Mass General Brigham develops AI to predict cancer survival from facial photos

Researchers at Mass General Brigham have developed an artificial intelligence system capable of estimating a cancer patient's mortality risk and biological age using a single facial photograph. The study, published in The Lancet Digital Health, introduces the FAHR-Face (Foundation Artificial Intelligence for Health Recognition) model, which was trained on over 40 million facial images across diverse ethnicities and ages.

The system utilizes two complementary tools: FAHR-FaceAge, which estimates biological age, and FAHR-FaceSurvival, which predicts survival outcomes. In clinical testing, the survival model demonstrated a significant ability to distinguish risk groups. Among patients in the highest-risk group, 52.5 percent died within 90 days of the photograph being taken, compared to only approximately 2 percent in the lowest-risk group.

Furthermore, the biological age tool revealed that patients whose faces appeared five or more years older than their chronological age faced a 21 percent higher risk of death. Conversely, those appearing five years younger had a 17 percent lower risk. These tools aim to assist clinicians in personalizing treatments, such as surgery or radiotherapy, by providing a quantifiable measure of a patient's physiological state.