To assess the utility of axial T2-weighted spinal cord images at multiple sclerosis (MS) diagnosis and in other inflammatory myelopathies.

MRI international guidelines from 2021 recommended including >=2 sagittal MRI spinal cord T2-weighted (T2) sequences for MS diagnosis but axial T2-images were considered optional, given a paucity of data on their additive benefit.  

This population-based study included incident people with MS (pwMS) from Olmsted County (01/01/2003-12/31/2018) with brain and cervico/thoracic spine MRI (both sagittal and axial-T2 images) at diagnosis. We compared to 90 other myelopathies (NMOSD, 20; MOGAD, 20; sarcoid, 20; spondylotic, 15; infarct, 15). The presence and number of cord lesions were independently evaluated on sagittal/axial T2, and sagittal short tau inversion recovery (STIR) by two blinded raters and compared using chi-squared or Wilcoxon signed-rank tests. Inter-rater agreement was measured with Cohen’s Kappa. 

We included 136 pwMS. There was substantial agreement on presence of spinal cord lesions between the two raters (Cohen’s kappa range=0.70-0.77; agreement, 88-93%). Among pwMS, ≥1 spinal cord lesion was noted on: T2 (axial, 103/136[75%]; sagittal, 76/136[56%]; STIR, 72/109[66%]). T2-lesion detection rates showed an advantage of axial-T2 versus sagittal-T2 (p=0.001) but not sagittal STIR (p=0.10) in MS but not in other myelopathy groups (p range 0.43-1.00). The mean[standard deviation] number of cord T2-lesions/patient was higher on axial-T2 (2.1[2.5]) than sagittal-T2 (1.1[1.6], p<0.001) and sagittal STIR (1.4[1.7], p<0.001). Overall, axial-T2 detected 280 lesions vs 147 on sagittal-T2 (+90%) and 226 vs 154 on STIR (+47%). In 31 pwMS, spinal cord MRI was needed to fulfill dissemination in space, and axial-T2 images led to fulfillment of MS criteria in 3/31 additional patients (10%) than sagittal images alone.

Axial-T2 images increase the sensitivity for spinal cord lesion detection beyond sagittal-T2 and STIR in MS, but not in the other myelopathies.