Spatial correlation and segregation of multimodal MRI abnormalities in multiple system atrophy.

Abstract

The variability of the severity and regional distribution of pathological process in basal ganglia (BG) and brainstem-cerebellar systems results in clinical heterogeneity and represents the motor subtype of multiple system atrophy (MSA). This study aimed to quantify spatial patterns of multimodal MRI abnormalities in BG and stem-CB regions and define structural MRI findings that correlate with clinical characteristics. We simultaneously measured R2*, mean diffusivity (MD), and volume in the subcortical structures (BG, thalamus, brainstem-cerebellar regions) of 39 probable MSA and 22 control subjects. Principal component analysis (PCA) and structural equation modeling (SEM) were performed to show a model consisting of multiple inter-dependencies. Structural MRI alterations were found to be significantly interrelated within BG as well as brainstem-cerebellar regions in MSA patients. PCA extracted four factors: three factors reflected alterations in R2*, MD and volume of the BG region including the caudate nucleus, putamen, and pallidum, and the remaining one factor represented degenerative changes in MD and volume of stem-CB region. In SEM, a latent variable reflecting brainstem-cerebellar degeneration did not show a significant correlation with the other latent variables associated with BG degeneration. Putaminal MD values and a PCA-driven factor reflecting MD values in the BG showed a significant correlation with UPDRS and UMSARS scores. Multimodal structural MRI abnormalities in MSA appear to be segregated into BG and stem-CB-related factors that can be associated with the clinical phenotype and motor severity.