AI Predicts Cancer Survival from a Single Selfie

A simple selfie might soon help doctors make life-saving treatment decisions for cancer patients.

Researchers at Mass General Brigham in Massachusetts have created an AI system that can predict a cancer patient's chances of survival from just one photograph of their face. The technology estimates biological age (how old someone looks versus their actual age) and directly predicts survival outcomes for people diagnosed with cancer.

The AI learned from more than 40 million facial photographs across different ethnicities and ages. It developed two complementary tools: one that estimates biological age to predict survival, and another that predicts survival directly from facial features.

The system looks at different parts of the face for each prediction. For biological age, it focuses on creases running from the nose to mouth corners, the forehead above the eyes, and the temples. For survival predictions, it examines the region below the eyes and around the nasal bridge.

The results were striking. Among cancer patients receiving radiotherapy in the US, people who looked five or more years older than their actual age had a 21% higher risk of dying during the four-year follow-up period. Those who appeared five or more years younger had a 17% lower risk of death.

Dr. Raymond Mak, a radiation oncologist at Mass General Brigham, explained the practical value. "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," he said.

The AI remained remarkably consistent across real-world photo variations including different poses, lighting, facial expressions, and makeup. It even stayed stable across most plastic surgery procedures, with only rhinoplasty causing a small shift in readings.

Researchers discovered that lifestyle factors showed up in facial aging patterns. Heavy alcohol use, drug use, and smoking all correlated with looking biologically older in AI analysis.

The Bright Side

This breakthrough offers a low-cost, easy way to support personalized treatment decisions without expensive tests or invasive procedures. The foundation-model approach means doctors can develop multiple facial health biomarkers rather than building each tool from scratch, making the technology scalable and accessible.

The tool is designed to support, not replace, physician judgment in helping patients navigate difficult choices about their care.

While more research is needed before widespread use, this technology represents a major step forward in making healthcare more personalized and precisely tailored to each patient's unique situation.