Abstract
BACKGROUND: The reliable and efficient measurement of spinal cord atrophy is of growing interest in monitoring disease progression in multiple sclerosis (MS). METHODS: We compared T1- and T2-weighted MRI for measuring cervical spinal cord volume in 31 patients with MS and 18 age-matched controls (NC) from T1-weighted gradient recalled echo and T2-weighted fast spin-echo 1.5 T axial acquisitions. The two sequences were matched on slice thickness, signal averages and voxel size. An active surface software tool determined the normalized mean cervical cord cross-sectional area. RESULTS: T1-derived cord areas were higher than T2 areas in the whole cohort (estimated mean difference = 7.03 mm(2) (8.89%); 95% Confidence Interval (CI): 5.91, 8.14; p < 0.0001) and in both groups separately. There were trends for lower spinal cord areas in MS vs. NC with both sequences. For the T1 cord area, the mean difference was 3.7 mm(2) (4.55%) (95% CI: -1.36, 8.78; p = 0.15). For the T2 cord area, the difference was larger [mean difference 4.9 mm(2) (6.52%) (95% CI: -0.83, 10.67); p = 0.091]. The T1 and T2 cord areas showed similar weak to moderate correlations with measures of clinical status and T2 spinal cord lesion volume in the MS group. Superficial spinal cord T2 lesions had no apparent confounding effect on the outlining tool. The mean intra-rater and inter-rater coefficients of variation ranged from 0.27 to 0.91% for T1- and 0.66 to 0.99% for T2-derived cord areas. CONCLUSION: T2-weighted images may prove efficient for measuring cervical spinal cord atrophy in MS, with the added advantage of lesion detectability.