Purpose: To explore the usefulness of histogram analysis of mean diffusivity (MD) derived from diffusion-weighted imaging of large infratentorial structures to distinguish parkinsonian syndromes.
Materials and methods: Local research ethics committee approval and informed consent were obtained. Ten patients with Parkinson disease (PD), nine with the parkinsonian variant of multiple system atrophy (MSA-P), seven with the cerebellar variant of MSA (MSA-C), 17 with progressive supranuclear palsy-Richardson syndrome (PSP-RS), and 10 healthy subjects were recruited. Histograms of MD values were generated for all pixels in the whole infratentorial compartment and separately for the whole brainstem, vermis, and cerebellar hemispheres. To assess the differences in MD values among groups, the Kruskal-Wallis test was used, followed by the Mann-Whitney U test for pairwise comparisons. All P values resulting from pairwise comparisons were corrected with the Bonferroni method.
Results: MSA-P and MSA-C groups had higher median MD values (P < .01) in the brainstem and cerebellum when compared with other groups; this finding was in line with the known consistent neurodegenerative damage in posterior cranial fossa structures in these diseases. Median MD values from cerebellar hemispheres were used to discriminate patients with MSA-C and those with MSA-P from patients with PD and those with PSP-RS (P < .01; sensitivity, specificity, and positive predictive value equaled 100%). Furthermore, patients with PSP-RS had significantly higher MD values in the vermis than did healthy subjects (P < .05) and patients with PD (P < .001).
Conclusion: These findings support the clinical usefulness of diffusion imaging in the differential diagnosis of parkinsonism, suggesting that the minimally operator-dependent histogram analysis of the infratentorial structures and particularly of the whole cerebellar hemispheres can be used to distinguish patients with MSA-P and those with MSA-C from patients with PSP-RS and those with PD.