Magnetic resonance imaging markers of Parkinson's disease nigrostriatal signature

One objective of modern neuroimaging is to identify markers that can aid in diagnosis, disease progression monitoring and long-term drug impact analysis. In this study, Parkinson-associated physiopathological modifications were characterized in six subcortical structures by simultaneously measuring quantitative magnetic resonance parameters sensitive to complementary tissue characteristics (i.e. volume atrophy, iron deposition and microstructural damage). Thirty patients with Parkinson's disease and 22 control subjects underwent 3-T magnetic resonance imaging with T(2)*-weighted, whole-brain T(1)-weighted and diffusion tensor imaging scans. The mean R(2)* value, mean diffusivity and fractional anisotropy in the pallidum, putamen, caudate nucleus, thalamus, substantia nigra and red nucleus were compared between patients with Parkinson's disease and control subjects. Comparisons were also performed using voxel-based analysis of R(2)*, mean diffusivity and fractional anisotropy maps to determine which subregion of the basal ganglia showed the greater difference for each parameter. Averages of each subregion were then used in a logistic regression analysis. Compared with control subjects, patients with Parkinson's disease displayed significantly higher R(2)* values in the substantia nigra, lower fractional anisotropy values in the substantia nigra and thalamus, and higher mean diffusivity values in the thalamus. Voxel-based analyses confirmed these results and, in addition, showed a significant difference in the mean diffusivity in the striatum. The combination of three markers was sufficient to obtain a 95% global accuracy (area under the receiver operating characteristic curve) for discriminating patients with Parkinson's disease from controls. The markers comprising discriminating combinations were R(2)* in the substantia nigra, fractional anisotropy in the substantia nigra and mean diffusivity in the putamen or caudate nucleus. Remarkably, the predictive markers involved the nigrostriatal structures that characterize Parkinson's physiopathology. Furthermore, highly discriminating combinations included markers from three different magnetic resonance parameters (R(2)*, mean diffusivity and fractional anisotropy). These findings demonstrate that multimodal magnetic resonance imaging of subcortical grey matter structures is useful for the evaluation of Parkinson's disease and, possibly, of other subcortical pathologies.