Approved research

Machine learning to assess biological age and predict disease from multimodality medical imaging

Lay summary

The goal of this study is to find a new way to assess biological age and the risk of disease, based on medical images. The premise is intuitive -- when meeting a new acquaintance, we reflexively size up that person's age and health based on their face, mannerisms and gait. In this study we will train a machine learning model - a type of artificial intelligence - to similarly gauge biological age and risk of disease based on advanced medical imaging that looks inside the body, head and eyes. This promises to give us new insights into aging and disease. It may also yield practical tools to inform decisions about lifestyle, prevention, and screening. Accurate and generalizable machine learning models require a large amount of data. So this will be an ongoing project with a rolling 3-year duration to allow for inclusion of all UK Biobank participants who have imaging.