Does high-field MR imaging improve cortical lesion detection in multiple sclerosis?

Objective Cortical lesions in multiple sclerosis (MS) are notoriously difficult to visualize with standard MR imaging (MRI) techniques. However, the use of higher field-strengths with intrinsically higher signal-to-noise, which can partly be used to increase spatial resolution, may improve cortical lesion detection. Therefore, in this post mortem study, the sensitivity of high fieldstrength MRI (4.7 T) for cortical lesions was investigated, and compared to that of standard field-strength (1.5 T). Methods At 1.5 T, dual-echo T2-weighted spin-echo, as well as 3D-FLAIR images of seventeen formalin-fixed coronal MS and four control hemispheres were acquired. At 4.7 T, the same specimens were imaged with a mainly proton-density (PD)- weighted sequence. Proteolipid protein (PLP)-stained tissue sections (10 mu m) of the same brain slices were matched to the corresponding MR images, and cortical lesions were scored on all three MR sequences (blinded to histology) and in tissue sections (blinded to MRI). Sensitivity of the sequences for four cortical lesion types was calculated. Additionally, an unblinded, retrospective MR scoring was performed. Results Sensitivity for purely intracortical lesions (histological lesion types II, III, and IV; n = 128) was below 10 % for both 1.5 T and 4.7 T MRI, while mixed gray matter-white matter (type I) lesions (n = 5) were detected in four out of five cases. All lesion counts increased upon retrospective (unblinded) scoring. However, up to 80% of the intracortical lesions still remained undetected. Conclusions MRI sensitivity for post mortem detection of cortical lesions is low, even when a higher field-strength was used. It varies, however, for different subtypes of cortical lesions.