Face aging rate quantifies change in biological age to predict cancer outcomes

Why how fast we age matters

Two people born in the same year can look—and feel—very different decades later. One may seem vibrant and youthful, while the other appears worn down by illness and time. Doctors know that age influences cancer outcomes, but calendars do not tell the whole story. This study asks a simple but powerful question: can we measure how quickly a person is aging just by looking at their face over time, and can that help predict how they will fare with cancer?

Reading health in the face

Our faces quietly record the passage of time: skin thins, muscles slacken, and bone structure shifts. These outward changes mirror deeper biological shifts linked to disease and survival. The researchers built on an artificial intelligence system that estimates a person’s “biological age” from a single facial photograph. This earlier work showed that cancer patients who look much older than their actual age tend to have worse survival. In the new study, the team moved beyond a one-time snapshot to track how faces change between two points during a patient’s cancer care, turning the face into a moving gauge of aging rather than a fixed reading.

Aging rate as a moving target

The key measure in this work is called the Face Aging Rate, or FAR. Instead of asking, “How old does this person look today?” FAR asks, “How quickly has their apparent age changed between two visits?” To calculate it, the researchers applied their age-predicting AI to two routine ID photos taken before separate courses of radiation therapy. They then divided the change in estimated facial age by the time elapsed between photos. A value above one means the face is aging faster than expected over that period; a value below one suggests slower-than-usual aging. Because patients returned for radiation at different times—ranging from just days to about four years apart—the team grouped them into short, medium, and long intervals and set different cutoffs that best separated people with better or worse outcomes in each group.

What the numbers revealed

The study analyzed 2,276 adults treated with radiation therapy for various cancers, most with advanced or metastatic disease. Across all time-interval groups, patients whose faces aged faster between photos were consistently more likely to die sooner than those whose faces aged more slowly. This pattern held even after accounting for usual factors such as sex, race, cancer type, and the time between photographs. In other words, the speed of facial aging carried independent information about survival that regular clinical details did not fully capture. When the researchers combined FAR with how old patients looked at their first photo, they found that those who both looked older than their age at the start and then aged quickly over time had the highest risk of death.

Why the pace of aging beats a single snapshot

By visualizing the interplay between initial appearance and later aging rate, the researchers showed that the pace of change becomes especially important over longer intervals. In patients followed for one to four years, differences in FAR overshadowed differences in how old they looked at the beginning. This suggests that rapidly worsening facial appearance may be a visible sign of accelerating biological wear and tear—processes such as cell damage, reduced tissue repair, and the impact of cancer treatments. In technical tests that compare how well different measures predict outcomes, FAR outperformed single time-point face-based age measures, particularly when there was more time between photos.

From camera to clinic

Because the method relies on standard ID photographs already taken during routine care, FAR could, in principle, be added to existing risk tools without extra scans or blood tests. For patients with advanced cancer, a high FAR might flag those who need closer monitoring, gentler treatment plans, or more emphasis on comfort and quality of life. Beyond oncology, repeated facial photos could allow near real-time tracking of an individual’s aging pathway, potentially revealing when lifestyle changes or therapies slow or speed biological aging. At the same time, the authors stress important caveats: their study was mainly in older, mostly White patients at a single hospital; photo quality varied; and the system’s fairness across different racial and age groups must be rigorously tested. They also highlight privacy and ethical concerns around using facial images in medicine, and the need for transparent, well-governed AI tools.

What this means for patients

This work suggests that how quickly a person’s face appears to age during cancer treatment reflects much more than surface changes. A faster facial aging rate tracks with poorer chances of survival, while a slower rate signals relatively better outlooks, independent of calendar age. In simple terms, the camera may be capturing the body’s internal struggle with disease and treatment in ways doctors can quantify. If confirmed in future, more diverse, and prospective studies, FAR could become a low-cost, non-invasive gauge to help tailor cancer care and, more broadly, to monitor how our bodies are truly aging beneath the skin.

Citation: Haugg, F., Lee, G., He, J. et al. Face aging rate quantifies change in biological age to predict cancer outcomes. Nat Commun 17, 3487 (2026). https://doi.org/10.1038/s41467-025-66758-w

Keywords: facial aging, biological age, cancer prognosis, artificial intelligence in medicine, radiation therapy