Researchers at Mass General Brigham say an experimental AI tool may someday help doctors judge how sick a patient is from nothing more than a single photo of their face.
The tool, called FaceSurvival, was trained on millions of photos of healthy faces and 35,000 photos of cancer patients, and estimates survival odds by closely analyzing features near the eyes and the bridge of the nose. In preliminary findings published in The Lancet Digital Health, the researchers say the pattern held across many cancer types and severities, in patients ranging from children to people in their 90s, drawn from the United States and the Netherlands.
As the population ages, physicians need better ways to estimate physical reserve and frailty before recommending intensive treatments such as radiation, surgery or systemic therapy. Our goal is to make face photograph-based health assessment more quantitative, consistent and scalable so it can support personalised treatment decisions rather than replace physician judgment.
Dr. Raymond Mak, Mass General Brigham
A related tool, FaceAge, estimates biological rather than calendar age by examining the forehead, temples, cheeks and skin around the mouth — areas shaped over time by sun exposure, smoking, alcohol and drug use. Mak said FaceAge held up consistently across ordinary real-world photos despite differences in lighting, pose and makeup, with only a small shift seen after rhinoplasty.
Both tools remain far from clinical use. Prof. Miroslaw Bober, a machine-learning researcher at the University of Surrey who was not involved in the work, called it "a promising proof of concept" that still needs testing against doctors' own judgment across different hospitals and patient groups. "Predicting an outcome is not the same as explaining it," he said. The researchers themselves caution that they need far more data, particularly from a wider range of ethnic groups, before the tools could be relied upon. A patent application is pending, and the team is inviting members of the public to submit their own photos for a personalized biological-age estimate.