High field (9.4 Tesla) magnetic resonance imaging of cortical grey matter lesions in multiple sclerosis

Abstract

Multiple sclerosis is an inflammatory, degenerative disease of the central nervous system. The most obvious pathological change in multiple sclerosis is multifocal demyelination of the white matter, but grey matter demyelination may be of equal or even greater importance for its clinical manifestations. In order to assess the pathogenetic role of lesions in the grey and white matter, and to explore the association between demyelinated and non-lesional brain tissue, tools are needed to depict each of these tissue components accurately in vivo. Due to its sensitivity in detecting white matter lesions, T(2)-weighted magnetic resonance imaging at 1.5 T is important in the diagnosis of multiple sclerosis. However, magnetic resonance imaging at 1.5 T largely fails to detect grey matter lesions. In this study, we used T(2)-weighted magnetic resonance imaging at 9.4 T to detect grey matter lesions in fixed post-mortem multiple sclerosis motor cortex. Furthermore, we produced T(1), T(2) and magnetization transfer ratio maps, and correlated these indices with quantitative histology [neuronal density, intensity of immunostaining for myelin basic protein (reflecting myelin content) and phosphorylated neurofilament (reflecting axonal area)] using t-tests and multivariate regression. In 21 tissue samples, 28 cortical grey matter lesions were visible on both T(2)-weighted magnetic resonance imaging and sections immunostained for myelin basic protein, 15/28 being mixed white and grey matter and 11/28 subpial cortical grey matter lesions; 2/28 cortical grey matter lesions involved all layers of the cortex. Compared with non-lesional cortex, cortical grey matter lesions showed reduction of neuronal density (98/mm(2), SD = 34/mm(2;) versus 129/mm(2), SD = 44; P < 0.01), phosphorylated neurofilament (1/transmittance = 1.16; SD = 0.09 versus 1.24; SD = 0.1; P < 0.01) and magnetization transfer ratio (31.1 pu; SD = 11.9 versus 37.5 pu; SD = 8.7; P = 0.01), and an increase of T(2) (25.9; SD = 5 versus 22.6 ms; SD = 4.7; P < 0.01). Associations were detected between phosphorylated neurofilament and myelin basic protein (r = 0.58, P < 0.01), myelin basic protein and T(2) (r = -0.59, P < 0.01), and neuronal density and T(1) (r = -0.57, P < 0.01). All indices correlated with duration of tissue fixation, however, including the latter in the analysis did not fundamentally affect the associations described. Our data show that T(2)-weighted magnetic resonance imaging at 9.4 T enables detection of cortical grey matter lesion in post-mortem multiple sclerosis brain. The quantitative associations suggest that in cortical grey matter T(1) may be a predictor of neuronal density, and T(2) of myelin content (and-secondarily-axons). Successful translation of these results into in vivo studies using high field magnetic resonance imaging (e.g. 3 T and 7 T) will improve the assessment of cortical pathology and thereby have an impact on the diagnosis and natural history studies of patients with multiple sclerosis, as well as clinical trial designs for putative treatments to prevent cortical demyelination and neuronal loss.