Automated Segmentation of Neuromelanin and Nigrosome for Parkinson's Disease Diagnostic Assessment

Motivation: Parkinson's disease involves loss of nigrosome and neuromelanin in the substantia nigra, along with iron deposition. We evaluated how combining quantitative markers improves diagnosis accuracy and analyzed correlation between neuromelanin and standardized uptake value ratio (SUVR). Goal(s): We aimed to evaluate whether combining these quantitative markers could improve diagnostic accuracy and to assess their relationship with SUVR. Approach: We prospectively collected data to evaluate the diagnostic performance of the model and to assess the correlation with the SUVR. Results: The combined model achieved an AUC of 0.956 in PD diagnosis, establishing neuromelanin as a useful marker and confirming its significant relationship with SUVR. Impact: Combining neuromelanin data with nigrosome volume information enhances its utility for disease classification. Neuromelanin demonstrates a meaningful correlation with the standardized uptake value ratio of basal ganglia subregions.