AI facial analysis may help doctors estimate cancer survival risk

Researchers at Mass General Brigham have developed an AI system that analyzes facial photographs to help estimate survival risk in cancer patients, with those in the highest-risk group showing over three times the mortality rate of the lowest-risk group. The system, called FAHR-FaceSurvival, was built on an AI model trained on more than 40 million faces and fine-tuned using 34,363 photos of cancer patients. It does not diagnose cancer but searches for facial clues linked to survival, with findings validated in a separate patient group. A second tool, FAHR-FaceAge, estimates biological age, and the gap between estimated and actual age was independently tied to survival. The two tools focused on different facial regions, indicating they capture distinct health information.