AI tool analyzes "face age" from selfies to predict cancer risk

A new study shows that an algorithm can look at an ordinary photograph and estimate how fast a body is aging – an insight that could change the nature of cancer care. The research was led by a team of scientists from Mass General Brigham, who constructed FaceAge, an AI tool.

The experts trained the tool on almost 60,000 images of healthy individuals and then tested it on more than 6,000 patients who were starting radiotherapy.

The team found that the typical cancer patient appeared about five years older than their birth certificate suggested, and that each additional year shortened their life expectancy.

From selfie to biomarker

FaceAge is a deep-learning network. It studies fine details – skin texture, muscle tone, and eye shape – then translates those patterns into a single number: your biological age.

The process is automatic. If FaceAge receives a head-and-shoulders photo, it can predict an age that reflects the body’s wear and tear better than the chronological age.

How old a person seems has long guided doctors informally. Frail features can steer therapy toward gentler options; youthful vigor can justify aggressive treatment.

Yet such judgments are subjective. FaceAge puts a more objective figure on that impression.

AI, face age, and cancer prediction

In the study, patients whose biological age topped 85 fared worst, even after the authors adjusted for sex, tumor site, and chronological age.

“We can use artificial intelligence (AI) to estimate a person’s biological age from face pictures, and our study shows that information can be clinically meaningful,” said Hugo Aerts, the director of the Artificial Intelligence in Medicine (AIM) program at Mass General Brigham.

“This work demonstrates that a photo, like a simple selfie, contains important information that could help to inform clinical decision-making.”

“How old someone looks compared to their chronological age really matters – individuals with FaceAges that are younger than their chronological ages do significantly better after cancer therapy.”

FaceAge captures hidden signals

Predicting how long a terminal patient has left is difficult. The researchers asked ten clinicians and scientists to view one hundred portraits of people who were receiving palliative radiotherapy, and to guess whether each person would be alive within months.

Even when the panel knew chronological age and cancer type, their accuracy barely beat chance.

Adding FaceAge shifted the odds. With the AI number in hand, the group’s predictions improved markedly, suggesting FaceAge captures hidden signals that physicians miss.

Teaching AI to see face age

The investigators began with public image banks that held 58,851 faces, each tagged with an age. Those photos came from everyday contexts, so the network first learned to recognize normal aging.

Next the team applied FaceAge to clinical snapshots taken during routine treatment setup. Linking those pictures to medical records let the algorithm discover how appearance, disease, and outcome connect.

FaceAge still needs validation in larger and more diverse populations. The study cohort came from only two centers, and lighting or camera angles could skew results.

Cosmetic surgery, heavy makeup, or cultural differences in skin care might also confuse the model. The team plans to follow patients over time to see whether the FaceAge number changes as therapy progresses.

An early detection system

“This opens the door to a whole new realm of biomarker discovery from photographs, and its potential goes far beyond cancer care or predicting age,” said co-senior author Ray Mak, a faculty member in the AIM program at Mass General Brigham.

“As we increasingly think of different chronic diseases as diseases of aging, it becomes even more important to be able to accurately predict an individual’s aging trajectory.”

“I hope we can ultimately use this technology as an early detection system in a variety of applications, within a strong regulatory and ethical framework, to help save lives.”

FaceAge illustrates how artificial intelligence can turn everyday data into medical guidance. A single selfie may soon complement blood tests and scans, giving oncologists a faster, less biased picture of a patient’s resilience.

If further trials confirm the findings, clinics could upload photos and receive instantaneous biological age estimates that fine-tune treatment plans.

FaceAge: Beyond cancer care

Aging underlies heart disease, diabetes, dementia, and more. An image-based biomarker could therefore aid many specialties by identifying people who need lifestyle changes or preventive therapy years before symptoms appear.

The key will be strict oversight. Algorithms trained on limited datasets risk embedding bias, and patient consent will be vital when personal images feed predictive models.

For now, FaceAge remains a research AI tool. Yet its promise is clear: the face you present to the camera may hold clues to how your body is coping with illness and time. Harnessed responsibly, that knowledge could guide decisions that extend life and improve its quality – one snapshot at a time.

The study is published in The Lancet Digital Health.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–