Jacqueline Chen1, Mark Lowe1, Ken Sakaie1, Kenneth Baker2, Andre Machado3, and Stephen Jones1
1Cleveland Clinic Imaging Institute, Cleveland, OH, United States, 2Cleveland Clinic Lerner Research Institute, Cleveland, OH, United States, 3Cleveland Clinic Neurological Institute, Cleveland, OH, United States
Synopsis
Accurate white-matter tractography maps can be a
useful clinical tool for assessing neurological disorders, however, incorrect
assumptions within tractography algorithms can yield non-physiological results.
We have developed a tractography methodology based on probability theory that
uses both local and global information to improve accuracy, and standard
partial differential equation solvers for fast whole-brain mapping. In this
abstract we demonstrate: 1) the accuracy of the method by comparing a topographical map of the corpus callosum (CC) generated from a symmetrized human data phantom
to published maps; 2) how differences in CC topography may be associated with stroke
location and functional disability.
Introduction
Fast, whole-brain maps of
white-matter (WM) tracts can serve as a useful clinical tool for assessing
structurally intact or damaged pathways present in patients suffering from
neurological diseases, including: stroke, multiple sclerosis and epilepsy. The
use of diffusion MRI to estimate brain WM tracts has been recently reviewed,1 emphasizing the need to limit biased results and strive
for biological accuracy. To address these concerns, we present a tractography
methodology based on probability theory that uses both local and global
information to improve accuracy, and standard partial differential equation
solvers for fast whole-brain mapping. First, we demonstrate the accuracy of the
method, by comparing a topographical map of the corpus callosum (CC) generated
from a symmetrized human data phantom to published maps from pathological and diffusion
tractography investigations. Second, we demonstrate how differences in CC topography
may be associated with the location and functional disability of patients who
suffered from stroke.Methods
The method is based on our previous work.2 Our
probabilistic approach includes both local voxel information from the high
angular diffusion imaging data, and the global conditions that require tracks to
start at a seed and end at a target without intersecting a non-WM boundary.
Demonstrating accuracy
A symmetrical phantom was computed from human data. Using
our tractography method with the Freesurfer3 segmentation, the
flux from right hemisphere cortex (RHC) to left hemisphere cortex (LHC) was calculated
(Fig. 1), and seed voxels from 35 RHC regions were used for tracking to the LHC.
A unified topographical CC map was defined based on transcallosal track clusters
and compared to published maps.
Demonstrating
differences in CC topography in relation to stroke location and functional
disability
We analyzed 7 tesla anatomical and diffusion MRI data
from 2 patients (Table). Patient A was imaged 30
months after a stroke in right frontal and parietal WM and exhibited
dysfunction of their affected upper limb. Patient B was imaged 22
months after a stroke in the left putamen and exhibited better function in both
affected and unaffected upper limbs compared to Patient A. Using our
tractography method with the Freesurfer segmentation, the flux from RHC to LHC
was calculated, and seed voxels from right precentral cortex were used for tracking to
the LHC. Maps of the intersection of the transcallosal tracks with the CC were
generated and compared.
Results
Demonstrating accuracy- Anterior CC (Fig. 2, A) contained
tracks from the most
anterior and inferior regions of prefrontal cortex, consistent with findings
from diffusion tractography,4 pathology,5 and T1 relaxometry.6
-
Mid-Anterior CC (Fig. 2, B) contained tracks mostly from
superior and middle frontal cortical regions, consistent with findings from diffusion
tractography.4, 7
- Central and Mid-Posterior CC (Fig. 2, C) contained tracks from cortical
regions proximal to the central sulcus (precentral, paracentral, postcentral)
and lateral sulcus (insula, transverse temporal, superior temporal,
supramarginal), consistent with findings from pathology.5
- Posterior CC (Fig. 2, D) contained tracks from posterior parietal,
inferior temporal and occipital lobes, consistent with findings from diffusion
tractography,4, 7 and pathology.5
Demonstrating
differences in CC topography in relation to stroke location and functional
disabilityWhen considering the intersection of the transcallosal
fibers connecting right and left precentral cortices with the CC, we found that
Patient A’s fibers extended a shorter
distance along the CC (Fig. 3, B) compared to
Patient B (Fig. 3, C) and
the symmetrized human data phantom (Fig. 3, A).
Discussion
We have developed a tractography methodology based
on probability theory that uses both local and global information to improve
accuracy, and standard partial differential equation solvers for fast
whole-brain mapping. The speed of the method lends itself to clinical use,
allowing for probing of multiple pathways to evaluate injury. To demonstrate
accuracy of the method, we derived a topographical map of the CC generated from a
symmetrized human data phantom, and compared it to published CC maps. Our
method showed consistency with published findings from diffusion tractography,
pathology and T1 relaxometry studies. To demonstrate the clinical relevance of
mapping the transcallosal pathways, we evaluated 2 patients with chronic stroke
injury. The patient with more pronounced upper-limb deficit exhibited a shorter
extent of the CC occupied by transcallosal fibers connecting right and left
precentral cortices compared to the patient with less upper-limb deficit and
compared to the symmetrized human data phantom.Conclusion
Our fast WM tractography method generates topographical maps of the CC that are consistent with published findings from diffusion
tractography, pathology, and T1 relaxometry studies. Analysis of clinical data
suggests that quantification of transcallosal tracks may be an important metric
of brain injury.Acknowledgements
No acknowledgement found.References
- Jeurissen B,
Descoteaux M, Mori S, Leemans A. Diffusion MRI fiber tractography of the brain.
NMR Biomed. 2019 Apr;32(4):e3785.
- Zhang M, Sakaie KE, Jones SE. Logical
foundations and fast implementation of probabilistic tractography. IEEE Trans
Med Imaging. 2013 Aug;32(8):1397-410.
- Laboratory for Computational Neuroimaging
at the Athinoula A. Martinos Center for Biomedical Imaging. FreeSurfer. [cited 2020 18 Jun]; Available from: http://surfer.nmr.mgh.harvard.edu/.
- Park HJ, Kim JJ, Lee SK, Seok JH, Chun
J, Kim DI, Lee JD. Corpus callosal connection mapping using cortical gray
matter parcellation and DT-MRI. Hum Brain Mapp. 2008 May;29(5):503-16.
- Aboitiz F, Montiel J. One hundred
million years of interhemispheric communication: the history of the corpus
callosum. Braz J Med Biol Res. 2003 Apr;36(4):409-20.
- Lee BY, Zhu XH, Li X, Chen W.
High-resolution imaging of distinct human corpus callosum microstructure and
topography of structural connectivity to cortices at high field. Brain Struct
Funct. 2019 Mar;224(2):949-60.
- Hofer S,
Frahm J. Topography of the human corpus callosum revisited--comprehensive fiber
tractography using diffusion tensor magnetic resonance imaging. Neuroimage.
2006 Sep;32(3):989-94.