AI Reads Cancer Survival Odds from a Single Face Photo

Researchers at Mass General Brigham have developed FAHR-Face, an AI trained on over 40 million face images that estimates cancer patients' survival odds from a single photo, potentially quantifying doctors' instinctive frailty assessments. The tool, described in The Lancet Digital Health on October 19, predicts survival directly and biological age separately, with combined predictions outperforming either alone. In a study of US cancer patients tracked for four years, those appearing five or more years older than their real age faced a 21% higher death risk, while those looking younger had a 17% lower risk. Lifestyle factors like smoking and heavy drinking correlated with looking biologically older, and the age tool remained robust across poses, lighting, and most cosmetic surgeries. The AI is not yet ready for clinical use, lacking trial validation and treatment-decision thresholds, and raises unresolved privacy concerns since face photos are widely accessible outside healthcare privacy laws. Mass General Brigham has filed a patent on facial health algorithms and runs a public research website where users can upload photos, highlighting the need for clear consent rules before such tools leave research settings.