Researchers at Mass General Brigham and Harvard Medical School have developed an artificial intelligence tool called FaceAge that analyzes a patient’s facial features to estimate their biological age and predict cancer survival time—helping doctors personalize treatment plans.
The deep-learning algorithm was trained on over 58,000 photos of healthy individuals and 6,000 cancer patients, revealing that cancer patients’ facial age averaged five years older than their chronological age. Patients who appeared older also tended to have worse outcomes.
While doctors often rely on visual assessments (the “eyeball test”), the study, published in The Lancet Digital Health, found that human predictions of short-term survival were only slightly better than chance. However, accuracy improved when clinicians used FaceAge’s estimates.
“We hypothesize that FaceAge could be used as a biomarker in cancer care to quantify a patient’s biological age and help doctors make tough decisions,” said co-senior author Raymond Mak, citing cases where the tool influenced treatment intensity based on perceived frailty or resilience.
Though promising, FaceAge requires further testing across diverse populations before clinical use. Researchers believe it could eventually monitor patient health before, during, and after treatment, offering a new way to assess risks and outcomes.
AI News Monitor 
Leave a comment