US researchers have developed an AI cancer survival tool that can predict a cancer patient’s chances of survival from a single facial photograph, according to a study published in The Lancet Digital Health.
The system, called FAHR-Face, was built by scientists at Mass General Brigham in Massachusetts. It analyses facial features to estimate biological age and mortality risk, a capability its creators say could one day help doctors make better treatment decisions.
Two tools, one photograph
FAHR-Face (Foundation Artificial Intelligence for Health Recognition) is a foundation model trained on more than 40 million facial images. From that base, the team built two tools.
The first, FAHR-FaceAge, estimates a patient’s biological age, or how old the body looks compared with the birth certificate. The second, FAHR-FaceSurvival, predicts survival outcomes in people with cancer directly from a face photograph, without using age as an intermediate step.
The survival model was trained on images of 34,363 patients and tested on another 3,820. Its results were then independently validated using data from 4,892 cancer patients in the Netherlands.
The numbers are striking. Patients placed in the highest-risk group faced more than three times the mortality risk of those in the lowest-risk group, even after accounting for clinical factors. In one analysis, 52.5 per cent of high-risk patients died within 90 days of assessment, compared with about 2 per cent of those in the lowest-risk band.
Patients whose faces appeared at least five years older than their actual age carried a 21 per cent higher mortality risk, while those who looked at least five years younger had a 17 per cent lower risk. Combining the two models improved prediction accuracy further.
Reading different parts of the face
The researchers found that the two tools pick up different signals. The survival model focuses mainly on the area beneath the eyes and around the bridge of the nose, while the age-estimation model examines features around the mouth, forehead and temples.
The tools proved robust across variations in pose, lighting, facial expression and makeup, as well as most plastic surgery procedures. Heavy alcohol use, drug use and smoking were associated with looking biologically older.
Hugo Aerts, director of the Artificial Intelligence in Medicine programme at Mass General Brigham and co-senior author of the study, called it “the first demonstration” that an AI model can directly predict cancer survival from a photograph.
Raymond Mak, a radiation oncologist at the Mass General Brigham Cancer Institute and the study’s corresponding author, said doctors need better ways to gauge 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,” he said.
Why this AI cancer survival tool matters
The clinical problem the AI cancer survival tool addresses is a familiar one in oncology. Chronological age alone is a poor guide to whether a patient can tolerate aggressive treatment, and doctors currently rely on a mix of tests, scans and clinical judgment to assess frailty.
A photograph-based risk score would not replace those assessments, the researchers stress. But as a fast, non-invasive supplement, it could help flag patients who need a closer look before a gruelling course of therapy. That is what makes the “first demonstration” claim significant: survival prediction has never before been extracted directly from a face image.
Mass General Brigham has filed a US provisional patent application on the next generation of facial health algorithms and is expanding the research through a public online study approved by an institutional review board.
Not ready for the clinic yet
The researchers are clear about the limits. Further prospective validation is needed before the technology can be used in routine clinical practice. The study’s authors position the tool as a supplement to existing prognostic workups, not a replacement.
Still, the direction of travel is clear. AI is moving from analysing scans and lab reports to reading subtler signals, and health technology continues to attract serious research investment. As with AI-generated content labels on music platforms, the technology is advancing faster than the rules around it.
Frequently asked questions
What is the AI cancer survival tool?
FAHR-FaceSurvival is an AI model developed by Mass General Brigham researchers that predicts cancer patients’ survival outcomes directly from a single facial photograph. It was described in a study published on October 19 in The Lancet Digital Health.
How accurate is it?
Patients in the highest-risk group had more than three times the mortality risk of those in the lowest-risk group. In one test, 52.5 per cent of high-risk patients died within 90 days, versus about 2 per cent of low-risk patients.
What does the AI look at in the face?
The survival model focuses on the area beneath the eyes and around the bridge of the nose. A companion tool, FAHR-FaceAge, estimates biological age from features around the mouth, forehead and temples.
Is the tool available for doctors to use?
Not yet. The researchers say further prospective clinical validation is needed before the technology can be widely used in practice. Mass General Brigham has filed a provisional patent and is expanding the research.
