FaceAge: AI Predicts Cancer Survival by Reading Your Face

Mass General Brigham’s researchers have introduced FaceAge, an AI tool that estimates a person's biological age from a single facial photograph—and it could help doctors better predict cancer survival outcomes.

What is FaceAge?

FaceAge is a deep learning system trained on over 58,000 facial images from healthy individuals aged 60+. It translates subtle features like skin texture and muscle tone into a biological age—a more insightful measure of health than chronological age.

In a study of more than 6,000 cancer patients, researchers found that patients appeared, on average, about 5 years older biologically than their actual age. Those with higher FaceAge scores had significantly worse survival outcomes, even after adjusting for sex, cancer type, and other clinical variables.

Why it matters

In physician testing, FaceAge risk scores improved the accuracy of 6-month survival predictions—from 74% to 80%—when added to standard clinical data.

Even more striking: FaceAge scores correlated with expression of a gene linked to cellular aging, suggesting it captures biological processes that go beyond surface appearances or calendar years.

"This work demonstrates that a photo, like a simple selfie, contains important information that could help inform clinical decision-making," said Dr. Hugo Aerts, co-senior author.

What’s next?

Though promising, researchers note the training data may not yet reflect global diversity in age and ethnicity. Additional validation is needed to confirm FaceAge’s performance across different populations and diseases.

Still, FaceAge represents a fascinating frontier: turning something as ordinary as a selfie into a clinically actionable biomarker. The face, it turns out, might really be a window to your biological clock.

🧠 Read the full study in The Lancet Digital Health: https://www.thelancet.com/journals/landig/article/PIIS2589-7500(25)00042-1/fulltext