Atrophied Brain Lesion Volume: A New Imaging Biomarker in Multiple Sclerosis

Abstract

BACKGROUND AND PURPOSE: Lesion accrual in multiple sclerosis (MS) is an important and clinically relevant measure, used extensively as an imaging trial endpoint. However, lesions may also shrink or disappear entirely due to atrophy. Although generally ignored or treated as a nuisance, this phenomenon may actually be an important stand-alone imaging biomarker. Therefore, we investigated the rate of brain lesion loss due to atrophy (atrophied lesion volume) in MS subtypes compared to baseline lesion volume and to new and enlarging lesion volumes, and evaluated the independent predictive value of this phenomenon for clinical disability. METHODS: A total of 192 patients (18 clinically isolated syndrome, 126 relapsing-remitting MS, and 48 progressive) received 3T magnetic resonance imaging at baseline and 5 years. Lesions were quantified at baseline, and new/enlarging lesion volumes were calculated over the study interval. Atrophied lesion volume was calculated by combining baseline lesion masks with follow-up SIENAX-derived cerebrospinal fluid partial volume maps. Measures were compared between disease subgroups, and correlations with disability change (Expanded Disability Status Scale [EDSS]) were evaluated. Hierarchical regression was employed to determine the unique additive value of atrophied lesion volume. RESULTS: Atrophied lesion volume was different between MS subtypes (P =.02), and exceeded new lesion volume accumulation in progressive MS (298.1 vs. 75.5 mm3). Atrophied lesion volume was the only significant correlate of EDSS change (r =.192 relapsing, r =.317 progressive, P