Advanced diffusion-weighted imaging models better characterize white matter neurodegeneration and clinical outcomes in multiple sclerosis.

Abstract

White matter (WM) atrophy is relevant in multiple sclerosis (MS), but the methods of analysis currently used are not specific for microstructural changes. The aims of this study were to assess the use of advanced diffusion-weighted imaging (DWI) techniques proposed as measures of baseline and longitudinal WM atrophy in MS and to analyze whether these measures helped explain MS clinical disability (including cognitive impairment) better than volumetric and diffusion tensor (DT)-derived measures. 3DT1-weighted and DWI sequences were applied to 86 MS and 55 healthy controls (HC) at baseline and after one-year. Intra-cellular volume (vic) maps were computed from neurite orientation dispersion and density imaging model. Voxel-wise fiber-bundle cross-section (FCS) atrophy in MS compared to HC was estimated. Maps of fractional anisotropy and mean diffusivity were also obtained from DWI for a comparison with the proposed advanced DW-derived measures (vic and FCS). Both at baseline and after 1-year, only FCS measure showed a significant atrophy in relapsing-remitting (RR) MS compared to HC and in progressive MS compared to RRMS, mainly located in specific WM tracts (corticospinal tract, splenium of the corpus callosum, left optic radiation, bilateral cingulum, middle cerebellar peduncle and anterior commissure, p value < 0.05). Global baseline FCS and vic were the selected predictors of clinical (R-sq = 0.33, p = 0.007) and cognitive scores (R-sq = 0.29, p = 0.0014) in a linear regression model. Voxel-based FCS was able to detect WM tracts atrophy in MS clinical phenotypes with greater anatomical specificity compared to other measures (volumetric and DT-derived measures of WM damage). FCS and vic measured at baseline in the WM were the best predictors of clinical disability and cognitive impairment.