Reconstruction and visualization of neuronal pathways from diffusion tensor data [Elektronische Ressource] = Rekonstruktion und Visualisierung neuronaler Bahnen aus Diffusionstensor-Daten / vorgelegt von Dorit Merhof

Sujets

• Gesundheit

Informations

• Publié par: friedrich-alexander-universitat_erlangen-nurnberg
• Publié le: 01 janvier 2008
• Nombre de lectures: 2
• Langue: English
• Poids de l'ouvrage: 13 Mo

Extrait

ReconstructionandVisualization
ofNeuronalPathways
fromDiffusionTensorData
RekonstruktionundVisualisierung
neuronalerBahnenausDiffusionstensor-Daten
DerTechnischenFakultätder
UniversitätErlangen–Nürnberg
zurErlangungdesGrades
DOKTOR–INGENIEUR
vorgelegtvon
Dipl.–Inf. DoritMerhof
Erlangen—2007AlsDissertationgenehmigtvon
derTechnischenFakultät
derUniversitätErlangen–Nürnberg
TagderEinreichung: 29.06.2007
TagderPromotion: 08.10.2007
Dekan: Prof. Dr.–Ing. J.Huber
Berichterstatter: Prof. Dr. G.Greiner
Prof. Dr.–Ing. B.PreimRevision1.0
c 2007,CopyrightDoritMerhof
AllRightsReserved
AlleRechtevorbehalteni
Abstract
Fordiagnosisandsurgicalplanning,magneticresonanceimaging(MRI)hasbecomean
important source of medical data. About a decade ago, a novel MRI technique called
diffusion tensor imaging (DTI) evolved. Due to its ability to reflect the location and
structure of fibrous tissue such as white matter in vivo, this technique has gained in-
creasinginterestindifferentresearchdisciplines.
Forneurosurgery,DTIdataisofhighvaluesinceinformationaboutthelocationand
the course of white matter tract systems is provided, thus supporting the anatomical
information obtained from MRI. White matter tracts, i.e. motor or sensory pathways,
areimportantstructureswithinthehumanbrain. Inordertoavoidneurologicaldeficits
afterbrainsurgery,thesefibertractsmustremainintact.
However, the reconstruction of neuronal structures from DTI data is a non-trivial
task due to the complextensor information that is captured pervoxel. For this reason,
extensive research has been conducted in recent years in order to develop techniques
fortheprocessingandvisualizationofDTItensordata.
Thisworkcontributes tocurrentresearchandintroducesnewtechniquesforthere-
construction and visualization of white matter tracts. The different approaches were
developedincollaboration with neurosurgeons andare intendedtosupport preopera-
tiveplanningandintraoperativeguidanceinsurgicalinterventions. Forthispurpose,a
DTI toolbox comprising dataset processing, tensor reconstruction, filtering techniques,
fibertrackingandconnectivityanalysis,hullalgorithms anddifferentvisualizationap-
proacheshasbeendeveloped.
In the future, the research currently conducted in the field of DTI will contribute
to the further improvement of planning in neurosurgery and to the reduction of the
inherentriskofpostoperativeneurologicaldeficitsforthepatients.ii Abstractiii
Contents
Abstract i
Contents iii
ListofFigures vii
ListofTables ix
ListofCodeExamples xi
Acknowledgements xiii
I Introduction 1
1 Backgroundandoverview 3
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Neurosurgery: History,diseasepatterns,andoperativeprocedures 9
2.1 Historyofneurosurgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Diseasepatternstreatedinneurosurgery . . . . . . . . . . . . . . . . . . . 10
2.3 Neurosurgicalprocedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Clinicalsetupintheoperatingroom . . . . . . . . . . . . . . . . . . 11
2.3.2 Navigationinimage-guidedneurosurgery . . . . . . . . . . . . . . 12
2.4 Neuroanatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Imagingtechniquesappliedinneurosurgery . . . . . . . . . . . . . . . . . 16
2.5.1 Physicalbackground . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.5.2 Magneticresonanceimaging(MRI) . . . . . . . . . . . . . . . . . . 18
2.5.3 Functionalmagneticresonanceimaging(fMRI) . . . . . . . . . . . 19iv Contents
2.5.4 Diffusiontensorimaging(DTI) . . . . . . . . . . . . . . . . . . . . . 20
2.5.5 Imagingparametersandimagedata . . . . . . . . . . . . . . . . . . 20
II Reconstructionofneuronalstructures 23
3 Tensorreconstructionandprocessing 25
3.1 TheStejskal-Tannerequation . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Diffusiontensorreconstruction . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Quantifyingtensorshape . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3.1 Tensoreigensystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3.2 Diffusionellipsoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.3.3 Scalarmetrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Tensorinterpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5 Tensorfiltering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.5.1 Gaussianfilter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.5.2 Two-classGaussianweighting . . . . . . . . . . . . . . . . . . . . . 33
3.5.3 Best-neighborfilter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4 Fibertracking 39
4.1 Basicsabouttrackingalgorithms . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1.1 Streamlinepropagationandtensordeflection . . . . . . . . . . . . . 40
4.1.2 Tract-selectiontechniques . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2 Localregularizationfilters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.1 Localfilterwindow . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2.2 Gaussianlocalregularizationfilter . . . . . . . . . . . . . . . . . . . 46
4.2.3 Movingleast-squaresfilter . . . . . . . . . . . . . . . . . . . . . . . 47
4.2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 Fibertrackingusingevenlyspacedstreamlines . . . . . . . . . . . . . . . . 50
4.3.1 Standardapproachforevenlyspacedstreamlines . . . . . . . . . . 51
4.3.2 Adaptivecontroloftheseparatingdistance . . . . . . . . . . . . . . 53
4.3.3 ROIbasedtrackingusingevenlyspacedstreamlines . . . . . . . . 55
4.4 DirectionalVolumeGrowing . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.4.1 Algorithmfordirectionalvolumegrowing . . . . . . . . . . . . . . 56
4.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5 Connectivityanalysisbetweenfunctionalregions 63
5.1 Previouswork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.2 Connectivityanalysisbasedonpathfinding . . . . . . . . . . . . . . . . . . 64
∗5.2.1 A algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64Contents v
5.2.2 Domainrestriction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.2.3 Gridforpartitioningthesearchspace . . . . . . . . . . . . . . . . . 67
5.2.4 Costfunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.2.5 Minimumcostpath. . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.2.6 Multiplestartandgoalvoxels . . . . . . . . . . . . . . . . . . . . . . 70
5.2.7 Heuristic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.2.8 Efficientimplementation . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.3.1 Qualityanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.3.2 Computingtime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6 Generationofhulls 77
6.1 Tetrahedralizationand3DAlphaShapes. . . . . . . . . . . . . . . . . . . . 78
6.1.1 3DDelaunayTetrahedralization . . . . . . . . . . . . . . . . . . . . 78
6.1.2 3DAlphaShapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.1.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.2 Isosurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.2.1 Rasterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.2.2 Volumefiltering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.2.3 SurfaceExtraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.2.4 Surfacefiltering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.2.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.3 Constructionofasafetymargin . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7 Distortioncorrection 93
7.1 DistortionartifactsofDTI-EPIdata . . . . . . . . . . . . . . . . . . . . . . . 93
7.2 Previouswork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.3 Non-linearregistration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.3.1 Similaritymeasure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.3.2 Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.3.3 3Dfree-formdeformation . . . . . . . . . . . . . . . . . . . . . . . . 96
7.3.4 Registrationframework . . . . . . . . . . . . . . . . . . .