Researchers at Mass General Brigham in Boston have developed an artificial intelligence system that analyzes a single facial photograph to estimate a person's biological age and predict mortality risk in patients with cancer. Detailed in a validation study published online October 19, 2026, in The Lancet Digital Health, the technology aims to give physicians an objective tool to assess physical frailty before recommending intensive medical treatments.

According to Mass General Brigham, the research team built a base foundation model called Foundation Artificial Intelligence for Health Recognition-Face (FAHR-Face), trained on more than 40 million publicly available facial images. From that foundation, scientists created two specialized algorithms: FAHR-FaceAge, which calculates how old a body appears physiologically rather than chronologically, and FAHR-FaceSurvival, which directly estimates survival odds for oncology patients.

"To our knowledge, this is the first demonstration that an AI model can directly predict cancer survival from a simple photograph," Dr. Hugo Aerts, director of the Artificial Intelligence in Medicine program at Mass General Brigham and co-senior author of the study, said in the institution's announcement.

How Facial Features Signal Underlying Health

Doctors have long relied on visual impressions—often called the "eyeball test"—to gauge how well a patient might tolerate harsh therapies like surgery, chemotherapy, or radiation. The Mass General Brigham team sought to make these subjective observations measurable and consistent.

Although both tools stem from the same base system, they examine entirely different parts of the face and capture distinct health signals. Speaking with CNN, study author Dr. Raymond Mak, a radiation oncologist at the Mass General Brigham Cancer Institute, explained that FAHR-FaceAge focuses on the temples, the forehead above the eyes, and the nasolabial folds—the lines running from the nose to the corners of the mouth.

By contrast, FAHR-FaceSurvival evaluates the bridge of the nose and the region directly beneath the eyes. Mak noted that these specific zones closely mirror the clinical observations made around 400 B.C. by the ancient Greek physician Hippocrates, who wrote that a sharp nose, sunken eyes, and collapsed temples signaled severe muscle wasting and impending mortality. Despite having no access to medical texts, the algorithm independently identified those same facial zones as key indicators of survival risk.

Lifestyle habits also left clear marks on the facial age estimates. According to voz.us, the study found that current smokers generally registered a higher facial biological age than former smokers, who in turn looked biologically older to the algorithm than people who had never smoked.

What the Study Data Reveals About Survival Risk

While the foundational FAHR-Face system learned general facial features from over 40 million images, BBC coverage notes that the cancer survival model was further trained using 35,000 cancer-patient photographs. The researchers then tested the tools across large patient groups, with news outlets highlighting different statistical thresholds from the data.

In an evaluation of more than 38,000 cancer patients treated with radiotherapy in the United States, The Boston Globe and The Indian Express report that patients whose estimated biological age was at least five years older than their actual chronological age faced a 21 percent higher risk of death over an average follow-up of four years, even after adjusting for age, sex, and cancer type. Conversely, patients who registered at least five years younger than their chronological age had a 17 percent lower risk of death.

Looking at shorter-term outcomes, CNN reports a separate comparison from the FaceAge analysis: 21 percent of 1,060 patients whose estimated biological age exceeded their chronological age by more than 20 years died within 90 days, compared with about 6 percent of 4,195 patients whose biological age was estimated at least 10 years younger.

When evaluating the FAHR-FaceSurvival algorithm directly in an analysis of 3,820 cancer patients, researchers assigned individuals a risk score between zero and one, according to International Business Times. Among the 122 patients placed in the highest-risk tier (scores of 0.9 to 1.0), 64 patients—or 52.5 percent—died within 90 days of their photograph being taken, as reported by CNN and Scientific American. In the lowest-risk group of 975 patients (scores of zero to 0.1), about 2 percent (20 patients) died within the same 90-day window. Combining both the biological age and survival models yielded stronger predictions than using either tool alone.

Clinical Limitations and What It Means for Patients

Both the study authors and independent medical experts emphasize that these algorithms are designed to support, not replace, a doctor's clinical judgment. The findings represent statistical risk associations rather than definitive individual prognoses.

Pratik Shah, a professor of pathology at the University of California, Irvine, who was not involved in the research, told The Boston Globe that while the large foundation model is innovative, the results rely on historical patient data rather than a prospective clinical trial. Shah pointed out that the facial-age model showed significant survival associations in 18 of 28 cancer categories, indicating that the technology may not perform consistently across every type of cancer.

Is the AI tool available to the public?

Only the biological age estimator is accessible to the general public for research purposes; the cancer survival algorithm is strictly limited to clinical research. According to The Boston Globe, members of the public can join the FaceAge Research Study online by signing a consent form, answering questions about sleep, exercise, and medical history, and uploading three to 10 selfies. More than 10,000 people have enrolled since the portal quietly opened in March, helping researchers study how lifestyle habits affect aging across more diverse populations.

How might doctors use this technology in the future?

Before the tools can be used in routine medical care, they must be validated in active clinical settings. Mass General Brigham has launched two clinical trials to evaluate how well the technology predicts cancer survival and surgical outcomes. If proven effective, the software could help oncologists identify frail patients who might struggle with aggressive chemotherapy or radiation, allowing care teams to tailor treatment intensity to a patient's true physical reserve.