Brain damage in primary-progressive MS assessed by diffusion tensor MRI

Introduction The pathological mechanisms underlying the progressive course of multiple sclerosis (MS) are still not well defined, even thought the presence and the severity of macroand micro-scopic tissue damage are thought to be important factors for the development of clinical disability. Diffusion tensor magnetic resonance imaging (DT-MRI) (1) is sensitive to pathological processes which result in loss of restricting barriers to water molecular motion and tissue anisotropy. Measures derived from DT-MRI include: a) the mean diffusivity (MD), a directionally averaged measure of diffusion, which is affected by cellular size and integrity; and b) the fractional anisotropy (FA), which reflects the degree of alignment of cellular structures within fiber tracts, as well as their structural integrity. MD values are increased and FA values are decreased in MS lesions and normalappearing white matter (NAWM) (2-3). In addition, MD histogramderived metrics of the brain well correlate with the patients' clinical status. In this study, we performed DT-MRI in patients with primary progressive (PP) MS to assess the value of this technique in defining the features of MS pathology in the brain of these patients. DT-MRI findings were also compared with those obtained from secondary progressive (SP) MS patients with similar levels of disability. Methods We studied 91 patients with PPMS, 41 with SPMS and 44 healthy subjects. The following scans were obtained from each subject, during a single session: a) dual-echo turbo spin echo (TSE) (TR/TE/NEX = 3300/16-98/1; number of slices: 24, contiguous, 5 mm thick); b) T1weighted conventional spin-echo (CSE) (TR/TE/NEX=768/15/2; ETL=5; number of slices: 24, contiguous, 5 mm thick); c) pulsedgradient spin-echo echo-planar (EPI) pulse sequence (inter-echo spacing = 0.8, TE = 123; number of slices. 10, contiguous, 5 mm thick), with diffusion-weighting applied in 8 non-collinear directions. (maximum b factor = 1044 s mm-2). For the EPI scans, the ten slices were acquired positioning the second-last caudal slice to match exactly the central slice of the dual-echo. This brain portion was chosen since these central slices are less-affected by B0 distortions. MS lesions were marked on hardcopies by an experienced observer, unaware to whom the scans belonged, and outlined on a computer display by a trained technician. After correction for eddy current-induced distortion, the diffusion tensor was estimated linearly for every voxel, assuming a mono-exponential relationship between the signal attenuation and the elements of the tensor matrix. Next, MD and FA were derived for each voxel. After interpolation to same matrix size as the dual-echo, the b=0 step of the EPI was co-registered with the T2-weighted scans of the dual echo. The same transformation parameters were then used to register MD and FA maps. FA and MD maps of the normal appearing brain tissue (NABT) were created by superimposing the lesion outlines and nulling out the corresponding regions on the co-registered images. NAWM and normal appearing gray matter (NAGM) were segmented from NABT images by an automated technique based on FA thresholding (4). Whole brain MD and FA, NABT, NAWM and NAGM MD histograms were created as previously described (5). Results Average (range) T2 lesion volume was 19.3 (0.7-53.3) ml for PPMS and 22.7 (6.1-87.7) ml for SPMS; average (range) T1 lesion volume was 8.8 (0.0-50.5) ml for PPMS and 12.8 (0.3-57.1) ml for SPMS. T1 lesion volume was significantly correlated with EDSS (r=-0.4, p=0.03) in PPMS patients. Average lesion MD and FA were respectively 1.11 (0.87-1.38)x10-3 mm2/s and 0.27 (0.21-0.35) for PPMS, and 1.19 (1.02-1.43) x10-3 mm2/s and 0.24 (0.19-0.28) for SPMS patients. These quantities were not significantly different between the two groups. In PPMS, average brain MD was significantly correlated with both T2 lesion volume (r=-0.7, p<0.001) (such correlation was still present when considering NAGM and NAWM separately), and T1 lesion volume (r=0.6, p<0.001). Table 1 reports the significant comparisons between controls and PPMS; Table 2 reports the significant comparisons between PPMS and SPMS. Discussion This study confirms a widespread damage of the brain tissue which appears normal on conventional MRI scans in patients with PPMS. The extent and severity of NAWM and NAGM damage in PPMS are likely to be among the major factors leading to irreversible disability in these patients. References 1. Basser P., et al; J. Magn Reson B, 103, 247-254 (1994). 2. Werring D.J., et al.; Neurology 52, 1626-1632 (1999). 3. Filippi M., et al.; Arch. Neurol 57, 1017-1021 (2000). 4. Cercignani M., et al; JNNP (2000) (in press). 5. Cercignani M. et al; Neurology 54, 1139-1144 (2000).