New AI tool can predict cancer survival by a single patient selfie, study suggests

A groundbreaking AI tool developed by experts at Mass General Brigham, a nonprofit healthcare system in Massachusetts, can predict a cancer patient's chances of survival based on a single photograph of their face. The AI system, called FAHR-FaceAge and FAHR-FaceSurvival, was trained on over 40 million facial images representing various ethnicities and ages.

The technology estimates a person's biological age and directly predicts cancer survival outcomes, offering a promising new approach to assessing patients' health and guiding medical decisions. The FAHR-FaceAge model focuses on specific facial features, such as creases near the nose and mouth, the forehead, and the temples, while FAHR-FaceSurvival analyzes areas below the eyes and around the nasal bridge to assess survival risk.

The study published in the Lancet Digital Health journal found that cancer patients who appeared five or more years older than their actual age through FAHR-FaceAge had a 21% higher risk of dying during a four-year follow-up compared to those who looked their chronological age. Conversely, patients who appeared five or more years younger had a 17% lower risk.

Researchers believe that biological age estimates may be closely tied to physical health and treatment outcomes, potentially aiding doctors in making informed choices about treatments like radiotherapy, surgery, or systemic therapy. The AI tool could help predict the risk of death or complications after major treatments for cancer, offering a more quantitative and consistent method to support personalized treatment decisions.

The study also discovered that lifestyle factors, such as heavy alcohol use, drug use, and smoking, are linked to looking biologically older in facial photographs. Additionally, FAHR-FaceAge demonstrated remarkable stability across various real-world photo variations, including pose, lighting, facial expression, and makeup, and showed a small shift after rhinoplasty.