Periventricular lesions correlate with cortical thinning in multiple sclerosis.

Abstract

It has been suggested recently that cortical pathology in multiple sclerosis (MS) may, at least partly, be caused by factors in cerebrospinal fluid (CSF). We thus hypothesized that MS-related tissue changes in compartments close to the CSF, such as periventricular lesions, might correlate with cortical pathology. We investigated a cohort of 160 patients, comprising 91 with a clinically isolated syndrome (CIS) and 69 with relapsing-remitting MS (RRMS; mean age: CIS: 31.4 ± 9.0; RRMS: 33.0 ± 8.7 years; mean disease duration: CIS: 7.2 ± 15 months; RRMS: 8.0 ± 6.5 years, Expanded Disability Status Scale (median, min-max): CIS: 1, 0-3.5; RRMS: 1.25, 0-4) with 3.0T magnetic resonance imaging. MS lesions were segmented semiautomatically on fluid-attenuated inversion recovery images. To quantify periventricular lesion load (PV-LL), we generated ventricle masks and dilated them by a voxel factor of 3. Lesions within the dilated ventricle margin were classified as periventricular. Cortical thinning was assessed by cortical mean thickness (CMT) and compared to data from 58 healthy controls (HCs; mean age: 29.1 ± 7.4 years). Compared to HC, CIS and (even more so) RRMS patients demonstrated significantly reduced CMT. Even after controlling for ventricular volume and total lesion load, increased periventricular lesion occupancy (percentage of PV-LL) significantly correlated with decreased CMT in RRMS (r = -0.295; p = 0.015), but not in CIS (r = 0.032; p = 0.768) patients. The correlation between increased periventricular lesion burden and decreased CMT indicative of subpial cortical pathology supports the concept that common CSF-mediated factors might play a role in the accumulation of damage to gray and white matter in MS.