ANALYZING GRAY MATTER INVOLVEMENT IN SYSTEMIC LUPUS ERYTHEMATOSUS BY VOXEL-BASED MORPHOMETRY

Background: Although clinical assessment is still the cornerstone in the diagnosis of neuropsychiatric SLE, the diagnosis is often difficult and remains presumptive in some patients. Magnetic resonance (MR) imaging is known to be more sensitive than computed tomography (CT) for the detection of anatomic brain abnormalities in patients with neuropsychiatric SLE because of the higher quality of the anatomic images due to excellent soft-tissue contrast and the ability to acquire multiplanar images at MR imaging.Measurements of brain volume are sensitive to both neuronal and axonal loss. Total and regional brain atrophy can be accurately assessed from conventional T1-wheighted images by means of computational methods allowing automatic or semiautomatic measurements of cerebral volumes. Voxel-based morphometry (VBM) is a automated method for characterizing regional cerebral volume and tissue concentration differences in structural MRI.Objectives: To investigate gray matter abnormalities in systemic lupus erythematosus (SLE) patients using voxel-based morphometry (VBM).Methods: A total of 75 SLE patients and 44 healthy age and sex-matched controls were enrolled in this study. MRI was performed in a 2T scanner (Elscint Prestige) and a T1-weighted gradient-echo sequence with 1 mm thickness (TR=22ms, TE=9ms, flip angle=35o, matrix=256x22) was used for VBM analysis. VBM was performed using SPM 99 (www fil.ion.ucl.ac.uk). Images were normalized to the standard space using 12 linear parameters and 7x8x7 nonlinear basis functions, using a brain mask. Spatially normalized images were re-sliced to isotropic voxels of 1.5mm and underwent segmentation of white matter and modulation of the estimated concentration of tissue based on the spatial deformations encountered during normalization. Finally the images were convolved with an Isotropic Gaussian Kernel of 10 mm in order to minimize inter-individual gyral variability. The resulting images were then compared voxel-by-voxel using t-test, to determine differences in white matter between controls and patients. Contrasts were defined in order to estimate the probability of each voxel being gray matter. This analysis included grand mean scaling and proportional threshold masking (set to 0.4) and implicit masking.Results: All SLE patients had a significant reduction in gray matter voxel number in the region corresponding to frontal, dorsolateral and medial temporal lobe (p<0.001) when compared to controls. Patients with longer disease duration and white matter abnormalities had significant less voxel number than patients with shorter disease. No relation to disease activity, total corticosteroid dose and the presence of antiphospholipid antibodies was noted.Conclusion: These data confirm substantial gray matter loss in SLE patients, especially in the frontal, dorsolateral and medial temporal lobe. Disease duration and white matter abnormalities were associated with reduction of voxel number in this region. Our conclusion relies on accurate segmentation of white matter in MRI scans without operator bias. Automated methods of MRI analysis may be useful in investigating brain abnormalities in SLE.Citation: , volume , supplement , year 2004, page Session: SLE – Clinical aspects