A cross-sectional multicentre study of multishell diffusion MRI in multiple sclerosis

Abstract

Diffusion weighted imaging; MRI; Multiple sclerosis

Imatges ponderades per difusió; Ressonància magnètica; Esclerosi múltiple

Imágenes ponderadas por difusión; Resonancia magnética; Esclerosis múltiple

Background and objectives White matter (WM) microstructural properties from advanced multishell diffusion MRI (dMRI) have been linked to clinical disability in multiple sclerosis (MS). This multicentre study used multishell dMRI to compute WM metrics and test for differences between people with MS (pwMS) and healthy controls (HCs). Methods We included multishell dMRI data from 251 pwMS or clinically isolated syndrome (CIS) (mean age 40.7 years, 72.4 % women, 88.8 % relapsing remitting MS) at six MAGNIMS centres and 543 HCs. Eleven scalar metric maps were estimated from multishell dMRI sequences, based on diffusion tensor imaging (DTI) and restriction spectrum imaging (RSI). The maps were analysed using tract-based spatial statistics (TBSS). The diffusion output was submitted to paired sampled t-tests to test for case-control differences and linear regression models to test for associations with Expanded Disability Status Scale (EDSS) scores, while accounting for confounders. In a sub-sample from Oslo, we tested for correlations between EDSS and dMRI metrics within WM lesions. Results Significant group differences were found in nine out of eleven dMRI metrics. Linear regression models revealed significant correlations between EDSS and fractional anisotropy (FA) fast (β=-4.54, p = 0.01) and apparent diffusion coefficient (ADC) fast (β=10.92, p = 8.7 × 10−3). Conclusions Diffusion MRI based on clinically feasible multishell sequences uncovers WM group differences between pwMS and HCs, but only a selection of the advanced multishell parameters were sensitive to disability, and no statistically significant correlations with disability remained after Bonferroni correction.

This work was supported by the South Eastern Regional Health Authority of Norway (project number 2017114; HFH), by the German Research Council (Deutsche Forschungsgemeinschaft – DFG; RC-TR-128;SG).