Poster Abstract #15: MRI Measurement of Whole Spinal Cord Volume at 3T: Reproducibility and Relevance to Multiple Sclerosis

Abstract

Objective Determine the reproducibility and clinical relevance of measuring whole spinal cord volume in healthy controls and patients with multiple sclerosis (MS) using 3T MRI. Background Spinal cord atrophy may be especially relevant to disability in MS. The advent of 3T MRI brings the potential to assess spinal cord volume more precisely than conventional platforms. Design/Methods Eight patients with MS and five healthy subjects underwent 3T MRI of the whole spinal cord using axial fast spin-echo T2-weighted images. The upper cervical cord volume was calculated at C2–3. The cord volume was acquired using an edge-finding tool based on local thresholding and was normalized by the craniocaudal extent of the cord. Results The whole spinal cord volume in the MS group was 8.2% lower vs. healthy subjects ( p = 0.13; effect size, Cohen's d = 0.89). Whole spinal cord volume normalized by number of slices was 5% lower in patients with MS ( p = 0.16; d = 0.79) and whole spinal cord volume normalized by intracranial volume (ICV) was lower by 4.7% in MS vs. controls ( p = 0.16; d = 0.63). Normalizing the whole cord volume with number of slices and ICV showed a 1.5% difference in MS vs. controls ( p = 0.79; d = 0.50). Upper cervical spinal cord volume was 7% lower in MS vs. controls ( p = 0.33; d = 0.51). The whole cord volume scan–rescan, intrarater, and inter-rater coefficient of variation ranged from 0.5% to 0.7%; those for upper cervical cord cross-sectional volume ranged from 1.3% to 3.8%. Conclusions In this pilot study, whole spinal cord volume shows larger effect sizes than upper cervical cord volume in differentiating mildly disabled patients with MS from normal controls. Normalization of spinal cord volume by intracranial volume reduced such effect sizes and therefore may not be desirable. Whole spinal cord volume showed higher reproducibility than measurement of upper cervical cord volume. Study supported by: National Multiple Sclerosis Society (RG3705A1; RG3798A2); National Institutes of Health (NINDS 1R01NS055083-01).