Mass General Brigham Researchers Train FAHR-Face on 40 Million Images
Mass General Brigham researchers unveiled an artificial intelligence model trained on more than 40 million face images to estimate biological age and predict cancer survival outcomes. Published on Oct. 19, 2026, in The Lancet Digital Health, the study details the development of FAHR-Face (Foundation Artificial Intelligence for Health Recognition) and its two derived clinical tools: FAHR-FaceAge and FAHR-FaceSurvival.
The foundation model approach allows researchers to scale facial health biomarkers from a single general representation rather than training individual models from scratch. “The key technical advance is that FAHR-Face learns a general representation of the face from more than 40 million images, which can then be adapted to distinct health outcomes such as biological age and survival,” Hugo Aerts, co-senior author and Director of Artificial Intelligence in Medicine at Mass General Brigham, said in a news release.
FAHR-FaceAge Outperforms Benchmark Models Across the Human Lifespan
In the evaluation metrics reported by the research team, FAHR-FaceAge outperformed existing benchmark models while maintaining consistent accuracy across the full human lifespan. The model links biological age—a measure of cellular function and overall health that often diverges from chronological age—to observable lifestyle habits. Researchers observed that heavy alcohol use, drug use, and smoking correlated with looking biologically older.
FAHR-FaceSurvival Predicts Cancer Mortality to Guide Treatment
Beyond estimating age, FAHR-FaceAge surpassed previous estimation models in cancer survival prognostication, while FAHR-FaceSurvival accurately predicted mortality in patients with cancer. Clinicians emphasize that these diagnostic aids remain in development and are not yet ready for use in hospitals.
Dr. Raymond Mak, a radiation oncologist at the Mass General Brigham Cancer Institute and corresponding author of the study, noted the growing clinical need for these metrics. “As the population ages, physicians need better ways to estimate physical reserve and frailty before recommending intensive treatments such as radiation, surgery or systemic therapy,” Mak said in the release. “Our goal is to make face photograph-based health assessment more quantitative, consistent and scalable so it can support personalized treatment decisions rather than replace physician judgment.”
How AI Uses Facial Photos to Predict Cancer Survival RiskCan a photograph of someone's face reveal information about their chances of surviving cancer?Researchers at Mass General Brigham have developed FAHR-Face, an AI system trained on more than 40 million facial images.… pic.twitter.com/QNydUbkFME
Frequently Asked Questions About Facial AI Health Tools
When was the FAHR-Face study published and where?
The study was published on Oct. 19, 2026, in The Lancet Digital Health by researchers at Mass General Brigham.
How many facial images were used to train the AI model?
Researchers trained the foundational FAHR-Face model on a dataset consisting of over 40 million face images to establish a general facial representation.
Are FAHR-Face tools currently available for use in hospitals?
No, these tools are not yet ready for use in hospitals, as researchers are continuing development to guide future treatment options safely.