AI Estimates Cancer Survival From a Face Photo

Researchers at Mass General Brigham developed AI tools that estimate biological age and cancer survival from a facial photograph, according to The Guardian. The study was published in Lancet Digital Health.

The underlying FAHR-Face model was trained on more than 40 million facial images covering different ages and ethnicities. FAHR-FaceAge estimates biological age and uses it to predict survival, while FAHR-FaceSurvival predicts survival directly from a photograph.

Among US cancer patients receiving radiotherapy, those estimated to look at least five years older than their chronological age had a 21 percent higher risk of death during follow up that averaged four years. Patients estimated to look at least five years younger had a 17 percent lower risk.

Researchers said facial analysis could support assessments of frailty and physical reserve before radiation, surgery or systemic therapy. They cautioned that more research is needed before the tools can enter widespread clinical use.