AI tool predicts cancer survival from a single selfie photograph, study finds
Researchers at Mass General Brigham in Massachusetts have built an AI system that estimates a cancer patient's biological age and predicts survival chances from a single facial photograph, according to a study published in Lancet Digital Health.
AI tool predicts cancer survival from a single selfie photograph, study finds
Doctors at Mass General Brigham, a nonprofit academic healthcare system in Massachusetts, have custom-built an artificial intelligence system for ai cancer survival prediction selfie by face: it estimates a person's biological age and directly predicts cancer survival from an ordinary photograph of the face, according to a study published in the Lancet Digital Health journal.
Two tools from one model
The model, called FAHR-Face, was trained on more than 40 million facial photographs across a range of ethnicities and ages. From it, the researchers built two complementary tools: FAHR-FaceAge estimates a patient's biological age - how old a person looks rather than their chronological age - and predicts cancer survival, while FAHR-FaceSurvival reads a facial photograph and produces a survival prediction on its own.
- FAHR-FaceAge reads the smile lines between nose and mouth, the strip of forehead just above the eye sockets, and the temples.
- FAHR-FaceSurvival draws its assessment from the area under the eyes and the bridge of the nose.
- Built from one shared model, the two tools nonetheless picked up on separate, mutually reinforcing health signals.
Hugo Aerts, a co-senior author of the study, said the key technical advance is that FAHR-Face learns a general face representation from over 40 million images that can then be adapted to distinct health outcomes.
What the study found
FAHR-FaceAge produced a clear survival split among US cancer patients treated with radiotherapy: over an average four-year follow-up, dying was 21% more likely in the group the tool judged to look at least five years beyond their real age, while looking at least five years younger than their real age came with a 17% lower chance of dying.
The study suggests biological age estimates are closely linked to physical health, potentially reflecting a person's ability to survive certain treatments. Heavy drinking, drug use and smoking all showed up in the AI's face analysis as features of people judged biologically older.
How researchers say it could be used
The tool could help better predict the risk of death or complications after major surgery or other cancer treatments, and assist doctors in treatment decisions. Aerts said FAHR-FaceSurvival advances facial health recognition AI by learning survival risk directly from cancer patients' photographs.
"To our knowledge, this is the first demonstration that an AI model can directly predict cancer survival from a simple photograph."
Dr Raymond Mak, a radiation oncologist at Mass General Brigham, said physicians need better ways to estimate physical reserve and frailty before recommending intensive treatments as the population ages. He noted how consistent FAHR-FaceAge was across common real-world photo variations, including pose, lighting, facial expression and makeup, and said the tool was stable across most tested plastic surgery procedures, with only a small shift observed after rhinoplasty.
"Our goal is to make face photograph-based health assessment more quantitative, consistent and scalable so it can support personalised treatment decisions."
Aerts added that the foundation-model approach offers a scalable way to develop multiple facial health biomarkers rather than training each model from scratch. Doctors cautioned, however, that more research is needed before the tool can be widely used, though they say single facial photographs could offer an easy, low-cost way to support treatment decisions.
