Date of Award
Summer 8-15-2020
Degree Name
Doctor of Philosophy (PhD)
Degree Type
Dissertation
Abstract
The diffusion of water molecules in biological tissues in vivo or ex vivo is not free, impeded by the presence of obstacles of the environment, e.g., macromolecules, fibers and membranes. Diffusion magnetic resonance imaging (MRI) leverages the effect of water diffusion pattern through diffusion-weighting. It has gained wide application in imaging brain white matter tracts owning to the highly anisotropic diffusion of water molecules in these structures. Diffusion MRI detects white matter tract abnormalities in the absence of T2W lesions. Thus, it has been widely employed to assess disease progression of multiple sclerosis (MS). Current diffusion MRI images water signals from both intra- and extra-axonal compartments representing a weighted average of the diffusion effects of the two compartments. Thus, current diffusion MRI could not differentiate complexity of diffusion MRI signals in the presence of vasogenic edema, cell infiltration, and nerve damages. Therefore, to accurately reflect development of MS, it is necessary to develop a noninvasive neuroimaging method to detect nerve injury without the confounding effects from the surrounding extra-fiber pathologies.
In this thesis work, we propose a new diffusion MRI model improving diffusion basis spectrum imaging (DBSI) by inclusion of an intra-axonal compartment (DBSI-IA) to eliminate the overwhelming impacts of extra-axonal compartment in the presence of inflammation and tissue loss. DBSI-IA maintains the advantage of DBSI in resolving crossing fibers using low-b-value diffusion weighting while quantifying diffusivities of intra-/extra-axonal compartment water. Thus, DBSI-IA may be used to quantify axonal injury (via intra-axonal axial diffusivity), axon loss (via intra-axonal volume), demyelination (via extra-axonal radial diffusivity), edema and inflammation (via isotropic diffusion spectrum) with high precision. Through the multiple-tensor modelling of diffusion MRI signals, DBSI-IA has shown the potential to detect axonal injury in MS missed by conventional diffusion tensor imaging (DTI) and DBSI.
We first examined the validity of DBSI-IA using Monte-Carlo simulation (Chapter 4), observing the impact of extra-axonal water diffusion on DBSI derived axonal injury metrics including axial diffusivity and fiber fraction. DBSI-IA derived intra-axonal diffusivity and intra-axonal volume fraction is immune to the effect of extra-axonal water compartment. To further validate the suitability of DBSI-IA for analyzing diffusion MRI data, we applied this new model to analyze a data set from previously published in vivo DBSI of optic nerves from EAE mice (Chapter 5). We found that DBSI-IA derived pathological metrics closely correlated with immunohistochemistry identified optic nerve pathologies. DBSI-IA was then applied to previously published DBSI data from MS patients (Chapter 6). We successfully assessed normal appearing corpus callosum axonal injury in MS patients that have been missed by both DTI and DBSI.
In conclusion, DBSI-IA derived axonal integrity metrics, such as the intra-axonal fiber fraction and intra-axonal axial diffusivities (AD), can accurately reflect axonal injury missed by DTI or DBSI. DBSI-IA retains the sensitivity of DBSI to demyelination and cell infiltration in MS. The application of DBSI-IA provided a new perspective in developing a more effective diffusion MRI model by separating the extra- and intra-axonal water compartments.
Language
English (en)
Chair
Sheng-Kwei (Victor) Song
Committee Members
Joseph J. Ackerman, William Spees, Hong Chen, Daniel W. Moran,