PhD defence Jakob Weiß

Keywords: Opportunistic screening, Risk assessment, Artificial intelligence, Medical imaging

"Improving Personalized Risk Estimation Using Routine Medical Imaging Studies: An Opportunistic Screening Approach From Diagnostic Imaging to Predictive Analytics: A Paradigm Shift in Risk Estimation"

This thesis explores how artificial intelligence (AI) can turn routine medical scans, such as chest X-rays and CT scans, into powerful tools for predicting future health risks. While these scans are often used mainly to check for immediate symptoms, they actually contain far more information that remains currently unused in clinical care. By developing AI models to automatically extract and quantify this hidden data, the thesis demonstrates that routine imaging can be used to estimate risks of future health conditions such as heart disease, lung cancer, and overall mortality more accurately than traditional methods. These predictions are based on both established and novel imaging markers, including liver fat, muscle loss, and coronary artery calcium. Overall, the work highlights how AI could transform ordinary medical imaging into a novel form of risk prediction—one that can guide prevention, personalize treatment, and ultimately improve patient outcomes.

Click here for the live stream.