Learning clinical features from patterns on PET scan in patients with dementia

Dementia affects every corner of the globe and has an estimate prevalence between 40-50 million people around the world. It affects an estimate of 211208 Australian according to a recent study. The diagnosis is based on clinical features supported by appropriate investigations. PET scans based on FDG has been used in the diagnosis of dementia in selected cases. Patterns of glucose hypometabolism has been associated with diagnosis of dementia but not the correlation with clinical features. In this project, the aim is to apply machine learning tool to relate patterns of glucose hypometabolism to clinical features among patients with dementia.