Change in AI Face Age Between Photos Linked to Cancer Survival

The work appeared in Nature Communications in 2026. It is an observational study of 2,276 adults with cancer, all over age 20, who each had at least two separate courses of radiation therapy. The photos were taken as part of routine care. An observational study can show that two things go together. It cannot show that one causes the other.

The tool is called FaceAge. It estimates biological age, meaning how old a person looks in health terms, from a facial photograph. The team used two photos per patient to calculate a Face Aging Rate (FAR): the change in FaceAge divided by the time between the photos. A FAR above 1 means faster-than-expected aging. Below 1 means slower.

Patients had a median age of 63.4 at the first course. About half were women (50.5%). Most were White (85.1%). Cancer had spread in 62.9% at the first course and in 78.7% at the second. The median gap between photos was 286 days, and the median followup was 35.7 months.

Higher FAR was linked to worse overall survival. The researchers split patients by the time between photos. After adjusting for time between photos, sex, race and diagnosis, the adjusted hazard ratios were 1.25 for gaps of 10 to 365 days, 1.37 for 366 to 730 days, and 1.65 for 731 to 1,460 days. A hazard ratio above 1 means a higher risk of death over the study period. The mid-range interval's confidence range (1.00 to 1.86) only just cleared 1.

Median survival differed too. It was 4.1 versus 6.5 months in the shortest-gap group, 6.4 versus 12.5 months in the middle group, and 15.2 versus 36.5 months in the longest-gap group, comparing faster with slower aging. The cut-off for "fast" differed by group (FAR above 20, 10 and 1). The authors explain that short gaps magnify small measurement noise, so short-interval values swing widely.

FAR added information beyond a single FaceAge reading. Patients who looked older than their years at the first photo and also had a high FAR faced the highest risk. The link also held in patients with metastatic cancer. Chronological age alone was not significantly tied to survival in any of the three groups.

There are clear limits. Everyone in the study had cancer and was receiving radiation, and many had advanced disease. Most were White. The study does not show that a face age score means anything for healthy people. It also does not test whether changing habits or treatment alters the rate. The authors call FAR a non-invasive prognostic marker. They do not present it as a diagnosis. The weakest point is the cut-offs: they differed by interval group, which makes the groups hard to compare.

Smoking is one visible thread. The paper notes that the team's earlier FaceAge work found links with health behaviors such as smoking, alcohol and drug use. The ASCO Post, reporting a separate Lancet Digital Health study from the same group, says smoking had the strongest association with looking biologically older in that work. That article quotes co-corresponding author Raymond H. Mak, a radiation oncologist at Mass General Brigham, who said the goal is to support treatment decisions "rather than replace physician judgment." That second study, which used a different model, is not the one analyzed here.

What to watch next

Whether FAR holds up in other patient groups and in people without cancer. The ASCO Post reports that the team is expanding its work through a FaceAge Research Study open to the public.

Two-Year Fish Oil Trial Finds No Memory or Brain-Shrinkage Benefit Detailed Data Confirm Personalized mRNA Melanoma Vaccine Cut Recurrence Risk 42% Blood Signatures of Fitness Linked to Lower Death Risk in UK Biobank Change in AI Face Age Between Photos Linked to Cancer Survival Untreated Sleep Apnea Linked to 69% Faster Cognitive Decline Over 10 Years Over 80 Nutrition Experts Urge White House to Release Ultra-Processed Food Definition Beetroot Juice and Postmenopausal Heart Oxygen Balance After Hard Exercise: A Small Trial Topical Urolithin A Lowered Eye Wrinkle Depth in Randomized Studies Short Breathing Exercises Blunted Heart Rate and Anxiety Under Stress Aerobic and Resistance Training Ended Up Equal for Type 2 Diabetes Plant and Dairy Protein Shakes Left Blood Chemistry Alike in Adults Over 50 Intensity Relative to Your Own Fitness Better Explained Fitness in Midlife Women Blood Flow Restriction Training Showed No Overall Advantage for Artery Health Cigarette Smoke Disrupts Eye Cell Mitochondria, Pointing to Macular Degeneration Risk Ultra-Processed Food Intake Leaves Distinct Blood Fat and Metabolite Signatures Gut-Made Kidney Toxins Linked to Tiny Brain Bleeds in Mice and Patients Flo App Survey of 17,494 Women Finds Gaps in Perimenopause Knowledge New Aging Framework Names Epigenetic Change the Dominant Driver of Aging Exercise Helps Blood Cancer Patients During Treatment, but Diet Evidence Lags Review: Protein Supplements Often Unnecessary Outside the Lab for Active Adults Ear Vagus Nerve Stimulation Linked to Less Inflammation After Surgery Fasting and Lean Mass Over 45: Short Trials, Small Differences Fragmented Sleep Tracked by Wrist Monitors Linked to More Brain Amyloid at Autopsy Blood Test for p-tau217 Predicted Decline Over 12 Years Steady Aerobic Exercise Raised Fitness on Average in Heart Disease Patients Rising Arterial Stiffness Linked to Progression of Heart-Kidney-Metabolic Disease Brain-Wide Gene Therapy Restored Memory in Alzheimer's and Old Mice Men With Heart and Metabolic Disease Show Faster Epigenetic Aging AI Spotted Heart Disease From Face and Tongue Photos in Small Test