Quantitative assessment of Nigrosome-1 volume and susceptibility in Parkinson’s Disease

Motivation: Parkinson's disease diagnosis, in early stages, still strongly relies on qualitative clinical evaluation, rather than quantitative data, often resulting in misdiagnosis. Goal(s): The goal was to identify PD patients when Nigrosome-1 is still visible on MRI imaging Approach: We extracted quantitative structural data of Nigrosome-1 and trained a Machine Learning model to perform a classification task Results: Quantitative features, Volume of Nigrosome-1 in particular, proved to be a good feature to differentiate PD from HC, performing 0.87 accuracy, and 0.94 AUC-ROC Impact: These results support the need to integrate visual assessment of N1 with a quantitative assessment of its structure and susceptibility properties to better characterize PD pathology