AI Tool Predicts Cancer Patients’ Survival Chances from Facial Photographs, Study Suggests

The study, published in The Lancet Digital Health, found that patients whose faces appeared older than their actual age had a higher risk of death during follow-up.

Two AI tools developed by researchers at Mass General Brigham estimate biological age and predict survival in cancer patients using facial photographs.

The study, published in The Lancet Digital Health, found that patients whose faces appeared older than their actual age had a higher risk of death during follow-up.

Doctors have developed an artificial intelligence (AI) tool that can estimate a cancer patient’s biological age and predict survival chances using a photograph of their face.

The system was developed by experts at Mass General Brigham, a nonprofit academic healthcare system in Massachusetts, US. The findings were published in The Lancet Digital Health.

The tool was trained on more than 40 million facial photographs of people from different ethnic and age groups. The researchers developed two tools from the same model: FAHR-FaceAge, which estimates biological age and uses it to predict cancer survival, and FAHR-FaceSurvival, which predicts survival directly from a photograph without using biological age as an intermediate step.

Biological age refers to how old a person appears, rather than their actual age in years. The two tools analyse different facial areas.

FAHR-FaceAge focuses on lines running from the sides of the nose to the corners of the mouth, the forehead above the eye sockets and the temples. FAHR-FaceSurvival examines the area below the eyes and around the bridge of the nose.

“To our knowledge, this is the first demonstration that an AI model can directly predict cancer survival from a simple photograph,” said Hugo Aerts, a co-senior author of the study.

The study found that among cancer patients treated with radiotherapy in the US, those assessed by FAHR-FaceAge as looking five or more years older than their actual age had a 21% higher risk of dying during an average follow-up of four years. Patients assessed as looking five or more years younger had a 17% lower risk.

Aerts said the tool learns a general representation of the face from millions of images, which can then be adapted to different health outcomes. “This foundation-model approach gives us a scalable way to develop multiple facial health biomarkers rather than training each model from scratch,” he said.

The researchers said the findings could help doctors assess patients’ physical health and ability to cope with intensive treatments. The technology could potentially support decisions involving radiation, surgery and drug therapy, as well as help predict the risk of death or complications following treatment.

Dr Raymond Mak, a radiation oncologist at Mass General Brigham, said doctors need better ways to assess patients’ strength and frailty before recommending intensive treatments as the population ages.

“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,” he said.

The AI analysis also linked heavy drinking, drug use and smoking with looking biologically older. Mak said the team was surprised by how consistent the results remained despite differences in everyday photographs, including pose, lighting, facial expression and makeup.

The tool also produced stable results after most plastic surgery procedures, although a small change was observed after a nose job.

The researchers cautioned that further research is needed before the technology can be used widely. For now, the tool could offer a low-cost way to support clinical assessment, but it is not intended to replace a doctor’s judgement.

Stay tuned for more such updates on Digital Health News