AI model estimates cancer survival from facial photos

Researchers at Mass General Brigham have developed an artificial intelligence model that estimates biological age and predicts cancer survival from facial photographs. The research-stage system, trained on more than 40 million face images, produces two measures: FAHRFaceAge estimates biological age, while FAHRFaceSurvival predicts survival directly from a photograph. The researchers reported that the tools capture different information. In an external cancer-patient cohort, combining the biological-age and direct-survival measures improved the fit of a survival model. The study did not test whether using the scores changes treatment decisions, improves patient outcomes or is safe for clinical use. In an analysis of roughly 38,000 cancer patients, those whose estimated biological age was at least five years older than their chronological age had a 21% higher risk of death over an average four-year follow-up. Patients estimated to look at least five years younger had a 17% lower risk, according to the study. A separate reported test group included 3,820 cancer patients. Of the 122 patients with the highest survival-risk scores, 64 died within 90 days of having their photograph...