Patterns of fractional anisotropy changes in white matter of cerebellar peduncles distinguish spinocerebellar ataxia-1 from multiple system atrophy and other ataxia syndromes

Abstract

Aim To determine prospectively if qualitative and quantitative diffusion tensor imaging (DTI) metrics of white matter integrity are better than conventional magnetic resonance imaging (MRI) metrics for discriminating cerebellar diseases. Methods Conventional MRI images from 31 consecutive patients with ataxia and 12 controls were interpreted by a neuroradiologist given only a clinical indication of ataxia. An expert ataxologist, blinded to radiological findings, determined the clinical diagnosis, as well as ataxia severity and asymmetry for each patient. For qualitative analysis, a comparison of the cerebellar white matter in ataxic vs. control patients was made by visual inspection of directionally encoded color (DEC) images. For quantitative analysis, segmentation of the cerebellar white matter in the inferior, middle, and superior cerebellar peduncles (ICP, MCP, and SCP) was attempted using three methods: a region of interest method, a deterministic DTI tractography (DDT) method, and a probabilistic DTI tractography (PDT) method. A statistical comparison of the average fractional anisotropy (FA) in these tracts was made between subject groups, and correlated to clinical diagnosis, severity, and asymmetry. Results Of the 31 consecutive patients with ataxia, the two largest subgroups had a clinical diagnosis of multiple system atrophy (cerebellar subtype; MSA-C), and spinocerebellar ataxia-1 (SCA1). Conventional MRI features, such as degree of pontocerebellar atrophy, correlated with ataxia severity, but were neither sensitive nor specific for the ataxia subtypes. PDT was the most accurate and least variable method of the three methods used for determining FA, especially in the ICP. Average FA in all ataxic patients was significantly decreased in the MCP, SCP and ICP and this decrease correlated to disease severity. Asymmetric ataxia correlated to proportionately larger contralateral MCP, ICP and SCP FA values. MCP, ICP, and SCP FA difference values formed distinct clusters that distinguished MSA-C from SCA-1, and other ataxia syndromes. Conclusions Qualitative and quantitative reductions in DTI metrics of white matter integrity in the cerebellar peduncles correlated better to clinical features of patients with sporadic and hereditary ataxias than conventional structural MRI measures of pontocerebellar atrophy.