Automated Segmentation of Substantia Nigra and Red Nucleus using Quantitative Susceptibility Mapping Images: Application to Parkinson’s Disease
University of Canterbury
Time & Place
Mon, 23 Nov 2020 15:00:00 NZDT in E16 - Engineering Core
All are welcome
Accurate segmentation of substantia nigra (SN) and red nucleus (RN) is challenging, yet important for understanding diseases like Parkinson’s disease (PD). We derived an algorithm to segment SN and RN from quantitative susceptibility mapping (QSM) MRI and used the results to investigate PD. Algorithm-derived segments (based on level set method and watershed transform) were compared to expert manually-derived segmentations in 40 participants. Using Bayesian regression models, we compared QSM values between PD and control groups, and investigated relationships with global cognitive ability and motor severity in PD. The proposed algorithm produced high quality segmentations, validated against expert manual segmentation. We showed moderate evidence of increased QSM values in SN in PD relative to controls, with moderate evidence for association between QSM, global cognitive ability, and motor impairment in the SN in PD. We suggest an improved midbrain segmentation algorithm may be useful for monitoring iron-related disease severity in Parkinson’s.
Dibash Basukala is a PhD student in the Computer Science and Software Engineering Department working under the supervision of Professor Ramakrishnan Mukundan. His research interests include MRI image segmentation, image feature extraction, and medical image analysis. Currently, he is finalizing his thesis.