MRI detection of hypointense brain lesions in patients with multiple sclerosis: T1 spin-echo vs. gradient-echo

Abstract

Objective Compare T1 spin-echo (T1SE) and T1 gradient-echo (T1GE) sequences in detecting hypointense brain lesions in multiple sclerosis (MS). Background Chronic hypointense lesions on T1SE MRI scans are a surrogate of severe demyelination and axonal loss in MS. The role of T1GE images in the detection of such lesions has not been clarified. Design/methods In 45 patients with MS [Expanded Disability Status Scale (EDSS) score (mean ± SD) 3.5 ± 2.0; 37 relapsing-remitting (RR); 8 secondary progressive (SP)], cerebral T1SE, T1GE, and T2-weighted fluid-attenuated inversion-recovery (FLAIR) images were acquired on a 1.5 T MRI scanner. Images were re-sampled to axial 5 mm slices before directly comparing lesion detectability using Jim (v.7, Xinapse Systems). Statistical methods included Wilcoxon signed rank tests to compare sequences and Spearman correlations to test associations. Results Considering the entire cohort, T1GE detected a higher lesion volume (5.90 ± 6.21 vs. 4.17 ± 4.84 ml, p < 0.0001) and higher lesion number (27.82 ± 20.66 vs. 25.20 ± 20.43, p < 0.05) than T1SE. Lesion volume differences persisted when considering RR and SP patients separately (both p < 0.01). A higher lesion number by T1GE was seen only in the RR group (p < 0.05). When comparing correlations between lesion volume and overall neurologic disability (EDSS score), T1SE correlated with EDSS (Spearman r = 0.29, p < 0.05) while T1GE (r = 0.23, p = 0.13) and FLAIR (r = 0.24, p = 0.12) did not. Conclusion Our data suggest that hypointense lesions on T1SE and T1GE are not interchangeable in patients with MS. Based on these results, we hypothesize that T1GE shows more sensitivity to lesions at the expense of less pathologic specificity for tissue destruction than T1SE.