Title: Visualization of Fibers at Risk for Neuronal Tract Injury in Early MS by Streamtube Diffusion Tracto
1Visualization of Fibers at Risk for Neuronal
Tract Injury in Early MS by Streamtube Diffusion
Tractography at 3 TeslaJack H Simon 1David E
Miller 1Mark Brown 1John R Corboy 1Jeffrey L
Bennett 1Song Zhang 2 David H Laidlaw 2
1University of Colorado Health Sciences Center
2Brown University
Supported by the National MS Society (RG
3307-A-1), GE Medical Systems, and the National
Science Foundation (CCR-0086065).
2Background Neuronal Tract Degeneration in Early
MS
- Axonal injury in inflammatory MS Lesions
- Trapp et al 1998 Ferguson et al 1997
- Empty myelin cylinders 2 to and distant from
focal lesion - Bjartmar et al 2001
- Major axonal loss (and atrophy) in corpus
callosum - Evangelou et al 2000
- Neuronal Tract Degeneration Patterns in Early MS
Bjartmar et al 2001
Transcallosal Bands
Corticospinal Tract Degeneration
Simon et al Neurology 2001
Simon et al Neurology 2000
3Goals/Question
- Develop a process to identify and then
interrogate neuronal fibers that are anatomically
related to focal demyelinating lesions - Can we detect and measure neuronal fiber
injury/degeneration in vivo ?
4Strategy
- Earliest MS
- Identify focal lesions at the time of a CIS
- Identify fibers transiting the focal lesion and
reaching a distant structure (corpus callosum) - Interrogate fibers (in corpus callosum) by MRI
distant from focal lesion - MTR, Diffusion tensor (ltDgt FA)
- Longitudinal MRI follow-up
53T Pulse Parameters
- Diffusion Tensor Imaging
- Echo-planar
- B-value 1000, 25 gradient directions
- Axial plane - 3 sets of overlapping slices
- 5.1 mm thick with 1.7mm shifts
- FOV 28 x 28 cm, matrix 160 x160, 2 excitations
- TR/TE 6075/69.8-minimum
- Magnetization Transfer Imaging
- 3D gradient-echo
- MT pulse
- Fermipulse applied middle 30 of the phase
encode steps (SAR) - Pulse width of 8 ms
- Peak B1 field of 9.24 µT, flip angle of 670
- 1200 Hz from the center frequency
- TR/TE/flip angle 50/2.4(fr)/15
- FOV 22x22, matrix 256 x192, one excitation
- Sagittal plane
- 3 mm partitions
6Patient Characteristics
7Outline of Streamtube Tractography Strategy
Calculate Diffusion tensor - non-linear
approach
- Generate streamtubes and streamsurfaces
- Westin et al 2002 Zhang et al 2003
- coherent white matter tracts fastest diffusion,
principal eigenvector
- Terminate streamtube
- low linear diffusion data boundary curvature
- 500,000 streamtubes
- Cull 1 length anisotropy redundancy
- 2,000 streamtubes
- Cull 2 - seed points within the 3D coordinates
of T2-lesions - Cull 3 - intersection with corpus callosum.
Register data sets- affine transformations
optimized by mutual information algorithm
Visualize in optimal projection quantitative
analyses
8Visualization -Streamtubes Streamsurfaces
Segmentation Registration
- AAMI (Analysis Application for Medical Imaging)
- David E. Miller - University of Colorado HSC
- Visualization Tools
- Song Zhang and David Laidlaw - Brown University
9Streamtubes culled to lesion Streamsurfaces not
utilized
10Streamtubes culled to lesion corpus callosum
11Case 1
Streamtubes and Streamsurfaces
Culled to Lesion
Culled to lesion corpus callosum
12Case 2
Culled to Lesion
Culled to Lesion Corpus Callosum
13Case 2
14Case 3
Culled to lesion corpus callosum
15Fibers at Risk (FAR)
Case 1 Registration of culled streamtubes to
sagittal representation of corpus callosum
16CIS 1
CIS 2
CIS 3
17Conclusions
- Viable strategy and methodology to identify a new
class of tissue called fibers at risk (FAR) - NAWM
- AAWM
- FAR
- FAR makes up a formidable percentage of corpus
callosum with minimal overlap with AAWM - Strategy applicable to any neuronal tract
- Future studies to determine injury profile over
time, factors associated with more severe
neuronal tract injury - Assay of MS neurodegeneration to evaluate
treatment ?
18Supported by The National Multiple Sclerosis
Society (RG 3307-A-1)GE Medical Systems (3T
diffusion tensor MRI)National Science Foundation
(CCR-0086065)University of Colorado 3T Research
Instrument (ONDCP)University of Colorado Brain
Imaging Research Laboratory (BIRL) Rebecca Leek
and MaryJoel Meyer BIRL and Brain Imaging Center
for Drug Abuse Research Deb Singel