Can an AI-based, non-invasive imaging framework, using high-dimensional MRI data from the UK Biobank, accurately assess early signs of organ health decline across multiple organs (e.g., lungs, heart, liver) and identify novel, actionable biomarkers that extend beyond existing prognostic factors to enable preventive and regenerative health strategies? This research question directly addresses the aim to create an automatic framework that can detect organ health decline relative to peers or adjacent organs, which is potentially actionable for early intervention. By focusing on discovering novel biomarkers and functional insights into organ health, the study highlights a clear scientific rationale for using AI to advance preventive healthcare. We envision a fully automatic framework that extracts high-dimensional image features, capable of running in the background of a workstation for opportunistic screening.
The organ health derived from MRI features will be validated against clinical and demographic data available in the UK Biobank, including genetic profiles, medical histories, and lifestyle factors such as diet and physical activity. We aim to investigate the relationships between organ health, these established prognostic health metrics, and various clinical outcomes, such as cardiovascular disease and liver disorders, to explore the additive information gain achieved by non-invasive opportunistic screening techniques.
The clinical and statistical analyses will involve regression models, survival analysis, and machine learning techniques, among others. By integrating imaging-based organ health with patient-level data, this project seeks to contribute to the growing field of precision medicine. The ultimate goal is to provide clinically relevant insights into how MRI-derived health metrics correlate with disease risk and overall health outcomes, supporting predictive models for early intervention and personalized treatment pathways.