Cortical Lesions in MS: Assessment at 7T

Introduction Multiple sclerosis is generally thought of as a white matter (WM) disease [1], yet there is limited correlation between white matter lesion (WML) load and neurological deficits found clinically [2]. Cortical damage was identified early [3,4] and has recently been correlated with the neurological episodes of MS patients [5]. Pathologically, cortical lesions (CLs) are classified into four types: leukocortical, (I), intracortical (II), subpial (III), and lesions that encompass all six layers of the cortex (IV), Histologically, subpial (III) lesions are the most numerous totaling 60-75% of all lesions, followed by types II, I, IV at 16-17%, 4-15%, and 4-8%, respectively. CL imaging at standard field strengths has had limited success [8], but recent studies at 7T using 2D T2* weighted imaging look promising [9,10,11]. The objective of this study was to evaluate the potential of 7T high resolution 3D imaging using a white matter attenuated turbo field echo (WHAT-TFE) and 3D T2*/susceptibility weighted magnitude and phase imaging (SWI) for detecting CLs. Materials and Methods With IRB approval, seven patients (4 RRMS and 3 progresive MS) and healthy controls were scanned with a 7.0 Tesla MRI scanner using a 16 channel receive coil. The WHAT-TFE sequence is an inversion recovery magnetization preparation sequence with TI selected to null the WM signal (TI = 550ms, shot interval TS = 3700ms, TR/TE/flip angle = 4.1/1.6/8, voxel size 0.4x0.4x0.7 mm). SW images were acquired using a non-spoiled gradient echo sequence with TR/TE/flip angle = 23/12/2-8, voxel size 0.23x0.23x0.7 mm). Phase images were calculated from complex raw data using high-pass filtering. For comparison, 5 patients had pre-, and 2 patients had preand post-contrast T1weighted IR-TFE scans (TI/TS = 4000/2000ms, 0.23x0.23x1.4 mm) for assessment of lesion enhancement. CLs were marked by 3 readers; two were previously trained for cortical lesion detection [12]. Results WHAT showed excellent GM/WM and WM/WML contrast. CLs were hyperintense compared to normal-appearing cortical gray matter (NACGM). Lesions were also brighter than NACGM on SWI magnitude images, but with diminished contrast. CLs were dark in T1-w IR-TFE (Fig. 1). In the post-contrast IR-TFE images, small enhancing veins were visualized adjacent to non-enhancing lesions. Localized in vivo cortical contrast changes that were identified as lesions were confirmed by specimen MR and histological comparisons (Fig. 2). Results of lesion counting are listed in Table 1. Regions with contrast inconsistency due to RF inhomogeneity were excluded from the cortical lesion assessment. Prior specimen training did not appear to affect significantly the ability to identify cortical lesions, and the absolute numbers were similar for the two readers who counted all cases. Reader DP counted only 5/7 cases, although the percentages of the lesions found are similar to the other two readers. Fewer lesions were seen in SWI overall due to lowered image