Thalamic atrophy and its substrates in multiple sclerosis by ultra-high field MRI

Introduction: Thalamic atrophy appears early in MS and serves as a biomarker for neurodegeneration. Some studies have linked MS thalamic atrophy to white matter lesion volume, but its relationship with local and cortical pathology was less investigated. Ultra-high field MRI demonstrates high sensitivity to cortical and thalamic lesions. Aim: We used 7T and 3T MRI to characterize the presence of lesional and neurodegenerative thalamic pathology in a heterogenous MS cohort. We further investigated the contribution of local and distant lesion pathology and cortical atrophy to thalamic atrophy. Methods: In 83 MS patients (57 RRMS, 26 SPMS) and 44 age-matched healthy controls, 7T T2*-weighted images were acquired for lesion segmentation, 3T T1-weighted scans for cortical thickness and thalamic volume estimation. Thalamic volume was normalized to total intracranial volume. Regression analysis was used to identify predictors of thalamic atrophy using age as a covariate. Results: 38/83 MS patients presented with thalamic lesions. Overall, normalized thalamic volume was smaller in patients relative to healthy controls (mean±SD = 3×10-3±5.7×10-4 mm3 vs. 3.4×10-3±5.1×10-4 mm3; p = 0.0002, by linear regression), including phenotypes (RRMS: 3.2×10-3±5.4×10-4 mm3; p = 0.017 vs. SPMS: 2.6×10-3±4.6×10-4 mm3; p < 0.0001, by linear regression). Thalamic volume, however, did not differ between groups of patients with or without thalamic lesions (p= 0.426). Regression analysis demonstrated that thalamic atrophy in MS patients was explained by white matter atrophy (p < 0.0001) and cortical thickness (p = 0.002) and unrelated to thalamic and cortical lesion load. Discussion: The data indicate that thalamic atrophy in MS is independent from local lesion pathology, and suggest that disconnection mechanisms, from white matter lesions and cortex, might play a greater role than local thalamic pathology in inducing thalamic atrophy in the disease.