Medical Student Washington University School of Medicine St. Louis, Missouri, United States
Introduction: Diffusion basis spectrum imaging (DBSI) is a non-invasive quantitative imaging modality that may improve understanding of cervical spondylotic myelopathy (CSM) pathology through detailed evaluations of spinal cord microstructural compartments. We previously employed diffusion basis spectrum imaging (DBSI) derived axial diffusivity to assess spinal cord white matter axonal injury in CSM. However, DBSI-derived axial diffusivity may be confounded by extra-axonal signal weighting due to vasogenic edema. Therefore, the aim of this study was to enhance DBSI by including an intra-axonal water compartment (DBSI-IA).
Methods: In DBSI-IA, we model intra-axonal and extra-axonal space of white matter bundles with three compartments: 1) anisotropic diffusion of intra-axonal water; 2) anisotropic diffusion of extra-axonal water; and 3) isotropic diffusion of extra-axonal water. Monte-Carlo simulation and previously published in-vivo MRI data from experimental autoimmune encephalomyelitis (EAE) mice (eight 10-week old female C57BL/6 mice actively immunized with MOG35-55 peptide) were analyzed using the newly developed DBSI-IA model.
Results: In Monte-Carlo simulation, DBSI-IA parameters, including fiber fraction, intra-axonal axial diffusivity, and extra-axonal axial diffusivity, more accurately estimated ground truth values when compared to traditional DBSI. Both DBSI-IA and DBSI restricted fraction estimations were stable and similarly close to ground truth. On analysis of in-vivo data from EAE mouse optic nerves, decreased DBSI-IA heatmap intensity at the post-onset optic neuritis timeframe suggests axon loss and axonal injury. DBSI-IA distinguished and further quantified the extent of demyelination and inflammation in this mouse animal model.
Conclusion : Compared to the original DBSI model, the DBSI-IA model performed better in detecting and quantifying axonal injury and loss, while maintaining comparable performance on isotropic diffusion metrics. Our results suggest that the DBSI-IA model provides improved accuracy in the evaluation of white matter tracts compared to traditional DBSI, with promising applications to further assess CSM pathology.
