrator and relay of sensory information to the cortex and has over 50 individual nuclei, each with its own specific function. Classic divisions are the anterior, medial, ventrolateral, and posterior groups of nuclei. Not among these larger categories are the midline, intralaminar, and reticular nuclei. The histologic appearance of each of the lobes is relatively similar, with variations depending on specific functions ( Fig. 2-10B ). Thalamic neurons consist of two main types: large projection neurons with axons that exit the thalamus (75% of the neuronal population), and smaller, inhibitory (GABAergic) interneurons. Each large projection neuron extends its process to the cerebral cortex through the internal capsule.

Hippocampus
The hippocampal formation consists of the hippocampus proper, subiculum, and dentate gyrus (a.k.a. dentate fascia) and is intimately associated with the entorhinal cortex ( Fig. 2-11 ). The entorhinal cortex occupies most of the parahippocampal gyrus, is the largest source of input into the hippocampus, and contains a distinctive six-layered cortical architecture. Near its surface in layer II are numerous large round neuronal clusters that can be seen as small protrusions on the brain’s surface. Deeper are the remainder of its six layers that contain pyramidal cells and a diverse array of smaller neurons. The subiculum sits at the base of the hippocampus and is a field populated largely by pyramidal neurons with an allocortical arrangement that transitions from the three-layered cortex of the hippocampal cornu ammonis 1 (CA1) subdivision at one end to the six-layered entorhinal cortex at the other. The hippocampus is divided into CA1, CA2, and CA3 subfields, each having distinct arrangements of pyramidal neurons and selective vulnerabilities to disease. CA3 emerges from the hilum (a.k.a. CA4) of the dentate gyrus and contains the largest pyramidal neurons. Pyramidal neurons of CA2 form a narrow band that runs between CA1 and CA3. Transition to CA1 is characterized by a wider band of slightly smaller pyramidal neurons that are more dispersed. CA1 (a.k.a. Sommer’s sector ) is generally much more sensitive to hypoxia, toxicants, seizures, and degenerative diseases than other subfields. The molecular layer of the hippocampal CA fields faces the dentate gyrus, and its white matter tracts form the alveus that runs along the space between the CA neurons and the lateral ventricle. White matter tracts of the alveus converge to form the fimbria of the hippocampus, which continues as the fornix , traveling around the peripheral portions of the septum pellucidum of the lateral ventricles to find their way to the hypothalamus and mamillary bodies ; this anatomic tract is also known as the circuit of Papez and forms an important portion of the limbic system.
Figure 2-11 Hippocampal formation. The hippocampus proper consists of CA1, CA2, and CA3 sectors of pyramidal neurons. CA1 continues as the subiculum (SUB) at the base of the hippocampal formation. The dentate fascia (DF) contains a narrow, densely populated band of granular cells (stratum granulosum), which surrounds the hilum (H), or CA4. The major white matter tract emerging from the hippocampus is the alveus (ALV), located between hippocampal pyramidal fields and the lateral ventricle.
The densely packed smaller granular cells that form the C-shaped structure of the dentate gyrus is the stratum granulosum. Hilar (or CA4) neurons that occupy the inner space within the C-shaped structure formed by the dentate fascia are a heterogeneous population of neurons including large pyramidal and smaller interneurons.

Pineal Gland
The pineal gland has a unique morphology, unlike any other region in the CNS ( Fig. 2-12 ). At low magnification, it has a lobulated arrangement with a prominent intralobular fibrovascular and glial stroma . The cellularity of the normal pineal is greater than most regions, which together with the unusual architecture, can lead to misinterpretation of this structure as a neoplasm. The dominant cell type, the pineocyte is a medium-sized specialized neuronal cell with round, regular “neuroendocrine” nuclei and delicate stippled chromatin. Pineocytes contain moderate amounts of pale pink cytoplasm with short processes and form small clusters and linear arrays. They stain avidly with synaptophysin and neurofilament protein antibodies, the latter of which often demonstrates small club-shaped swellings. At the periphery of nests and surrounding blood vessels is a higher density of fibrillarity. Interspersed among the nests and within the perivascular region are the less common pineal astrocytes, which are highlighted with GFAP stains.
Figure 2-12 Pineal gland. A, The pineal gland (arrow) is located in the midline, posterior and superior to the midbrain tectum (asterisk) . B, It consists of loose lobules of pineocytes arranged in small clusters and linear arrays and separated by glial and fibrovascular septae.

Pituitary Gland
The relationship between normal pituitary histology and disease is also covered in part in Chapter 18 . The pituitary gland is composed of anterior, intermediate, and posterior lobes ( Fig. 2-13 ). The anterior and intermediate lobes ( adenohypophysis ) have differing embryology, functions, and microscopic appearances from that of the posterior lobe ( neurohypophysis ). The adenohypophysis is not of neuroectodermal origin, but rather is derived from oral ectoderm which invaginates superiorly as Rathke pouch to eventually find its place within the sellar compartment. Notwithstanding their non-CNS origin, diseases of the sellar space often affect neurologic function.
Figure 2-13 Pituitary gland. A, B, The anterior pituitary gland consists of tightly packed acini of acidophils (pink) , basophils (blue) , and chromophobes (amphophilic) separated by a fine fibrovascular stroma. C, The posterior pituitary (neurohypophysis) is formed by the axonal projections of neurons from the hypothalamus together with primary glial cells and pituicytes, which are most commonly located in a perivascular distribution. D, Eosinophilic axonal dilations that store neurosecretory peptides (Herring bodies, arrow) can be seen distributed throughout the posterior gland. E, The intermediate lobe is small and often shows mild fibrosis, along with cysts (arrow) lined by flattened, Rathke-type epithelium.
The pituitary gland is connected to the more superior hypothalamus by the pituitary stalk , which is composed of the infundibulum , a superior extension of the neurohypophysis, and the pars tuberalis , an extension of the anterior gland. The stalk also carries a functionally vital vascular supply between hypophyseal and hypothalamic compartments. Arterial supplies to the pituitary are the inferior and superior pituitary arteries, which branch from each internal carotid artery. These give rise to a network of capillary loops within the gland (gomitoli), which in turn lead to a substantial network of venous sinuses that drain back to the hypothalamus, carrying vital hormonal feedback. Thus, the vascular network of the pituitary is extensive and critical to endocrine function.
The anterior pituitary accounts for over 75% of the sellar volume. It is composed of variably sized nests, or acini, interrupted by stromal and vascular septa ( Fig. 2-13A and B ). Most are filled with cellular elements and lack appreciable lumina. Only occasionally are glands with central spaces noted, some containing mucinous or colloid content. Stroma surrounding individual acini can be highlighted by reticulin stains, a helpful adjunctive test for establishing a normal glandular arrangement and ruling out an adenoma. Individual cells of the anterior lobe are classified as acidophils (40%), basophils (10%), and chromophobes (50%) based on their H&E staining. These staining patterns are not absolutely specific for endocrine function or hormone production (see Chapter 18 for general patterns). Rather, glandular cells are more often classified based on their immunohistochemical staining properties as lactotrophs (prolactin), thyrotrophs (thryrotrophic hormone, TSH), somatotroph s (growth hormone, GH), corticotrophs (adrenocorticotrophic hormone, ACTH), or gonadotrophs (follicle-stimulating hormone, FSH or leutinizing hormone, LH). Whereas gonadotrophs are diffusely spread throughout the gland with even density, other hormone-producing cells show regional variation. Corticotrophs and thyrotrophs are located in highest density within the central portion of the gland, and lactotrophs and somatotrophs are in highest density laterally.
The thin intermediate lobe of the pituitary is derived from the posterior Rathke cleft. In humans, it is not well developed and contains only glandular and colloid-filled cystic remnants within a slightly fibrous stroma. Individual cells are cuboidal or columnar, some with apical cilia, others containing cytoplasmic mucin. When large or clinically symptomatic, these cystic spaces are termed Rathke cleft cysts ( Fig. 2-13E ).
As distinguished from the glandular anterior lobe, the posterior pituitary , or neurohypophysis, is an extension of the CNS and has a “neural” histology ( Fig. 2-13C ). It is composed of neuronal processes that extend from their cell bodies in the hypothalamus down the pituitary stalk (infundibulum) to occupy the posterior portion of the sella and terminate near blood vessels. Scattered in the neuropil are Herring bodies —subtle esosinophilic axonal dilations that are filled with lysosomes and neurosecretory granules containing vasopressin and oxytocin ( Fig. 2-13D ). The most prominent nucleated cells of the neurohypophysis are pituicytes: GFAP-expressing spindle or stellate glial cells that abut the basal lamina of blood vessels. Their cytoplasm engulfs the nerve terminals and regulates the release of hormones into the bloodstream. The overall histology of the neurohypophysis is complex, with a seemingly disorganized cell arrangement that includes sweeping axonal processes punctuated by more cellular perivascular regions, causing occasional confusion with neoplastic disease.

Cerebellum
Although the circuitry of the cerebellar cortex is exceedingly intricate, its histologic appearance is homogeneous and relatively simple throughout ( Fig. 2-14 ). Outermost is the molecular layer , a rich neuropil network containing abundant axonal and dendritic processes, but only a few small neuronal cell bodies. The Purkinje cell layer is at the junction of the molecular layer and the deeper granular cell layer (see Figs. 2-1E and 2-3G arrows ). Purkinje cells are large neurons that have widely arborizing dendritic trees that extend into the molecular layer, serving as synaptic input for the parallel fibers of the granular cells. Purkinje cell axons are the main output of the cerebellar cortex, and a majority terminate on the neurons of the dentate nucleus. Granular cells of the cerebellum are the most common neuronal cell in the CNS and are present in a high-density region central to Purkinje cells. Each granular cell sends an axon to the molecular layer, which then bifurcates to form the parallel fibers that synapse with numerous Purkinje cell dendritic trees.
Figure 2-14 Cerebellum. The cerebellar cortex contains a sparsely cellular molecular layer (ML), a Purkinje cell layer (PCL), a granular cell layer (GCL), and white matter (WM).
The deep cerebellar nuclei are set on either side of the midline cerebellum in the midst of the white matter tracts of the medullary center that are entering and leaving cerebellar cortex. These are seen as thin, undulating ribbons of gray matter containing large and small neurons. Within the gray matter ribbon is a central zone of white matter tracts that projects out of the cerebellum. The largest and most lateral nucleus is the dentate nucleus , which has both developmental ties and morphologic similarity to the inferior olive of the medulla. It is the source of most efferent signals traveling out of the cerebellum via the superior peduncle. The other deep cerebellar nuclei, from lateral to medial, are the emboliform, globose, and fastigial nuclei.

Brainstem
Throughout the brainstem, anatomic regions are broadly subdivided, from ventral to dorsal, as base, tegmentum, and tectum. The base is located ventrally and consists mostly of long white matter tracts (cerebral peduncles, basis pontis, and medullary pyramids). The tegmentum lies dorsal to the base and ventral to the cerebral aqueduct or fourth ventricle. Among other structures, it contains the reticular formation, an area of centrally located, relatively uniform gray matter that lacks strict organization and boundaries but is critical to the control of basal bodily activities, including cardiovascular tone, respiration, and consciousness. The tectum is the area located dorsal to the brainstem ventricular compartments, serving as their roof. Together, the tectum and tegmentum house most of the integrative and cranial nerve nuclei components of the brainstem.
The locations of cranial nerve nuclei display the same general pattern throughout the brainstem. Nuclei are located in the dorsal tegmentum in the vicinity of the fourth ventricle. Motor nuclei are located medially, sensory nuclei are located laterally, and the autonomic nuclei are found between them.

Midbrain
At the most ventral aspect of the midbrain are the large cerebral peduncles. These dense white matter bundles occur on both sides of the midbrain and are composed predominantly of inferiorly projecting corticospinal and corticopontine fibers. Immediately dorsal is the substantia nigra (SN), a thin strip that extends laterally and dorsally from the midline and contains large, heavily pigmented dopaminergic neurons ( Fig. 2-1G , 2-15A ). In the midline between the right and left SN is the ventral tegmental area, where there is a functionally discreet population of pigmented dopaminergic neurons. The red nuclei are paired, round gray matter structures dorsal to the SN in the rostral midbrain. Around the ependymal-lined cerebral aqueduct is the periaqueductal gray matter , a collection of neurons involved in pain modulation. The midbrain tectum is almost entirely composed of inferior and superior colliculi, and the tegmentum contains predominantly white matter structures.
Figure 2-15 Brainstem. A, In the midbrain, the substantia nigra is composed of the reticulata (SNr), which resembles basal ganglia histologically, and the compacta (SNc), which contains a high density of large pigmented dopaminergic neurons. Ventral to the SNr is the cerebral peduncle (CP), and dorsal to the SNc in the superior midbrain is the red nucleus (RN). B, The base of the pons contains numerous, prominent pontine crossing fibers (arrow) that intertwine with pontine nuclei (asterisk) and descending fibers (arrowhead) , including corticospinal tracts (Luxol fast blue stain). C, The medulla contains the inferior olivary nucleus (arrow) , which like the related dentate nucleus in the cerebellum, is made of a thin ribbon of undulating gray matter surrounding a white matter hilum (Luxol fast blue/PAS stain).

Pons
The pons is dominated by its large base ( basal pons or basis pontis ) and by its large white matter connections to the cerebellum: the superior, middle, and inferior cerebellar peduncles. The basal pons consists of both transversely and longitudinally oriented white matter tracts ( Fig. 2-15B ). Longitudinal fibers include corticospinal tracts that continue as the medullary pyramids and corticopontine tracts that terminate on the interspersed pontine nuclei. The eye-catching transverse fibers represent white matter bundles arising from pontine nuclei, crossing the midline, and entering the cerebellum via the middle cerebellar peduncle . Near the fourth ventricle on each side of the pons is the locus ceruleus (“blue spot”), a small nucleus containing a high density of pigmented, noradrenergic neurons that project diffusely throughout the CNS. Near the midline throughout the brainstem, but concentrated mostly in the dorsal pons, are the midline raphe nuclei (midline “seam”). These nuclei contain large serotonergic neurons that project extensively throughout the brain.

Medulla
Anterior in the medulla are the paired medullary pyramids , which carry corticospinal tracts to their decussation at the medullary-spinal junction, then continue as the lateral corticospinal tracts in the spinal cord. Posterior to the pyramids in the midline is the medial lemniscus , a white matter tract projecting from the contralateral cuneate and gracile nuclei. More lateral in the rostral medulla are the dominant olivary nuclei , seen as bulges (olives) on the anterolateral medullary surface ( Fig. 2-15C ). This ovoid structure consists of a ribbon of convoluted gray matter with large pyramidal-type neurons surrounding a hilus of outwardly projecting white matter tracts that extend to the contralateral cerebellar peduncle. The olivary nucleus is developmentally and functionally related to the cerebellar dentate nucleus and resembles it histologically. Posteriorly, the fasciculus gracilis and cuneatus are continuations of the posterior columns and terminate in the nucleus gracilis and nucleus cuneatus, respectively. The medial longitudinal fasciculus is a white matter tract that rides the midline dorsally, while the spinothalamic tract maintains its anterolateral position in the brainstem, immediately dorsal to the olive in the medulla.

Spinal Cord
The spinal cord has the same basic histologic organization throughout its length, with unique features superimposed at specific spinal levels ( Fig. 2-16 ). On cross section the cord contains central gray matter in the shape of an H and surrounding white matter tracts. The white matter tracts are functionally diverse and precisely organized in terms of sensory and motor function. Nonetheless, they are fairly uniform in histologic cross sections, showing mostly bundles of myelinated and unmyelinated fibers traveling in the superior–inferior direction with scattered oligodendrocytes and fibrillary astrocytes. Anterior horns are the ventral extensions of the H-shaped gray matter and contain the large anterior horn cells ( lower motor neurons ) and smaller gamma motor neurons, which innervate muscle spindles. Anterior horns are the largest and contain the greatest number of lower motor neurons at the cervical and lumbar enlargements due to their output to the arms and legs. The posterior horn contains large projection neurons and smaller interneurons. The substantia gelatinosa is a posteriorly located portion of the posterior horn that is distinguished by its lack of myelinated fibers, giving rise to its pale appearance. It continues dorsally into Lissauer’s tract , another poorly myelinated region of white matter.
Figure 2-16 Spinal cord. Cross section of the thoracic spinal cord shows anterior horns (AH), posterior horns (PH), intermediolateral cell columns (IML), white matter (WM), the ependymal-lined central canal (CC), substantia gelatinosa (SG), Lissauer’s tract (LT), the fasciculus gracilis (FG) of the dorsal columns, anterior spinal artery (ASA), ventral roots (VR) and dorsal roots (DR).
The gray matter region between anterior and posterior horns contains cells of the autonomic nervous system. Between levels T1 and L3 is located the intermediolateral cell column, which extends off the central gray matter as a lateral horn. It contains the cell bodies of preganglionic sympathetic neurons, which project out through the ventral roots. The intermediate zone from S2 to S4 contains mostly a parasympathetic neuronal population. Lastly, Clarke’s nucleus is a medial extension of the intermediate gray matter found from spinal levels T1 to L2. It contains large neurons important to sensory processing with the cerebellum.

Meninges
The dura mater has a histologic appearance unlike any other region of the nervous system ( Fig. 2-17 ). It consists of a thick, monotonous layer of dense fibrous connective tissue composed mostly of layered collagen with only scattered interspersed flattened fibroblasts. Because its appearance is so consistent, its identification within a histologic section ensures the pathologist that the location of the surgically sampled lesion was superficial (i.e., near or involving the dural covering); nonetheless, some caution is warranted since markedly fibrotic leptomeninges may occasionally approach the thickness of dura. Normally, however, the arachnoid membranes that traverse the space between dura and underlying brain contains the arachnoid trabeculae, which are a delicate meshwork of thin connective tissue containing flattened fibroblast-like cells, scattered meningothelial cells, and rare melanocytes (see Figs. 2-7 , 2-17B ). The most superficial layer of cells (arachnoid cap cells) forms a continuous lining that is tethered to the overlying dura and forms a restrictive barrier to the flow of fluids between the subarachnoid space and the dura. Normally, there is in fact no subdural space per se, but these relatively weak attachments between arachnoid and dura are easily disrupted or pealed back by hemorrhage (e.g., subdural hematoma) or “unnatural” forces, such as the prying hands of a surgeon or pathologist. The pial layer is found on the surface of the brain as a delicate fibrous coating that is slightly eosinophilic compared with the underlying cortex and contains only rare small flattened cells ( Fig. 2-17C ). It extends peripherally to fuse with the overlying arachnoid trabeculae to form a continuous pia–arachnoid network.
Figure 2-17 Meninges. A, The dura mater is a thick, dense, fibrous connective tissue covering for the brain with low cellularity. B, The arachnoid membranes are delicate fibrous bands (arrow) that traverse the subarachnoid space (asterisk) , embed subarachnoid vessels, and have attachments to both underlying pia and overlying dura. C, The pia mater (arrow) is a thin, fine coating on the surface of the brain that is brightly eosinophilic and merges with the arachnoid.

Peripheral Nerve, Schwann Cells, and Dorsal Root Ganglia
Within millimeters of their exit from the CNS, both cranial nerves and spinal nerve roots transition from a central to a peripheral nerve morphology and myelinating pattern (with the exception of cranial nerve VIII, which transitions at the internal auditory meatus) (see also Chapter 22 ). Schwann cells are the glial cell equivalents of the peripheral nervous system that provide an insulating coat of myelin around axons to improve conduction speeds ( Fig. 2-18A ). Larger nerves (e.g., sural nerve) typically have multiple subunits known as fascicles, which appear rounded on cross section and consist of ensheathed bundles of myelinated and unmyelinated axons , along with Schwann cells, small blood vessels, and stromal support. Together with peripheral nerve fibroblasts and the collagen-rich network of endoneurium (within the fascicle), perineurium (surrounding individual fascicles), and epineurium (surrounding the entire nerve), Schwann cells provide a structural support to their underlying axonal processes. In contrast to oligodendrocytes, the myelin-rich cytoplasm of a single Schwann cell is flattened and concentrically laminated around a segment of a large axon at specific intervals between nodes of Ranvier. In standard H&E-stained tissue sections Schwann cells are the most numerous cell bodies within peripheral nerves and are seen in longitudinal sections as elongated, spindled cells containing cigar-shaped hyperchromatic nuclei. On cross section of nerve, their myelin-rich coating is seen as a clear, donut-shaped ring around a central, tiny, eosinophilic axon ( Fig. 2-18B ). Stains for myelin (Luxol fast blue) dramatically improve the visibility of the myelin sheath.
Figure 2-18 Dorsal root ganglia and peripheral nerve. A, Peripheral nerve in longitudinal plane showing bundles of axons (arrow) , which are only barely visible within their thicker, clear myelin sheath. Schwann cells have elongate nuclei with slightly bulbous ends and are oriented along the length of the axon to provide its myelination. B, On transverse section of a peripheral nerve, the clear ring of bubbly myelin is seen surrounding a central zone occupied by the axon (arrow) . C, Each large neuronal cell body (ganglion cell) of the dorsal root ganglion is surrounded by satellite cells—a specialized Schwann cell population.
Dorsal root ganglia are located near the spinal exit foramina, invested within a dural sheath, and are the home of cell bodies for spinal afferent sensory neurons. Individual cell bodies of ganglion cells are large, with abundant cytoplasm, Nissl substance, prominent vesicular nuclei, large nucleoli, and variable quantities of cytoplasmic lipofuscin pigment ( Fig. 2-18C ). Peripheral extensions terminate in transducing sensory receptors that give rise to incoming signals. Large, long processes extend centrally via the dorsal roots into the spinal cord, with the largest myelinated tracts becoming the ascending posterior columns. Around the perimeter of each ganglion cell body are slender satellite cells (specialized Schwann cells), which most likely serve a support role and provide a committed stem cell source for repopulation of their more peripheral progeny.

Features of Infancy and Childhood
The germinal matrix is a neural stem cell population that is adjacent to the lateral ventricles as a subependymal layer and gives rise to sequentially differentiated neuronal and glial precursors that migrate to their homes in the cerebrum ( Fig. 2-19A ). The germinal matrix is prominent in early brain development and does not begin to thin out until the 26th week of gestation. The matrix persists as scattered cell islands and perivascular nests until term. After birth, most of the germinal matrix disappears except for a portion called the ganglionic eminence, which is located between the thalamus and caudate. It fragments and diminishes in size throughout the first year of life.
Figure 2-19 Features of development and infancy. A, The germinal matrix (GM) is a periventricular precursor cell population located directly adjacent to the ependyma (E) of the lateral ventricles (LV). Although heavily populated by neural precursors during fetal development, it diminishes and eventually disappears in the first year of postnatal life (germinal matrix of 20-week-gestation fetus). Neural precursors migrate away from the germinal matrix to eventually populate the cerebral hemispheres with mature neuronal and glial populations ( arrow ). B, The cerebral cortex undergoes gradual lamination during fetal development, with individual layers emerging in the fifth month of gestation. The cortex of a 30-week-gestation fetus shows a clearly formed molecular layer (ML), initial separation of cerebrocortical layers (CC), and demarcation of the cortex from white matter (WM). C, The fetal and infant cerebellum contains an external granular cell layer (EGCL, arrow ), which is a precursor cell population that migrates inward through the molecular layer (ML) to form the internal granular cell layer (IGCL) (cerebellum of 6-week-old infant). WM, white matter.
The cerebral cortex derives from neuroblasts that migrate outwardly along radial glia from the germinal matrix. The inner-most neurons of the cortex are the first to arrive and are subsequently joined by neuroblasts migrating to progressively more superficial regions. By the fifth month of fetal development, the cortex shows a superficial molecular layer and a deeper, densely cellular band ( Fig. 2-19B ). From the latter, a six-layered cortex gradually emerges starting in the sixth month. Cortical layering results from the maturation of cortical laminar neurons, the selective cell death of neuronal populations, and expansion of the neuropil due to the growth of dendritic fields.
Cerebellar cortical development occurs along two major pathways. The Purkinje cells form early in embryonic life after migrating to their final location from the alar plate. Granular cells develop from the rhombic lip. They first form a precursor population as the external granular cell layer , which is located at the surface of the cerebellar folia, superficial to the molecular cell layer. External granular cells are actively dividing and give rise to inwardly migrating cells that form the internal granular cell layer —the granular cell population that persists in adulthood ( Fig. 2-19C ). Although the external granular cell layer begins to diminish at 2 to 3 months after birth, it does not totally disappear until 12 months.

Features of the Aging Nervous System
A wide range of histologic features may be encountered in the aging nervous system, often becoming most prominent in elderly patients; they are summarized in Box 2-1 and illustrated in Figures 2-20 and 2-21 . Recognition of these structures is critical in order to avoid misinterpreting them as pathologic. In the case of neurofibrillary tangles ( Fig. 2-21A ) and neuritic plaques ( 2-21B ), the distinction between normal aging and disease becomes a matter of quantity and location. Small numbers of tangles in the mesial temporal lobe are considered part of normal aging, but widespread neocortical involvement is a sign of Alzheimer’s disease (AD), and extensive subcortical deposits are characteristic of other neurodegenerative disorders, such as progressive supranuclear palsy (see Chapter 25 ). In fact, the precise number and type of neuritic plaques (and tangles) needed for a definitive diagnosis of AD has been a topic of great debate over the years, although fortunately most cases of advanced disease contain numerous widespread plaques and tangles, making the diagnosis relatively straight-forward. Similarly, while corpora amylacea ( Fig. 2-20A ) are extremely common and are considered totally innocuous, similar structures may be seen within neuronal cells or their processes in rare disorders such as Lafora’s disease and adult polyglucosan body disease.
Figure 2-20 Findings of normal aging. A, Corpora amylacea (arrow) are spherical basophilic polyglucosan bodies that accumulate as astrocytic inclusions during the aging process. Their highest density is around blood vessels, under the pial surface, and adjacent to the ventricles—locations where astrocytic foot processes are most common. These eye-catching laminated bodies are not always recognized as being intracellular, and they can accumulate to striking densities. B, Perivascular mineralization of the large vessels of the globus pallidus (arrow) is a common aging process and can begin as early as childhood. C, Microvascular mineralization also occurs with increasing age and is seen most frequently in the hippocampus and the basal ganglia (arrow) . D, The arachnoid membranes become thicker and more fibrous with age. Fibrous plaques are thick, densely hyalinized forms of fibrosis that occur in the most superficial layer of the arachnoid membranes. These are noted most often over the median aspects of the superior frontal and parietal lobes and the covering of the spinal cord.
Figure 2-21 Findings that may occur in limited fashion in normal aging. A, Neurofibrillary tangles are slightly basophilic, crystalline inclusions that fill the neuronal cytoplasm, generally taking the shape of a flame (arrows) . B, Amyloid plaques represent the extracellular accumulation of β-amyloid that deposits as part of aging or Alzheimer disease (arrow) . C, Granulovacuolar degeneration consists of small cytoplasmic vacuoles and basophilic granules and is noted most often in the hippocampal pyramidal cells of elderly individuals (arrows) .

Box 2-1     Features of Aging
Neurons
Neurofibrillary tangles
Granulovacuolar degeneration
Hirano bodies
Neuritic plaques
Ferrugination
Marinesco bodies
Lipofuscin accumulation
Pigment incontinence of substantia nigra

Glia
Corpora amylacea adjacent to ependyma, subpial regions, and vasculature

Meninges
Fibrous thickening
Hyaline plaques
Arachnoid granulation collagenization (Pacchionian bodies)
Meningothelial hyperplasia (reactive proliferation)
Psammoma body formation

Pituitary Gland
Squamous cell metaplasia of pars tuberalis
Adenohypophyseal fibrosis

Other
Perivascular mineralization, globus pallidus
Micronodular mineralization, globus pallidus and hippocampal molecular layer
Choroid plexus mineralization and cystic change
Pineal mineralization and cystic change

Suggested Readings

Nolte, J. The Human Brain: An Introduction to Its Functional Anatomy, 5th ed. St. Louis: Mosby, 2002.
Fuller, G. N., Burger, P. C. Central nervous system. In Mills S.E., ed.: Histology for Pathologists , 3rd ed, Philadelphia: Lippincott Williams & Wilkins, 2007.
Ortiz-Hildago, C., Weller, R. O. Peripheral nervous system. In Mills S.E., ed.: Histology for Pathologists , 3rd ed, Philadelphia: Lippincott Williams & Wilkins, 2007.
Lopes, M. B.S., Pernicone, P. J., Scheithauer, B. W., . Pituitary and sellar region. In Mills S.E., ed.: Histology for Pathologists , 3rd ed, Philadelphia: Lippincott Williams & Wilkins, 2007.
Kandel, E.R., Schwartz, J.H., Jessell, T.M. Principles of Neural Science, 4th ed. New York: McGraw-Hill, 2000.
Friede, R.L. Developmental Neuropathology, 2nd ed. Berlin: Springer-Verlag, 1989.
Nelson, J.S., Mena, H., Parisi, J.E., Schochet, S.S. Principles and Practice of Neuropathology, 2nd ed. New York: Oxford University Press, 2003.
3
Intraoperative Consultation and Optimal Processing
Gregory N. Fuller

Types of Neurosurgical Specimens   35
Intraoperative Cytologic Preparations as a Complement to Frozen Tissue Sections   36
Fixation and Staining Options for Intraoperative Cytologic Preparations   36
Frozen Sectioning of Central Nervous System Tissue   36
Artifacts   37
Iatrogenically Introduced Hemostatic and Embolic Agents   42
The Bottom Line: What Does the Surgeon Need to Know?   42
Intraoperative consultation is unquestionably one of the most important and often most challenging tasks for the surgical pathologist. Of all the organ systems, neurosurgical specimens appear to be particularly problematic. 1 Each surgical specimen and clinical setting offers its own unique challenge. Nevertheless, some time-tested principles can be applied in virtually all situations to provide the most reliable diagnostic interpretation. In this chapter the various aspects of intraoperative surgical neuropathology will be explored, with an emphasis on practical techniques and approaches.

Types of Neurosurgical Specimens
The physical size of tissue specimens submitted for frozen section evaluation varies greatly ( Box 3-1 ). Specimen size is directly related to the goal of the surgical procedure. Endoscopic and stereotactic biopsies are typically performed for diagnostic purposes only, whereas at the other end of the spectrum, hemispherectomies and lobectomies are often performed for surgical cure under the best of circumstances. Between these two extremes lies a large range of specimen sizes encompassed by open biopsies , partial resections, and gross total resections . Despite this variation in specimen size, some general principles are more or less universally applicable ( Box 3-2 ). The two most helpful rules are (1) perform a cytologic preparation as a complement to the frozen tissue section, and (2) never freeze all of the lesional tissue ; that is, always save some tissue for formalin-fixed paraffin-embedded (FFPE) permanent sections. A rare exception to the latter rule occurs when the surgeon had intended to obtain additional tissue for optimal preservation, but was unable due to intraoperative complications that arose.

Box 3-1     Surgical Specimen Size, Generally Arranged from Smallest to Largest
Endoscopic biopsy
Stereotactic biopsy
Open biopsy
Partial resection
Gross total resection
Lobectomy
Hemispherectomy

Box 3-2     General Principles for Handling Intraoperative Consultation Tissue Specimens
Perform a cytologic preparation as a complement to the frozen tissue section.
Don’t freeze all of the tissue.
Don’t submit the entire specimen for paraffin embedding (save some in glutaraldehyde or formalin).
For minute biopsy specimens, cut unstained slides from the frozen section block and order unstained sections from the paraffin blocks up front in order to avoid loss of diagnostic tissue when refacing the block.
Sometimes, definitive tumor is identified on the frozen section but unusual or unexpected morphologic features are apparent. In such cases it is prudent, if sufficient tissue is available, to place a small representative fragment in glutaraldehyde for ultrastructural examination, if needed. If glutaraldehyde is not available, retaining a small amount of tissue in formalin is the next best option. Retrieving tissue from paraffin blocks followed by postfixation and processing for electron microscopy is possible, but often yields suboptimal ultrastructural detail.
The pathologist is sometimes faced with a situation in which a minute stereotactic biopsy tissue fragment has been entirely frozen and reveals the presence of neoplasm or other lesional tissue, but immunostaining will likely be required for further classification or grading. If the surgeon is reluctant to take additional tissue cores due to the risk of intracranial hemorrhage (a rare scenario when good preoperative planning has happened and open communication exists with the pathologist), it is advisable to have unstained tissue sections cut from the frozen section block before removing it from the cryostat, since there is no guarantee that lesional tissue will still be present in the paraffin block after processing and refacing. Similarly, if small biopsy cores are provided by the surgeon for permanent sections, it is prudent to request unstained sections initially (i.e., cut at the same time as the section for hematoxylin and eosin (H&E) staining) in order to avoid loss of lesional tissue secondary to block refacing.
An additional safeguard in the setting of scant lesional tissue is the preparation of several unstained cytologic touch preparations from the fresh tissue before submitting for processing.

Intraoperative Cytologic Preparations as a Complement to Frozen Tissue Sections
At some institutions, historical practices have dictated that only frozen sections are performed for intraoperative consultation; at others, intraoperative diagnoses are made primarily by cytologic preparation, with frozen sections being only rarely employed. However, a large number of surgical neuropathologists worldwide routinely use a combination of cytologic preparations and frozen sections to render intraoperative diagnoses, and these two procedures are viewed as complementary, with the cytologic preparation providing exquisite nuclear and cytoplasmic detail free of freeze artifact and distortion, and the frozen section providing architectural details, including the relationship between the disease process and the host tissue (e.g., solid versus infiltrative tumor). 2 – 7
Several types of technical procedures are available for generating the cytologic preparation and the choice depends on a number of factors, including the disease type and the consistency of the tissue submitted for intraoperative evaluation ( Box 3-3 ). For example, touch (imprint) preparations are generally optimal for pituitary adenoma. The concept is to take advantage of an integral aspect of adenoma pathobiology—the clonal expansion of loosely cohesive adenoma cells together with the attendant effacement and loss of the fibrovascular septa that compartmentalize normal adenohypophyseal tissue into acini. Adenoma cells tend to shed profusely on touch preparations compared with normal adenohypophysis. Touch preparations also work well on other hypercellular, loosely cohesive or dishesive tumors, such as lymphoma and melanoma. However, for more cohesive neoplasms and disease processes, the touch preparation is frequently hypocellular and often consists primarily of uninformative red blood cells. For this reason, many pathologists prefer the smear (squash, crush) preparation for soft tissue specimens, including most primary and metastatic central nervous system (CNS) tumors.

Box 3-3     Types of Intraoperative Cytologic Preparations
Touch (imprint)
Smear (squash, crush)
Drag
Scrape
Cytologic preparations are of particular value for specific problematic tissue specimens ( Table 3-1 ). The pathologist sometimes receives a minute biopsy specimen that is too small to divide into separate portions for cytologic smear and frozen tissue section. In such cases, the tissue fragment can be first dragged across a slide to yield a cytologic preparation ( cytologic drag preparation ) before freezing. Another problematic specimen is the extensively cauterized tissue fragment. Although some of these may be beyond salvage, an attempt at diagnosis can be made by bisecting the tissue and performing a cytologic drag preparation before freezing the tissue. Fibrous or desmoplastic tissue samples often do not smear well. In such situations, a scrape preparation may be optimal. With this procedure, the tissue is grasped with forceps and a scalpel blade repeatedly drawn across the surface. The collected debris is then spread across a glass slide. Another challenging situation is presented by a Petri dish full of grossly necrotic tissue fragments. Arbitrarily choosing one or two fragments for frozen section often yields only nonspecific necrosis. In this situation, it is advantageous to sample as much tissue as possible for any viable tumor cells or cell clusters, yet preparing 10 or 20 frozen sections is impractical. One solution is to perform a cytologic drag preparation on multiple tissue fragments . With this procedure, one fragment after another is quickly dragged across the same, single slide; 5, 10, 15 or more fragments can be sampled in a matter of only seconds. If any viable tumor cells are present, this procedure is an efficient and cost-effective way to maximize the chances of their detection. Occasionally the pathologist is confronted with a request for “frozen section” on a specimen consisting of bony fragments , without an identifiable soft tissue component. An attempt to obtain a diagnosis can be made by performing a cytologic drag preparation of the bony fragments; this may yield a diagnostic preparation for some pathologic processes, such as metastatic carcinoma.

Table 3-1
Problematic Frozen Section Specimens Tissue Recommended Handling Minute sample (endoscopic or stereotactic biopsy) Perform a cytologic touch or drag preparation before freezing the tissue fragment. Cauterized tissue fragment Bisect with scalpel and perform a cytologic drag preparation of the freshly cut surface before freezing the tissue. Fibrous or desmoplastic tissue Perform a cytologic scrape preparation before freezing the tissue. Petri dish full of necrotic tissue fragments Perform a cytologic drag preparation using multiple (10-15) tissue fragments on the same slide to maximize sampling for viable tumor cells. Bony tissue fragments submitted for “frozen section” Perform a cytologic drag preparation.

Fixation and Staining Options for Intraoperative Cytologic Preparations
Depending on the background, experience, and personal preference of the pathologist, cytologic preparations may either be air-dried or immersed immediately in 95% ethanol to avoid drying artifact. For gliomas in particular, the latter method provides superior preservation since nuclear cytology is particularly critical and air drying often produces marked artifacts of size, shape, and chromatin density. Similarly, stain preference varies among pathologists, although many prefer routine H&E. At some institutions, the preference is to perform an initial rapid one-step Diff-Quick stain on the first cytology slide, followed by routine H&E staining of a second preparation.

Frozen Sectioning of Central Nervous System Tissue
Many pathologic processes that involve the CNS, including diffuse gliomas and metastatic tumors, elicit marked vasogenic edema of brain parenchyma that is most pronounced in the white matter. This increase in water content of an already very soft and lipid-rich tissue can lead to pronounced freeze artifact , often severe enough to significantly impede or even prevent diagnosis. In general, the optimal conditions for cryostat sectioning of brain tissue may not be the same as those used for other tissue types. Experimentation with block temperature and section thickness may reduce artifact. In addition, cutting two or three serial sections onto the slide can be helpful. An additional strategy to reduce the formation of large ice crystals is to make sure the specimen freezes rapidly. Snap freezing is the most common means for achieving such rapid freezing and is performed by first immersing the specimen (surrounded by a small amount of optimal cutting temperature [OCT] compound on a metal chuck) into an isopentane cryobath. Once the chuck is transferred to the cryostat, additional OCT is added and a metal heat extractor is quickly placed on top, simultaneously creating a more rapid freeze and a flat cutting surface. Frozen tissue sections of CNS tissue can be obtained in a quality that approaches FFPE tissue sections in quality, but optimizing the procedure may require practice, experimentation, and close collaboration between pathologist and histotechnician.

Artifacts
A number of artifacts can interfere with frozen section diagnosis ( Box 3-4 ). Some artifacts can impede or even prevent interpretation, whereas others can mislead the pathologist into rendering an inaccurate diagnosis. Chief among the artifacts that can make reliable diagnosis impossible are freeze and cautery artifacts. The former can be minimized as described in the prior section. Cautery artifact , on the other hand, is out of the pathologist’s control. Some specimens, usually minute tissue fragments, are completely uninterpretable ( Fig. 3-1A ); others may be salvaged by judicious selection of the least cauterized areas for freezing. Even in areas where the tissue is less severely cauterized, however, artifacts may lead to misinterpretation. For instance, nuclei appear thinner and more elongate, often polarizing in the same direction ( Fig. 3-1B ). This can give a false sense of astrocytic, fibrous, or schwannian cytology with associated palisading or pseudopalisading. As a general rule, whenever cauterized vessels are identified by their characteristic smudgy purple walls and dark pyknotic nuclei, the surrounding brain parenchyma must be interpreted with extreme caution because minor, misleading distortions of nuclear and cellular morphology may not appear to be artifact ( Fig. 3-1C ).
Figure 3-1 Cautery artifact. Cautery artifact ranges from severe and uninterpretable ( A ) to mild and misleading ( B, C ). A, Severe cautery artifact in which the distortion has rendered the tissue uninterpretable. B, Moderate cautery artifact in which the nuclei are artifactually elongated and tend to polarize in the same direction. A cauterized vessel is seen at the bottom of the field. C, Mild cautery artifact with artifactually elongated nuclei. The obviously cauterized vessels serve as a warning to the pathologist to interpret the surrounding tissue with caution. Compare with an uncauterized area of the same surgical specimen shown in D, which reveals the nuclei to be rounded, without the artifactual elongation seen in cauterized areas of the tissue.

Box 3-4     Artifacts
Freeze
Cautery
Crush
Pseudocalcification (“bone dust”)
Pseudonecrosis (see Box 3-5 )
“Blue sponge”
Formalin precipitate
Formic acid treatment
Dark cell change
Freeze artifact is also high on the list of misleading artifacts. Similar to cautery artifact, a spectrum of tissue distortion from freezing can be seen, and may be virtually uninterpretable in its most severe form ( Fig. 3-2 ). Nuclear distortion and irregularity that result from freezing in oligodendroglioma is a very common situation. The morphologic hallmark of this distinctive diffuse glioma, as instantly recognized on FFPE sections, is a monotonous population of cells with uniformly regular round nuclei surrounded by perinuclear halos. On frozen sections, however, the perinuclear halos, which are themselves an artifact, albeit a useful one, resulting from FFPE processing, are not present, and freezing also induces a varying degree of nuclear irregularity and pleomorphism that is not seen in FFPE sections. The end result is that frozen sections of oligodendroglioma tend to look like astrocytoma ( Fig. 3-3A and B ). Misinterpretation of this artifact is a common cause of diagnostic discrepancy between frozen section interpretation and final diagnosis. What can be done to avoid this pitfall? The cytologic preparation may be helpful because it preserves oligodendroglioma features, with regular round nuclei and delicate chromatin pattern beautifully displayed without the distortion induced by freezing. But does the surgeon really need to know whether a diffuse glioma displays astrocytic, oligodendroglial, or mixed oligoastrocytic differentiation at the time of intraoperative consultation? Will the answer change what the surgeon does? The answer in most instances is “no.” It should be sufficient to convey that the disease process is a diffuse glioma, and if high-grade features are present, that information can be conveyed as well; otherwise, the precise classification and grading of a diffuse glioma should await the superior morphology of FFPE tissue sections and the results of adjunctive special studies. One exception to this rule arises when the preoperative imaging studies and clinical information strongly suggest glioblastoma (e.g., magnetic resonance imaging [MRI] shows a ring-enhancing mass lesion crossing the corpus callosum). Although a number of entities can manifest with these imaging features (glioblastoma, lymphoma, abscess, demyelinating pseudotumor, etc.), once the pathologist has identified the presence of diffuse glioma on the frozen section, the likelihood that the surgical specimen is representative should be assessed; specifically, can the histologic features seen account for the contrast enhancement (i.e., microvascular proliferation) and central hypodensity (i.e., necrosis) on the imaging studies? If not, additional tissue should be requested.
Figure 3-2 Freeze artifact. Similar to severe cautery artifact, severe freeze artifact can also render brain tissue so distorted as to be virtually uninterpretable.
Figure 3-3 Freeze artifact. Freeze artifact distorts nuclear morphology such that oligodendroglioma looks more like astrocytoma. A, B, In this frozen section ( A ) of an oligodendroglioma, the artifactual nuclear distortion produces irregular contours reminiscent of fibrillary astrocytoma. However, the formalin-fixed paraffin-embedded permanent section ( B ) shows round uniform nuclei.
The neurosurgeon may occasionally introduce an artifact of particles of “ bone dust ” into the specimen. These microscopic fragments of cranial bone are generated during the drilling of burr holes and the turning of craniotomy flaps and find their way into the surgically resected tissue. When artificially trapped between tissue fragments that have been pushed up against each other, bone dust fragments mimic dystrophic calcification and can be mistaken as part of the disease process, perhaps leading to an inaccurate conclusion that the disease is more chronic in nature ( Fig. 3-4A ). Simple awareness of bone dust as a common contaminant in neurosurgical specimens can significantly mitigate the chance of misinterpretation. Surgical re-excision specimens may contain bone dust from the prior operation that has become embedded in reactive scar tissue, often with an associated foreign body-type giant cell reaction ( Fig. 3-4B ).
Figure 3-4 Bone dust. A, Microscopic fragments of cranial bone are seen artifactually surrounded by resected tissue, which can convey the mistaken impression of dystrophic calcification and lesion chronicity. B, At subsequent repeat surgery, bone dust fragments can be seen incarcerated in reactive fibrous tissue, often accompanied by a foreign body-type giant cell reaction. Also seen in this field of resected dura are suture fibers from a previous operation.
Necrosis is a key diagnostic finding in many disease processes. The misidentification of necrosis in a specimen can lead to serious inaccuracies in diagnosis. A number of artifacts can lead to “ pseudonecrosis ” ( Box 3-5 ). Among the most common is a simple technical staining error yielding focal absence of hematoxylin staining, often seen at the end of the slide in smear preparations or frozen tissue sections. The resultant bright eosin staining of tissue protein without hematoxylin staining of nuclear nucleic acid mimics the dissolution of nucleic acid seen in necrotic tissue ( Fig. 3-5 ).
Figure 3-5 Pseudonecrosis. A common cause of pseudonecrosis is absence of hematoxylin staining (usually near the end of the slide) due to technical error. In this smear preparation, the right side of the smear preparation is brightly eosinophilic, mimicking necrosis. At high power, no necrosis was seen, only lack of hematoxylin staining.

Box 3-5     Pseudonecrosis Artifacts
No hematoxylin
Hemorrhage/Fibrin
Cerebellar molecular layer
Degenerating hemostatic agent (especially microfibrillar collagen)
Cavitational ultrasonic surgical aspirator (CUSA, Cavitron) artifact
Severe freeze artifact
Formic acid overdigestion
Focal hemorrhage or fibrin deposition within CNS tissue can sometimes be misinterpreted as necrotic foci and result in, for example, overgrading of a diffuse astrocytoma as glioblastoma ( Fig. 3-6 ). The absence of karyorrhectic nuclei is one clue favoring a conclusion of pseudonecrosis.
Figure 3-6 Red blood cell pooling and fibrin deposition mimicking necrosis. Hemorrhage with red blood cell pooling and fibrin deposition within a diffuse glioma can mimic necrosis to the unwary, potentially leading to overgrading. The obvious intact red blood cells seen at the far edge of this photomicrograph will not always be present to alert the pathologist to the artifact.
One normal CNS region that often mimics necrosis on smear preparations and frozen sections is the cerebellar cortex . The brightly eosinophilic, hypocellular molecular layer stands out in sharp chiaroscuro when juxtaposed against the densely hypercellular, hematoxylinophilic granular cell layer. This appearance has fooled more than one pathologist into an impression of necrosis, sometimes with attendant misinterpretation of the small granular cell layer neurons as metastatic small cell carcinoma or lymphoma. This mimicry is further heightened by freeze artifact ( Fig. 3-7 ).
Figure 3-7 Cerebellar cortex molecular layer mimicking necrosis. The cerebellar cortex is striking in its juxtaposition of the hypocellular outer molecular layer and hypercellular granular cell layer, as exemplified in this frozen section. The sharp contrast in cellularity and staining characteristics can lead to the mistaken interpretation of the molecular layer as necrosis on smear preparations and frozen sections, with the “small cell” features of the granular cell layer potentially being mistaken for metastatic small cell carcinoma, lymphoma, or medulloblastoma.
One iatrogenically introduced form of pseudonecrosis artifact is caused by degenerating hemostatic agent . Foreign material introduced by the surgeon or interventional radiologist as a source of tissue artifact is discussed in more detail in the following section.
Another source of pseudonecrosis is the artifact produced in tissue fragments that are obtained through ultrasound-induced tissue cavitation followed by suction removal ( cavitational ultrasonic surgical aspiration, or CUSA). The resulting specimen, usually received in a saline-filled suction collection bag, is a mixture of intact and mechanically disrupted or partially autolytic tissue fragments, bone dust, blood, and hemostatic material. The distorted tissue fragments, altered by both the suction process and saline immersion, can simulate necrosis (“CUSA artifacts”) ( Fig. 3-8 ). The pathologist must be careful not to overinterpret such specimens, but rather carefully search for relatively intact fragments that can be reliably interpreted.
Figure 3-8 Pseudonecrosis. Pseudonecrosis is a common artifact seen in tissue specimens obtained by cavitational ultrasonic surgical aspiration (CUSA). This field from a CUSA specimen nicely illustrates the distorted nuclei of pseudonecrosis in comparison with an adjacent fragment of relatively preserved brain parenchyma.
Because of its soft consistency, brain specimens are much more susceptible to mechanical distortions during surgery (e.g., crush artifacts from forceful use of surgical instruments), during handling by the pathologist (e.g., crush artifacts from forceps), or even during tissue processing. In terms of the last circumstance, a common artifact results from tissue placed into cassettes between blue sponges in order to prevent loss of small fragments of tissue (“ blue sponge artifact ”) ( Fig. 3-9A ). Given the finger-like projections of these sponges, the permanent sections often reveal triangular “clear holes” in the tissue, with significant crush artifact at the edges of these holes. Similarly, biopsy embedding bags can parcellate small brain biopsy specimens into a checkerboard of small tissue squares ( Fig. 3-9B ). Particularly with small specimens, such as those retrieved in stereotactic needle biopsies, these processing techniques may leave little well-preserved tissue to examine. Therefore, careful wrapping of small brain specimens in wax paper or the use of cassettes with very small windows are preferable methods for minimizing tissue loss during routine processing.
Figure 3-9 Embedding artifacts. A, “Blue sponge” artifact severely distorts the tissue, often leaving large roughly triangular shaped holes with crush artifact at the edges. B, Similarly, delicate brain tissue placed in biopsy embedding bags can be mechanically parcellated into a grid of microscopic tissue squares.
Other artifacts are less common, but worth mentioning due to their potential for misinterpretation. Formalin precipitate may occasionally obscure cytologic details or be mistaken for either hemosiderin or infectious organisms ( Fig. 3-10 ). The latter mistake is made even worse when formic acid is improperly used for treating biopsy specimens when Creutzfeldt-Jacob disease is a possibility ( Fig. 3-11 ). For example, when formic acid is applied either before formalin fixation or for an excessively long time by an inexperienced pathologist or technician, the tissue can essentially dissolve, leaving behind amorphous fragments of anucleate brain resembling necrosis ( Fig. 3-11A ) or tangled balls of blood vessels that may look remarkably similar to fungal hyphae ( Fig. 3-11B ). Lesser degrees of formic acid overtreatment generally lead to a hypereosinophilic state with variable loss of nuclear detail ( Fig. 3-11C ). An additional artifact is called “laking” of red blood cells. These look like amorphous aggregates of eosinophilic material ( Fig. 3-11D ). Lastly, although less common in surgical specimens than with autopsy material, dark cell change in neurons may mimic ischemic changes ( Fig. 3-12 ). These darkly stained neurons are thought to result from manual handling of the specimen; they differ from “red dead neurons” in that they lack the hypereosinophilia of the latter and have relatively well-preserved, rather than pyknotic, nuclei.
Figure 3-10 Formalin precipitate. Formalin precipitate can mimic pigment or microscopic organisms, in this case deposited around blood vessels ( A ) and neurons ( B ).
Figure 3-11 Formic acid-associated artifacts. Severe digestion mimicking fungal hyphae. Severe overdigestion with formic acid can partially dissolve the brain parenchyma, leaving either amorphous fragments of eosinophilic brain mimicking necrosis ( A ) or only skeletonized vasculature, which resembles fungal hyphae ( B ; a higher magnification of right tissue edge seen in A ). C, Moderate overdigestion with tissue distortion. Lesser degrees of formic acid overdigestion can produce significant distortion of cytologic detail, including nuclei and other structures, such as copora amylacea as seen in this field. D, An additional formic acid artifact is red blood cell “laking” as seen here in an appropriately digested specimen, which appears as homogeneous eosinophilic intravascular material.
Figure 3-12 Dark cell change. This artifactual contraction is associated with increased basophilia of neuronal cell bodies subjected to mechanical compression, such as that resulting from overly aggressive tissue handling with forceps. Dark cell change mimics the effects of hypoxia, but the characteristic eosinophilia of the latter (“red dead neurons”) is lacking.

Iatrogenically Introduced Hemostatic and Embolic Agents
A number of agents used by interventional radiologists and neurosurgeons to control bleeding may be seen by the surgical pathologist in tissue sections, including intraoperative frozen sections. The most common of those in current clinical use are listed in Box 3-6 . Some are employed as preoperative embolic agents to minimize blood loss during subsequent surgery; these include polyvinyl alcohol (PVA), gelatin foam (Gelfoam), acrylic microspheres (Embospheres), and ethylene-vinyl alcohol with tantalum (Onyx) ( Fig. 3-13 ). Gelfoam and other agents, such as oxidized cellulose (Surgicel, Oxycel) and microfibrillar bovine collagen (Avitene), are resorbable hemostatic agents left in the surgical cavity to prevent postoperative bleeding and may appear in tissue sections from repeat surgeries in various stages of degeneration and resorption ( Fig. 3-14 ). Any of these agents may elicit a pronounced inflammatory reaction during the resorption process, which, with the accompanying edema, can mimic recurrent tumor; such mass lesions are called textilomas or gossypibomas . Microfibrillar collagen (Avitene) textilomas are sometimes accompanied by a marked eosinophilic infiltrate ( Fig. 3-14D ). Another artificial material sometimes encountered by the pathologist in repeat operation specimens of brain tumors that have been previously treated by surgical resection followed by intracavitary chemotherapy are Gliadel wafers, which are biopolymer wafers impregnated with a chemotherapeutic agent (carmustine, BCNU) ( Fig. 3-15 ).
Figure 3-13 Preoperative embolic agents. A, Polyvinyl alcohol (PVA) is a very common agent used for preoperative embolization of highly vascular lesions. B, Gelatin foam (Gelfoam), another popular preoperative embolic agent, has a very distinctive morphology that is easily recognizable in tissue sections. C, Acrylic microspheres (Embospheres) are also quite distinctive and can be seen as a single brightly eosinophilic sphere occluding a vessel in cross-section, or as chains of spheres in vessels sectioned longitudinally, as illustrated here. D, Onyx is composed of an ethylene-vinyl polymer as the occlusive agent, to which black tantalum powder has been added. The latter provides for easy visibility when preparing the agent for injection, facilitates identification of “Onyx casts” on postembolization scans, and also allows quick and unambiguous identification by the pathologist on tissue sections as the only black-colored embolic agent.
Figure 3-14 Intraoperative hemostatic agents. A, Gelatin foam (Gelfoam) is available in both particulate and sheetlike forms and is widely employed as an embolic agent and as an intraoperative hemostatic. It is frequently seen in the fresh state in surgical specimens and cavitational ultrasonic surgical aspiration (CUSA) specimens, and in various states of degradation and resorption in re-excision specimens. B, Surgicel, like all of the other embolic and hemostatic agents, displays a unique, distinctive morphology, with intertwined bundles of filaments cut in cross-section and longitudinal section. C, Bovine collagen (Avitene) is unique among the embolic/hemostatic agents in being of animal origin. In tissue sections, it appears as thick strands of amphophilic material that show a tendency to “fray” at the ends into smaller fibers. D, Eosinophilic reaction accompanying Avitene. The bovine source of Avitene likely accounts for the frequently robust eosinophil component of the accompanying inflammatory reaction.
Figure 3-15 Biopolymer wafers (Gliadel). A, Magnetic resonance imaging (MRI) study. Gliadel wafers are a polymer in which chemotherapeutic agents are embedded. They are seen as sharply rectangular profiles lining a previous resection cavity in this fluid-attenuated inversion recovery (FLAIR) MRI scan. B, Gliadel wafers in tissue section. Portions of two translucent Gliadel wafers are seen in this field. Note the Surgicel that is also present. Gliadel wafers are placed along the walls of a resection cavity to deliver chemotherapy to the adjacent parenchyma; a hemostatic agent, such as Surgicel or Gelfoam, is typically used to hold the wafers in place along the cavity wall.

Box 3-6     Common Iatrogenically Introduced Hemostatic and Embolic Agents Seen in Neurosurgical Tissue Specimens
Gelfoam
Surgicel
Avitene
Polyvinyl alcohol (PVA)
Embospheres
Onyx

The Bottom Line: What Does the Surgeon Need to Know?
Much has been written over the years regarding the most appropriate indications for intraoperative histopathologic consultation . At the time of frozen section, it is important to ask “What does the surgeon really need to know?” Another way to pose this question is: “Will the information conveyed in the consultation affect the course of the operation?” The answer to these questions varies for each clinical scenario and often requires experience to address with certainty. However, in many instances, the frozen section interpretation will dramatically alter the course of the operation ( Box 3-7 ). For many procedures, such as a stereotactic biopsy, a definitive diagnosis will essentially end the operation, whereas an equivocal or nondiagnostic report will require its continuation.

Box 3-7     Some Common Valid Indications for Intraoperative Consultation
Is adequate, representative lesional tissue present?
Do the histopathologic features suggest an infectious process?
If the disease is a neoplasm, is it of a type for which the optimal treatment is gross total resection?
Is adequate, representative lesional tissue present ? “ Adequate tissue ” implies sufficient tissue to perform any and all anticipated ancillary tests, such as special stains, immunohistochemistry, fluorescence in situ hybridization (FISH), or polymerase chain reaction (PCR) testing. “ Representative tissue ” implies that in the judgment of the pathologist, having reviewed the clinical history and, most importantly, the preoperative imaging findings, the morphologic features seen in the frozen section are likely to be representative of the disease process and can account for the imaging abnormalities. For example, if the preoperative MRI scans indicate the presence of a ring-enhancing mass, but the frozen section shows features suggestive of a low-grade glioma or, even worse, only rare atypical cells suggestive of tumor, then it is highly likely that the tissue will not be fully representative on permanent sections, and the surgeon should be so informed. In the eventuality of either insufficient quantity or quality of tissue, additional tissue should be requested for histopathologic evaluation if safely feasible.
Do the frozen section findings suggest an infectious process ? A “yes” answer to this question most often implies that the surgeon should submit additional sterile tissue samples directly from the operating room to the microbiology laboratory.
Is the disease process a neoplasm of a type that is amenable to gross total resection ? For some brain tumor types, complete surgical resection is not a therapeutic goal. Examples include primary CNS lymphoma and small cell carcinoma, which are treated with radiation therapy or chemotherapy (or both). For other tumors, gross total resection (or even an aggressive subtotal resection or debulking) is the optimal therapy. Obviously, identification of one of these tumor types at intraoperative consultation is of critical importance to the surgeon, who would in most such cases proceed with the resection. A list of some of the most common brain tumors for which surgical resection is the treatment of choice at initial clinical presentation is shown in Box 3-8 . Likewise, the surgeon will probably abort any “heroic attempt” toward aggressive resection if a suggested lymphoma or other tumor for which resection is not the optimal treatment is diagnosed on frozen section ( Box 3-9 ).

Box 3-8     Representative Brain and Spinal Cord Tumors Treated by Surgical Resection
Angiocentric glioma
Cerebellar liponeurocytoma
Central neurocytoma
Chordoid glioma
Circumscribed and dorsally exophytic brainstem gliomas
Desmoplastic cerebral astrocytoma of infancy
Dysembryoplastic neuroepithelial tumor
Dysplastic gangliocytoma of the cerebellum (Lhermitte-Duclos disease)
Ependymoma
Ganglioglioma
Hemangioblastoma
Meningioma
Myxopapillary ependymoma
Papillary tumor of the pineal region
Paraganglioma of the filum terminale
Pilocytic astrocytoma
Pleomorphic xanthoastrocytoma
Subependymoma

Box 3-9     Representative Brain and Spinal Cord Tumors Treated with Radiation Therapy or Chemotherapy Rather Than Aggressive Surgical Resection
Diffuse gliomas
Lymphoma (primary or metastatic)
Metastatic small cell carcinoma
Germinoma
Are there other valid indications for intraoperative consultation that provide the surgeon with needed critical information while the patient is still in the operating theater? In general, to reiterate the principal dictum, any histopathologic information that will influence the surgeon’s decision to terminate, extend, or alter the surgical procedure constitutes a valid indication. For example, under some circumstances the surgeon may alter the attempt to aggressively resect a lesion based on the identification of high-grade features. Any help the pathologist can provide ultimately benefits the patient. A recently added indication has been the use of Gliadel therapy, in which chemotherapeutic wafers are placed into the newly formed surgical cavity immediately after resection. Since this approach is approved only for glioblastoma and not other types of gliomas, the surgeon may require a specific diagnosis at frozen section, rather than the more general one of “high-grade glioma.”
Lastly, it is worth discussing the “ curiosity frozen section .” Bear in mind that the main purpose of the frozen section is to guide intraoperative management. Therefore, if only a small sample is submitted and the surgeon informs the pathologist that this is all the tissue that they will receive, there is no justification for jeopardizing the chances of reaching an accurate final diagnosis by expending precious tissue on a frozen section based solely on a “desire to know the diagnosis ahead of time.” On the other hand, there is no harm in freezing a small fraction of the tissue from a larger specimen to provide a preliminary diagnosis that will help the surgeon and oncologist begin postoperative treatment planning, facilitate an informed discussion with the patient and family members, and help the pathologist plan for any anticipated special studies. As is always the case, the best way to approach intraoperative consultation is through direct, open communication with the surgeon about the needs and implications of a frozen section diagnosis within the context of the individual circumstances. As in all aspects of medicine, the patient’s welfare is always of primary importance.

Suggested Readings

Adesina, A. M. Intraoperative consultation in the diagnosis of pediatric brain tumors. Arch Pathol Lab Med. . 2005; 129:1653–1660.
Burger, P. C. Go to the operating room. Am J Surg Path. . 1988; 12:961. [[Editorial]].
Powell, S. Z. Intraoperative consultation, cytologic preparations, and frozen section in the central nervous system. Arch Pathol Lab Med. . 2005; 129:1635–1652.
Ribalta, T., McCutcheon, I. E., Neto, A. G., . Textiloma (gossypiboma) mimicking recurrent intracranial tumor. Arch Pathol Lab Med. . 2004; 128:749–758.

References

1. White, V. A., Trotter, M. J. Intraoperative consultation/final diagnosis correlation: relationship to tissue type and pathologic process. Arch Pathol Lab Med. . 2008; 132:29–36.
2. Brainard, J. A., Prayson, R. A., Barnett, G. H. Frozen section evaluation of stereotactic brain biopsies. Arch Pathol Lab Med. . 1997; 121:481–484.
3. Brommeland, T., Lindal, S., Straume, B., . Does imprint cytology of brain tumours improve intraoperative diagnosis? Acta Neurol Scand. . 2003; 108:153–156.
4. Burger, P. C. Use of cytologic preparations in the frozen section diagnosis of central nervous system neoplasia. Am J Surg Pathol. . 1995; 9:344–354.
5. Folkerth, R. D. Smears and frozen sections in the intraoperative diagnosis of central nervous system lesions. Neurosurg Clin North Am. . 1994; 5:1–18.
6. Olasode, B. J., Ironside, J. W. The brain smear, a rapid affordable intraoperative diagnostic technique for brain tumors appropriate for Africa. Trop Doct. . 2004; 34:223–225.
7. Yachnis, A. T. Intraoperative consultation for nervous system lesions. Semin Diagn Pathol. . 2002; 19:192–206.
4
Neuroradiology

The Surrogate of Gross Neuropathology
Franz J. Wippold, II

Basic Noninvasive Diagnostic Imaging Techniques   47
Advanced Noninvasive Diagnostic Imaging Techniques   51
Basic Invasive Diagnostic Imaging Techniques   52
Advanced Invasive Therapeutic Techniques   53
Imaging Patterns in Neuroradiology   53
Pattern 1—The Intracranial Mass  54
Pattern 2—The Calcified Mass  58
Pattern 3—The Cystic Mass  58
Pattern 4—The Hemorrhagic Mass  58
Pattern 5—The Rim-Enhancing Mass  59
Advanced Strategies of Lesion Analysis   59
Conclusion   61
In 1895, the German physicist Wilhelm Konrad Röntgen (1845–1923) published his groundbreaking research on a newly appreciated and mysterious form of electromagnetic radiation, which he designated the “x-ray.” To his amazement, these x-rays could penetrate the body’s soft tissues and reveal the bones concealed from the naked human eye. Recognized with the first Nobel Prize in physics in 1901, Röntgen launched an era of medical discovery that in little more than a century has resulted in astounding in vivo anatomic and physiologic imaging capabilities.
Although conventional radiography using exposed film viewed with a light box was long the mainstay of medical imaging, the typical armamentarium for the imaging specialist has vastly expanded in the past few decades. Computed tomography (CT) and magnetic resonance imaging (MRI) have now emerged as the prime noninvasive methods for exploring normal and diseased brains. Moreover, hardware and software developments have transformed these initially tedious and labor-intensive methods into ones that can be quickly performed and readily obtained in most community hospitals. These modalities typically display the bony and soft tissue anatomy in slice format, but can also reconstruct digital data into stunning three-dimensional (3-D) displays. Advanced cross-sectional techniques are moving beyond mere anatomic display and are now exploring functional aspects of the brain. Functional MRI is able to record brain activation as patients perform simple tasks and MR spectroscopy (MRS) can analyze brain metabolites. Additionally, injectable radioisotope pharmaceuticals reveal nuances of brain metabolism in the technique known as positron emission tomography (PET). Molecular imaging uses biomarker probes coupled with imaging tools, such as PET, to explore various molecular pathways in the brain implicated in preclinical and disease states. Optical imaging , still in its infancy, uses the absorption and scattering of visible or infrared light to analyze the chemical composition and physiologic processes of the brain. Despite advances in digital cross-sectional and functional imaging, older techniques remain essential. For example, invasive catheter angiography portrays the vascular anatomy in great detail and remains a prime method of nonsurgical treatment of vascular conditions such as aneurysms. Interestingly, both noninvasive CT and MR vascular imaging have advanced sufficiently to replace catheter angiography in many diagnostic applications. All of these methods complement and supplement the traditional physical examination and have become fundamental steps in the evaluation of neurologic disease. As the frontier of neuroimaging approaches the tissue and cellular level, the neuroradiologist and neuropathologist are finding much common ground in charting a mutually rewarding partnership.

Basic Noninvasive Diagnostic Imaging Techniques
Conventional radiography was the primary method of examination for many decades until the advent of computer-assisted cross-sectional imaging in the 1970s. In conventional radiography, a penetrating x-ray beam traverses a body part. The energy of the beam is differentially absorbed depending on the density of the tissue through which it passes. The exiting beam then exposes a film or electronic detector. The differential absorption of the x-ray beam is known as attenuation. The different attenuations account for the varied shades of gray on the exposed film or, more recently, the display monitor. The gray scale corresponds to the tissues imaged. For example, bone has a high attenuation and typically appears white or light gray, whereas air has virtually no attenuation and appears black. 1, 2 Conventional radiography best reveals bony anatomy and is still used in screening trauma patients. Because of the relatively minimal differences in attenuations of the intracranial tissues, conventional radiography has limited application in neuroradiology. Calcified lesions may appear dense in conventional radiography; however, noncalcified brain tumors are virtually invisible on a skull film. Historically, infusions of air into the subarachnoid space, in a process known as pneumoencephalography , and radiopaque contrast material, in the technique known as myelography , have been used to increase anatomic information. Although myelography is still widely used, especially when coupled with cross-sectional imaging, pneumoencephalography has been abandoned in favor of CT and MRI.
Godfrey N. Hounsfield (1919–2004) and Allan M. Cormack (1924–1998) shared the 1979 Nobel Prize in physiology and medicine for the development of CT, a truly revolutionizing tool that transformed neuroimaging. The method of image creation in CT relies on a beam of x-rays passing through the patient similar to conventional radiography. Rather than collecting the information in the form of film exposures as required in conventional radiography, a computer analyzes the exiting beam and reconstructs an anatomic slice in CT. CT boasts excellent spatial resolution; that is, it reliably displays small anatomic parts exceedingly well ( Fig. 4-1 ). CT not only displays bony anatomy well, but also can differentiate soft tissue densities much better than conventional radiography. Therefore, CT is the technique of choice in evaluating the skull base and calvarium for subtle signs of hyperostosis, bony erosion, and remodeling, or foraminal enlargement ( Fig. 4-2 ). CT also depicts cerebral tissue moderately well, especially comparatively hyperdense material such as blood or hypodense material such as edema ( Figs. 4-3 , 4-4 ). Lesions containing calcium, such as meningiomas, are also fairly well detected with CT. However, CT falters when attempting to differentiate soft tissues with similar densities, such as gray matter and white matter. For example, infiltrating non-necrotic, noncalcified masses with little vasogenic edema or early infarctions with subtle cytotoxic edema may be poorly seen with CT ( Fig. 4-5 ).
Figure 4-1 Normal temporal bone anatomy, 0.3-mm axial computed tomography (CT) scan. High-resolution CT of the right temporal bone at the level of the internal auditory canal displays exquisitely detailed anatomy, including the modialis of the cochlea ( arrowhead ), malleus ( short arrow ), and incus ( long arrow ). High-resolution CT is excellent for imaging the bony anatomy of the temporal bone. Note that the cerebellum and right temporal lobe soft tissue detail is quite poor using the bone algorithm.
Figure 4-2 Acute myloid leukemia (granulocytic sarcoma). Twenty-three-year-old man with acute myeloid leukemia in the skull base. A, High-resolution bone algorithm computed tomography (CT) scan of the skull base reveals erosion of the right mastoid bone ( arrowhead ) with soft tissue filling the mastoid air cells. Soft-tissue fullness is also identified in the right parapharyngeal space ( arrow ). B, Axial CT scan at lower slice level with soft-tissue windows reveals a contrast-enhancing mass in the right parapharyngeal space ( arrowheads ), encasing the carotid artery. C, Contrast-enhanced magnetic resonance imaging (MRI) slice demonstrates a contrast-enhancing mass in the right parapharyngeal space ( arrowheads ) D, Positron emission tomography (PET) scan fused with the corresponding CT slices reveals hypermetabolism within the right skull base mass ( arrowheads ). Multiple modalities are helpful in evaluating lesions. The CT scan best displays detailed bony anatomy, whereas the MRI scan demonstrates soft-tissue changes. PET adds further information regarding the metabolic activity of the lesion.
Figure 4-3 Acute epidural hematoma. Three-year-old child who sustained head trauma. A, Computed tomography (CT) scan reveals an ellipsoid-shaped dense hematoma over the right cerebral convexity. The lateral ventricles are shifted to the left, and the right lateral ventricle is effaced. B, Bone window examination demonstrates a fracture of the parietal bone ( arrowhead ). Acute blood is typically dense on CT images and progressively becomes less dense with time.
Figure 4-4 Mycobacterial abscesses. Forty-five-year-old man with acquired immunodeficiency syndrome, pulmonary Mycobacterium kansasii infection, and brain abscesses that responded to appropriate antibiotics. A, Axial unenhanced computed tomography (CT) image shows multiple ring-configured lesions surrounded by hypodense vasogenic edema. Mass effect effaces the right occipital horn and partially effaces the left frontal horn B, Unenhanced T1-weighted magnetic resonance imaging (MRI) demonstrates hypointense bilateral masses with surrounding edema corresponding to the CT image. C, Axial T2-weighted fluid-attenuated inversion recovery (FLAIR) image demonstrates hyperintense vasogenic edema surrounding the ring-configured masses. D, Following administration of intravenous contrast medium, the enhanced T1-weighted axial MRI reveals multiple ring-configured lesions. The differential diagnosis would include metastases or abscesses. E, Diffusion-weighted images reveal hyperintense central portions of the lesions, suggesting abscesses rather than necrotic tumors. Note that the enhanced T1-weighted image greatly increases conspicuousness of lesions compared with the unenhanced T1-weighted image and that the FLAIR sequence optimally demonstrates the edema.
Figure 4-5 Cerebral infarction. Eighty-year-old woman presenting with right hemiparesis. A, Computed tomography (CT) scan demonstrates subtle left hemisphere cytotoxic edema manifested by loss of definition of the lateral margin of the basal ganglia ( arrowhead ) and relative effacement of the left cerebral sulci. B, Diffusion-weighted magnetic resonance imaging (MRI) demonstrating hyperintensity in the distribution of the left middle cerebral artery ( arrowheads ) confirms the presence of a subacute left cerebral infarct. CT may not differentiate soft tissues in the early phase of infarction. MRI is more sensitive than CT to subtle changes caused by edema.
Administration of intravenous iodinated contrast material improves the visibility of cerebral vessels. Contrast material also improves the conspicuity of many brain lesions by taking advantage of the disrupted blood–brain barrier often associated with these lesions. The leakage of contrast material into the extracellular space increases the density of the lesion compared with the surrounding brain and is termed enhancement. 3, 4 A major advantage of CT is its rapid scanning speed. Helical or spiral technique using multirow detector technology can image the entire brain in seconds with the option of high-quality multiplanar reformations based on the raw data. CT angiography (CTA) uses CT-acquired data to display intracranial and extracranial vessels ( Fig. 4-6 ). The resolution is inferior to that of catheter angiography and requires administration of an iodinated contrast medium. Calcifications adjacent to or within vessel walls may pose difficulties in interpretation.
Figure 4-6 Computed tomographic (CT) angiogram. Using contrast-enhanced CT data acquired without invasive intravascular catheter techniques, the circle of Willis can be displayed in 3D. Using a computer workstation, the vessels can be rotated to optimally reveal pathology such as aneurysms and stenoses.
CT does have shortcomings. The images produced in older machines are usually limited to axial (transverse) and sometimes coronal display. Changing the plane of section requires the patient to adjust position. Newer scanners can produce high-resolution data sufficient to allow for multiplanar reformations from the original axially oriented acquisitions. Beam hardening, a technical artifact seen especially in older scanners in which streaks obscure the anatomy, poses significant problems in the posterior fossa. Nevertheless, CT remains a valuable method in neuroradiology.
Paul C. Lauterbur (1929–2007) and Peter Mansfield (b. 1933) shared the 2003 Nobel Prize in physiology or medicine for discoveries that led to the development of MRI. MRI exploits the properties of the atomic nucleus and generates images using an entirely different method from CT. In MRI, a magnetic field many times more powerful than the earth’s and measured in teslas (T) teams with precisely tuned radio waves to produce cross-sectional images. Hydrogen nuclei are usually targeted in clinical practice because of the abundance of hydrogen in body water. The hydrogen nuclei absorb applied energy in a process called excitation. The nuclei then achieve baseline equilibrium by releasing this energy in the form of radiowaves that are detected by antennae called coils. Relaxation consists of so-called T1 and T2 components. By adjusting the images to emphasize either the T1 properties or the T2 properties of relaxation, tissues take on different gray-scale values, which can be used when analyzing the images ( Fig. 4-7 ). No x-rays or ionizing radiation are used. The resulting MRIs resolve subtle differences in the water content of tissues, such as the gray matter and white matter. Although MRI lacks some of the spatial resolution of CT, new and promising equipment and techniques, such as the introduction of powerful 3-T magnets, which improve signal compared with 1.5-T magnets, continue to push the frontier forward. 5
Figure 4-7 Central neurocytoma. Thirteen-year-old boy with diplopia and headaches. A, Unenhanced T1-weighted magnetic resonance imaging (MRI) reveals a slightly hypointense intraventricular mass near the foramina of Monro obstructing the right lateral ventricle. Cerebrospinal fluid (CSF) is hypointense on this sequence. The gray matter is less intense than the white matter. B, The mass becomes heterogeneously hyperintense on the T2-weighted sequence. Note that cystic components of the mass and CSF are extremely hyperintense. Gray matter is more intense than white matter. C, On the T2-weighted FLAIR image, CSF is hypointense. The slightly hyperintense right frontal horn suggests elevated protein content in this obstructed portion of the ventricle. Note that the transependymal edema adjacent to the occipital horns is hyperintense ( arrowheads ) and better appreciated when compared with the traditional T2-weighted image ( B ). D, The mass becomes hyperintense following administration of intravenous gadolinium on this enhanced T1-weighted image.
As in CT, MRI portrays the brain’s anatomy in slice format. MRI depicts this anatomy in multiple planes of orientation without the need for repositioning the patient. Also, many annoying artifacts commonly encountered with CT, such as streaking in the posterior fossa caused by beam hardening or poor x-ray penetration, do not occur with MRIs. Therefore, MRI is the technique of choice for studying the brain and has become indispensible in the evaluation of brain tumors.
The typical MRI examination varies from institution to institution, but usually includes T1-weighted images , which are excellent for anatomic portrayal. T2-weighted and fluid attenuated inversion recovery (FLAIR) T2-weighted images emphasize free water changes that are observed in pathologic conditions such as tumors, infection, and edema. Water, such as cerebrospinal fluid (CSF) and edema, tend to be hypointense on T1-weighted images and hyperintense on T2-weighted images 6 (see Fig. 4-7 ). Depending on the medical indication, sequences may be obtained in the axial, coronal, or sagittal planes. Slice thickness and anatomic region of interest can also be adjusted.
MRI also uses intravenous contrast material to enhance detectability of lesions (see Figs. 4-4 , 4-7 ). Gadolinium-containing compounds are used with T1-weighted images, and the enhancement they provide is superior to that obtained with the iodine contrast medium used in CT. Newer agents such as gadobenate dimeglumine are improving lesion conspicuity beyond that of equivalent doses of older MRI contrast agents. 7 Moreover, the safety margin of gadolinium compounds generally exceeds that of iodine agents, although a rare condition known as nephrogenic systemic fibrosis/nephrogenic fibrosing dermatopathy (NSF/NFD) has been reported in several patients with renal failure, who have received gadolinium. 8 – 10
As in CT, MRI also has shortcomings. Because of the strong magnetic field, patients with ferromagnetic or electrical medical appliances such as some cerebral aneurysm clips, dorsal column stimulators, certain heart valves, cardiac pacemakers, cochlear implants, ocular metallic foreign bodies, and stapes prostheses are necessarily excluded from study because of potential appliance motion, dislodgement, and malfunction. Deaths from aneurysm clip motion within the MRI device have been reported.
MRI often fails to adequately image calcifications and cortical bone, and therefore calcified lesions, fractures, and other bony pathology may be overlooked. MRIs also require more time to complete than CT. Therefore, patient motion may degrade the scans. Elderly, restless, anxious, or critically ill patients may have difficulty tolerating prolonged periods within the magnet when they are required to remain motionless.
Low-field so-called open MRI units have become popular in recent years, but the anatomic detail of the scans produced by these machines may be insufficient to diagnose subtle lesions. Alternatively, some manufacturers are now producing short-bore 1.5-T magnets, which reduce the reported subjective sensation of claustrophobia while maintaining the superior quality of standard field strength machines.
MRI also offers angiographic examination of the large intracranial and extracranial arteries and veins. Unlike CT angiography, magnetic resonance angiography (MRA) does not require the administration of contrast material, although newer protocols are successfully using gadolinium agents ( Fig. 4-8 ). Resolution is inferior to that from catheter angiography. Nevertheless, MRA may add significant screening information about the vertebrobasilar arterial system and carotid bifurcations.
Figure 4-8 Magnetic resonance angiography (MRA). MRA can be used to noninvasively generate images of the vascular tree, such as the circle of Willis in this study.

Advanced Noninvasive Diagnostic Imaging Techniques
Diffusion-weighted MRI (DWI) produces images that are sensitized to the random motion of water molecules. 11 – 14 Processes that impede random motion of water, such as cytotoxic edema and cellular swelling in early cerebral infarction, tend to produce hyperintense diffusion-weighted regions on the scan. Other lesions that restrict diffusion include epidermoids and cerebral abscesses. 14 Diffusion information has also been useful in evaluating tumors. 15 – 17 Current applications use DWI in tumor grading, in analyzing peritumoral edema, in white matter tracking, and in determining postoperative injury.
The use of DWI in tumor grading is dependent on the tendency of higher grade tumors to have increased cellularity. 18 – 25 Cell packing increases the amount of intracellular relative to extracellular space for a given volume of brain tissue. The observed diffusion becomes progressively restricted with higher grade tumors, and the signal can be analyzed on images using DWI parameters 26, 27 ( Fig. 4-9 ). Unfortunately, overlap of signal intensity demonstrated with tumor grades and the signal arising from tumors in which grading is not directly related to cellularity, such as in atypical meningioma or lymphoma, have limited the specificity of this application of DWI. 28 Using the rationale that hypercellularity restricts diffusion, DWI has also been used to evaluate peritumoral edema. 18, 29, 30 Vasogenic edema distends the extracellular compartment with water and actually facilitates diffusion. Vasogenic edema associated with noninfiltrating tumors such as metastases and meningiomas tends to demonstrate increased diffusion, whereas infiltrating tumors such as gliomas usually restrict diffusion due to hypercellularity within the vasogenic edema. Examination of the diffusion characteristics of peritumoral edema may therefore be helpful in determining the presence of cellular infiltration.
Figure 4-9 Oligodendroglioma. Fifty-nine-year-old man with oligodendroglioma. A, Fluid-attenuated inversion recovery (FLAIR) magnetic resonance imaging (MRI) demonstrates a heterogeneously hyperintense intra-axial predominantly right occipital lobe mass. B, Corresponding diffusion-weighted image reveals heterogeneous hyperintensity within the tumor bed, indicating cell packing. The anterior portion of the tumor is especially hyperintense ( arrowhead ), suggesting active proliferation of cells. C, A relative cerebral blood volume map demonstrates hyperperfusion in the anterior portion of the mass ( arrowheads ) corresponding to the area of restricted diffusion. Caution in interpreting diffusion and perfusion data in oligodendrogliomas is warranted because low-grade tumors may display striking abnormalities reminiscent of high-grade lesions.
Venous infarcts or injury in the immediate postoperative period may also restrict diffusion. 18 On follow-up scanning, enhancement may then be observed as the infarcts evolve. This enhancement may simulate recurrent tumor, but the correct diagnosis is usually suggested by referring to previous scans that demonstrate the nature of the original injury. Also, treatment with topoisomerase inhibitors and antibodies to vascular endothelial growth factor (VEGF) may reduce levels of VEGF and cause restricted diffusion within the tumor bed. This finding should not be mistaken for tumor progression. Finally, a variation of diffusion imaging known as diffusion tensor imaging (DTI) contains directional information and can be used to track white matter fibers. 18, 23, 24, 31 – 33 Tractography may reveal white matter deflection or infiltration from adjacent tumors and can be helpful in surgical planning and post-treatment evaluation. 31, 34 – 36
Magnetic resonance spectroscopy (MRS) noninvasively measures and graphically displays tissue metabolites. Ratios of various metabolites may aid in the preoperative grading of tumors. 18, 19, 21, 37, 38 Clinical application has primarily focused on hydrogen spectra. Choline (Cho), a precursor to phosphatidylcholine, is involved in membrane turnover. A relative elevation of the Cho spectral peak within tissue implies cellular turnover and may serve as a potential tumor marker. N -acetyl-aspartate is a neuronal marker and is typically depressed in infiltrating tumors. Lactate is produced in hypoxia-driven nonoxidative glycolysis, and an elevated lactate peak may mark necrotic portions of tumors. Lipid peaks may occur in areas of membrane destruction and have been implicated in high-grade gliomas. MRS has shown promise in the metabolic mapping of tumors and may complement anatomic data from conventional imaging ( Fig. 4-10 ).
Figure 4-10 Glioblastoma. Elderly man with a left frontal glioblastoma. A, Enhanced coronal T1-weighted magnetic resonance imaging demonstrates a hyperintense left frontal lobe mass. B, Corresponding spectroscopy reveals an elevated choline (Cho) peak ( arrow ) and a depressed N -acetyl aspartate peak (NAA) ( arrowhead ), suggesting a neoplasm. Cr and Cr2, creatine peaks.
Perfusion MRI uses magnetic resonance techniques to derive information on tumor angiogenesis and capillary permeability. 15, 18 – 21 , 39 – 42 One of the most intriguing parameters of perfusion imaging is the relative cerebral blood volume (rCBV) calculated for a region of interest within the brain. For example, the tumor grade in diffuse astrocytomas has correlated with rCBV, with high grades demonstrating increased rCBV, but the precise histologic explanation for the phenomenon is uncertain 43 – 47 ( Fig. 4-11 ). Moreover, tumor grade in general cannot always be correlated with rCBV because lower grade oligodendrogliomas and meningiomas may demonstrate high rCBV ( Figs. 4 – 9 , 4-12 ). Perfusion information can also be derived from CT. 48
Figure 4-11 Recurrent glioblastoma. Fifty-eight-year-old woman with possible tumor recurrence following surgery and radiation for glioblastoma. A, Enhanced T1-weighted axial magnetic resonance imaging (MRI) demonstrates a left frontal lobe rim-enhancing mass. B, Corresponding perfusion map of relative cerebral blood volume shows increased perfusion in the peripheral part of the mass ( arrowhead ), suggesting recurrent tumor. C, Enhanced T1-weighted axial MRI inferior to ( A ) shows irregular rim enhancement. D, Perfusion is diminished in this portion of the mass ( arrowhead ), suggesting radiation necrosis.
Figure 4-12 Meningioma. Fifty-one-year-old woman with a right convexity meningioma. A, T2-weighted axial magnetic resonance imaging shows a heterogeneously hyperintense extra-axial right convexity mass that partially effaces the right lateral ventricle and displaces the midline structures to the left. B, The mass heterogeneously enhances following the administration of intravenous contrast material. C, Map of relative cerebral blood volume reveals markedly increased perfusion of the mass. D, Selective right external carotid artery catheter angiogram demonstrates hypertrophied branches of the middle meningeal artery feeding and the densely staining mass ( arrowheads ).
Functional MRI relies on the relative MR signatures of oxyhemoglobin and deoxyhemoglobin in activated cortex compared with nonactivated cortex during specific tasks such as word generation or finger tapping in order to portray areas of the brain at work 20 ( Fig. 4-13 ). This technique has been useful in mapping the motor and language cortex prior to tumor resection.
Figure 4-13 Ganglioglioma. Seventy-year-old woman with new onset of seizures. A, Enhanced T1-weighted axial magnetic resonance imaging (MRI) demonstrates an irregular enhancing mass in the medial temporal lobe. B, Functional MRI reveals activation of the motor strips indicated by the color activation map. Functional MRI is helpful in locating motor and speech areas prior to surgery.

Basic Invasive Diagnostic Imaging Techniques
Probably one of the most mature techniques currently used by neuroradiologists to portray the vascular anatomy is catheter angiography . In this invasive test a femoral artery (or in some cases, a suitable substitute such as one of the brachial arteries) is cannulated with a catheter through which iodinated contrast material is injected directly into the arteries of the neck and head. Images are made when x-rays pass through the patient and expose an image intensifier; data are then used to produce digital images. This form of angiography is considered the reference standard for all noninvasive modalities examining the blood vessels. Catheter angiography is still used to evaluate vascular patterns that may suggest a specific diagnosis, to define the feeding arteries and draining veins prior to surgery, and to plan for possible endovascular intervention ( Fig. 4-14 ). Although catheter angiography displays the blood vessels exquisitely, it does not image the surrounding soft tissues. Therefore, some of the relationships of the vessels to brain tissue and nerves must be inferred from anatomic proximity of the vessels rather than by direct visualization of the soft tissues. With the innovative strides of noninvasive imaging, catheter angiography is now used much less frequently and is almost never used as a screening test.
Figure 4-14 Glioblastoma. Sixty-five-year-old man with progressive somnolence. A, Enhanced T1-weighted axial magnetic resonance imaging (MRI) demonstrates a right temporal lobe-enhancing mass. B, Frontal catheter angiographic projection and ( C ) lateral catheter angiogram projection of a right common carotid artery injection demonstrating an intense stain ( arrowheads ) fed by internal carotid artery branches corresponding to the right temporal lobe tumor identified on MRI.

Advanced Invasive Therapeutic Techniques
Catheter angiography can also be used as a therapeutic tool. Angioplasty, balloon occlusion, therapeutic drug infusion, the placing of arterial stents, and vascular obliteration can be performed for a variety of indications such as aneurysm and hypervascular tumors. Occasionally therapeutic angiography can obviate the need for surgery. Other invasive techniques include CT-guided and MRI-guided biopsies. In these techniques, the superior anatomic information provided by cross-sectional imaging is used to guide the precise placement of the biopsy needle. CT and MRI may also be useful in planning radiation therapy. Conventional x-ray fluoroscopy, in which x-rays serve as a guidance tool, can also be used to direct needles and catheters for both diagnosis and therapy.

Imaging Patterns in Neuroradiology
To the pathologist, gross anatomic appearances found either postmortem or with large resection specimens and microscopically examined histologic patterns of cellular architecture on stained tissue preparations may imply a specific tissue diagnosis. Neuroradiologists similarly rely on pattern recognition based primarily on the gray-scale variations revealed in images (see pattern recognition tables at the front of this textbook). Unlike pathologists, neuroradiologists are usually confronted with the imaging slices of the entire brain. Specific signs of pathology on imaging may suggest specific diagnoses, although imaging signs (just like individual histopathology patterns) are rarely pathognomonic of a single disease process. Therefore, most neuroradiologists suggest a family of lesions that may have similar imaging appearances ( Box 4-1 ). Families or groups of diagnoses form differential diagnostic possibilities and are often referred to as gamuts . For example, a rim-enhancing mass may suggest an abscess, glioblastoma, metastasis, or even tumefactive demyelination. Although these diagnoses may be quite distinct clinically and pathologically, they nevertheless may appear virtually identical on cross-sectional imaging. Subtle neuroimaging clues may further refine a gamut. For example, the appreciation of a thinned margin of enhancement adjacent to the ventricular wall may suggest an abscess rather than a tumor. In addition to the imaging appearance of a lesion, a gamut may be further characterized by clinical information such as the age of the patient at presentation and the location of the lesion. For example, a rim-enhancing mass in an elderly patient with a clinical decline measured in months suggests a primary glial tumor rather than an abscess or tumefactive demyelination. Although the neuroradiologist is daily confronted with numerous and often complex imaging patterns and an exhaustive discussion is beyond the scope of this chapter, several recurrent basic patterns are useful to review.

Box 4-1     Major Imaging Patterns
Intra-axial lesion

(location, age of patient, and special features such as calcification useful in further refining differential diagnosis) -->
Primary glial neoplasm
Astrocytoma
Oligodendroglioma
Ependymoma
Subependymoma
Choroid plexus papilloma
Primary nonglial or mixed neoplasm

Ganglioglioma
Dysembryoplastic neuroepithelial tumor (DNET)
Central neurocytoma
Primitive/embryonal tumor (e.g. AT/RT)
Hemangioblastoma
CNS lymphoma
Metastatic disease
Tumefactive demyelination
Cerebral abscess
Intracerebral hematoma
Acute infarction

Extra-axial lesion

Meningioma
Hemangiopericytoma
Solitary fibrous tumor
Hemangioblastoma
Sarcomas
Schwannoma
Metastasis
Melanoma/melanocytoma
Secondary lymphoma/leukemia/plasmacytoma
Paraganglioma
Pituitary Adenoma
Sarcoidosis/granulomatous diseases
Inflammatory pseudotumors
Calcifying pseudotumor of neuraxis
Primary bone tumor
Histiocytosis (e.g., Rosai-Dorfman disease)

Intraventricular masses

(location within ventricular system may vary with age and tumor type)
Choroid plexus tumors
Ependymoma
Subependymoma
Meningioma
Central neurocytoma
Oligodendroglioma
Astrocytoma
Xanthogranuloma
Medulloblastoma

Rim-enhancing mass

Glial neoplasm
Abscess
Infarct
Metastasis
Multiple sclerosis
Lymphoma (in immunocompromised patients)
Radiation necrosis
Resolving hematomas

Intracranial cystic lesion

Arachnoid cyst
Colloid cyst
Parasitic cyst
Dermoid/epidermoid cyst
Rathke cleft cyst
Dandy-Walker cyst
Juvenile pilocytic astrocytoma
Hemangioblastoma
Metastasis
Pineal cyst
Porencephaly

Calcified mass

Oligodendroglioma
Ependymoma
Astrocytoma
Mucinous adenocarcinoma metastasis
Vascular malformation
Tuberculoma
Meningioma
Choroid plexus tumor
Craniopharyngioma
Germ cell tumor
Chondrosarcoma
Chordoma
Aneurysm (especially giant aneurysm)
Teratoma
Pineocytoma
Cortical tuber
Sturge-Weber malformation

Hemorrhagic mass

Neoplastic hemorrhage -->
Primary tumor
Metastasis (especially choriocarcinoma, melanoma, renal cell carcinoma, lung carcinoma, and breast carcinoma)
Non-neoplastic hemorrhage

Hypertension
Hemorrhagic infarct
Aneurysm
Vascular malformation
Vasculitis
Trauma
Coagulopathy
Infection
Amyloid angiopathy

Fat-containing lesion

Dermoid
Lipoma
Meningioma
Teratoma

Gyral enhancement

Reperfusion of ischemic infarct
Migraine headache
Posterior reversible encephalopathy syndrome
Seizures

Pachymeningeal (Dural) enhancement

Meningioma
Postoperative meningeal enhancement
Intracranial hypotension
Metastatic disease (especially breast and prostate)
Secondary CNS lymphoma
Sarcoid/granulomatous disease

Leptomeningeal enhancement

Meningitis
Meningeal carcinomatosis (or lymphomatosis, gliomatosis, melanomatosis, meningiomatosis, etc.)

Pattern 1—The Intracranial Mass
An intracranial mass is usually detected by appreciating a distortion of the normal anatomy of the brain or by noting an abnormal density pattern on CT or a signal intensity abnormality on MRI. The first task for the neuroradiologist when confronted with a mass lesion is to determine whether it is intra-axial or extra-axial. 20 An intra-axial mass arises from the brain parenchyma itself, whereas an extra-axial mass arises from the skull or meninges and secondarily compresses the brain. The importance of this fundamental determination resides in formulating the appropriate gamut of possible diagnoses; the gamut for an intra-axial mass is quite different from that for an extra-axial mass. Moreover, decisions involving further diagnostic tests and therapy hinge on the suggested gamut of possible causes.
The differential diagnosis for extra-axial masses is somewhat limited and includes relatively common neoplasms such as meningioma and metastasis and, rarely, hemangiopericytoma. Sarcoidosis and granulomatous infection may also present as an extra-axial mass. Intra-axial lesions include a much wider range of possibilities, such as the many glial and nonglial neoplasms. Tumefactive demyelination, infection in the form of cerebritis and abscess, and intracerebral hematoma are just a few of the many non-neoplastic entities that can manifest as an intra-axial mass. Further characterization may depend on the location of the mass, whether or not other masses are identified, and on the imaging features of the lesion. For example, the discovery of multiple masses may suggest metastases or possibly disseminated infection rather than a primary glial neoplasm. Unfortunately, many of these diagnoses have virtually identical imaging features, and biopsy is often the only method for establishing a definitive diagnosis.
Imaging signs for extra-axial lesions include a broad attachment of the mass to the inner skull table or dura, adjacent calvarial changes such as erosion or hyperostosis, buckling of the gray matter and white matter subjacent to the mass, widening of the ipsilateral subarachnoid space, a cleft of CSF separating the mass from the brain parenchyma, and deviated cortical vessels between the mass and the brain 20 (see Fig. 4-12 ). Imaging signs of intra-axial lesions include brain parenchyma surrounding the lesion, cortical vessels compressed against the inner table of the skull between the mass and the skull, and thinned cerebral cortex between the mass and the inner table of the skull (see Fig. 4-9 ). Some lesions are difficult to classify as either intra-axial or extra-axial. For example, a superficial oligodendroglioma may appear to have a broad dural contact resembling the dural attachment of a meningioma. Other lesions such as meningioangiomatosis may demonstrate both intra-axial and extra-axial features 49 ( Fig. 4-15 ). Finally, extremely large lesions may obscure some of the imaging signs due to compression of brain and obliteration of otherwise recognizable features such as cortical blood vessels.
Figure 4-15 Meningioma and meningioangiomatosis. Eight-year-old boy with seizures. A, Enhanced T1-weighted and T2-weighted ( B ) magnetic resonance imaging demonstrates an enhancing left frontal lobe mass. Portions appear extra-axial with dural thickening and apparent displacement of the subjacent cortex; however, the medial portion of the mass thickens the cortex and appears intra-axial. At surgery, both intra-axial and extra-axial components were confirmed. Pathology examination revealed meningioma, grade 1, with a meningoangiomatosis-like pattern of tumoral spread along the perivascular Virchow-Robin spaces of the adjacent cortex. (See Chapter 10 ).
Once the location of the mass is established as intra-axial or extra-axial, additional imaging signs may be used to further characterize its histology. For example, the presence of calcium, hemorrhage, cystic spaces, or fat may point to a specific diagnosis. For instance, teratomas may have significant calcium and fat deposits. Lesions with high nucleus-to-cytoplasm ratios and decreased water content, such as lymphoma, may appear dense on CT scans and display hypointensity on T2-weighted images. CT and MRI may also noninvasively reveal the effects of the mass on the brain, such as the presence of obstructive hydrocephalus, hemorrhage, ischemia, and vasogenic edema. 20 For the pathologist, inspection of preoperative imaging is especially valuable because the spatial relationships cannot typically be appreciated on the biopsy specimen alone.

Pattern 2—The Calcified Mass
The presence of calcium may be very useful in establishing a differential diagnosis or gamut. 6 Because calcium may not be readily evident on gross anatomic specimens and furthermore may be difficult to detect on limited biopsy specimens, the demonstration of calcium on imaging can provide a useful sign for the pathologist. Calcium is best detected on CT scans as hyperdensity, but can also be detected on MRI scans as signal voids on all imaging sequences ( Fig. 4-16 ). As a cautionary note, deposits of minerals such as iron and manganese may resemble calcium, and careful analysis of all of the imaging evidence is crucial for correct characterization of the finding. Once the presence of calcium is established, pinpointing its location may be a useful clue in the differential diagnosis. For example, supratentorial intra-axial calcified masses include oligodendroglioma, ependymoma, some astrocytomas, and metastatic mucinous adenocarcinomas. Arteriovenous malformations and tuberculomas may also calcify. Meningioma is the most common calcified extra-axial mass. A calcified mass in the suprasellar region may represent a craniopharyngioma, meningioma, germ cell tumor, chondrosarcoma, chordoma, or even aneurysm, whereas a calcified mass in the pineal region may suggest a teratoma or pineocytoma. An intraventricular location may imply the presence of a choroid plexus papilloma, ependymoma, meningioma, neurocytoma, oligodendroglioma, astrocytoma, or xanthogranuloma. Diffuse or focal calcification within basal ganglia without discrete mass formation may suggest a metabolic condition such as Addison’s disease, hypothyroidism, hyperparathyroidism, lead toxicity, and infection, or hereditary conditions such as tuberous sclerosis and neurofibromatosis.
Figure 4-16 Glioblastoma with oligodendroglial features. Twenty-two-year-old man with seizures. A, Nonenhanced axial computed tomography (CT) scan reveals a mass in the left frontal lobe with dense peripheral calcifications (a common feature of oligodendroglial neoplasms, though not specific) and surrounding vasogenic edema. The left frontal horn is effaced, and the midline structures are shifted to the right. B, Corresponding unenhanced T1-weighted axial magnetic resonance imaging (MRI) reveals the hypointense left frontal mass. However, the isointense calcifications are virtually invisible. C, The mass is heterogeneously hyperintense on T2-weighted axial MRI. D, The mass intensely enhances on the enhanced T1-weighted axial MRI. Note that the calvarium is dense on CT. The inner and outer tables of skull have no signal on the MRI sequences.

Pattern 3—The Cystic Mass
Cross-sectional imaging also provides an excellent method of evaluating intracranial cystic lesions. 6 Often limited biopsy specimens or even postmortem evaluations fail to adequately demonstrate cystic lesions in the context of the intact brain. For that reason, imaging may be very useful in determining the appropriate gamut. Benign non-neoplastic cysts, such as the arachnoid cyst and Dandy Walker cyst, usually demonstrate no contrast enhancement and a homogenous fluid content on both CT and MRI ( Fig. 4-17 ). Neoplastic cysts, such as those associated with juvenile pilocytic astrocytoma and hemangioblastoma, often demonstrate enhancement within the wall or within a tumor mural nodule ( Fig. 4-18 ). The density on CT or the intensity on MRI may not parallel that of cerebral spinal fluid in neoplastic cysts, and debris may be present.
Figure 4-17 Arachnoid cyst. Sixty-six-year-old woman with a large extra-axial left hemispheric cystic lesion. A, Unenhanced T1-weighted axial magnetic resonance imaging demonstrates a large extra-axial left hemisphere cystic lesion, which displaces the lateral ventricles to the right. B, Axial T2-weighted image reveals hyperintense signal within the cystic fluid. C, Following administration of contrast material, the cyst fails to enhance on the T1-weighted images. D, Diffusion-weighted imaging demonstrates unrestricted water diffusion properties. E, The fluid-attenuated inversion recovery (FLAIR) T2-weighted image further demonstrates suppression of the signal of the fluid. On all sequences, the cyst fluid is identical in intensity to cerebrospinal fluid. An epidermoid cyst would have demonstrated restricted diffusion on the diffusion-weighted image and nonwater signal on the FLAIR image.
Figure 4-18 Hemangioblastoma. Forty-four-year-old man with a cystic lesion in the posterior fossa. A, T1-weighted axial magnetic resonance imaging (MRI) shows a large intra-axial cystic lesion in the right cerebellar hemisphere displacing the fourth ventricle. B, T2-weighted axial MRI demonstrates hyperintense fluid signal. C, Enhanced T1-weighted axial image reveals the small enhancing mural nodule ( arrowhead ) that differentiates this tumor from a developmental cyst.
Central necrosis within a high-grade primary tumor, metastasis, or abscess may mimic a cyst; however, close inspection of all MRI sequences usually reveals a high protein content that differentiates the contents from the CSF-like water present in a true cyst. DWI may further demonstrate diffusion restriction in an abscess cavity, whereas a necrotic tumor cavity typically does not restrict water diffusion (see Fig. 4-4 ).

Pattern 4—The Hemorrhagic Mass
One of the important features of both CT and MRI is the ability to detect hemorrhage. This feature has obviously important ramifications in the treatment of clinically ill patients and can also assist in establishing a differential diagnosis. Hemorrhage is usually identified easily on CT scans as hyperdensity. Differentiation from calcium, which may also be hyperdense, is usually aided by the clinical context and subtle clues in the image appearance. On MRI, the appearance of hemorrhage is more complex. The signal intensity on T1-weighted and T2-weighted images may suggest the age of the hemorrhage ( Fig. 4-19 ). A difficult issue often confronting the neuroradiologist is whether hemorrhage is due to benign causes, such as hypertension, or to an underlying neoplasm. Often, this determination cannot be confidently made until the blood products resolve. Differentiation from nontumoral hemorrhage can be suggested by location. Hemorrhagic metastases often favor the gray matter/white matter junction. 50 Tumoral hemorrhage tends to be heterogeneous in MRI signal and surrounded by extensive edema compared with simple hematoma. Discovery of a coexisting second lesion with typical nonhemorrhagic signal may also be useful. Benign hematomas tend to evolve from the periphery of the clot centripetally to the center, whereas tumoral hemorrhage evolves in a much more poorly organized fashion. Also, tumoral hemorrhages tend to recur and produce blood products of various ages on CT and MRI. Hemorrhagic neoplasms tend to enhance after the administration of intravenous contrast material. Although benign hematomas may also enhance during the resolution stage, this feature is usually much less apparent than the enhancement in neoplasm.
Figure 4-19 Intracerebral hematoma. Seventy-six-year-old man presenting with left hemiparesis. A, T1-weighted magnetic resonance imaging demonstrates a deep frontal lobe hyperintense mass. B, On the T2-weighted image, the mass is also hyperintense with a thin hypointense rim. This pattern of signal suggests that the predominant blood product is extracellular methemoglobin with a thin rim of developing hemosiderin.

Pattern 5—The Rim-Enhancing Mass
One of the most intriguing imaging signs is the rim-enhancing mass. The intravenous contrast agents used in CT and MRI identify blood–brain barrier breakdown. Contrast material is sequestered in the extracellular space due to leaking capillaries. 3, 4 An avascular and therefore nonenhancing central component implies necrosis or avascular debris. The rim-enhancing pattern embraces several prominent diagnoses, such as high-grade astrocytoma and metastasis, as well as infections, such as abscess (see Figs. 4-11 , 4-16 ). Infarcts, radiation necrosis, resolving hematomas, and tumefactive demyelination may also produce a ring-configured pattern with the administration of intravenous contrast agent, and therefore careful correlation with the patient’s history and past imaging is required ( Fig. 4-20 ). Diffusion imaging is very useful in that bacterial abscesses tend to manifest restricted diffusion as hyperintensity on DWI, whereas necrotic cavities of most tumors tend to be associated with relatively unrestricted diffusion (see Fig. 4-4 ).
Figure 4-20 Tumefactive demyelination. Thirty-year-old woman with seizure and headaches. A, Enhanced T1-weighted axial magnetic resonance imaging demonstrates an incomplete enhancing rim in the right temporal lobe with additional enhancement in the right occipital lobe. Mass effect partially effaces the right lateral ventricle and shifts the midline structures to the left. B, T2-weighted fluid-attenuated inversion recovery imaging shows extensive vasogenic edema in the right temporal lobe and occipital lobe. The incomplete rim-configured enhancement is suggestive of tumefactive demyelination and differentiates this lesion from an aggressive tumor. C, A sagittal T2-weighted image of the cervical spine reveals faint hyperintensity in the cervical spinal cord at the C7 level ( arrowheads ). The finding of an additional lesion in the spinal cord is also suggestive of demyelination.

Advanced Strategies of Lesion Analysis
Having established the basic pattern of imaging findings, advanced methods can be used to further refine the differential diagnosis. 19, 21 In general, extra-axial masses have a limited list of possible differential diagnoses and rarely require the application of advanced imaging techniques. However, intra-axial masses may pose significant challenges to the neuroradiologist. The response to administration of intravenous gadolinium should be thoroughly analyzed. High-grade gliomas, intra-axial metastases, tumefactive demyelination, lymphomas, and abscesses typically enhance, whereas many low-grade neoplasms and encephalitides fail to enhance. 51, 52
For enhancing lesions, diffusion data are often helpful. Diffusion as demonstrated on DWI tends to be facilitated in tumefactive demyelination, whereas lesions with restricted diffusion, such as lymphoma and abscess, can usually be differentiated on the basis of their enhancement pattern. Lymphomas (in immunocompetent patients) usually demonstrate uniform enhancement, whereas abscesses display rim enhancement. Increased perfusion as demonstrated on perfusion MRI usually indicates a high-grade tumor or metastatic deposit, whereas abscesses and tumefactive demyelination usually do not have increased perfusion. MRS may also aid in differentiating tumor from abscess and in suggesting grade of tumor. Certainly exceptions to these guidelines do exist. For example, low-grade oligodendrogliomas may demonstrate increased perfusion and MRS may fail to adequately demonstrate convincing spectra due to numerous technical issues such as the location of the tumor.

Conclusion
In summary, cross-sectional imaging has revolutionized the diagnostic approach to brain tumors. CT and MRI are now widely available and greatly increase the detectability and differentiation of brain tumors. Emerging advanced imaging techniques such as MRS, functional MRI, and perfusion MRI are further aiding the diagnosis and grading of lesions. The once well-demarcated roles of the neuropathologist and the neuroradiologist are becoming somewhat less distinct as available diagnostic tools are proving beneficial to both specialties.

Selected Readings

Al-Okaili, R. N., Krejza, J., Wang, S., Woo, J. H., Melhem, E. R. Advanced MR imaging techniques in the diagnosis of intraaxial brain tumors in adults. Radiographics . 2006; 26(Suppl 1):S173–189.
Al-Okaili, R. N., Krejza, J., Woo, J. H., . Intraaxial brain masses: MR imaging-based diagnostic strategy—initial experience. Radiology . 2007; 243:539–550.
Cha, S. Update on brain tumor imaging: from anatomy to physiology. AJNR Am J Neuroradiol. . 2006; 27:475–487.
Inoue, T., Ogasawara, K., Beppu, T., Ogawa, A., Kabasawa, H. Diffusion tensor imaging for preoperative evaluation of tumor grade in gliomas. Clin Neurol Neurosurg. . 2005; 107:174–180.
Young, R. J., Knopp, E. A. Brain MRI: tumor evaluation. J Magn Reson Imaging . 2006; 24:709–724.

References

1. Huda, W., Slone, R. M. Analog X-ray imaging. In Review of Radiologic Physics , 2nd ed, Philadelphia: Lippincott Williams & Wilkins; 2003:49–67.
2. Huda, W., Slone, R. M. X-ray production. In Review of Radiologic Physics , 2nd ed, Philadelphia: Lippincott Williams & Wilkins; 2003:16–33.
3. Gado, M. H., Phelps, M. E., Coleman, R. E. An extravascular component of contrast enhancement in cranial computed tomography. Part I. The tissue-blood ratio of contrast enhancement. Radiology . 1975; 117:589–593.
4. Gado, M. H., Phelps, M. E., Coleman, R. E. An extravascular component of contrast enhancement in cranial computed tomography. Part II. Contrast enhancement and the blood-tissue barrier. Radiology . 1975; 117:595–597.
5. Krautmacher, C., Willinek, W. A., Tschampa, H. J., . Brain tumors: full- and half-dose contrast-enhanced MR imaging at 3.0 T compared with 1.5 T—Initial experience. Radiology . 2005; 237:1014–1019.
6. Bird, C. R., Dean, B. Pathology of brain neoplasms. Top Magn Reson Imaging . 1989; 1:6–15.
7. Knopp, M. V., Runge, V. M., Essig, M., . Primary and secondary brain tumors at MR imaging: bicentric intraindividual crossover comparison of gadobenate dimeglumine and gadopentetate dimeglumine. Radiology . 2004; 230:55–64.
8. Lauenstein, T. C., Salman, K., Morreira, R., . Nephrogenic systemic fibrosis: center case review. J Magn Reson Imaging . 2007; 26:1198–1203.
9. Kanal, E., Broome, D. R., Martin, D. R., Thomsen, H. S. Response to the FDA’s May 23, 2007, nephrogenic systemic fibrosis update. Radiology . 2008; 246:11–14.
10. Partain, C. L. On the potential causal relationship between gadolinium-containing MRI agents and nephrogenic systemic fibrosis. J Magn Reson Imaging . 2007; 25:879–880.
11. Bammer, R. Basic principles of diffusion-weighted imaging. Eur J Radiol. . 2003; 45:169–184.
12. Neil, J. J. Measurement of water motion (apparent diffusion) in biological systems. Concept Magn Reson. . 1997; 9:385–401.
13. Neil, J. J. Diffusion imaging concepts for clinicians. J Magn Reson Imaging . 2008; 27:1–7.
14. Castillo, M., Mukherji, S. K. Diffusion-weighted imaging in the evaluation of intracranial lesions. Semin Ultrasound CT MR . 2000; 21:405–416.
15. Provenzale, J. M., Mukundan, S., Barboriak, D. P. Diffusion-weighted and perfusion MR imaging for brain tumor characterization and assessment of treatment response. Radiology . 2006; 239:632–649.
16. Higano, S., Yun, X., Kumabe, T., . Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis. Radiology . 2006; 241:839–846.
17. Murakami, R., Sugahara, T., Nakamura, H., . Malignant supratentorial astrocytoma treated with postoperative radiation therapy: prognostic value of pretreatment quantitative diffusion-weighted MR imaging. Radiology . 2007; 243:493–499.
18. Cha, S. Update on brain tumor imaging: from anatomy to physiology. AJNR Am J Neuroradiol. . 2006; 27:475–487.
19. Al-Okaili, R. N., Krejza, J., Woo, J. H., . Intraaxial brain masses: MR imaging-based diagnostic strategy—initial experience. Radiology . 2007; 243:539–550.
20. Young, R. J., Knopp, E. A. Brain MRI: tumor evaluation. J Magn Reson Imaging . 2006; 24:709–724.
21. Al-Okaili, R. N., Krejza, J., Wang, S., . Advanced MR imaging techniques in the diagnosis of intraaxial brain tumors in adults. Radiographics . 2006; 26(suppl 1):S173–S189.
22. Inoue, T., Ogasawara, K., Beppu, T., . Diffusion tensor imaging for preoperative evaluation of tumor grade in gliomas. Clin Neurol Neurosurg. . 2005; 107:174–180.
23. Field, A. S., Alexander, A. L. Diffusion tensor imaging in cerebral tumor diagnosis and therapy. Top Magn Reson Imaging . 2004; 15:315–324.
24. Field, A. S., Alexander, A. L., Wu, Y. C., . Diffusion tensor eigenvector directional color imaging patterns in the evaluation of cerebral white matter tracts altered by tumor. J Magn Reson Imaging . 2004; 20:555–562.
25. Guo, A. C., Cummings, T. J., Dash, R. C., Provenzale, J. M. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology . 2002; 224:177–183.
26. Kono, K., Inoue, Y., Nakayama, K., . The role of diffusion-weighted imaging in patients with brain tumors. AJNR Am J Neuroradiol. . 2001; 22:1081–1088.
27. Gauvain, K. M., McKinstry, R. C., Mukherjee, P., . Evaluating pediatric brain tumor cellularity with diffusion-tensor imaging. AJR Am J Roentgenol. . 2001; 177:449–454.
28. Stadnik, T. W., Chaskis, C., Michotte, A., . Diffusion-weighted MR imaging of intracerebral masses: comparison with conventional MR imaging and histologic findings. AJNR Am J Neuroradiol. . 2001; 22:969–976.
29. Krabbe, K., Gideon, P., Wagn, P., . MR diffusion imaging of human intracranial tumours. Neuroradiology . 1997; 39:483–489.
30. Lu, S., Ahn, D., Johnson, G., Law, M., . Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology . 2004; 232:221–228.
31. Pierpaoli, C., Jezzard, P., Basser, P. J., . Diffusion tensor MR imaging of the human brain. Radiology . 1996; 201:637–648.
32. Jellison, B. J., Field, A. S., Medow, J., . Diffusion tensor imaging of cerebral white matter: a pictorial review of physics, fiber tract anatomy, and tumor imaging patterns. AJNR Am J Neuroradiol. . 2004; 25:356–369.
33. Melhem, E. R., Mori, S., Mukundan, G., . Diffusion tensor MR imaging of the brain and white matter tractography. AJR Am J Roentgenol. . 2002; 178:3–16.
34. Roberts, T. P., Liu, F., Kassner, A., . Fiber density index correlates with reduced fractional anisotropy in white matter of patients with glioblastoma. AJNR Am J Neuroradiol. . 2005; 26:2183–2186.
35. Lazar, M., Alexander, A. L., Thottakara, P. J., . White matter reorganization after surgical resection of brain tumors and vascular malformations. AJNR Am J Neuroradiol. . 2006; 27:1258–1271.
36. Smith, J. S., Cha, S., Mayo, M. C., . Serial diffusion-weighted magnetic resonance imaging in cases of glioma: distinguishing tumor recurrence from postresection injury. J Neurosurg. . 2005; 103:428–438.
37. Stadlbauer, A., Gruber, S., Nimsky, C., . Preoperative grading of gliomas by using metabolite quantification with high-spatial-resolution proton MR spectroscopic imaging. Radiology . 2006; 238:958–969.
38. Sibtain, N. A., Howe, F. A., Saunders, D. E. The clinical value of proton magnetic resonance spectroscopy in adult brain tumours. Clin Radiol. . 2007; 62:109–119.
39. Le Bihan, D., Douek, P., Argyropoulou, M., . Diffusion and perfusion magnetic resonance imaging in brain tumors. Top Magn Reson Imaging . 1993; 5:25–31.
40. Taylor, J. S., Tofts, P. S., Port, R., . MR imaging of tumor microcirculation: promise for the new millennium. J Magn Reson Imaging . 1999; 10:903–907.
41. Choyke, P. L., Dwyer, A. J., Knopp, M. V. Functional tumor imaging with dynamic contrast-enhanced magnetic resonance imaging. J Magn Reson Imaging . 2003; 17:509–520.
42. Cha, S. Perfusion MR imaging of brain tumors. Top Magn Reson Imaging . 2004; 15:279–289.
43. Law, M., Brodsky, J. E., Babb, J., . High cerebral blood volume in human gliomas predicts deletion of chromosome 1p: preliminary results of molecular studies in gliomas with elevated perfusion. J Magn Reson Imaging . 2007; 25:1113–1119.
44. Law, M., Oh, S., Johnson, G., . Perfusion magnetic resonance imaging predicts patient outcome as an adjunct to histopathology: a second reference standard in the surgical and nonsurgical treatment of low-grade gliomas. Neurosurgery . 2006; 58:1099–1107. [discussion 1099–1107].
45. Law, M., Yang, S., Wang, H., . Glioma grading: sensitivity, specificity, and predictive values of perfusion MR imaging and proton MR spectroscopic imaging compared with conventional MR imaging. AJNR Am J Neuroradiol. . 2003; 24:1989–1998.
46. Young, R., Babb, J., Law, M., . Comparison of region-of-interest analysis with three different histogram analysis methods in the determination of perfusion metrics in patients with brain gliomas. J Magn Reson Imaging . 2007; 26:1053–1063.
47. Shin, J. H., Lee, H. K., Kwun, B. D., . Using relative cerebral blood flow and volume to evaluate the histopathologic grade of cerebral gliomas: preliminary results. AJR Am J Roentgenol. . 2002; 179:783–789.
48. Hoeffner, E. G., Case, I., Jain, R., . Cerebral perfusion CT: technique and clinical applications. Radiology . 2004; 231:632–644.
49. Moreno, A., de Felipe, J., Garcia Sola, R., . Neuronal and mixed neuronal glial tumors associated to epilepsy. A heterogeneous and related group of tumours. Histol Histopathol. . 2001; 16:613–622.
50. Williams, K. D., Drayer, B. P., Bird, C. R. Magnetic resonance imaging in the diagnosis of intracerebral hematoma. BNI Q. . 1989; 5:16–23.
51. Runge, V. M., Muroff, L. R., Jinkins, J. R. Central nervous system: review of clinical use of contrast media. Top Magn Reson Imaging . 2001; 12:231–263.
52. Watanabe, M., Tanaka, R., Takeda, N. Magnetic resonance imaging and histopathology of cerebral gliomas. Neuroradiology . 1992; 34:463–469.
5
Astrocytic and Oligodendroglial Tumors
Daniel J. Brat,
Arie Perry

Introduction and Brief Historical Overview   63
Diffuse Astrocytomas   63
Pilocytic Astrocytoma   82
Subependymal Giant-Cell Astrocytoma   88
Pleomorphic Xanthoastrocytoma   91
Oligodendroglioma   93
Oligoastrocytoma   99

Introduction and Brief Historical Overview
The infiltrative, or “diffuse,” gliomas include astrocytomas, oligodendrogliomas, and mixed oligoastrocytomas. 1, 2 Taken together, these are the most frequent neoplasms of the central nervous system (CNS). 3 Bailey and Cushing first classified these lesions by their cellular differentiation patterns in 1926, and this approach was improved upon over the ensuing decades by using histopathologic features to grade and prognosticate. 4 A variety of schemes for classifying and grading the diffuse gliomas have been employed, including those of Kernohan, Zülch, Ringertz, Burger, and St. Anne-Mayo. The World Health Organization (WHO) classification is the continually updated international standard that will be referred to in this chapter. 1 The WHO classification divides infiltrating gliomas into astrocytomas, oligodendrogliomas, and oligoastrocytomas and includes criteria for their grading. Infiltrative astrocytoma (WHO grade II), anaplastic astrocytoma (WHO grade III), and glioblastoma (GBM; WHO grade IV) form a malignancy continuum for the diffusely infiltrating astrocytomas. The oligodendrogliomas and mixed oligoastrocytomas include WHO grade II and III designations, with the mixed subtype including a grade IV variant as well (GBM with an oligodendroglial component). These diffuse gliomas are nearly as clinically devastating as they were when they were first classified almost 100 years ago. In large part, this is because of their widespread invasiveness, their strong tendency toward biologic progression, and the resistance of many of these neoplasms to conventional adjuvant therapies. A significant challenge that remains for neuropathologists is the reproducible diagnosis of the diffuse gliomas, which still hinges primarily on the histopathologic analysis of hematoxylin and eosin (H&E) slides.
The term “astrocytoma” has also been applied more broadly and includes tumors with astrocytic differentiation that are lower grade and better circumscribed than the diffuse astrocytomas. These include the pilocytic astrocytoma, pleomorphic xanthoastrocytoma (PXA), and subependymal giant-cell astrocytoma (SEGA). Each of these has a distinct localization, histology, and natural history, which is generally more favorable than that for the diffuse astrocytomas.

Diffuse Astrocytomas


Definitions and Synonyms
The infiltrative or “diffuse” forms of astrocytoma occur throughout the CNS, generally in adults, and are most frequent in the cerebral hemispheres. These lesions are composed of individual tumor cells that infiltrate widely throughout the brain parenchyma with a cellular density and degree of anaplasia that increase with tumor grade. These tumors have an inherent tendency to progress to a higher grade, with glioblastoma representing the most malignant form. Because of the inability to totally resect these tumors and their resistance to conventional therapies, they are almost universally fatal, with length of survival depending on many factors including tumor grade.

Incidence and Demographics
Diffuse astrocytic tumors (WHO grades II–IV) comprise roughly 60% of primary intracranial tumors, with an annual incidence of 5 to 7 per 100,000 person-years. 3 There is a slight male predominance (1.3:1). These tumors can arise at any age in children and the very elderly, although incidence increases substantially with advancing age. Older patients are also more likely to have higher grade gliomas, especially glioblastoma.
Grade II tumors (diffuse astrocytomas) account for roughly 5 to 10% of the infiltrative astrocytomas, with a peak incidence between 30 and 40 years of age. Grade III tumors (anaplastic astrocytomas) account for approximately 10% and have a peak incidence in the mid-40s. Glioblastoma (GBM; grade IV) is the most frequent neoplasm in this category, accounting for approximately 80% of the diffusely infiltrative astrocytomas. GBMs can progress from lower grade (II or III) astrocytomas (progressive GBMs, 10%) or can arise as de novo tumors without a clinically apparent precursor lesion (90%). 5 GBMs are most common between the ages of 45 and 75, with a median age of diagnosis of 60.

Clinical Manifestations and Localization
Like other diseases of the CNS, the clinical presentation of the diffuse astrocytomas varies according to the sites of involvement and the rate of growth. The most common site is within the cerebral hemispheres , with a slight predilection for the frontal and parietal lobes, and a lower frequency in the occipital lobes. These lesions are most often centered in the subcortical white matter, but have a tendency to infiltrate widely and include the cerebral cortex, deep gray structures, and even the contralateral hemisphere. In children, and also less commonly in adults, the diffuse forms of astrocytoma can arise in the brainstem , thalamus , and basal ganglia . These tumors may also occur in the spinal cord of adults , but the cerebellum is a highly unusual site.
The most common clinical symptoms are new-onset seizures, change in behavior, motor deficits, and signs/symptoms of increased intracranial pressure (e.g., headaches, nausea and vomiting, papilledema). High-grade (WHO grade III–IV) astrocytomas tend to have short histories with rapid progression, whereas low-grade (WHO grade II) examples are more indolent, often with insidious onset and a long, protracted clinical course.

Radiologic Features and Gross Pathology
Grade II astrocytomas are most commonly seen on magnetic resonance imaging (MRI) as ill-defined, deep-seated, or predominantly subcortical cerebral hemispheric masses. They are typically nonenhancing lesions, best appreciated on T2-weighted or fluid-attenuated inversion recovery (FLAIR) MRI sequences, where the signal hyperintensity reflects vasogenic edema generated in response to diffuse infiltration by individual tumor cells ( Fig. 5-1 ). 6, 7 Due to their infiltrative nature, microscopic disease is almost always present beyond the tumor margins defined by MRI ( Fig. 5-2 ). Secondary signs of mass effect include midline shift, ventricular compression, and sulcal effacement. Anaplastic astrocytomas (grade III) may be radiographically identical or may show faint, punctate, or irregular foci of contrast enhancement ( Fig. 5-3A ). Glioblastomas commonly show a “ring-enhancing” or “rim-enhancing” pattern with a central low-density region of necrosis surrounded by an irregular, variable-thickness rim of contrast enhancement ( Fig. 5-4A ). Importantly, this rim-enhancing component is always surrounded by T2- or FLAIR signal hyperintensity that represents an associated diffusely infiltrating neoplasm. Those GBMs that cross the corpus callosum are often referred to as “butterfly lesions” due to involvement of the white matter in the centrum semiovale bilaterally. Occasionally, diffusely infiltrative astrocytomas have multiple small separate foci of contrast-enhancement and might be considered by some as “ multifocal glioma .” In fact, these do not usually represent multiple distinct tumors, but are instead multiple foci of high-grade progression to GBM within a single widely disseminated tumor.
Figure 5-1 Infiltrative astrocytoma, WHO grade II on magnetic resonance imaging (MRI). A, T1-weighted MRI with gadolinium reveals a subtle area of low signal intensity in the right frontoparietal basal ganglia, white matter, and cortical regions. B, This same region is much better visualized by its increased signal intensity on T2-weighted MRI. Note the ill-defined margins.
Figure 5-3 Anaplastic astrocytoma, WHO grade III. A, Postcontrast T1-weighted magnetic resonance imaging showing modest, focal contrast enhancement within a much larger region of hypodensity in the left basal ganglia and temporal lobe. B, Anaplastic astrocytomas are cytologically and immunohistochemically similar to WHO grade II astrocytomas, but are generally more cellular and have mitotic figures ( arrows ). C, The MIB-1 (Ki-67) immunostain shows an increased labeling index, commonly reaching over 5%.
In treated gliomas, foci of radiation necrosis are radiographically and grossly very similar to GBM and may produce considerable neurologic worsening due to mass effect ( Fig. 5-5A ). Therefore, the differential of tumor recurrence and progression versus radiation necrosis is a common clinical dilemma (see Chapter 19 for greater detail). Special techniques, such as positron emission tomography (PET), single-photon emission computerized tomography (SPECT), diffusion-weighted imaging (DWI), and perfusion studies may be useful for further distinguishing these two diagnostic considerations, yet each has its shortcomings and a biopsy may ultimately be necessary. 8
Figure 5-5 Radiation necrosis in previously treated diffuse gliomas. A, Radiation necrosis in a treated glioblastoma. The postcontrast magnetic resonance image shows a heterogeneously enhancing mass, which can be difficult to distinguish from tumor recurrence. B, C, Radiation of glioblastoma causes large zones of coagulative and fibrinoid necrosis ( B ) with parenchymal rarefaction and vascular changes, including telangiectasias, hyalinization, and hemorrhage ( C ).
Grossly, the diffuse astrocytomas are ill-defined and subtly discolored, with secondary mass effects identical to those seen radiographically. As described earlier, there are typically foci of microscopic disease beyond the grossly suspected borders that are invisible to the naked eye. GBMs are classically variegated, with foci of necrosis and hemorrhage.

Histopathology
Diffuse astrocytomas may have a wide spectrum of cell types in pure or mixed forms (see Histologic Variants and Grading section), though the fibrillary variant is the prototype, composed of elongate, irregular hyperchromatic nuclei, often with no discernable cytoplasm ( “ naked nuclei ” ) , but embedded in a dense fibrillary matrix (see Figs. 5-2 through 5-4 ). Alternatively, these cells may have a modest number of visible eosinophilic cytoplasmic processes that blend with surrounding neuropil. In establishing the diagnosis of a diffuse astrocytoma, critical attention must be given to defining an astrocytoma as infiltrative before applying grading criteria to a biopsy or resection specimen in which an astrocytic neoplasm is present. Noninfiltrative forms of astrocytomas such as pilocytic astrocytoma, subependymal giant-cell astrocytoma, and pleomorphic xanthoastrocytoma (see Differential Diagnosis section) are relatively compact and discrete compared with diffusely infiltrative astrocytomas, yet they cause diagnostic challenges because of overlapping morphologic features. Infiltration of CNS parenchyma by astrocytic tumor cells is readily identified when neoplastic cellularity is low and normal elements of the brain (i.e., neuronal cell bodies or axons) are seen in the background. A low- to high-cellularity gradient of astrocytoma cell density from regions of non-neoplastic brain tissue is also helpful in documenting infiltration. When biopsy material consists of a high density of tumor cells that obscure background CNS microarchitecture, infiltration can be difficult to appreciate on H&E sections. In these instances, identification of tumor cells between axons, highlighted either by silver stains or immunohistochemistry for neurofilament protein, can assist in establishing infiltration ( Fig. 5-2E ).
The lowest grade infiltrative or diffuse astrocytomas (WHO grade II) are characterized by a modestly increased cellularity as a result of the presence of neoplastic cells percolating through CNS parenchyma. 9 – 11 Smear preparations can be helpful in demonstrating fibrillar tumoral processes, which defines their glial differentiation ( Fig. 5-2A ). Nuclei are enlarged , hyperchromatic , and have irregular contours compared with non-neoplastic astrocytes ( Fig 5-2A–D ). Their slightly oblong shape and irregular contours contrast with the round regular nuclei of classic oligodendroglioma. Architectural distortion of involved CNS parenchyma in the form of edematous splaying of neuropil or microcyst formation often accompanies low-grade infiltrative astrocytomas (see Fig. 5-2D ). Microcalcifications are also occasionally noted. In other cases, histologic findings can be subtle, with only mild hypercellularity and slight disturbance of neuropil microarchitecture. Secondary structures of Scherer are less commonly encountered in astrocytomas than they are in oligodendrogliomas, but include subpial condensation, perineuronal satellitosis, and perivascular aggregation (see Major Patterns in the Introduction of this book).
Anaplastic astrocytoma (AA, WHO grade III) is distinguished from grade II infiltrating astrocytoma by the presence of mitotic activity ; additionally, there is generally increased cellularity and a higher degree of nuclear pleomorphism and atypia ( Fig. 5-3B ). 11 Conceptually, it represents the transition from grade II infiltrative tumors to GBM. Correlation of histopathologic findings with the neuroimaging finding is always recommended, because the biopsy specimen may contain a nonrepresentative portion of a larger and more aggressive lesion (i.e., GBM). Histologic differences between a grade II and grade III astrocytoma can be subtle, and distinction occasionally hinges on the finding of only occasional mitoses (see Histologic Grading and Variants section). 12
Glioblastoma is the highest grade form of infiltrating astrocytoma and often harbors diagnostically useful genetic alterations ( Fig. 5-6 ; see Genetics section). In addition to the histopathologic findings of AA, either microvascular hyperplasia (a.k.a., microvascular proliferation, endothelial proliferation, or endothelial hyperplasia) or necrosis (often both) is seen (see Fig. 5-4 ). 13 Necrosis can be in the form of pseudopalisading , with a dense packing of neoplastic cells around a necrotic center, generally with their long axis perpendicular to necrosis ( Fig. 5-4G ), or it can be broad, infarct-like areas ( Fig. 5-4H ). Within short distances of pseudopalisades, there is usually evidence of microvascular hyperplasia in a pattern that mirrors the margins of the pseudopalisading cells. 14 Microvascular hyperplasia can also be seen at the infiltrating edge of GBMs and consists predominantly of rapidly dividing endothelial cells that form tufted microaggregates. Occasionally, these aggregates form multiple lumens similar to renal glomeruli and are referred to as “glomeruloid” ( Fig. 5-4C–F ). 15 Some mistakenly use the term “microvascular hyperplasia” to refer only to an increased number of blood vessels. Although increased numbers of vessels may indeed be present in high-grade astrocytomas, for the purpose of grading astrocytic neoplasms, microvascular hyperplasia takes on a specific meaning and refers to the morphologic finding of proliferating endothelial and perivascular cells forming tufted aggregates that emerge from larger parent vessels.

Frozen Section Diagnosis of Diffuse Gliomas
The role of a frozen section diagnosis of diffuse gliomas is to guide the neurosurgeon at the time of the operation, to ensure that diagnostic tissue has been obtained, and to give the most accurate intraoperative diagnostic interpretation, acknowledging limitations of sampling and of the technique. Interpretation should be in the context of clinical history, radiographic features, and neurosurgical findings. Frozen sections are not an optimal technique for detecting cytologic features of an infiltrating glioma, especially those that distinguish oligodendrogliomas from astrocytomas. In particular, the features of oligodendrogliomas, such as perinuclear halos, delicate chromatin pattern, and nuclear regularity, are not as evident in frozen tissue. These nuclear features are better appreciated on smears (see Chapter 3 ), but in most instances, the distinction between oligodendroglial and astrocytic differentiation at frozen section is not critical, and the diagnosis of “infiltrating glial neoplasm” is sufficient for guiding intraoperative management. Definitive classification and grading of glial neoplasms is most accurate following examination of all tissue submitted for permanent sections, as tissue examined at frozen section may not represent the entire disease process. Nonetheless, a general degree of histologic differentiation and grade can usually be derived by assessing the cellular density, nuclear anaplasia, mitotic activity, microvascular hyperplasia, or necrosis.
The process of freezing tissue for intraoperative diagnosis introduces artifacts that remain in permanent sections and can limit their interpretation. Most notably, nuclei appear more hyperchromatic and atypical in previously frozen tissue; perinuclear halos of oligodendroglioma are not as evident; and the overall cytologic resolution is lower. Therefore, a portion of sampled tumor tissue should be reserved for permanent sections without freezing. If it is not clear at the time of frozen section whether additional tissue will be available for permanent sections, it is prudent to freeze only a portion of the tissue submitted for frozen section.

Histologic Variants and Grading
Although the majority of diffuse astrocytomas, including GBMs, have “fibrillary” morphology and are composed of tumor cells with elongated processes on cytologic and histologic preparations, there is a great deal of histomorphologic diversity (see Figs. 5-7 through 5-14 ) ( Box 5-1 ). Descriptors such as gemistocytic , small cell , protoplasmic , sarcomatous , epithelioid , granular cell , and giant cell have been applied to specific variants of the infiltrative astrocytomas. Whether each category of neoplasm behaves in a distinct manner, either in terms of biologic aggressiveness or response to therapy, is not clear in all cases.

Box 5-1     Diffuse Astrocytoma Variants

Fibrillary Astrocytoma (see Figs. 5-2 through 5-4 ) -->
Figure 5-4 Glioblastoma, WHO grade IV. A, Postcontrast magnetic resonance imaging of glioblastoma showing deep-seated irregular rim-enhancing mass in the frontal lobe. B, Intraoperative sampling from a viable portion of a glioblastoma produces a cellular smear with cytologic features of fibrillary astrocytoma (compare with Fig. 5-2A) . C, Viable portion of a glioblastoma, showing marked cellularity, mitotic activity, and typical microvascular hyperplasia characterized by small multilayered vessels with plump (hyperplastic) endothelia. D, “Glomeruloid” bodies represent larger tufts of microvascular hyperplasia with multiple lumens. However, the multilayering is the more important diagnostic feature, rather than the multiple lumens resembling glomeruli. E, F, Immunohistochemical studies show that the microvascular proliferation consists not only of CD34-positive endothelia ( E ) but also smooth muscle actin-positive cells ( F ). G, Pseudopalisading necrosis in a glioblastoma. H, Infarct-like necrosis in glioblastoma. I, Occasional glioblastomas show foci with perivascular pseudo-rosette-like structures. This should not be misconstrued as ependymal differentiation if other features are classic for GBM. J, Immunostaining for WT1 is often more coarse and widespread than in lower grade astrocytic neoplasms (compare with Fig. 5-2J ).
Figure 5-2 Infiltrative astrocytoma, fibrillary subtype, WHO grade II. A, Intraoperative smear preparation reveals cells with elongate to irregular, hyperchromatic nuclei, some of which display long cytoplasmic processes. B, C, At the infiltrative edge, scattered individual or clustered mildly enlarged, elongated, irregular and hyperchromatic “naked nuclei” ( arrowheads ) are visible, along with native non-neoplastic cells, such as cortical neurons ( arrows ). D, Even in regions composed predominantly of tumor cells, entrapped linear axons are evident ( arrows ). Microcystic changes are also seen. E, The infiltrative nature is further supported by the presence of numerous entrapped axons on neurofilament protein immunostain. F, A glial fibrillary acidic protein stain highlights irregular clusters of neoplastic cells with variable quantities of cytoplasm, along with a few coarse processes. G, In contrast, a comparison case of gliosis demonstrates evenly spaced reactive astrocytes with numerous radially arranged processes, some of which terminate on blood vessels (astrocytic endfeet). H, Immunohistochemistry for p53 is often strong in subsets of neoplastic nuclei. I, The Ki-67 labeling index is often low, but staining of enlarged atypical nuclei provides additional support for diagnosis of astrocytoma, rather than a reactive process. J, The presence of wispy WT1-positive cytoplasmic processes also supports a neoplastic astrocytic process, since reactive astrocytes are usually immunonegative.
Prototype of diffuse astrocytomas
Cells with inconspicuous cytoplasm embedded in a densely fibrillary matrix (“naked nuclei”) or with thin GFAP-positive cytoplasmic processes
Enlarged elongate hyperchromatic nuclei with irregular nuclear contours
Hypocellular examples may be difficult to distinguish from reactive gliosis
Gemistocytic Astrocytoma (see Fig. 5-7 )
Figure 5-7 Gemistocytic astrocytomas. Common features include perivascular lymphocytic cuffing ( A ) and pleomorphic mono- to multinucleate gemistocytic astrocytes displaying eccentric bellies of eosinophilic cytoplasm and nuclei ranging from irregular and hyperchromatic to oval and vesicular with prominent nucleoli ( B ). Most examples show strong nuclear immunoreactivity for p53 protein ( C ). Gemistocytic astrocytomas may present with features of WHO grade II, III, or IV neoplasms, although the low-grade examples often progress to grade III or IV more rapidly than their fibrillary astrocytoma counterparts.

Eccentrically placed eosinophilic, GFAP-positive cytoplasm with short, polar cytoplasmic processes, round, variably hyperchromatic nuclei, and asymmetrical distribution of tumor cells (unlike reactive gemistocytes)
Associated perivascular inflammation
Rare mitoses, but high-grade examples frequently associated with proliferating small-cell astrocytes
High rate of malignant progression to glioblastomas
Granular Cell Astrocytoma (see Fig. 5-8 )
Figure 5-8 Granular cell astrocytomas. These rare tumors are characterized by large granular to clear foamy cells resembling macrophages ( A, B ). Clues to the correct diagnosis include intervening cells that resemble fibrillary astrocytoma ( A ) and frequent mitoses ( B ; arrows ). As with other lysosome-rich cells, they are positive with both periodic acid–Schiff with diastase ( C ) and CD68 ( D ) stains. Note in the latter however, that the smaller true histiocytes are more intensely positive. In contrast, more specific histiocytic markers, such as CD163 are typically negative in the tumor cells, but positive in infiltrating macrophages ( E ). Tumor cells also express glial fibrillary acidic protein ( F ). This variant may have features of astrocytomas of WHO grades II, III, or IV, although clinically, the behavior is very aggressive regardless of the grade assigned.

Large lysosome rich cells, imparting a granular or macrophage-like appearance
Deceptively bland, but often has fibrillary astrocytoma cells in background
High-grade radiologic and histologic features are common
Clinical behavior is aggressive (GBM-like) regardless of histologic grade
Granular cells are PAS+, CD68±, CD163-, GFAP+
Protoplasmic Astrocytoma

Thin, wispy processes creating a cobweb-like growth pattern
Oval, mildly hyperchromatic nuclei
Less widely infiltrative than other variants
Usually low-grade (WHO grade II)
Not accepted as a distinct variant by all pathologists, due to some overlapping features with oligodendrogliomas and pilocytic astrocytomas
Giant-Cell Astrocytoma/GBM (see Fig. 5-9 )
Figure 5-9 Giant-cell glioblastoma (GBM). Common features include superficial location and relatively sharp demarcation on postcontrast T1-weighted magnetic resonance imaging ( A ), large bizarre mono- and multinucleate tumor cells with intervening smaller astrocytoma cells ( B, C ), glial fibrillary acidic protein immunoreactivity ( D ), and increased Ki-67 labeling predominantly in, but not limited to the smaller tumor cells ( E ). In other areas, this tumor had classic features of GBM, including pseudopalisading necrosis.

Mono- and multinucleate giant cells with abundant cytoplasm and bizarre nuclei
Often deceptively circumscribed grossly
Usually WHO grade IV
Previously known as “monstrocellular glioblastomas”
Small subset of patients with more favorable survival
Gliosarcoma (see Fig. 5-10 )
Figure 5-10 Gliosarcomas. A-D, At low magnification, this gliosarcoma with a leiomyosarcoma-like component is relatively demarcated, but has blurred borders consistent with at least focal infiltration of adjacent brain parenchyma ( A ). The glial element is strongly glial fibrillary acidic protein (GFAP) positive ( B ), whereas the sarcomatous portion is reticulin-rich ( C ) and expresses smooth muscle actin ( D ). E-G, Another example displays intermixed features resembling fibrosarcoma ( E ; left ) and fibrillary astrocytoma ( E ; right ), the former highlighted on reticulin stain ( F ; outer portions reticulin-rich and inner glial element reticulin-poor), while the latter is highlighted on a GFAP stain ( G ; glial elements positive, sarcomatous elements negative). H, Yet another example displays mixed features of gemistocytic astrocytoma and chondrosarcoma. In other areas, these tumors had classic features of GBM, including pseudopalisading necrosis.

WHO grade IV tumor with both astrocytic and sarcomatous components
Thought to represent a form of mesenchymal metaplasia in gliomas, analogous to carcinosarcoma arising in epithelial neoplasms
Sarcomatous element is usually pleomorphic fibrosarcoma, but may show bone, cartilage, fat, or muscle differentiation
Often superficial and deceptively circumscribed
Prognosis and response to therapy does not differ significantly from conventional glioblastoma
Adenoid (Epithelioid) GBM or Gliosarcoma (see Fig. 5-11 )
Figure 5-11 Adenoid (epithelioid) glioblastomas (GBMs) and gliosarcomas. GBMs and gliosarcomas may have foci ranging from a mere resemblance of metastatic carcinoma (epithelioid) to true epithelial metaplasia, the latter being either squamous or glandular in nature. A, GBM with glandlike epithelioid nests. B, small foci of squamous metaplasia in a gliosarcoma. C, strong CK7 immunoreactivity in epithelioid nests within an adenoid GBM (other keratins showed a similar pattern). D, glial fibrillary acidic protein positivity in the glial elements of an adenoid GBM.

Foci of classic GBM or gliosarcoma
Foci of epithelioid cytology or
Foci of true glandular or squamous differentiation
Expression of cytokeratins, including CK7 in some cases
Glial portions typically GFAP-positive
Does not alter standard prognosis or therapy for GBM
Small-Cell Astrocytoma/GBM (see Fig. 5-12 )

Small cells with minimal cytoplasm and oval mildly hyperchromatic nuclei
Deceptively bland chromatin and perinuclear haloes mimics oligodendroglioma
Brisk mitotic/proliferative indices
Usually WHO grade IV
Subset present as nonenhancing masses and appear WHO grade III on biopsy specimen, but typically behave like GBM nonetheless
EGFR amplification in ≈70% and 10q deletion in >90%
GBM with an Oligodendroglioma Component (see Fig. 5-13 )

High-grade mixed oligoastrocytoma with foci of tumor necrosis (WHO grade IV)
The oligodendroglioma cells should have classic cytology in order to avoid misconstruing a focus of small-cell GBM as oligodendroglial (see Fig. 5-12 )
Mucin-rich microcystic spaces are common, as are foci resembling lower grade mixed oligoastrocytoma
Better prognosis than conventional (purely astrocytic) GBM
Worse prognosis than anaplastic mixed oligoastrocytoma
Malignant Glioma with PNET-like Foci (see Fig. 5-14 )

Hypercellular nodules of small blue cells within an otherwise infiltrative glioma
Homer Wright rosettes in a subset
Anaplastic/large-cell features common in the PNET-like areas (pleomorphism, cell wrapping, large nuclei, prominent nucleoli, aggregates of apoptotic cells)
Glial element most often resembles astrocytoma, GBM, or gliosarcoma
PNET-like areas variably express multiple neuronal markers
Most are strongly p53-immunoreactive with Ki-67LI > 80% in the primitive element
High incidence of CSF dissemination
Chromosome 10q deletions are common in both tumoral components
Either N- myc (most common) or c- myc gene amplifications are seen in the PNET-like foci in ≈40% of cases
Gemistocytic astrocytomas ( Fig. 5-7 ) are composed of cells with abundant glassy pink cytoplasm and prominent stout cellular processes, while granular cell astrocytomas ( Fig. 5-8 ) are populated by large round histiocyte-like tumor cells packed with eosinophilic granules that correspond to autophagic vacuoles at the ultrastructural level. 16, 17 Both gemistocytic and granular cell variants seem to have more aggressive clinical behavior, in that WHO grade II gemistocytic astrocytomas often progress to higher grades more rapidly than the fibrillary subtype, whereas patients with granular cell astrocytoma often die in less than a year, regardless of the histologic grade assigned.
Giant-cell glioblastoma ( Fig. 5-9 ) is regarded as a distinct clinicopathologic variant, presenting clinically as a peripheral cerebral hemispheric mass that is better circumscribed than other types of GBM. 18 Its slightly more favorable prognosis may be related to greater resectability. Tumor cells in this variant are extremely large, misshapen, multinucleated, and separated by a well-developed reticulin network. Nuclei are equally bizarre and often demonstrate markedly atypical mitoses. Gliosarcoma ( Fig. 5-10 ) is a morphologic form of GBM that is best regarded as a biphasic tumor consisting of malignant glial and mesenchymal components, the latter resulting from metaplasia of the former. 19, 20 These lesions are typically more firm and discrete because of the high content of reticulin and collagen in the sarcomatous component. There does not seem to be any major prognostic difference between gliosarcoma and typical GBM. Epithelial metaplasia ( Fig. 5-11 ) is less common, but may be encountered in either GBMs or gliosarcomas (often referred to as adenoid or epithelioid ). This phenomenon ranges from mere epithelioid cytology to true glandular or squamous differentiation, either type capable of mimicking metastatic carcinoma.
The small-cell GBM ( Fig. 5-12 ) is composed of a high density of monotonous, modestly sized, and highly proliferative tumor cells with minimal cytoplasm, oval nuclei, delicate chromatin, and only mild hyperchromasia, resulting in a deceptively bland cytology. 21, 22 These lesions manifest more often in the elderly and have a rapid clinical progression, even if microvascular hyperplasia and necrosis are not evident on biopsy material. In contrast, GBMs with an oligodendroglial component ( Fig. 5-13 ) display foci with more uniformly rounded nuclei. Although there is great morphologic overlap between these last two subtypes, the presence of mucin-rich microcystic spaces and high-grade oligodendroglial cytology (enlarged epithelioid cells with increased cytoplasm, large nuclei, open chromatin, and prominent nucleoli) favors GBM with an oligodendroglial component.
Diffuse, high-grade gliomas with distinct PNET-like foci have also recently been described ( Fig. 5-14 ). 23 These uncommon tumors are also sometimes confused with the small-cell variant of GBM, but typically display foci of classic diffuse astrocytoma (or other glioma subtype) with more discrete nodules of primitive cells containing oval to carrot-shaped hyperchromatic nuclei, anaplastic/large-cell features similar to those of medulloblastoma, or Homer Wright (neuroblastic) rosettes (see Fig. 5-14A–C ). The primitive component appropriately displays convincing evidence of neuronal differentiation ( Fig. 5-14D ), may show evidence of myc gene amplifications ( Fig. 5-14H ), and like medulloblastoma or CNS PNET in general, commonly seeds the CSF.
Gliomatosis cerebri is a clinicopathologic diagnosis that implies extensive concurrent involvement of multiple lobes (at least three) or brain compartments by an infiltrating glioma ( Fig. 5-15 ). 24 The large majority of examples are composed of neoplastic cells with astrocytic differentiation, yet tumors with oligodendroglial differentiation have also been described. Microscopically these tumors usually resemble fibrillary astrocytomas of grade II or III. Long, thin, mildly hyperchromatic “microglia-like” astrocytoma nuclei are typical ( Fig. 5-15C ). In many instances, secondary structures, including subpial or subependymal condensation, perivascular aggregates, and perineuronal satellitosis are prominent. Although gliomatosis cerebri is characterized by widespread involvement by diffusely infiltrating astrocytoma, there may be multiple smaller foci of dedifferentiation to GBM, either at the time of clinical presentation or later in the disease progression.
Figure 5-15 Gliomatosis cerebri. This clinicopathologically defined, extensively infiltrative variant requires radiographic demonstration of involvement of at least three lobes, along with pathologic documentation of diffuse glioma on biopsy. A, B, T2-weighted magnetic resonance imaging shows widespread and bilateral hyperintensities of the cerebral hemispheres, in this case extending from the superior frontoparietal white matter ( A ) to the inferomedial temporal lobes ( B ). C, Specimen from stereotactic needle biopsy demonstrates that the pathology was consistent with anaplastic astrocytoma, WHO grade III (mitoses present elsewhere). The prominence of these characteristic “rodlike” nuclei that are long, thin, twisted and “microglia-like” is common in cases of gliomatosis cerebri.
A number of grading schemes have been applied for the diffuse forms of astrocytoma. It should be stated that the application of grade is independent of the histologic differentiation patterns listed earlier (i.e., gemistocytic, small-cell, etc.). Most pathologists today utilize the WHO scheme . 1 Adopted from a simple St. Anne-Mayo approach, four criteria are utilized, the “AMEN criteria” of a typia, m itoses, e ndothelial hyperplasia, and n ecrosis. 25 Thus, according to the WHO, grade II lesions are characterized as infiltrating astrocytomas that have nuclear atypia only (see Fig. 5-2 ). Anaplastic astrocytomas (WHO grade III) additionally have mitotic activity (see Fig. 5-3 ). The number of mitoses necessary for a grade III designation has been debated, but one mitotic figure is generally enough in a small biopsy specimen (e.g., from stereotactic needle biopsy), whereas more (>1) are required in larger resection specimens. In addition to the histopathologic findings of AA, either microvascular hyperplasia or necrosis (or both), are required for the diagnosis of GBM (see Fig. 5-4 ). GBMs with microvascular hyperplasia but no necrosis have been shown to have similar behavior to GBMs with necrosis. However, in the majority of cases both of these findings are seen together.

Differential Diagnosis
The most common and most challenging alternatives for consideration in diagnosing diffuse astrocytomas are other diffuse gliomas, including oligodendrogliomas and mixed oligoastrocytomas . This is especially true among the low- and intermediate-grade gliomas. In one study, concordance among four experienced neuropathologists was only 50% in classifying and grading oligodendroglioma, astrocytoma, and oligoastrocytoma. 26 Agreement improved modestly to 70% after the pathologists reviewed the cases together and discussed diagnostic criteria to facilitate consistency. Other investigations have demonstrated poor intraobserver reproducibility and an unacceptably wide variation in the diagnostic frequency of astrocytoma, oligodendroglioma, and mixed oligoastrocytoma, even among the most experienced of neuropathologists.
The primary means of distinguishing diffuse astrocytoma from oligodendroglioma is based on cytologic features. Oligodendrogliomas contain rounded regular nuclei, bland chromatin, small nucleoli, and clear perinuclear haloes. 9 In contrast, the nuclei of astrocytoma are enlarged, elongate, irregular, and heterogeneously hyperchromatic. When the main differential alternative is oligodendroglioma, the finding of p53 immunoreactivity or trisomy 7 favors the diagnosis of astrocytoma in low-grade examples, whereas amplification of epidermal growth factor receptor (EGFR) or loss of chromosome 10q are more specific for high-grade astrocytoma. Codeletion of 1p and 19q is typical of oligodendroglioma.
Other conditions that need to be ruled out both at the time of frozen section and on permanent sections are macrophage-rich processes, such as demyelinating disease and stroke . Although biopsies are not typically performed for these diseases when the clinical presentation is classic, unusual clinical presentations, such as a masslike growth pattern or an atypical location often lead to neurosurgical biopsy. Both of these diseases are characterized by a prominent infiltrate of macrophages —a finding that should strongly discourage a neoplastic diagnosis. At the time of intraoperative consultation, a smear preparation is extremely helpful in the recognition of macrophages (see Chapter 22 ). Indeed, in this situation, smear preparations are often more valuable than frozen section slides, since a macrophage infiltrate can appear shockingly similar to an infiltrating astrocytoma in frozen material. On permanent sections of tissue that has not been previously frozen, macrophages are more evident. In many instances, it is prudent to perform immunohistochemistry for macrophages (CD68 or Ham56) in order to exclude or highlight the infiltrate.
In the setting of an intra-axial, poorly differentiated malignant neoplasm, the differential diagnosis usually includes GBM, metastatic carcinoma or melanoma , and primary CNS lymphoma . The latter tumors are GFAP-negative, being highlighted instead by epithelial, melanocytic and lymphoid markers, respectively. As stated earlier, neurofilament is also helpful for highlighting the infiltrative nature of astrocytomas as opposed to metastases, which tend to push axon-rich parenchyma to the side rather than infiltrate. Primary CNS lymphomas can have infiltrative patterns similar to the diffuse gliomas, and the cytologic features can also overlap. However, lymphomas are characterized by a distinct angiocentricity that is rare in high-grade astrocytomas. In post-treated astrocytomas, a common differential possibility is that of tumor recurrence or progression versus radiation necrosis . Histologically, there is often a combination of each, with some suggestion that the prognosis is improved when radiation necrosis predominates. Although it is not always possible to distinguish radiation necrosis from tumor necrosis, the former is typified by large geographic zones of infarct-like necrosis unassociated with nuclear pseudopalisading, with or without dystrophic calcification ( Fig. 5-5B ). Other radiation effects include parenchymal rarefaction and vascular changes, including telangiectasias, hyalinization, and fibrinoid necrosis of vessel walls ( Fig. 5-5C ).
In cases of low-grade (WHO II) astrocytomas with low cellularity and minimal atypia, the distinction from reactive gliosis may be difficult. Features supportive of neoplasia include radiographic features of diffuse glioma, nuclear enlargement and hyperchromasia (see Fig. 5-2A–C ), nuclear clustering (see Fig. 5-2C ), clustering of GFAP-positive cells (see Fig. 5-2D ) rather than evenly spaced astrocytes with radially arranged processes (see Fig. 5-2E ; reactive gliosis pattern), p53 immunoreactivity ( Fig. 5-2H ), and increased MIB-1 proliferative index, including some positivity in cytologically abnormal nuclei ( Fig. 5-2I ). Another potentially useful, recently identified marker is the Wilm’s tumor antibody, WT1. 27 In addition to the normal endothelial reactivity (positive internal control), the majority of astrocytomas of any grade display at least some wispy cytoplasmic immunoreactivity, whereas non-neoplastic astrocytes are typically negative ( Fig. 5-2J ).

Ancillary Diagnostic Studies
GFAP is an intermediate filament expressed by normal glial cells and by glial neoplasms, which can be identified reliably using immunohistochemistry ( Figs. 5-2F, 5-8F, 5-9D, 5-10B, 5-10G, 5-11D ). 28 Since GFAP can be detected in nearly all malignant gliomas and is negative in nearly all carcinomas, lymphomas, melanomas, and sarcomas, a positive GFAP stain of tumor cells supports the diagnosis of glioma in the setting of a malignant CNS neoplasm. GFAP is nearly 100% sensitive as a marker of glial differentiation. In some cases, however, it is difficult to distinguish a positive reaction in native non-neoplastic tissue or reactive astrocytes from that resulting from the tumor. GFAP is expressed more intensely and more frequently in astrocytomas than in oligodendrogliomas, including both their low- and high-grade forms. Without good evidence, however, GFAP has sometimes been regarded as a “marker” of astrocytic, rather than oligodendroglial, differentiation. More recent investigations have demonstrated that neoplastic oligodendroglioma cells, especially “minigemistocytes” and “gliofibrillary” oligodendrocytes, can also show GFAP staining (see Fig. 5-26D ). 29 Thus, GFAP is not a reliable marker for distinguishing oligodendrogliomas from astrocytomas. As mentioned in the prior section, astrocytes are typically evenly spaced and display radially oriented processes in reactive astrocytosis ( Fig. 5-2G ). In contrast, clustering of astrocytes with only rare irregular or course processes favors glioma (see Fig. 5-2F ).
Figure 5-25 Anaplastic oligodendroglioma, WHO grade III. This temporal lobe mass displayed extensive cortical and subcortical involvement, was bright on T2-weighted magnetic resonance imaging ( A ), and had patchy enhancement on postgadolinium T1-weighted images ( B ). Areas of cyst formation were bright on T2 and dark on T1 sequences ( arrowheads ), while foci of calcification showed signal voids on all sequences ( arrow ).
Caution must be taken in the interpretation of the immunohistochemical identification of cytokeratin expression in malignant neoplasms involving the CNS. 28, 30 Cytokeratin expression usually indicates epithelial differentiation and it might be assumed that its expression would support the diagnosis of metastastic carcinoma rather than GBM in the setting of a poorly differentiated malignancy. However, malignant gliomas often show immunoreactivity to cytokeratins, especially AE1/3 , which is a commonly used antibody preparation that recognizes numerous cytokeratin types (i.e., a pancytokeratin marker). This immunoreactivity is thought to represent cross-reactivity of the antibodies with GFAP. Epithelial differentiation in metastatic carcinoma is best documented using cytokeratin antibodies that are not positive in astrocytomas, such a CAM5.2 , or other epithelial markers, such as epithelial membrane antigen (EMA). Additionally, as mentioned previously, some rare GBMs and gliosarcomas have true epithelial metaplasia; not surprisingly, those foci will stain with a variety of cytokeratins ( Fig. 5-11C ).
Nearly all malignant gliomas express S-100 protein , making it a sensitive, but nonspecific, marker, since it wouldn’t distinguish them from other S-100–positive tumors such as melanoma. The diagnosis of metastatic melanoma should be supported (if necessary) with other markers of melanocytic differentiation such as HMB-45 , Melan-A , or microphthalmia transcription factor . The diagnosis of lymphoma is supported by tumoral expression of leukocyte common antigen (LCA, CD45) as well as B-cell and T-cell markers.
Neurofilament protein immunohistochemistry can be useful in the diagnosis of diffuse gliomas (see Fig. 5-2E ), especially among the astrocytic neoplasms, in which grade I, predominantly solid tumors (i.e., pilocytic astrocytoma, etc.) enter the differential. Neurofilament stains highlight entrapped axons of infiltrative lesions, supporting the presence of an infiltrative growth pattern. Other common noninfiltrative tumors include metastases and ependymomas. Just over half of WHO grade II astrocytomas harbor TP53 mutations, usually leading to p53 protein accumulation and nuclear immunoreactivity ( Fig. 5-2H, 5-14F ). Thus, immunoreactivity of tumor cell nuclei is occasionally useful for supporting a neoplastic designation in the differential diagnosis of grade II astrocytoma versus a reactive process, although a negative stain is less useful since it doesn’t exclude a diffuse astrocytoma (or other diffuse glioma such as oligodendroglioma). MIB-1 (Ki-67) is useful for estimating proliferative index, which is roughly proportional to histologic grade ( Figs. 5-2I, 5-7C, 5-9E, 5-12F, 5-14G ). 12 It is typically low in non-neoplastic processes, except in endothelial cells, macrophages, and other inflammatory cells which may also proliferate. Therefore, it is often helpful to add CD68 immunohistochemistry so that histiocytes and activated microglia can be “mentally subtracted” from the MIB-1 labeling index. However, this latter antibody simply recognizes lysosomes and is, therefore, also positive in a variety of lysosome-rich neoplasms, including the granular cell variant of GBM ( Fig. 5-8D ). CD163 represents a more specific marker of histiocytic lineage ( Fig. 5-8E ). Lastly, WT1 immunostaining can be particularly helpful in differentiating between astrocytoma and gliosis with “reactive atypia” (see prior section) (see Fig. 5-2J ) or in providing additional support of glial differentiation in rare high-grade examples (e.g., GBM), in which GFAP may be negative or difficult to interpret ( Fig. 5-4J ).

Genetics
TP53 mutations are among the most frequent genetic alteration in the grade II infiltrative astrocytomas, occurring in approximately 60%. 1, 31 These mutations are also noted in the GBMs that progress from them (progressive or secondary GBMs). More recently, IDH1 mutations have been found to be even more common in secondary GBMs and in all types of WHO grade II and III gliomas, including oligodendrogliomas with 1p/19q codeletions 32, 33 In contrast, they are rare in primary glioblastomas. As such, this alteration is now thought to be one of the earliest alterations in the tumorigenesis of the lower grade diffuse gliomas, including those that eventually progress to GBM secondarily.
The most common cytogenetic alteration in grade II astrocytomas is trisomy of chromosome 7q. Losses of chromosome 22q and 6, as well as gains on chromosome 8 have also been reported. Importantly, codeletion of chromosome 1p and 19q is exceedingly rare in low-grade astrocytomas (<3%).
Anaplastic astrocytomas are tumors with genetic alterations in transition between grade II astrocytomas and GBM. Thus, they have a similar frequency of TP53 mutations (50–60%). However, they also have a higher frequency of genetic alterations that are characteristic of tumor progression, included losses of chromosome 10q (40%–60%) and phosphatase and tensin homolog ( PTEN ) mutations (10%–20%), and p16(CDKN2A) losses (30%–50%). EGFR amplification occurs in approximately 10% of these tumors, although many of these likely represent undersampled glioblastomas.
In GBMs, numerous genetic alterations have been identified. Biologically important genetic alterations, whether amplifications, deletions or mutations, seem to involve three separate cellular signaling pathways that govern cell regulatory processes, such as proliferation, antiapoptosis, and invasion. 5, 34 According to the recent data for GBMs in The Cancer Genome Atlas project, these genetic families and their frequencies of involvement are (1) the p53 pathway, including TP53 mutations (35%), MDM2 amplifications (14%), and p14(ARF) deletion (49%); (2) the receptor tyrosine kinase pathway, including EGFR amplification or mutation (45%), PDGFR amplification (13%), ERBB2 mutation (8%), PTEN mutation or deletion (36%), PI 3-K mutation (15%), and NF1 mutation (18%); and (3) the Rb pathway, including p16(CDKN2A) deletion (52%), Rb deletion (11%), and CDK4/6 amplification (19%). Most GBMs have alterations involving each of these three pathways. Some of these genetic alterations are used in the diagnostic setting, either to provide assistance with pathologic classification or to provide independent prognostic information (see Fig. 5-6 ).

Epidermal Growth Factor Receptor
Amplifications of EGFR occur in 30% to 40% of all GBMs and roughly 70% of small-cell GBMs ( Fig. 5-6A ); in contrast, such amplifications are rare in giant-cell GBMs and gliosarcomas. They are not seen in low-grade astrocytomas and are considered to be relatively restricted to the primary or de novo form of GBM. Roughly half of cases with EGFR amplification express not only the wild-type form, but also a mutated form lacking exons 2 through 7, which results in a truncated cell-surface protein with constitutive tyrosine kinase activity (EGFRvIII). Neither EGFR amplification nor EGFRvIII rearrangement appears to be independently associated with shorter survival in patients with GBM. 35
EGFR amplifications are rare in oligodendroglial tumors and analysis of EGFR status has proven useful for distinguishing high-grade astrocytomas from anaplastic oligodendrogliomas in some instances. This is especially helpful in distinguishing “small-cell GBM,” which has a high frequency of EGFR amplification (69%) and chromosome 10 losses (97%), from anaplastic oligodendroglioma and mixed gliomas, which have high frequencies of 1p/19q deletions, but only rare EGFR amplifications and chromosome 10 losses. 21, 22
Therapies directed at the tyrosine kinase activity of EGFR in GBMs have been utilized in specific clinical trials and continue to be explored. 36 Therefore, it may become critical to establish the EGFR status of GBM as a part of the pathologic diagnosis in order to predict pharmacologic responses to EGFR inhibitors. In that study, only the small subset of tumors that expressed the constitutively active EGFRvIII mutant and retained PTEN immunoreactivity responded to therapy.

Loss of Chromosome 10
Genetic losses of chromosome 10, whether detected by traditional loss of heterozygosity (LOH) studies or by fluorescence in situ hybridization (FISH) are the most frequent genetic alterations in GBM, seen in over 80% of cases (see Fig. 5-6B ). Losses on 10q are seen in both de novo and secondary (progressive) GBMs, but losses of 10p appear to be more specific to de novo GBMs. Losses on chromosome 10 are much less common in low-grade astrocytomas and in oligodendrogliomas, making this a useful diagnostic marker for high-grade astrocytomas.

O 6 -Methylguanine-DNA Methyltransferase Testing
Many of the chemotherapeutic agents used to treat GBM, including temozolomide, are agents that crosslink DNA by alkylating at the O 6 of guanine. DNA crosslinking is reversed by the DNA repair enzyme O 6 -methylguanine-DNA methyltransferase (MGMT). 37, 38 Thus, tumors with low levels of MGMT expression would be expected to demonstrate an enhanced response to alkylating agents. The expression level of MGMT is determined in large part by the methylation status of the gene’s promoter. This “epigenetic silencing” of MGMT occurs in 20 to 40% of GBMs and can be assessed by its promoter methylation status on polymerase chain reaction (PCR)-based tests of genomic DNA. Epigenetic silencing of MGMT in GBM is associated with a longer survival period among patients treated with temozolomide and radiotherapy. Moreover, promoter methylation also seems to have a positive prognostic effect independent of therapy. Antibodies are available for the detection of the MGMT protein, but interpretation is challenging, the correlation of MGMT immunoreactivity with promoter status is poor, and an association with response to therapy or survival has not been found. 39 Since temozolomide has become a standard of care for the treatment of GBM, testing for MGMT methylation status is now an important component of a complete diagnostic workup.

Treatment and Prognosis
The biological behavior of diffuse astrocytic neoplasms varies greatly, although all are considered malignant. The two most powerful prognostic variables are patient age and histologic grade. Age is inversely proportional to survival time, such that younger patients live significantly longer than elderly patients with the same diagnosis. Average survival times by grade are 5 to 10 years, 2 to 3 years, and 1 year for grades II, III, and IV, respectively. Less powerful prognostic variables include Karnofsky performance status (degree of neurologic impairment) and extent of surgical resection. -->
Radiation therapy is a mainstay in the treatment of anaplastic astrocytoma and GBM. Treatment of GBM with temozolomide has become standard of care following the demonstration that it was able to extend survival. Those patients whose tumors display MGMT promoter methylation have a better prognosis following treatment with temozolomide. For grade II astrocytomas, radiation may be given at initial presentation or following tumor progression.

Pilocytic Astrocytoma


Definitions and Synonyms
The pilocytic astrocytoma (PA) is a circumscribed, well-differentiated neoplasm of childhood and young adults that can occur throughout the CNS, but is most common in the cerebellum, where it is often cystic. It is composed of bland, highly fibrillated tumor cells arranged in either a solid, microcystic, or biphasic pattern and typically has Rosenthal fibers and eosinophilic granular bodies (EGBs).
Numerous nonspecific clinical terms have been used to refer to pilocytic astrocytomas, including cerebellar astrocytoma , optic pathway glioma , tectal glioma , and dorsal exophytic brainstem or medullary glioma . These terms should be avoided, or at least clarified whenever possible, since diffuse astrocytomas may also involve those sites, albeit considerably less often.

Incidence and Demographics
Pilocytic astrocytoma (PA) accounts for only 2% of all CNS neoplasms and 6% of all gliomas. 3 The annual incidence rate is roughly 0.3 per 100,000 person-years. In children, PA is the most common form of glioma, with a median age of diagnosis being 13 years. Although most frequent in children, it may also be seen in adults of virtually any age, even into the 60s and 70s. In these older patients, it is generally assumed that tumors have been present in an asymptomatic form for decades. No gender or racial predilection has been appreciated.

Clinical Manifestations and Localization
Pilocytic astrocytomas can occur throughout the neuroaxis, but are predisposed to a smaller number of stereotypic sites, including the cerebellum ; the optic pathway including nerves, chiasm, and tracts; hypothalamus ; dorsal brainstem ; and spinal cord . 40, 41 PA accounts for approximately 10% of cerebral and 85% of cerebellar astrocytomas. These tumors are typically well-circumscribed , intra-axial masses that grossly and radiologically appear to displace adjacent brain rather than diffusely infiltrate it. Symptoms depend largely on the site. In the case of cerebellar tumors, the most frequent symptoms are headache, nausea, vomiting, and ataxia. Tumors that include the optic pathways present with slowly deteriorating vision or occasionally with proptosis if the orbit is involved. Tumors localized to the hypothalamus can cause endocrine dysfunction, including but not limited to obesity and diabetes insipidus, due to impingement on either the hypothalamus or pituitary gland. Those tumors that grow in the periventricular or periaqueductal regions, such as the thalamus or brainstem, often manifest with signs of hydrocephalus due to CSF obstruction. The majority of PAs are sporadic. However, NF1 patients are predisposed to this tumor type, particularly in the optic pathway and, to a lesser extent, the brainstem, and other sites (see Chapter 20 ). Fifteen percent of all PAs arise in this setting.

Radiologic Features and Gross Pathology
On neuroimaging, the majority of PAs are well circumscribed , with either a dominant cystic or solid pattern. The classic radiologic finding of a cerebellar PA is a cyst with an enhancing mural nodule ( Fig. 5-16A and B ). When PAs involve the optic nerves or chiasm, they are usually solid and variably enhancing ( Fig. 5-16C ). When tumors involve optic tracts, they typically expand posteriorly with a wedgelike growth pattern. In regions of the hypothalamus and brainstem, tumors are circumscribed, lobulated, or exophytic and intensely contrast-enhancing. In the spinal cord, PAs are intra-axial and enhancing and may be associated with a syrinx that extends over multiple spinal levels.
Figure 5-16 Pilocytic astrocytomas. A, B, T1-weighted magnetic resonance imaging (MRI) without ( A ) and with ( B ) contrast shows a well-demarcated tumor characterized by a cyst with an enhancing mural nodule within the left cerebellum. C, Postcontrast MRI demonstrates a nonenhancing pilocytic astrocytoma involving the left optic nerve ( arrow ).
On gross inspection of surgical resections or autopsy material, PAs typically appear well demarcated, though there may be some regions where tumor blends with adjacent brain. Cystic degeneration is common.

Histopathology
The classic PA has a biphasic appearance with alternating densely fibrillar and loose/microcystic components ( Fig. 5-17A and C ). 40 In some cases, only one of the two patterns is encountered. The dense regions often resemble fibrillary astrocytoma, except that the cytoplasmic processes are particularly long and hairlike (i.e., “piloid” as indicated by the tumor’s name). The latter is best appreciated on cytologic specimens, such as intraoperative smears, in which the hairlike process extends long distances, often across an entire low-magnification microscopic field ( Fig. 5-17B ). PAs also differ from fibrillary astrocytoma by their solid growth pattern, such that the dense fibrillarity noted in PAs is due almost entirely to the cellular processes of neoplastic cells. There are generally no entrapped axons or neuronal cell bodies within the central portions of the tumor and upon neurofilament immunohistochemistry infiltrated axons are absent. However, nearly all cases show at least limited infiltration at the periphery, and rare examples show extensive invasion ( Fig. 5-18C ).
The presence of corkscrew-shaped brightly eosinophilic Rosenthal fibers (RFs) strongly favors the diagnosis of PA rather than diffuse astrocytoma, although this is not absolute ( Fig. 5-17E ). PAs can also contain cellular arrangements that are less densely fibrillated, with cells that are more loosely disposed and look remarkably similar to oligodendroglioma, including round nuclei and perinuclear cytoplasmic clearing ( Fig. 5-17D ). Even in these regions, however, the long thin cellular processes typical of PA can be highlighted with a GFAP immunostain ( Fig. 5-18B ). Additionally, this component often harbors mulberry-shaped eosinophilic granular bodies (EGBs), which may be numerous ( Fig. 5-17F ) or rare. In the latter scenario, a PAS with diastase stain may help to draw attention to these structures ( Fig. 5-18A ).
Most often, the nuclei of PAs are oval with bland chromatin, but significant atypia (perhaps degenerative in nature) may occasionally be seen and is generally unassociated with increased mitotic or proliferative activity ( Fig. 5-17G ). True multinucleated forms are often appreciated with tight clustering of nuclei in a horseshoe-like configuration, often described as having a “pennies on a plate” arrangement ( Fig. 5-17G ), wherein the nuclei resemble a stack of pennies splayed out peripherally on a plate (i.e., cytoplasm). Glomeruloid single-layered vessels with multiple lumens are also typical, especially around the cyst lining, and should not be mistaken for the multilayered endothelial hyperplasia of diffuse gliomas ( Fig. 5-17I ). Nevertheless, the latter may also be encountered in PAs and does not have the ominous prognostic significance it does in the diffuse gliomas. The presence of occasional mitotic figures is acceptable ( Fig. 5-18D ), although high mitotic index is uncommon. Bland infarct-like necrosis is encountered in roughly 5% of PAs and similarly, has no clinical significance ( Fig. 5-18G ). Likewise, extension into the subarachnoid space is quite common and does not alter the prognosis ( Fig. 5-17H ). In contrast, pseudopalisading necrosis and foci of hypercellularity with increased proliferative activity should prompt consideration of an alternative diagnosis or malignant transformation , an exceptionally rare complication in PAs, most examples being encountered after radiation therapy.

Histologic Variants and Grading
The pilomyxoid astrocytoma (PMA) is a tumor of early childhood or adolescence that has recently been designated as a grade II variant of PA by the WHO. 42, 43 PMAs most often arise in the hypothalamic region, where they are well circumscribed, generally solid, and homogenously contrast-enhancing midline masses ( Fig. 5-19A ). PMA consists of a hypercellular, monomorphous population of piloid cells ( Fig. 5-19B and C ) that are typically embedded within a rich myxoid matrix and often display an angiocentric arrangement (see Fig. 5-19C ). Like PA, the PMA has a relatively discrete architecture, with only a slight tendency for peripheral infiltration of adjacent brain. Individual tumor cells have elongate fibrillar processes, are moderate in size, and contain hyperchromatic nuclei with only modest nuclear pleomorphism. Mitotic figures can be noted but are not abundant. The diagnosis of PMA is made only when this tissue pattern is predominant, since focal myxoid or angiocentric cell arrangement may be noted in typical PA ( Fig. 5-18H ) or infiltrating astrocytoma. Unlike ordinary PA, PMAs typically lack a biphasic appearance, do not contain Rosenthal fibers, and only exceptionally contain eosinophilic granular bodies. PMAs are associated with a more aggressive clinical course than typical PAs, warranting a WHO grade II designation. This includes high recurrence rates, occasional CSF dissemination, and increased risk of patient death.
Figure 5-19 Pilomyxoid astrocytomas (PMA). Despite a cytologically bland histology, these tumors of infancy are often aggressive (WHO grade II). A, Postcontrast magnetic resonance imaging (MRI) demonstrating a large, relatively solid contrast-enhancing mass involving the optic chiasm and the hypothalamic region. B, On smear preparations, nuclear uniformity and thin piloid processes are evident, essentially being indistinguishable from conventional pilocytic astrocytoma (compare with Fig. 5-17B ). C, On sections, tumors are composed of monomorphic bipolar, highly fibrillar cells with a loose myxoid background and an angiocentric pattern resembling perivascular pseudorosettes. They lack the dense component, RFs, and EGBs of conventional pilocytic astrocytoma. D, Glial fibrillary acidic protein stain highlights the piloid cytoplasmic processes, some of which appear to radiate toward central blood vessels. E, F, This patient had a classic PMA during infancy and presented with a recurrence that appeared highly infiltrative on T2-weighted MRI several years later ( E ). Despite this alarming appearance on neuroimaging, the biopsy specimen showed decreased cellularity with features of conventional pilocytic astrocytoma, including a more dense or compact component and scattered Rosenthal fibers ( F ). Some have termed this phenomenon as “maturation,” although its clinical significance remains unclear.
Figure 5-17 Pilocytic astrocytomas. These histologically benign tumors are associated with a wide morphologic spectrum. A, In classic examples such as this one, the tumor appears sharply demarcated from the adjacent atrophic cerebellar cortex ( left ). Nonetheless, most examples show at least focal invasion into adjacent brain parenchyma. The loose microcystic component is in the middle and the dense element is on the right. B, Cytologic smear preparations demonstrate long thin “hairlike” or piloid processes. C, This area shows a biphasic appearance with intermixed dense and loose foci. D, The loose component may resemble oligodendroglioma, although thin glial fibrillary acidic protein-positive processes are typically seen on immunohistochemistry (see Fig. 5-18B ). E, Densely fibrillar regions with abundant Rosenthal fibers (RFs) are seen in some pilocytic astrocytomas such as this one, whereas RFs are rare to absent in others. In the absence of a loose component, “piloid gliosis” should be excluded. F, Abundant eosinophilic granular bodies (EGBs) are more often encountered in the loose regions. G, Multinucleated cells often show a horseshoe or “pennies on a plate” arrangement ( arrow ). The cytologic atypia has no prognostic bearing and is likely degenerative in nature. H, Extension into the subarachnoid space is common (note edge of two adjacent cerebellar folia in right upper and lower regions). I, Microvascular hyperplasia is typical, especially in regions of the cyst lining. In this example, the vessels are mostly glomeruloid with multiple lumens and a single cell lining. However, the endothelia also appear hypertrophic and multilayered focally, as one might see in glioblastomas. In the setting of pilocytic astrocytoma, this finding has no prognostic implications. J, Some pilocytic astrocytomas have tortuous and variably dilated vasculature with marked hyalinization, potentially mimicking cavernous angioma.
Figure 5-18 Special studies and “atypical features” in pilocytic astrocytomas. A, A periodic acid–Schiff with diastase stain is often useful for highlighting rare eosinophilic granular bodies when they are difficult to see on routine hematoxylin and eosin sections. B, In contrast to most diffuse gliomas, pilocytic astrocytomas are often diffusely glial fibrillary acidic protein-positive, and this stain highlights their long thin processes. C , Occasional pilocytic astrocytomas have a more diffusely infiltrative growth pattern as evidenced by the many entrapped neurofilament-positive axons in this example. The prognosis for such cases is no different from that of conventional counterparts when other classic features are present. D-F, Increased mitotic figures ( D ; arrow ) and/or Ki-67 labeling ( E ) is often worrisome, but similarly has little clinical significance if the case is otherwise classic for pilocytic astrocytoma. An additional stain for CD68 may be helpful in such cases, as it may reveal a surprising number of intratumoral macrophages or activated microglia ( F ), either of which can undergo cell division and artificially elevate the proliferative index. G, The presence of infarct-like necrosis is seen in a small subset of pilocytic astrocytomas, but should not be interpreted as a malignant feature in the absence of associated pseudopalisading and other worrisome features, such as hypercellularity and brisk mitotic activity. H, Some pilocytic astrocytomas display perivascular pseudorosettes focally. This finding is similar to pilomyxoid astrocytomas (see Fig. 5-19 ), but its potential significance is unclear in cases that display classic features of pilocytic astrocytoma elsewhere.

Differential Diagnosis
The most common differential diagnoses for PA include diffuse astrocytoma (discussed in prior section), oligodendroglioma (discussed in subsequent section), and reactive “piloid” gliosis . Given the occasional findings of nuclear atypia, endothelial hyperplasia, mitotic activity, and necrosis in PA (see Figs. 5-17 and 5-18 ), GBM may constitute an unsettling differential diagnosis. In most instances, the unique clinical, radiographic, and microscopic features of PA help distinguish it from these other considerations. Although neither RFs nor EGBs are entirely specific for PA, they generally suggest a benign or slowly evolving process, such as PA, pleomorphic xanthoastrocytoma (PXA), and ganglioglioma, all representing tumor types with a favorable prognosis, often with similar radiographic features. PXA has much greater nuclear pleomorphism, mesenchymal-like spindled elements, vacuolated or lipidized cells, and reticulin-rich foci that are not seen in PAs (see Fig. 5-23 ). Gangliogliomas have a well-differentiated neuronal component, characterized by dysmorphic ganglion cells, and often have a lymphocytic infiltrate. Dysembryoplastic neuroepithelial tumors (DNTs) most often resemble oligodendrogliomas, but may have areas resembling PA as well. Temporal lobe predilection, patterned mucin-rich nodules, and floating neurons serve to distinguish this entity. Lastly, RFs are also often encountered in a piloid form of reactive gliosis (similar appearance to Fig. 5-17E ), most often next to craniopharyngiomas, ependymomas, hemangioblastomas, developmental cysts, and syringomyelia. Piloid gliosis is typically less cellular and does not have a microcystic component. Additional sampling and attention to clinical and radiographic features generally allows one to avoid this pitfall.
Figure 5-23 Pleomorphic xanthoastrocytomas (PXAs). A, Regions where spindled cells predominate may resemble mesenchymal neoplasms, such as pleomorphic sarcoma. B, Pleomorphism with multinucleated giant-cell formation is common. Also note the perivascular lymphocytic cuffing ( lower right ), xanthic astrocytes ( arrowheads ) and the eosinophilic granular body (EGB; arrow ). C, Periodic acid–Schiff-positive EGBs. D, Dense intercellular reticulin network most often surrounding small groups of tumor cells. E, glial fibrillary acidic protein (GFAP) staining is highly variable, with at least subsets of positive cells in nearly all cases. F, GFAP may also be useful for identifying the lipid vacuoles of xanthic astrocytes ( arrow ). G, CD34-positive cells with ramified processes are common in PXA. H, A subset of PXAs express neuronal markers such as synaptophysin, often in cells that do not look obviously neuronal.

Ancillary Diagnostic Studies
In the majority of cases, the diagnosis of PA can be established on morphologic grounds alone. PAS with diastase and trichrome stains are occasionally helpful in highlighting rare EGBs and RFs that were not evident on H&E ( Fig. 5-18A ). PAs are typically strongly immunoreactive for GFAP ( Fig. 5-18B ) and show moderate staining for S-100 protein . In smaller biopsies, the distinction between an infiltrating glioma and PA can be aided using immunohistochemistry for neurofilament , which highlights infiltrated axons in diffuse gliomas, but usually only shows scattered entrapped axons at the edges of PA. A negative p53 stain also favors PA in this diagnostic situation, but is not conclusive since diffuse gliomas may also be negative. The MIB-1 labeling index of PA is generally low, ranging from <1% to 5%, yet a higher index does not exclude the diagnosis ( Fig. 5-18E ). Most studies have not shown any prognostic value for MIB-1 indices in this tumor type. However, in the case of subtotal resection, the proliferation index may be taken into account for clinical management in some instances. Although electron microscopy is rarely required, these studies generally show electron-dense cells with prominent cytoplasmic intermediate filaments corresponding to GFAP.

Genetics
Consistent genetic alterations in sporadic PAs have only recently been identified. The most common alteration involves gains on 7q34, often associated with duplications of the BRAF or HIPK2 genes (or both), or characteristic rearrangements and mutations of the former. 44, 45 Several of the recently identified alterations are thought to be associated with activation of the MAP kinase signaling pathway.
Multiple studies have also confirmed that the genetic alterations seen in the diffuse forms of astrocytoma, such as TP53 and PTEN mutations, EGFR amplification, and deletion of chromosomes 9p21, are not typical of PA. Comparative genomic hybridization studies have demonstrated a variety of nonspecific chromosomal gains and losses. Gains of chromosome 5 and 7 were the most frequent, followed by 6, 11, 12, 15, 17, 19, 20, and 22. Gene expression profiling studies have shown that PAs have distinct patterns compared with the diffuse gliomas. In particular, PAs have been shown to have extremely high expression of ApoD and galectin-3, compared with diffuse astrocytomas. 46
PAs that arise in neurofibromatosis type-1 (NF1) patients are genetically distinct from sporadic PAs in that they show allelic loss or mutation of the NF1 gene and reduced expression of neurofibromin (the protein product of NF1 gene), resulting in downstream activation of Ras and mTOR. Gene expression studies that compare NF1 associated to sporadic PAs have shown that the former have increased expression of CUGBP2, RANBP9, ITGAV1, and INFGR1. 47 One of the underexpressed genes in NF1 associated PAs, aldehyde dehydrogenase 1 family member L1 (ALDH1L1), was also reduced in clinically aggressive PAs and in PAs with atypical histologic features including necrosis.

Treatment and Prognosis
PAs are benign (WHO grade I) neoplasms treated primarily with surgery. The overall prognosis for this tumor type is excellent, with an estimated 80% 20-year survival. 41 As a group, the supratentorial PAs have a less favorable prognosis than those of the cerebellum. 41 PAs associated with NF1 generally have a better clinical outcome than sporadic PAs, especially those involving the optic tracts. 48 Often the latter grow early in childhood, only to stabilize or even regress spontaneously as the patient gets older.
A subset of PA is associated with significant morbidity and mortality, and it is often difficult to predict based on histology alone which cases will behave in a more aggressive fashion. Depending on the tumor location, subtotally resected cases may be irradiated for enhanced local control. Rare examples, particularly those arising from the hypothalamic region, may disseminate through the CSF. Some of these patients may have stable or slowly progressive disease despite metastatic meningeal deposits, presumably due to their slow growth.

Subependymal Giant-Cell Astrocytoma


Definitions and Synonyms
Subependymal giant-cell astrocytomas (SEGAs) are benign (WHO grade I), slow-growing discrete tumors that arise in the walls of the lateral ventricles and are composed of large, atypical-appearing astrocyte-like cells. Almost always, they occur in the vicinity of the foramen of Monro in children or young adults and are tightly linked with the tuberous sclerosis complex (TSC).

Incidence and Demographics
SEGAs are uncommon tumors and account for less than 1% of all intracranial masses. The incidence of tuberous sclerosis is 1/5000 in the U.S. population, and SEGAs occur in 5% to 15% of TSC patients. These tumors are important to recognize because of their strong association with TSC and because they can be confused with higher grade neoplasms of the CNS. Patients with SEGAs usually present clinically between ages 2 and 30 years, but the tumors occur most frequent in the early teen years , with a mean age at presentation of 13 years.

Clinical Manifestations and Localization
SEGAs are typically located in the lateral or third ventricles near the foramen of Monro . 49, 50 In most cases, the diagnosis of TSC has already been established and the symptoms of the SEGA may be related to worsening CNS manifestations of TSC, including epilepsy, infantile spasm, autistic withdrawal, or mental status changes. The mass itself may give rise to symptoms as well, causing increased intracranial pressure due to CSF obstruction, leading to nausea, vomiting, and lethargy.
Less commonly, a SEGA may be the first detected manifestation of TSC, and other characteristic features should be sought, including cortical tubers, “candle gutterings” (smaller masses along ventricular lining, resembling wax drippings), and gray matter heterotopias (see Chapter 21 ). The most common extracranial manifestations include facial cutaneous angiofibromas, renal angiomyolipoma, pulmonary lymphangioleiomyomatosis, subungual fibroma, cardiac rhabdomyoma, intestinal polyps, and visceral cysts. 51

Radiologic Features and Gross Pathology
By neuroimaging, these tumors are solitary and circumscribed within the lateral ventricles, ranging in size from less than 1cm to greater than 6cm ( Fig. 5-20 ). 52 They rarely occur bilaterally or extend into the third ventricle. SEGAs may fill the lateral ventricle and distort the adjacent brain due to mass effect, but do not usually show any invasive properties. Moreover, it is uncommon to note dissemination of these tumors into the CSF space. Following administration of contrast agents, SEGAs show intense enhancement . These tumors are frequently calcified, often heavily. Obstruction of CSF flow can result in hydrocephalus and transependymal edema. Candle gutterings are smaller versions that are more widely distributed along the ventricular surface; these structures along with the presence of tubers are essentially diagnostic of tuberous sclerosis (see Fig. 5-20 ).
Figure 5-20 Subependymal giant-cell astrocytoma. T2-weighted ( A ) and postcontrast T1-weighted ( B ) magnetic resonance imaging showing an enhancing intraventricular mass near the foramen of Monro with obstructive hydrocephalus. Note also the additional findings of subependymal “candle gutterings” ( arrowheads ) and cortical tubers ( arrows ), providing diagnostic features of tuberous sclerosis.
Grossly, SEGA is a solid, well-demarcated mass, often with zones of dense calcification.

Histopathology
Histologically, SEGA is a discrete mass composed of spindled , epithelioid , or gemistocyte-like cells arranged in sweeping fascicles ( Fig. 5-21A ), in nests separated by dense fibrillary septae ( Fig. 5-21B ), or surrounding blood vessels, mimicking the perivascular pseudorosettes of ependymoma ( Fig. 5-21C ). 49, 50 Most characteristic is the compact arrangement of large, astrocyte-like cells with abundant glassy cytoplasm , combined with large vesicular nuclei and prominent nucleoli similar to those of ganglion cells ( Fig. 5-21D ). Occasional tumor cells are atypical and binucleate, and these unusual features can give the mistaken impression of anaplasia. SEGAs are not infiltrative and generally show a well-demarcated border with adjacent brain ( Fig. 5-21E ). As their name implies, they grow directly under the ependymal surface and therefore a benign ependymal lining can be noted histologically at the surface of the tumors. Calcification is an almost constant feature and is often so extensive that the mass becomes extremely hard. Inflammatory infiltrates can also be seen, including scattered lymphocyte aggregates and individual mast cells within the tumor stroma ( Fig. 5-21H ). Despite the large size of tumor cells and the presence of occasional bizarre nuclei, these tumors have a benign clinical course. Mitotic figures and necrosis are uncommon, but when they are noted, they do not constitute a high-grade diagnosis.
Figure 5-21 Subependymal giant-cell astrocytomas (SEGAs). A, Tumors often display a “sweeping” growth pattern with spindled to epithelioid cells. B, Occasionally, tumors grow in clusters, separated by zones of dense fibrillarity. C, Orientation of cells around central vessels can mimic the perivascular pseudorosettes of ependymoma. D, Typical cytologic features include large ganglion cell-like nuclei with prominent nucleoli, but with pink glassy astrocyte-like cytoplasm. E, This neurofilament protein stain shows that native axons are pushed to the periphery, consistent with a solid growth pattern. F, SEGAs are variably glial fibrillary acidic protein-positive, often with strong expression in only a subset of tumor cells. G, Some SEGAs also express neuronal markers, such as synaptophysin, most often focally. H, SEGAs often contain numerous intratumoral mast cells, as highlighted with this c-kit immunostain.

Differential Diagnosis
The differential diagnosis of SEGA can be approached by location or by histology. Other tumors that occur in the ventricular system near the foramen of Monro include central neurocytoma , subependymoma , meningioma , choroid plexus tumors , and germ cell tumors . These tumors have histologic features that are quite distinct from SEGAs.
Histologically, SEGA can resemble gemistocytic astrocytoma , since both contain cells with abundant eosinophilic cytoplasm. Unlike gemistocytic astrocytoma, SEGAs display a solid growth pattern (see Fig. 5-21E ), express GFAP less uniformly ( Fig. 5-21F ), and present within the ventricle rather than the parenchyma.
The presence of perivascular orientation of tumor cells in a SEGA raises the possibility of ependymoma , though the latter are typically intraparenchymal, rather than intraventricular when they are supratentorial. Furthermore, the ganglion-like and gemistocyte-like cytology is not typical of ependymoma.
Ganglion cell tumors are best distinguished from SEGAs by the presence of true tumoral ganglion cells that display distorted triangular shapes and amphophilic cytoplasm containing Nissl substance, similar to large pyramidal cells of the CNS, rather than the occasional neuron-like nuclei scattered among tumor cells in SEGAs. Moreover, ganglion cell tumors do not usually occur within the ventricular system. In ganglioglioma, a portion of the tumor resembles low-grade astrocytoma, either pilocytic or fibrillary, which is also not typical of SEGAs. Also more characteristic of ganglion cell tumors are eosinophilic granular bodies, lymphocytic infiltrates, and collagen deposition.
Like SEGAs, pleomorphic xanthoastrocytoma (PXA) contain a solid arrangement of atypical, pleomorphic tumor cells with an astrocytic morphology and abundant pink or pale cytoplasm. Bizarre giant cells are present, but mitoses are unusual. Features of PXA that distinguish it from SEGAs include intra-axial location, xanthomatous cells, a higher degree of cellular and nuclear pleomorphism, EGBs, and a reticulin-rich stroma.

Ancillary Diagnostic Studies
Most tumor cells that display astrocytic differentiation morphologically show patchy immunoreactivity for GFAP ( Fig. 5-21F ), but the staining intensity may be surprisingly weak in some instances. 53 Staining for S-100 is more reliably strong in tumor cells. Other tumor cells more closely resemble neurons or have intermediate features with astrocytoma-like cytoplasm and neuronal-like nuclei and these may show staining for neurofilament protein , synaptophysin ( Fig. 5-21G ), and chromogranin . Ultrastructural studies typically reveal electron- dense cells with prominent cytoplasmic intermediate filaments corresponding to GFAP. Rarely, neuronal features such as dense core granules can be appreciated as well. Because of these limited neuronal characteristics, some prefer the term subependymal giant-cell tumor .

Genetics
SEGA is one of the major diagnostic criteria for TSC, an autosomal dominant disorder characterized by hamartomas and benign neoplasms of multiple organ systems, including the CNS (see Chapter 20 ). Two related genes are linked to TSC: TSC1 , found at chromosome 9q34, which encodes for the protein hamartin; and TSC2 , located on chromosome 16p, which encodes for the protein tuberin. 1 Patients with TSC1 mutations have slightly milder disease than those with TSC2 mutations when compared at similar ages. TSC1 mutations account for 10% to 15% of cases, while TSC2 mutations account for 55% to 65%. Tuberin and hamartin interact physically within the cell cytoplasm to form a tumor suppressor complex that inhibits the function of mTOR (mammalian target of rapamycin). Loss of function of either gene product results in up-regulation of mTOR and increased proliferative activity. The penetrance of TSC1 and TSC2 mutations is 100%, but phenotype is highly variable and genotype does not predict phenotype, even within families, indicating that nongenetic factors influence phenotype. Indeed, even monozygotic twins can have widely differing disease manifestations.

Treatment and Prognosis
SEGAs are benign tumors (WHO grade I) and likely represent hamartomatous rather than neoplastic proliferations. 54 The treatment of SEGAs is surgical resection , which may be necessary more than once in the unusual instance of tumor recurrence. However, the prognosis is generally excellent following complete resection and these tumors do not undergo malignant progression. Despite their benign nature, there are significant prognostic implications to the patients and their families, based on this tumor’s nearly universal association with tuberous sclerosis.

Pleomorphic Xanthoastrocytoma


Definitions and Synonyms
Pleomorphic xanthoastrocytoma (PXA) is a WHO grade II generally circumscribed cerebral cortical tumor of children and young adults that often manifests with seizures and is generally associated with a favorable clinical outcome, although 15% to 20% undergo malignant progression. 55, 56 It is composed of large, often bizarre tumoral astrocytes that are lipidized and embedded in a reticulin-rich stroma.

Incidence and Demographics
PXAs are rare, accounting for less than 1% of all CNS neoplasms. They are most common in the first two decades of life and have their peak incidence between ages 10 and 19 years. There is no gender or racial predilection.

Clinical Manifestations and Localization
PXAs occur most frequently in the superficial cortex of the cerebral hemispheres, particularly in the temporal lobe , and often involve the overlying leptomeninges . Many patients with these tumors present with seizure disorders of long duration. PXAs located elsewhere, including the cerebellum and spinal cord, generally have slowly progressive symptoms that depend on the site.

Radiologic Features and Gross Pathology
On CT and MRI, PXAs are most frequently noted as cysts with contrast-enhancing mural nodules 57, 58 or as partially cystic enhancing masses with minimal mass effect ( Fig. 5-22A and B ). They are typically centered in the cerebral cortex and extend superficially to involve the overlying meninges; they may even cause remodeling of the adjacent calvarium, a finding often associated with slow-growing superficial brain tumors ( Fig. 5-22C ). Occasionally, PXAs have only a solid, contrast-enhancing appearance, lacking a cyst. In these instances, the association with overlying dura can mimic a meningioma.
Figure 5-22 Pleomorphic xanthoastrocytomas (PXA). A, B, T2-weighted ( A ) and postcontrast T1-weighted ( B ) magnetic resonance imaging (MRI) showing a well-circumscribed, partially cystic, focally enhancing right medial temporal lobe mass with minimal surrounding edema and mass effect. C, T2-weighted MRI from a right parietal PXA showing superficial cortical localization and molding of the adjacent skull.
Grossly, PXAs appear well demarcated and firm. Cystic components and calcifications are seen in some.

Histopathology
Histologically, PXAs are hypercellular and composed of atypical, pleomorphic tumor cells with astrocytic or mesenchyme-like morphology and abundant pink or pale cytoplasm arranged in fascicles or a storiform pattern ( Fig. 5-23A ). 56, 59 Xanthomatous cells (“xanthoastrocytes”) with foamy, lipid-filled cytoplasm are also diagnostically helpful, but are only encountered in about a quarter of cases ( Fig. 5-23B ). Cells with large bizarre nuclei and multinucleation are more common, giving the tumor its pleomorphic appearance ( Fig. 5-23B ). However, in most instances, mitotic figures are scarce.
The relative lack of mitoses in the setting of marked pleomorphism is one of the clues to the diagnosis of PXA. Another key to recognizing these neoplasms is their superficial localization , which can be noted under the microscope by the incorporation of large-caliber, muscularized arteries, typically of the leptomeninges, into the solid component of the tumor. Adjacent to the solid neoplastic component and most often involving the cerebral cortex, individual tumor cells infiltrating the brain parenchyma can be noted, reminiscent of diffuse astrocytomas. A nearly defining feature of PXA is the deposition of intercellular reticulin , either diffusely or in a patchy fashion, which invests cells either individually, or more commonly, as cell clusters ( Fig. 5-23D ). Its pattern corresponds to basal lamina ultrastructurally and has led to speculation that PXAs arise from a specialized subpial astrocyte with basal lamina production.
Other findings that help establish the diagnosis are eosinophilic granular bodies (EGBs) and clusters of lymphocytes that are present within the tumor stroma and in a perivascular distribution ( Fig. 5-23B and C ). Rosenthal fibers may also be seen, particularly at the edges of the tumor, likely reflecting a reactive change by adjacent brain. Occasional PXAs are mixed with classic ganglioglioma, display intratumoral ganglion cells, or show lesser forms of neuronal differentiation such as synaptophysin or neurofilament positivity in otherwise astrocytic-appearing cells ( Fig. 5-23H ). 60

Histologic Variants and Grading
PXA is currently designated as a grade II neoplasm by the WHO. This is based primarily on the significant risk of subsequent recurrences and evidence of malignant progression in roughly 15% to 20% of cases, the latter often being difficult to predict in the initial resection specimen. Although most authorities acknowledge the existence of these more aggressive forms of PXA, the definition of anaplasia remains controversial. The presence of necrosis (especially if pseudopalisading) or excessive mitotic activity has been suggested to define a more aggressive tumor. One study found that a mitotic index greater than 5/10hpf was predictive of a more aggressive clinical course. 60 In some instances, these more aggressive variants actually have less pleomorphism than typical grade II PXAs and have an overall resemblance to high-grade diffuse astrocytomas. The WHO currently designates these tumors as “ PXAs with anaplastic features ,” although a formal grade has not been applied.

Differential Diagnosis
Due to the prominent nuclear pleomorphism and mesenchymal-like foci, the most common differential diagnoses are giant-cell GBM , gliosarcoma , and pleomorphic sarcoma . However, EGBs are not found in any of these entities. Foci of reticulin deposition also help to exclude GBM. There is significant clinical, radiologic, and histologic overlap between PXA and ganglioglioma . Indeed, in rare instances, both entities may combine to form a composite neoplasm. In general though, ganglioglioma has less pleomorphism, a more obvious neuronal component, and no lipidized astrocytes. Similarly, PAs may have bizarre, atypical or pleomorphic nuclei, but do not display deposits of intercellular reticulin.

Ancillary Diagnostic Studies
Despite showing predominantly astrocytic features, GFAP immunoreactivity varies from focal to strong and extensive ( Fig. 5-23E and F ). 61 Scattered staining with neuronal markers, including synaptophysin (see Fig. 5-23H ), neurofilament , and MAP2 , is encountered in roughly a fourth of cases, leading some to speculate that there may be a histogenetic relationship between PXA and ganglioglioma. This hypothesis is further strengthened by rare cases of combined PXA–ganglioglioma 61 and the frequent finding of stellate CD34 -positive cells ( Fig. 5-23G ) in both tumor types. The MIB-1 proliferation index has been reported to be less than 3% in the majority of tumors. However, with malignant transformation, proliferation indices can exceed 10%.

Genetics
The most extensive genetic analysis of PXAs involved the definition of chromosomal gains and losses by comparative genomic hybridization in which loss of chromosome 9 was uncovered as the most frequent abnormality, noted in 50% of cases. 1, 62, 63 Higher resolution mapping also indicated losses on chromosome 9 corresponding to the 9p21.3 locus that includes the CDKN2A/p14ARF/CDKN2B gene complex. Less common alterations include mutations of the TP53 gene (6%) and LOH of chromosome 17 (10%). PXAs do not harbor other genetic alterations, such as EGFR , CDK4 , and MDM2 amplifications; PTEN mutations; or LOH of chromosome 10q, which are typical of the diffusely infiltrative forms of astrocytoma.

Treatment and Prognosis
PXA generally has a favorable prognosis , with many cases being cured with surgery alone . 60 The postoperative survival rate at 5 years is 81% and 70% at 10 years. Nevertheless, it is recognized that 15% to 20% of PXAs will undergo malignant progression. Because of the potential for malignant degeneration and the less predictable long-term prognosis, PXA has been assigned a WHO grade of II. Some of the cases designated as “ PXA with anaplastic features ” are associated with aggressive behavior and shortened survival time, although others do surprisingly well, even with GBM-like features such as pseudopalisading necrosis.

Oligodendroglioma


Brief Historical Overview
In the original classification of glial neoplasms by Bailey and Cushing (1926) and later in the article “Oligodendrogliomas of the Brain” by Bailey and Bucey (1929), the authors described a unique cerebral hemispheric tumor of adults that had histologic features unlike those of other gliomas. 4, 64 Oligodendrogliomas were described as having nuclei that “are almost all perfectly round and of a fairly constant size” and “surrounded by a ring of cytoplasm which stains very feebly,” adding that they have a “network of fine capillaries” and “are prone to become calcified.” Our current concept of oligodendroglioma retains most of these essential features. 65 Some might argue that the diagnostic criteria have expanded to gradually encompass nonconventional morphologies. 66 However, early articles also foreshadowed some of the challenging diagnostic dilemmas that we continue to encounter: “There are also many cells which appear to be transitions between gigantic oligodendroglia and astrocytes. It is impossible definitely to classify them as belonging in either group” and “Practically every stage of gradual transition from typical oligodendroglia to typical astrocytes can be found”. 4, 64

Definitions and Synonyms
Oligodendrogliomas are infiltrating gliomas that occur most often in adults and typically involve the cerebral hemispheres. The neoplastic cells have the morphologic appearance of oligodendrocytes, are predisposed to involve the cerebral cortex, and often have combined genetic losses on chromosome 1p and 19q.

Incidence and Demographics
Oligodendrogliomas, including grade II and III tumors, account for 4% of primary brain tumors and 15% to 25% of the infiltrating gliomas. 3 Their annual incidence is roughly 0.6 per 100,000 person-years. The incidence has increased over the past three decades, but it remains unclear if this reflects a diagnostic trend among pathologists or a meaningful epidemiologic shift. Oligodendrogliomas typically affect young to middle-aged adults , with a peak incidence in the 30s and 40s. As with other diffuse gliomas, there is a slight male predominance (3:2). For reasons that are not clear, these tumors are distinctly uncommon in children .

Clinical Manifestations and Localization
The vast majority of oligodendrogliomas are cerebral hemispheric masses, most frequently involving the frontal lobe and followed in frequency by parietal and temporal lobe locations. 65, 67 They are often centered superficially in the brain and demonstrate extensive cortical involvement. The clinical presentation depends heavily on the location. Due to their corticotropism , seizure disorders are particularly common, accounting for nearly two thirds of initial clinical presentations. Focal neurologic deficits are sometimes noted, as are more generalized signs and symptoms of increased intracranial pressure, including headaches, nausea and vomiting, and papilledema. Rapid onset and progression suggests anaplastic oligodendroglioma (WHO grade III), whereas a protracted history (e.g., chronic seizure disorder) is more consistent with low-grade (WHO grade II) tumor. Brainstem, cerebellum, and spinal cord examples are distinctly unusual .

Radiologic Features and Gross Pathology
Radiologically, the low-grade (WHO grade II) cases are typically nonenhancing intra-axial masses that expand the involved brain, are hypointense on T1-weighted MRI, and show signal hyperintensity on T2 and FLAIR MRI sequences ( Fig. 5-24 ). Calcifications are best appreciated on CT. The diagnosis of oligodendroglioma based on neuroimaging will be favored when a pattern of microcalcifications follows the pattern of an expanded cortical ribbon . Anaplastic (WHO grade III) oligodendrogliomas are nearly always contrast-enhancing , most commonly in either an irregular or diffuse fashion ( Fig. 5-25 ).
Figure 5-24 Oligodendrogliomas, WHO grade II. A, Fluid-attenuated inversion recovery magnetic resonance imaging (FLAIR MRI) shows a hyperintense lesion centered in the cortex of the left frontal lobe, but also extending superficially into underlying white matter. The dark spaces likely represent areas of cyst formation or high mucin content, since they were bright on T2-weighted sequences (not shown). B, On T1-weighted MRI with gadolinium, the lesion is hypodense and nonenhancing.
Figure 5-26 Oligodendrogliomas, WHO grade II. A, On intraoperative smear, characteristic features include a monomorphous population of small rounded nuclei with little cytoplasm and a rich branching capillary network. B, “Chicken wire”-like branching capillary network, uniformly rounded nuclei, clear perinuclear haloes imparting a “fried egg” pattern, and mucin-rich microcystic spaces are evident on sections. C, Perineuronal satellitosis is often prominent in regions of cortical infiltration. D, Minigemistocytes have small rounded eccentric bellies of eosinophilic cytoplasm with nuclei identical to those of adjacent classic oligodendroglial cells. E, In well-preserved tissue, the cytology of oligodendroglioma includes rounded nuclear contour, delicate chromatin, small nucleoli, and crisp nuclear membranes. Cells with brightly eosinophilic granules are also seen occasionally. F, In some oligodendrogliomas, glial fibrillary acidic protein (GFAP) highlights entrapped reactive astrocytes, while tumor cells are negative. G, However, at least two cell types strongly express GFAP. Gliofibrillary oligodendrocytes are characterized by thin perinuclear rims of GFAP positivity, sometimes with a short tadpole-like cytoplasmic tail. Minigemistocytes have small eccentric bellies of immunoreactive cytoplasm ( arrows ). H, Although oligodendroglioma is generally considered a purely glial neoplasm, it is not uncommon to encounter some expression of neuronal markers. A paranuclear dotlike pattern of synaptophysin positivity is particularly common in morphologically classic oligodendrogliomas with 1p/19q codeletion.
Like other diffuse gliomas, oligodendrogliomas have ill-defined tumor borders with blurring of gray-white junctions, though some appear deceptively circumscribed grossly. There may also be zones of microscopic disease that are only notable grossly due to secondary mass effects, such as ventricular compression, midline shift, and sulcal effacement. Calcifications are grossly evident in some.

Histopathology
Current diagnosis of oligodendroglioma requires the identification of infiltrating glioma cells that have round, regular, and monotonous nuclei with sharply defined nuclear membranes (in well-preserved specimens), and only modest cell-to-cell variability ( Fig. 5-26 ). 9, 65 Tumor cell cytoplasm tends to swell during routine formalin-fixation and paraffin-embedding, resulting in cells with well-defined cell membranes, cytoplasm clearing, and a central spherical nucleus—a combination of factors that gives rise to the classic “ fried egg ” cell (also described as “honeycomb” and “woody plant” histology). 64
Though helpful in diagnosis, the fried egg appearance is a formalin fixation artifact that is neither necessary for diagnosis nor encountered in frozen sections or rapidly fixed specimens. Not all tumor cells in an oligodendroglioma show cytoplasmic clearing; in contrast, occasional tumors with classic astrocytic differentiation may contain perinuclear halos due to fixation artifact.
Taken literally, the term “oligodendroglioma” defines a glioma with few cellular processes . In tissue sections and especially on smear preparations, there is a paucity of glial processes emerging from oligodendrogliomas compared with the long, finely fibrillar processes that are in abundance in cytologic and histologic preparations of most astrocytic neoplasms. For reasons that are not entirely clear, a delicate, branching capillary network with a “chicken-wire” appearance is frequently noticed in oligodendrogliomas ( Fig. 5-26A and B ). Other common but nonspecific features include cortical involvement , microcalcifications , and microcysts filled with mucin ( Fig. 5-26B ), as well as secondary structures , such as perineuronal satellitosis ( Fig. 5-26C ), perivascular aggregation, and subpial condensation.
The morphologic spectrum of oligodendroglioma includes two strongly GFAP-positive cell types: mini- or microgemistocytes and gliofibrillary oligodendrocytes . 29 The former are gemistocyte-like cells with small bellies of eosinophilic cytoplasm, round nuclei resembling those of classic oligodendroglioma, and no cytoplasmic processes ( Fig. 5-26D, 5-26G ). Gliofibrillary oligodendrocytes are histologically identical to classic oligodendroglioma cells on H&E stains, but have a thin perinuclear rim of strong GFAP immunoreactivity ( Fig. 5-26G ).

Histologic Variants and Grading
Grading systems have typically divided oligodendrogliomas into two, three, or four grades depending on cellularity, cytologic atypia, mitotic activity, vascular proliferation, and necrosis. In contrast to astrocytomas, where most are high grade at the time of diagnosis, over half of patients with oligodendrogliomas present with grade II lesions. 68 Criteria for grading are not as clearly defined for oligodendrogliomas as they are for astrocytomas. The current WHO classification recognizes two grades: oligodendroglioma (grade II) and anaplastic oligodendroglioma (grade III) . 1 Grade II tumors vary from low to moderate cellularity. These tumors have a tendency to involve the cerebral cortex and as they progress, they often grow in a distinctly nodular or lobulated pattern ( Fig. 5-27B ). Nodular growth is compatible with a grade II lesion, but may represent a transition to a higher grade, especially once the nodules coalesce into regions of confluent hypercellularity. Grade II oligodendrogliomas can show occasional mitotic figures and cytologic atypia, but marked mitotic activity, microvascular proliferation , or necrosis is consistent with a WHO grade III, anaplastic oligodendroglioma ( Fig. 5-27 ). High-grade examples also commonly develop cytologic features that are more epithelioid , with increased cytoplasm, sharper cytoplasmic borders, and prominent nucleoli ( Fig. 5-27C ). Investigations of prognostic features in oligodendroglioma have identified: (1) endothelial hypertrophy ( Fig. 5-27D ); (2) necrosis ( Fig. 5-27E and F ); and (3) more than 6 mitotic figures/10 hpf as significant univariate markers of poor outcome, providing a framework for establishing the diagnosis of anaplastic oligodendroglioma (grade III). 68 The significance of focal versus diffuse anaplasia has yet to be elucidated. Previous multivariate analyses have suggested that necrosis is the single most important feature of aggressive clinical behavior. Pseudopalisading of tumor cells around necrosis is uncommon, but may occasionally be seen in anaplastic oligodendrogliomas; in the absence of a significant astrocytic component (see discussion of oligoastrocytoma), it does not convert the diagnosis to GBM or imply a grade IV designation ( Fig. 5-27F ).
Figure 5-27 Anaplastic oligodendrogliomas, WHO grade III. A, In this tumor that underwent tumor progression, the grade II component is on the right and the grade III component on the left . Note the higher cell density and increased cell size in the anaplastic region. B, A nested or lobulated growth pattern is common in anaplastic foci. C, Increased pleomorphism with epithelioid cytology and prominent nucleoli is also typical, as is increased mitotic activity ( arrow ). D, Endothelial hyperplasia is another criterion. E, F, Foci of necrosis are most often infarct-like ( E ), but may also be pseudopalisading ( F ). The presence of classic nuclear cytology and mucin-filled microcystic spaces ( arrow ) help to distinguish such tumors from small-cell glioblastomas (see Fig. 5-12 ).
Figure 5-12 Small-cell glioblastoma (GBM). Given the cellular uniformity, variable finding of clear perinuclear haloes, and “chicken wire”-like capillary proliferation, this variant often resembles oligodendroglioma at low magnifications ( A ). Common findings such as microcalcifications and perineuronal satellitosis add further to the mimicry ( B ). At higher magnifications, however, the cells are more oval than round ( B, C ) and display the odd combination of low-grade appearing cytology (i.e., delicate chromatin) admixed with frequent mitotic figures ( C ; arrowheads ). As with other glioblastomas, pseudopalisading necrosis is common ( D ); nonetheless, cases with only “WHO grade III” histology are not uncommon and still follow a GBM-like clinical course most of the time. Although minimal cytoplasm is seen on hematoxylin and eosin-stained sections, thin glial fibrillary acidic protein-positive processes are common ( E ). The Ki-67 labeling index is typically high ( F ). This variant is sometimes confused with high-grade oligodendroglial neoplasms (compare with Fig. 5-13 ) or with primitive neuroectodermal tumor-like elements (compare with Fig. 5-14 ). The majority of small-cell GBMs show EGFR gene amplification and chromosome 10 losses (see Fig. 5-6 ).
Figure 5-6 Characteristic fluorescence in situ hybridization findings in glioblastoma. A, Epidermal growth factor receptor (EGFR) amplification is seen in 40% of glioblastomas (GBM) and a greater percentage of small-cell GBMs. Chromosome 7 centromere = green (mostly two or three signals), EGFR gene = red (many signals). DAPI (4′,6-diamidino-2-phenylindole) nuclear counterstain (blue). B, Loss of chromosome 10 is the most frequent genetic alteration in GBM. PTEN gene (10q23) = green (mostly one signal), DMBT1 gene (10q25-q26) = red (mostly 1 signal).
Figure 5-13 A-C , Glioblastoma (GBM) with an oligodendroglial component. This example had foci that were identical to its purely astrocytic GBM counterparts ( A ), but at least focally, showed classic oligodendroglial cytology, including uniformly rounded nuclei ( B ) with delicate chromatin, small nucleoli, and crisp nuclear membranes ( C ). As with pure oligodendrogliomas, higher-grade oligodendroglial foci show similar cytologic features, but with larger, more epithelioid cells containing enlarged more pleomorphic, but generally rounded nuclei displaying vesicular chromatin and prominent nucleoli ( D ). GBM with an oligodendroglial component is synonymous with “grade IV oligoastrocytoma” and is associated with a somewhat more favorable prognosis than conventional GBM displaying purely astrocytic morphology.
Figure 5-14 Malignant gliomas with primitive neuroectodermal tumor (PNET)-like foci. These are thought to be clonally derived neoplasms where the primitive component forms relatively discrete nodules ( A ; right side ) arising from a diffuse glioma ( A ; left side ). This may occur in any type of diffuse glioma, although GBM and gliosarcoma are most common. The PNET-like component resembles central nervous system PNET or medulloblastoma and may form Homer Wright (neuroblastic) rosettes ( B ). In a further analogy to medulloblastoma, there are often large cell/anaplastic features ( C ), variably including increased cell size and pleomorphism, large nuclei with vesicular chromatin, prominent nucleoli, and cell wrapping ( arrows ). The PNET-like regions typically display immunoreactivity for one or more neuronal markers, such as synaptophysin ( D ), NeuN, neurofilament protein, chromogranin, and neuron-specific enolase. The diffuse glioma often displays extensive glial fibrillary acidic protein and vimentin immunoreactivity, whereas the primitive element shows only scattered positive tumor cells ( E ). Most cases show nearly diffuse p53 positivity ( F ) and the Ki-67 labeling index is markedly elevated in the PNET-like foci ( G ). Genetic alterations common to high-grade gliomas in general (e.g., 10q deletion) are often seen in both components, whereas N -myc or c -myc gene amplifications are found in roughly 40% of cases, being limited to the PNET-like regions ( H ; fluorescence in situ hybridization showing innumerable green N- myc and two to three red centromere 2 signals in most tumor nuclei). Cerebrospinal fluid dissemination of the primitive component is seen in up to 40% of cases.

Differential Diagnosis
Since the rounded nuclei and clear haloes of oligodendroglioma may be seen in other types of CNS neoplasms, the differential diagnosis is quite extensive. On a small biopsy specimen, dysembryoplastic neuroepithelial tumor (DNT) may be totally indistinguishable from oligodendroglioma if the tissue contains only the component that consists of bland, oligodendroglioma-like cells involving the cerebral cortex. In larger biopsy and resection specimens of DNT, the typical mucin-rich-patterned intracortical nodules and floating neurons can usually be identified. Similarly, the loose component of PA may look remarkably similar to oligodendroglioma, but usually has EGBs, RFs, a less infiltrative growth pattern, and GFAP-positive processes (see Figs. 5-17 and 5-18 ). The clinical presentation, radiologic appearance, and age of onset are often helpful in classifying a PA. Given that DNT and PA are benign, their accurate distinction from oligodendroglioma carries considerable prognostic and therapeutic impact.
The clear cell variant of ependymoma shares the haloes and rounded nuclei of oligodendroglioma, but in addition has a sharp interface with adjacent parenchyma that is more typical of ependymal neoplasms. Perivascular pseudorosettes may be less conspicuous in the clear cell variant than in other ependymomas, but are usually present on close inspection. Central neurocytoma is distinguished by its intraventricular location, neurocytic rosettes, and diffuse synaptophysin immunoreactivity. Extraventricular neurocytomas are less common, but are also diffusely synaptophysin-immunoreactive and display rosettes or neuropil-rich islands. Rare cases of oligodendroglioma with neurocytic differentiation and 1p/19q codeletions in both components have been reported recently, suggesting that there may be greater overlap between these two entities than previously thought. 69 Metastatic clear cell carcinomas , such as renal cell carcinoma, are discrete or noninfiltrative tumors and are positive for epithelial rather than glial markers. CNS lymphoma may show haloes, but is associated with greater nuclear irregularities than oligodendroglioma and is positive for lymphoid and B-cell markers. Clear cell meningioma is distinguished by both location and histology. It is an extra-axial rather than parenchymal tumor with a predilection for spinal cord and posterior fossa. As opposed to oligodendroglioma, clear cell meningioma demonstrates a pronounced vascular and interstitial collagenization, PAS-positive glycogen-rich cytoplasm, and EMA immunoreactivity.
The diagnostic overlap of oligodendroglioma with diffuse astrocytoma has already been discussed and is a common problem. In the particular case of anaplastic oligodendroglioma, the small-cell variant of GBM contains many overlapping features, including monomorphic deceptively bland nuclei, clear haloes, a chicken wire capillary network, and microcalcifications. 21 However, the nuclei of small-cell GBMs tend to be oval or elongate, rather than round; there is a surprisingly brisk proliferative index given the bland chromatin, mucin-filled microcystic spaces are generally absent; and GFAP often reveals long cytoplasmic processes in astrocytic tumor cells. Additionally, there may be more conventional fibrillary or gemistocytic astrocytic elements elsewhere in the tumor, and neuroimaging often reveals a rim-enhancing mass. Lastly, small-cell GBMs are genetically characterized by EGFR gene amplification (≈70%) and chromosome 10 losses (>90%), rather than the 1p and 19q codeletions that are typical of anaplastic oligodendroglioma ( Fig. 5-28 ).
Figure 5-28 Dual-color fluorescence in situ hybridizations. A, B, Oligodendroglioma with chromosome 1p ( A ) and 19q ( B ) deletions. A, 1p32 probe = green (mostly 1 signal per tumor nucleus), 1q42 = red (mostly 2 signals). B, 19p13 = green (mostly 2 signals), 19q13 = red (mostly 1 signal). 4′,6-Diamidino-2-phenylindole (DAPI) nuclear counterstain ( blue ). C, D, Relative 1p ( C ) and 19q ( D ) deletions have the same favorable implications for prolonged survival and chemosensitivity, but are most often seen in anaplastic oligodendrogliomas with a large subset of near-tetraploid cells containing four reference probe and two test probe signals.

Ancillary Diagnostic Studies
Immunohistochemistry for GFAP is usually negative in classic oligodendroglioma cells ( Fig. 5-26F ). However, GFAP should not routinely be used as a marker to distinguish oligodendrogliomas from astrocytomas. Indeed, two forms of oligodendroglioma cells are strongly positive: minigemistocytes, which have rounded eccentric cytoplasmic bellies, and gliofibrillary oligodendrocytes, which show a thin rim of perinuclear staining, with or without thin cytoplasmic tail ( Fig. 5-26G ). 29, 70 S-100 protein is typically positive, while stains for neurofilament highlight entrapped axons and clearly demonstrate the infiltrative growth pattern of these tumors. Immunoreactivity for p53 protein is usually negative in classic oligodendrogliomas, and this finding can be helpful in distinguishing these tumors from diffuse astrocytomas in some instances. Nevertheless, this association is far from perfect, with plenty of exceptions to this rule. The MIB-1 labeling index is highly variable and depends on the grade. Grade II tumors typically show proliferation indices between 1% and 6%, whereas anaplastic (grade III) variants range from 5% to more than 20%. Surprisingly, immunoreactivity for neuronal markers is relatively common, especially for synaptophysin . In fact, a paranuclear dotlike pattern of immunoreactivity seems to be particularly common in morphologically classic oligodendrogliomas containing 1p and 19q codeletions ( Fig. 5-26H ).
Ultrastructural studies have demonstrated electron-lucent, organelle-poor cells with minimal accumulations of intermediate filaments, except in minigemistocytes and gliofibrillary oligodendrocytes.

Genetics
The genetic characterization of oligodendrogliomas has provided important advances in the diagnosis and treatment of diffuse gliomas. A characteristic codeletion of chromosomal arms 1p and 19q is found in 50% to 90% of cases, particularly those with the most classic histology (see Fig. 5-28 ). 71 – 73 Furthermore, this genetic signature is associated with both prolonged survival time and a favorable response to procarbazine, CCNU, vincristine (PCV) and temozolomide chemotherapy or radiation therapy. 71, 74, 75 Based on these findings, ancillary testing for 1p and 19q status has become routine in many medical centers. The most commonly utilized techniques include FISH and LOH, each with its advantages and disadvantages. Similar to their astrocytic counterparts, p16 deletions are common progression-associated alterations. However, TP53 mutation, EGFR gene amplification, and chromosome 10 losses are uncommon and suggest the possibility of an astrocytic, rather than oligodendroglial, neoplasm.

Treatment and Prognosis
The prognosis for grade II oligodendrogliomas is significantly better than for grade II astrocytomas, with average survival times of 10 to 15 years. Patients with the genetically favorable (1p/19q deleted) variant may survive even longer. As with astrocytomas though, considerable individual variability exists in time to progression and overall survival. The average survival for anaplastic (grade III) oligodendrogliomas is 3 to 5 years, though similarly, some patients with the genetically favorable subset may survive 10 years or longer despite the high-grade histology. Age is a powerful predictor of prognosis as well, with survival being inversely related to age at diagnosis. Genetically favorable oligodendrogliomas with 1p/19q codeletion have been shown to be more likely to respond to chemotherapy and radiation. Patients are often stratified into therapeutic groups according to age, extent of resection, tumor grade, and 1p/19q status. Radiotherapy is used to treat most cases when there is subtotal resection, anaplasia, or the patient is older (e.g., >40 years). Chemotherapy is more often used in those oligodendrogliomas with chromosome 1p and 19q deletions, potentially reserving radiation for use at the time of recurrence given the long life expectancy and potential long-term morbidity associated with radiation damage. Therefore, genetic parameters are often used for both prognosis and for guiding management, though there is still much debate over the optimal therapeutic approaches and timing of various therapeutic modalities.

Oligoastrocytoma


Definitions and Synonyms
Oligoastrocytomas are diffusely infiltrating gliomas that consist of an admixture of tumor cells with oligodendroglial and astrocytic differentiation, either in geographically distinct zones (biphasic or compact variant) or, much more commonly, intermixed (intermingled or diffuse variant). These tumors occur most frequently in the cerebral hemispheres of adults and can correspond to WHO grades II, III, or IV. 1

Incidence and Demographics
The true incidence of mixed oligoastrocytomas (MOAs) can only be estimated, since this designation depends heavily on pathologic criteria used to establish the diagnosis, which varies considerably among pathologists. In large published series, MOAs account for 5% to 20% of all diffuse gliomas, amounting to an annual incidence rate of approximately 0.4 per 100,000 person-years. They occur mostly in adults, peaking in the 30s and 40s, and are slightly more common in males than females (3:2).

Clinical Manifestations and Localization
The clinical features of MOA are similar to those of pure astrocytomas and pure oligodendrogliomas of similar grade and depend on location and rate of growth. 77, 78 Among low-grade tumors, seizures are the most common clinical manifestation, followed by motor or sensory dysfunction, personality changes, and signs of increased intracranial pressure. Anaplastic MOAs and GBMs with an oligodendroglial component may have similar signs and symptoms, but the temporal evolution is more rapid.

Radiologic Features and Gross Pathology
No specific radiologic or gross pathologic features distinguish MOAs from pure oligodendrogliomas or diffuse astrocytomas. Thus, low-grade (WHO grade II) MOAs are T1-hypointense, T2- or FLAIR hyperintense, and do not enhance following the administration of contrast medium. The anaplastic (WHO grade III) MOAs show greater mass effect and nearly always show some degree of contrast enhancement. GBMs with an oligodendroglial component (“grade IV MOAs”) are often rim-enhancing masses.

Histopathology
The histologic features of the oligoastrocytomas are similar to those of oligodendrogliomas and astrocytomas, but both elements are present within the same neoplasm. 76, 77 The WHO recognizes two types of MOA: a biphasic (compact) variant with anatomically separate areas resembling oligodendroglioma and astrocytoma ( Fig. 5-29A ) and an intermixed (diffuse) variant in which the two elements are intermingled ( Fig. 5-29B ). The intermixed form is by far the more common and diagnostically challenging type, often displaying nuclei with intermediate or ambiguous features, as well as more classic types. The presence of mucin-rich microcystic spaces is a common, albeit nonspecific, feature ( Fig. 5-29C ). The recognition of an oligodendroglial component within an infiltrating glioma is significant, since its presence is associated with a longer survival time than astrocytomas of the same grade. 76, 77 The minimal percentage of each component required for the diagnosis of a mixed glioma has been debated. One study suggested that a single 100 × field filled with an oligodendroglioma component could be used as a threshold for mixed oligoastrocytomas. 78 This criterion identifies a subset with a better prognosis than astrocytoma and results in improved interobserver concordance among pathologists. In the intermixed variant at least some of the cells in the tumor should display classic oligodendroglial cytology (see Oligodendroglioma section).
Figure 5-29 Mixed oligoastrocytomas. A, Mixed oligoastrocytomas can contain spatially separate astrocytic ( bottom ) and oligodendroglial ( top ) components. However, this biphasic or “compact” form is quite rare. B, Tumor cells with intermingled astrocytic ( arrows ) and oligodendroglial ( arrowheads ) cytologic features are much more common (“diffuse subtype”). In this variant, cells with intermediate or indeterminate nuclear features are also common. C, Microcystic spaces are typical in well differentiated regions. D, Anaplastic oligoastrocytomas are characterized by increased nuclear pleomorphism, mitotic activity ( arrow ), and microvascular hyperplasia ( left side ).
Because the designation of MOA is sometimes mistakenly used for infiltrating gliomas with ambiguous morphologic features (i.e., it has become a “wastebasket”) rather than for truly mixed tumors, it is not surprising that MOAs are associated with the least diagnostic concordance and reproducibility of all glioma designations. Nonetheless, the MOA diagnosis remains a viable category based on pathologic features, and these tumors have intermediate prognosis between that encountered in the two pure gliomas of the same grade.

Histologic Variants and Grading
The grading of MOAs is similar to that utilized for oligodendrogliomas, with the two major categories being low grade (WHO grade II) and anaplastic (WHO grade III). 1 However, the 2007 WHO scheme now includes a grade IV variant as well, referred to as GBM with an oligodendroglial component. The presence of hypercellularity, pleomorphism, excessive mitotic activity, and endothelial hyperplasia are most consistent with the diagnosis of anaplastic oligoastrocytomas , WHO grade III ( Fig. 5-29D ). The threshold level of mitotic activity for a grade III designation has not been firmly established. However, the threshold established for pure oligodendrogliomas (>6/10hpf) has been used by some. 68
The grading and classification of anaplastic MOAs that have necrosis has been the subject of recent investigation. In a large series of patients with anaplastic MOAs, the presence of necrosis divided tumors into prognostically distinct groups, with the mean survival time being significantly shorter for patients with anaplastic oligoastrocytomas with necrosis than for patients whose tumors lacked necrosis. 79 The presence of endothelial proliferation was not prognostically significant among anaplastic mixed tumors. Based on these findings, the WHO now recognizes anaplastic mixed gliomas with necrosis as “ GBM with an oligodendroglioma component ” (GBMO) (see Fig. 5-13 ), although a smaller contingency of neuropathologists prefer the term “grade IV MOA.” 1

Differential Diagnosis
As discussed earlier, the most challenging considerations when establishing the diagnosis of a MOA are pure astrocytoma and pure oligodendroglioma. Of course, all of the other differential considerations associated with these two entities also come into play. These include DNT, PA, and neurocytic tumors, among others.

Ancillary Diagnostic Studies
The immunohistochemical tests used in the diagnosis of MOAs are similar to those used in the diagnosis of oligodendrogliomas and astrocytomas. Thus, GFAP is used to demonstrate glial differentiation, if necessary, while neurofilament can assist in establishing an infiltrative growth pattern. Immunohistochemistry for p53 positivity may be helpful in establishing a neoplastic diagnosis or in supporting the presence of an astrocytic element in a MOA. FISH studies of 1p/19q codeletion are used diagnostically to support the presence of an oligodendroglioma component and also predict a more favorable clinical course and response to chemotherapy. 80 MIB-1 (Ki-67) proliferative indices correlate with grade in the MOAs.

Genetics
Most cases of oligoastrocytoma are either genetically similar to pure astrocytomas (e.g., TP53 mutations, monosomy 10, EGFR gene amplification) or pure oligodendrogliomas (1p/19q deletions), typically throughout the tumor regardless of regional differences in histologic appearances (i.e., monoclonal). 81, 82 A subset of cases does not have any of these alterations. The same clonal aberrations are found throughout the entire tumor, regardless of the morphologic component being examined. Studies suggest that 25% to 50% of biphasic MOAs have 1p/19q codeletion, whereas the intermixed variant shows this pattern much less frequently (≈10%–20%).

Treatment and Prognosis
The prognosis for MOAs is generally intermediate between that of pure astrocytoma and oligodendroglioma of similar grade, though there is great individual variability. 76, 77 Average survival times are 5 to 10 years for grade II MOAs , 2 to 4 years for grade III MOAs , and roughly 22 months for GBMO . 79 Treatment recommendations are usually identical to those given to patients with pure oligodendroglioma of similar grade. Thus, gross total resection and tumor debulking are believed to be beneficial for these lesions. Radiotherapy is often used in cases of subtotal resection, anaplastic tumors, and for older patients (e.g., >40 years). Chemotherapy is also beneficial, especially in those tumors that are 1p and 19q codeleted.
A recent FISH study showed that MOAs with genetic alterations typically associated with astrocytoma progression (e.g., 10q deletion, EGFR amplification, 9p or p16 deletion) had poor survival times (average of 16 months), whereas those with 1p/19q codeletion, solitary 19q deletion, or none of the investigated genetic alterations had favorable survival times (averages of 7–8 years). Nevertheless, patient age and histologic grade remained the two most powerful predictors of outcome in this cohort. 80

Suggested Readings

Brat, D. J., Prayson, R. A., Ryken, T. C., Olson, J. J. Diagnosis of malignant glioma: role of neuropathology. J Neurooncol. . 2008; 89:287–311.
Burger, P.C., Scheithauer, B.W.American Registry of Pathology. Washington, DC: Armed Forces Institute of Pathology (U.S.): Tumors of the Central Nervous System. American Registry of Pathology in collaboration with the Armed Forces Institute of Pathology, 2007.
CBTRUSStatistical Report: Primary Brain Tumors in the United States, 2000–2004. Central Brain Tumor Registry of the United States, 2008. [Published by the].
Louis, D.N., Ohgaki, H., Wiestler, O.D., Cavenee, W.K.WHO Classification of Tumours of the Central Nervous System. Lyon: International Agency for Research, 2007.
Miller, C. R., Perry, A. Glioblastoma. Arch Pathol Lab Med. . 2007; 131:397–406.

References

1. Louis, D. N., Ohgaki, H., Wiestler, O. D., Cavenee, W. K.WHO Classification of Tumours of the Central Nervous System. Lyon: International Agency for Research, 2007.
2. Burger, P. C., Scheithauer, B. W.American Registry of Pathology. Washington, DC: Armed Forces Institute of Pathology (U.S.): Tumors of the Central Nervous System. American Registry of Pathology in collaboration with the Armed Forces Institute of Pathology, 2007.
3. CBTRUSStatistical Report: Primary Brain Tumors in the United States, 2000–2004. Central Brain Tumor Registry of the United States, 2008. [Published by the].
4. Bailey, P., Cushing, H.A Classification of Tumors of the Glioma Group on a Histogenetic Basis with a Correlated Study of Prognosis. Philadelphia: JB Lippincott, 1926.
5. Ohgaki, H., Kleihues, P. Genetic pathways to primary and secondary glioblastoma. Am J Pathol. . 2007; 170:1445–1453.
6. Burger, P. C., Nelson, J. S., Boyko, O. B. Diagnostic synergy in radiology and surgical neuropathology: radiographic findings of specific pathologic entities. Arch Pathol Lab Med. . 1998; 122:620–632.
7. Burger, P. C., Nelson, J. S., Boyko, O. B. Diagnostic synergy in radiology and surgical neuropathology: neuroimaging techniques and general interpretive guidelines. Arch Pathol Lab Med. . 1998; 122:609–619.
8. Hoffman, J. M. New advances in brain tumor imaging. Curr Opin Oncol. . 2001; 13:148–153.
9. Gupta, M., Djalilvand, A., Brat, D. J. Clarifying the diffuse gliomas: an update on the morphologic features and markers that discriminate oligodendroglioma from astrocytoma. Am J Clin Pathol. . 2005; 124:755–768.
10. Brat, D. J., Prayson, R. A., Ryken, T. C., Olson, J. J. Diagnosis of malignant glioma: role of neuropathology. J Neurooncol. . 2008; 89:287–311.
11. Perry, A. Pathology of low-grade gliomas: an update of emerging concepts. Neuro Oncol. . 2003; 5:168–178.
12. Giannini, C., Scheithauer, B. W., Burger, P. C., . Cellular proliferation in pilocytic and diffuse astrocytomas. J Neuropathol Exp Neurol. . 1999; 58:46–53.
13. Miller, C. R., Perry, A. Glioblastoma. Arch Pathol Lab Med. . 2007; 131:397–406.
14. Rong, Y., Durden, D. L., Van Meir, E. G., Brat, D. J. ‘Pseudopalisading’ necrosis in glioblastoma: a familiar morphologic feature that links vascular pathology, hypoxia, and angiogenesis. J Neuropathol Exp Neurol. . 2006; 65:529–539.
15. Brat, D. J., Van Meir, E. G. Glomeruloid microvascular proliferation orchestrated by VPF/VEGF: a new world of angiogenesis research. Am J Pathol. . 2001; 158:789–796.
16. Brat, D. J., Scheithauer, B. W., Medina-Flores, R., . Infiltrative astrocytomas with granular cell features (granular cell astrocytomas): a study of histopathologic features, grading, and outcome. Am J Surg Pathol. . 2002; 26:750–757.
17. Watanabe, K., Peraud, A., Gratas, C., . p53 and PTEN gene mutations in gemistocytic astrocytomas. Acta Neuropathol (Berl) . 1998; 95:559–564.
18. Peraud, A., Watanabe, K., Schwechheimer, K., . Genetic profile of the giant cell glioblastoma. Lab Invest. . 1999; 79:123–129.
19. Biernat, W., Aguzzi, A., Sure, U., . Identical mutations of the p53 tumor suppressor gene in the gliomatous and the sarcomatous components of gliosarcomas suggest a common origin from glial cells. J Neuropathol Exp Neurol. . 1995; 54:651–656.
20. Reis, R. M., Konu-Lebleblicioglu, D., Lopes, J. M., . Genetic profile of gliosarcomas. Am J Pathol. . 2000; 156:425–432.
21. Perry, A., Aldape, K. D., George, D. H., Burger, P. C. Small cell astrocytoma: an aggressive variant that is clinicopathologically and genetically distinct from anaplastic oligodendroglioma. Cancer . 2004; 101:2318–2326.
22. Burger, P. C., Pearl, D. K., Aldape, K., . Small cell architecture—A histological equivalent of EGFR amplification in glioblastoma multiforme? J Neuropathol Exp Neurol. . 2001; 60:1099–1104.
23. Perry, A., Miller, C. R., Gujrati, M., . Malignant gliomas with primitive neuroectodermal tumor-like components: a clinicopathologic and genetic study of 53 cases. Brain Pathol. . 2009; 19:81–90.
24. Kim, D. G., Yang, H. J., Park, I. A., . Gliomatosis cerebri: clinical features, treatment, and prognosis. Acta Neurochir (Wien) . 1998; 140:755–762.
25. Daumas-Duport, C., Scheithauer, B., O’Fallon, J., Kelly, P. Grading of astrocytomas. A simple and reproducible method. Cancer . 1988; 62:2152–2165.
26. Coons, S. W., Pearl, D. K. Mitosis identification in diffuse gliomas: implications for tumor grading. Cancer . 1998; 82:1550–1555.
27. Schittenhelm, J., Mittelbronn, M., Nguyen, T. D., . WT1 expression distinguishes astrocytic tumor cells from normal and reactive astrocytes. Brain Pathol. . 2008; 18:344–353.
28. Cosgrove, M., Fitzgibbons, P. L., Sherrod, A., . Intermediate filament expression in astrocytic neoplasms. Am J Surg Pathol. . 1989; 13:141–145.
29. Herpers, M. J., Budka, H. Glial fibrillary acidic protein (GFAP) in oligodendroglial tumors: gliofibrillary oligodendroglioma and transitional oligoastrocytoma as subtypes of oligodendroglioma. Acta Neuropathol (Berl) . 1984; 64:265–272.
30. Oh, D., Prayson, R. A. Evaluation of epithelial and keratin markers in glioblastoma multiforme: an immunohistochemical study. Arch Pathol Lab Med. . 1999; 123:917–920.
31. Hunter, S. B., Brat, D. J., Olson, J. J., . Alterations in molecular pathways of diffusely infiltrating glial neoplasms: application to tumor classification and anti-tumor therapy (Review). Int J Oncol. . 2003; 23:857–869.
32. Watanabe, T., Nobusawa, S., Kleihues, P., Ohgaki, H. IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol. . 2009; 174:1149–1153.
33. Yan, H., Parsons, D. W., Jin, G., . IDH1 and IDH2 mutations in gliomas. N Engl J Med. . 2009; 360:765–773.
34. TCGA. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature . 2008; 455:1061–1068.
35. Liu, L., Backlund, L. M., Nilsson, B. R., . Clinical significance of EGFR amplification and the aberrant EGFRvIII transcript in conventionally treated astrocytic gliomas. J Mol Med. . 2005; 83:917–926.
36. Mellinghoff, I. K., Wang, M. Y., Vivanco, I., . Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors. N Engl J Med. . 2005; 353:2012–2024.
37. Hegi, M. E., Diserens, A. C., Gorlia, T., . MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med. . 2005; 352:997–1003.
38. Stupp, R., Mason, W. P., van den Bent, M. J., . Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. . 2005; 352:987–996.
39. Rodriguez, F. J., Thibodeau, S. N., Jenkins, R. B., . MGMT immunohistochemical expression and promoter methylation in human glioblastoma. Appl Immunohistochem Mol Morphol. . 2008; 16:59–65.
40. Giannini, C., Scheithauer, B. W. Classification and grading of low-grade astrocytic tumors in children. Brain Pathol. . 1997; 7:785–798.
41. Forsyth, P. A., Shaw, E. G., Scheithauer, B. W., . Supratentorial pilocytic astrocytomas. A clinicopathologic, prognostic, and flow cytometric study of 51 patients. Cancer . 1993; 72:1335–1342.
42. Komotar, R. J., Burger, P. C., Carson, B. S., . Pilocytic and pilomyxoid hypothalamic/chiasmatic astrocytomas. Neurosurgery . 2004; 54:72–79. [discussion 79–80].
43. Tihan, T., Fisher, P. G., Kepner, J. L., . Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol. . 1999; 58:1061–1068.
44. Pfister, S., Janzarik, W. G., Remke, M., . BRAF gene duplication constitutes a mechanism of MAPK pathway activation in low-grade astrocytomas. J Clin Invest. . 2008; 118:1739–1749.
45. Deshmukh, H., Yeh, T. H., Yu, J., . High-resolution, dual-platform aCGH analysis reveals frequent HIPK2 amplification and increased expression in pilocytic astrocytomas. Oncogene . 2008; 27:4745–4751.
46. Hunter, S., Young, A., Olson, J., . Differential expression between pilocytic and anaplastic astrocytomas: identification of apolipoprotein D as a marker for low-grade, non-infiltrating primary CNS neoplasms. J Neuropathol Exp Neurol. . 2002; 61:275–281.
47. Rodriguez, F. J., Giannini, C., Asmann, Y. W., . Gene expression profiling of NF-1-associated and sporadic pilocytic astrocytoma identifies aldehyde dehydrogenase 1 family member L1 (ALDH1L1) as an underexpressed candidate biomarker in aggressive subtypes. J Neuropathol Exp Neurol. . 2008; 67:1194–1204.
48. Tibbetts, K. M., Emnett, R. J., Gao, F., . Histopathologic predictors of pilocytic astrocytoma event-free survival. Acta Neuropathol . 2009. [117:657–665].
49. Hirose, T., Scheithauer, B. W., Lopes, M. B., . Tuber and subependymal giant cell astrocytoma associated with tuberous sclerosis: an immunohistochemical, ultrastructural, and immunoelectron and microscopic study. Acta Neuropathol. . 1995; 90:387–399.
50. Shepherd, C. W., Scheithauer, B. W., Gomez, M. R., . Subependymal giant cell astrocytoma: a clinical, pathological, and flow cytometric study. Neurosurgery . 1991; 28:864–868.
51. Roach, E. S., Gomez, M. R., Northrup, H. Tuberous sclerosis complex consensus conference: revised clinical diagnostic criteria. J Child Neurol . 1998;; 13:624–628.
52. Kingsley, D. P., Kendall, B. E., Fitz, C. R. Tuberous sclerosis: a clinicoradiological evaluation of 110 cases with particular reference to atypical presentation. Neuroradiology . 1986; 28:38–46.
53. Lopes, M. B., Altermatt, H. J., Scheithauer, B. W., . Immunohistochemical characterization of subependymal giant cell astrocytomas. Acta Neuropathol. . 1996; 91:368–375.
54. Cuccia, V., Zuccaro, G., Sosa, F., . Subependymal giant cell astrocytoma in children with tuberous sclerosis. Childs Nerv Syst. . 2003; 19:232–243.
55. Kepes, J. J. Pleomorphic xanthoastrocytoma: the birth of a diagnosis and a concept. Brain Pathol. . 1993; 3:269–274.
56. Kepes, J. J., Rubinstein, L. J., Eng, L. F. Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis; a study of 12 cases. Cancer . 1979; 44:1839–1852.
57. Rippe, D. J., Boyko, O. B., Radi, M., . MRI of temporal lobe pleomorphic xanthoastrocytoma. J Comput Assist Tomogr. . 1992; 16:856–859.
58. Tien, R. D., Cardenas, C. A., Rajagopalan, S. Pleomorphic xanthoastrocytoma of the brain: MR findings in six patients. Am J Roentgenol . 1992; 159:1287–1290.
59. Giannini, C., Scheithauer, B., Burger, P., . Pleomorphic xanthoastrocytoma: what do we really know about it? Cancer . 1999; 85:2033–2045.
60. Perry, A., Giannini, C., Scheithauer, B. W., . Composite pleomorphic xanthoastrocytoma and ganglioglioma: report of four cases and review of the literature. Am J Surg Pathol. . 1997; 21:763–771.
61. Giannini, C., Scheithauer, B. W., Lopes, M. B., . Immunophenotype of pleomorphic xanthoastrocytoma. Am J Surg Pathol. . 2002; 26:479–485.
62. Paulus, W., Lisle, D. K., Tonn, J. C., . Molecular genetic alterations in pleomorphic xanthoastrocytoma. Acta Neuropathol (Berl) . 1996; 91:293–297.
63. Yin, X. L., Hui, A. B., Liong, E. C., . Genetic imbalances in pleomorphic xanthoastrocytoma detected by comparative genomic hybridization and literature review. Cancer Genet Cytogenet. . 2002; 132:14–19.
64. Bailey, P., Bucy, P. C. Oligodendrogliomas of the brain. J Pathol Bacteriol . 1929; 32:735–751.
65. Perry, A. Oligodendroglial neoplasms: current concepts, misconceptions, and folklore. Adv Anat Pathol. . 2001; 8:183–199.
66. Burger, P. C. What is an oligodendroglioma? Brain Pathol. . 2002; 12:257–259.
67. Shaw, E. G., Scheithauer, B. W., O’Fallon, J. R., . Oligodendrogliomas. The Mayo Clinic experience. J Neurosurg. . 1992; 76:428–434.
68. Giannini, C., Scheithauer, B. W., Weaver, A. L., . Oligodendrogliomas: reproducibility and prognostic value of histologic diagnosis and grading. J Neuropathol Exp Neurol. . 2001; 60:248–262.
69. Perry, A., Scheithauer, B. W., Macaulay, R. J., . Oligodendrogliomas with neurocytic differentiation. A report of 4 cases with diagnostic and histogenetic implications. J Neuropathol Exp Neurol. . 2002; 61:947–955.
70. Kros, J. M., Van Eden, C. G., Stefanko, S. Z., . Prognostic implications of glial fibrillary acidic protein containing cell types in oligodendrogliomas. Cancer . 1990; 66:1204–1212.
71. Perry, A., Fuller, C. E., Banerjee, R., . Ancillary FISH analysis for 1p and 19q status: preliminary observations in 287 gliomas and oligodendroglioma mimics. Front Biosci. . 2003; 8:a1–9.
72. Burger, P. C., Minn, A. Y., Smith, J. S., . Losses of chromosomal arms 1p and 19q in the diagnosis of oligodendroglioma. A study of paraffin-embedded sections. Mod Pathol. . 2001; 14:842–853.
73. Aldape, K., Burger, P. C., Perry, A. Clinicopathologic aspects of 1p/19q loss and the diagnosis of oligodendroglioma. Arch Pathol Lab Med. . 2007; 131:242–251.
74. Smith, J. S., Perry, A., Borell, T. J., . Alterations of chromosome arms 1p and 19q as predictors of survival in oligodendrogliomas, astrocytomas, and mixed oligoastrocytomas. J Clin Oncol. . 2000; 18:636–645.
75. Cairncross, J. G., Ueki, K., Zlatescu, M. C., . Specific genetic predictors of chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Cancer Inst. . 1998; 90:1473–1479.
76. Shaw, E. G., Scheithauer, B. W., O’Fallon, J. R. Supratentorial gliomas: a comparative study by grade and histologic type. J Neurooncol. . 1997; 31:273–278.
77. Shaw, E. G., Scheithauer, B. W., O’Fallon, J. R., Davis, D. H. Mixed oligoastrocytomas: a survival and prognostic factor analysis. Neurosurgery . 1994; 34:577–582. [discussion 582].
78. Coons, S. W., Johnson, P. C., Scheithauer, B. W., . Improving diagnostic accuracy and interobserver concordance in the classification and grading of primary gliomas. Cancer . 1997; 79:1381–1393.
79. Miller, C. R., Dunham, C. P., Scheithauer, B. W., Perry, A. Significance of necrosis in grading of oligodendroglial neoplasms: a clinicopathologic and genetic study of newly diagnosed high-grade gliomas. J Clin Oncol. . 2006; 24:5419–5426.
80. Fuller, C. E., Schmidt, R. E., Roth, K. A., . Clinical utility of fluorescence in situ hybridization (FISH) in morphologically ambiguous gliomas with hybrid oligodendroglial/astrocytic features. J Neuropathol Exp Neurol. . 2003; 62:1118–1128.
81. Mueller, W., Hartmann, C., Hoffmann, A., . Genetic signature of oligoastrocytomas correlates with tumor location and denotes distinct molecular subsets. Am J Pathol. . 2002; 161:313–319.
82. Maintz, D., Fiedler, K., Koopman, J., . Molecular genetic evidence for subtypes of oligoastrocytomas. J Neuropathol Exp Neurol. . 1997; 56:1098–1104.
6
Ependymomas and Choroid Plexus Tumors
Christine E. Fuller,
Sonia Narendra

Definitions and Synonyms   103
Brief Historical Overview   103
Ependymal Tumors   103
Choroid Plexus Tumors   113
Ependyma and choroid plexus play critical structural and biologic functions within the central nervous system (CNS), and the neoplasms that recapitulate these cell types are the most common intraventricular and intramedullary spinal cord tumors. This chapter covers the various categories of ependymal and choroid plexus neoplasms and presents a thoughtful approach to differentiating these tumors from possible diagnostic mimics in both adult and pediatric patients. Special attention is given to ancillary techniques that are useful in diagnostically challenging cases, as well as to the controversial issue of histologic grading.

Definitions and Synonyms
Ependymomas represent a group of gliomas with morphologic and ultrastructural evidence of predominantly or exclusively ependymal differentiation, as opposed to a growing list of tumors with only partial or limited ependymal features. The latter include astroblastoma, chordoid glioma, papillary tumor of the pineal region, and angiocentric glioma (covered elsewhere). The most recent rendition of the World Health Organization (WHO) classification system recognizes the following tumor categories: subependymoma (WHO grade I), myxopapillary ependymoma (WHO grade I), ependymoma (WHO grade II) including a number of variants, and anaplastic ependymoma (WHO grade III). Ependymoblastoma, an embryonal tumor typified by multilayered (ependymoblastic) rosettes, is now considered a form of primitive neuroectodermal tumor (PNET) with focal ependymal differentiation and is similarly discussed elsewhere. Intraventricular papillary neoplasms recapitulating choroid plexus epithelium include choroid plexus papilloma (WHO grade I), atypical choroid plexus papilloma (WHO grade II), and choroid plexus carcinoma (WHO grade III). 1

Brief Historical Overview
Dating back to early perspectives from Bailey and Cushing, 2 our concepts of ependymoma histogenesis have been related to embryology and the stages of normal ependymal cell development. This “stem cell” or “progenitor cell” theory of tumorigenesis proposes that radial glia, the multipotent neuroglial progenitor cells, give rise to multiple different populations of elongate uni- and bipolar cells termed tanycytes ; fetal ependymal tanycytes directly give rise to mature ependymocytes, whereas the more highly specialized ependymal cells of the circumventricular organs and choroid plexus are ultimately derived from this developmental pathway. 3 Choroid plexus tumors and ependymomas (including the various histologic subtypes) clearly recapitulate specific cell types found at various stages in this ontologic sequence.
In more recent years, our focus has turned to accurate tumor classification and the development of objective histologic grading criteria for ependymomas and choroid plexus tumors in an attempt to more effectively stratify patients into distinctive prognostic categories. We have made significant strides in both areas, with definitive diagnostic criteria set forth for clear cell and other ependymoma variants, as well as the previously poorly defined atypical choroid plexus papilloma. Despite these triumphs, ependymoma grading and the concept of focal anaplasia remain unresolved issues awaiting further clarification.

Ependymal Tumors
Incidence and Demographics
Ependymomas represent slightly less than 10% of all neuroepithelial tumors; they are the most common primary tumor of the spinal cord, and the third most common pediatric CNS tumor, accounting for up to 30% of intracranial tumors in children younger than 3 years of age. 1 They display a bimodal age distribution, with peak incidences at ages of 6 and 30 to 40 years, respectively. The vast majority of pediatric ependymomas arise intracranially, but more than 60% of adult ependymomas are centered in the spinal cord . 4 Ependymomas have an equal gender distribution, though they are nearly twice as frequent in caucasians than in African-Americans. Though most are sporadic, they may also be seen as part of neurofibromatosis type 2, nearly all of which are spinal (see Chapter 20 ). Subependymomas are often incidental autopsy findings in the brains of older adults and represent approximately 10% of all ependymal tumors. They are uncommon in children. 5 In contradistinction, anaplastic ependymomas are far more frequent in the pediatric age group. Arising predominantly in adults, only 10% to 20% of myxopapillary ependymomas manifest in children. They show a 2:1 male-to-female bias. 6, 7

Localization and Clinical Manifestations
In children, the most common site of involvement by ependymal tumors (predominantly WHO grade II and III categories) is the posterior fossa/ fourth ventricle followed by supratentorial localization, the latter showing an equal mix of primarily intraventricular and intraparenchymal tumors. 8 Supratentorial tumors (including ependymomas and subependymomas) more frequently involve the lateral ventricles than the third ventricle. However, many supratentorial ependymomas appear to be centered in the cortex or subcortical white matter. Therefore, a lack of ventricular involvement does not exclude ependymoma from the differential diagnosis. Intracranial ependymomas often become symptomatic when their growth results in blockage of cerebrospinal fluid (CSF) pathways, causing signs and symptoms related to hydrocephalus and increased intra-cranial pressure. These include ataxia, headache, nausea and vomiting, strabismus, irritability, and altered mental status; macrocephaly and bulging fontanelles may be encountered in affected infants. Clinical signs and symptoms of anaplastic ependymoma are similar to those of low-grade ependymoma, although they tend to develop in an accelerated fashion.
Ependymal tumors may arise at any level of the spinal cord , though certain histologic subtypes have preferred locations. For example, conventional ependymomas, including the tanycytic variant, typically manifest as central intramedullary tumors within the thoracic/cervicothoracic cord, 4, 9 whereas subependymomas more often arise within the cervical cord in an eccentric fashion. 10, 11 Spinal ependymomas are only rarely anaplastic. Myxopapillary ependymomas invariably arise in the region of the filum terminale ; infrequent sites of origin include other spinal cord levels, intracranial sites (both intraventricular and intraparenchymal), and subcutaneous sacrococcygeal areas. 12 – 19 As opposed to the typically benign behavior of intradural myxopapillary ependymomas, the soft tissue variant has a relatively high incidence of systemic metastasis, most often to the lung. Any of the spinal ependymal tumors may cause back pain and motor or sensory deficits, depending on their specific anatomic involvement. Rare extraneural sites for ependymomas include the ovaries, mediastinum, and sacrococcygeum.
Ependymomas and anaplastic ependymomas occasionally metastasize via subarachnoid spread to seed other spinal and intracranial locations; rare extracranial metastases have also been reported. 20 Pediatric myxopapillary ependymomas more often disseminate through the CSF pathways, a feature not typical of the adult counterpart. 6 Although they uncommonly recur, subependymomas do not otherwise show metastatic potential. 5

Radiologic Features and Gross Pathology
Conventional ependymomas commonly involve several contig- uous spinal segments (three, on average) and grow as sausage-shaped centrally situated intramedullary tumors with discreet margins.

  • Accueil Accueil
  • Univers Univers
  • Ebooks Ebooks
  • Livres audio Livres audio
  • Presse Presse
  • BD BD
  • Documents Documents