One hundred fifty-one MS patients and 69 age- and sex-matched healthy controls (HC) underwent brain and cSC 3T MRI acquisition and a comprehensive clinical assessment including evaluation of Expanded Disability Status Scale (EDSS), 9-hole peg test (9-HPT), finger tapping test (FTT), timed 25-foot walk test (T25FWT) and vibration detection threshold (VDT). Using random forest analyses, we investigated whether brain and cSC MRI measures (T2-hyperintense lesion volume [T2-LV] and total, gray matter [GM] and white matter [WM] cross-sectional areas [CSA] at C2-C3 level) were informative predictors of EDSS disability milestones (EDSS≥3.0, 4.0 and 6.0), impairment at pyramidal and sensory functional systems (P-FS and S-FS ≥2), as well as at 9-HPT, FTT and T25FWT (z-scores≤5th percentile of HC), and higher VDT.
Different combinations of brain and cSC MRI measures were informative predictors of EDSS milestones (out-of-bag [OOB]-area under the curve [AUC]=0.879-0.900), VDT (OOB-R2=0.194), and impairment at P-FS (OOB-AUC=0.820), S-FS (OOB-AUC=0.795), 9-HPT (OOB-AUC=0.793), FTT (OOB-AUC=0.740), T25FWT (OOB-AUC=0.825). For all these clinical outcomes, cSC GM CSA was identified as the most relevant predictor, except for 9-HPT, for which cSC T2-LV was the best predictor.
A multiparametric approach including brain and cSC MRI measures may explain a more severe clinical disability in MS with high accuracy. cSC GM atrophy represents an impactful pathological substrate specifically contributing to motor and sensory impairment in MS.