AI Image
AI Tool Predicts Cancer Survival by Analyzing a Patient's Selfie
Newser • • health
Key Points:
- Researchers at Mass General Brigham have developed an AI model called FAHR-FaceSurvival that predicts mortality risk in cancer patients using only a facial photograph, as reported in Lancet Digital Health.
- The model demonstrated that patients in the highest 10% risk category had about a 50% chance of dying within 90 days, while those in the lowest 10% risk category had roughly a 2% mortality rate.
- FAHR-FaceSurvival evolved from an earlier AI tool, FaceAge, which estimates a person's biological age, but the new model focuses on facial areas below the eyes and along the nose, though the exact predictive signals remain unclear.
- Experts find the approach intriguing but emphasize the need to validate that the AI provides additional insights beyond clinical assessments and to ensure it does not perpetuate racial or other biases.
