Predicting amyloid pathology from MRI and assessing its clinical utility using CRIS-linked SLaM Image Bank data

Dementia, including Alzheimer’s disease, causes major disability in older adults, but current tests (such as PET scans) to detect underlying amyloid (necessary for confirmation of Alzheimer’s) are invasive, expensive and rarely available in routine mental health services.
This project will use electronic health records and brain scans (MRI) from a large SLaM cohort of around 12,000 patients (SLaM Image Bank) to apply an existing AI model that estimates amyloid levels from MRI alone, and will test how well these estimates predict cognitive decline and clinical outcomes in real-world patients.