To explore clinically feasible imaging metrics for predicting disability in secondary progressive multiple sclerosis (SPMS)

Spinal cord pathology is known to play a strong role in disability in SPMS, however methods used for its assessment are often time-consuming or require advanced imaging techniques.

137 individuals with SPMS underwent motor testing (timed 25-foot walk, T25FW; 9-hole peg test, 9HPT), cognitive testing (symbol digit modalities test, SDMT), and a survey of walking impairment (multiple sclerosis waking scale, MSWS-12), as part of their routine care from 2018-2023. 108/137 had clinical cervical spine MRIs within 1 year of assessment that included sagittal STIR, axial T2, and axial gradient echo at 1.5 or 3T. An experienced neurologist reviewed each spine MRI and determined the total lesion count and which cord segments (i.e. C1-2, C2-3) had at least one lesion. To determine interrater reliability, two experienced neurologists each reviewed a subset of 20 scans. The correlation coefficient for percentage of cord segments lesioned was excellent (r=0.92). Relationships between clinical and MRI metrics were determined using nonparametric correlations.

Median spinal cord lesion count was 4 (range 0-11). Lesion count was not correlated with any disability measure. Median percentage of segments with lesions was 71% (range 0%-100%). Percentage of segments with lesions was correlated with T25FW (r=0.25, p=0.010) and MSWS12 (r=0.29, p=<0.003), but not 9HPT (p=0.57). When segments were analyzed individually, only lesion presence at C1-2 was associated with worse function across all motor outcomes: T25FW (r=0.22, p=0.020), NHPT (r=0.26, p=0.008), and MSWS12 (r=0.24, p=0.012). Spinal cord lesion burden was not correlated with SDMT.

The presence of lesions at specific spinal cord segments (i.e. C1-2) and the percentage of cervical spinal cord segments containing lesions may carry more prognostic weight for motor outcomes than total lesion count and are more feasible to incorporate into routine clinical practice.