An AI system developed at Mass General Brigham can estimate biological age and forecast survival for cancer patients using a single facial photograph, according to research published in Lancet Digital Health. Trained on more than 40 million images, the model underpins two tools: one that infers survival via biological age and another that predicts risk directly from facial features. Patients who appeared at least five years older than their chronological age faced a 21% higher mortality risk over four years, while those who looked five years younger had a 17% lower risk. The authors say the approach could help guide treatment intensity and surgical decisions as populations age, though they caution that further validation and safeguards around privacy and bias are needed before clinical deployment.
Related articles:
– Dermatologist-level classification of skin cancer with deep neural networks
– Clinically applicable deep learning for diagnosis and referral in retinal disease
– DeepGestalt: AI-assisted facial phenotyping for genetic syndrome diagnosis
