Harvard researchers announced on May 21 that artificial intelligence can help predict cancer outcomes by analyzing how old patients appear in digital photographs. The research, published in two studies, found that both looking younger than one’s chronological age and showing slower facial aging during treatment are linked to better survival rates for cancer patients.
The findings highlight the potential of medical artificial intelligence tools like FaceAge, which uses simple photos to estimate biological age and inform care decisions. Researchers say this approach could improve screening and guide doctors in tailoring treatments based on a patient’s biological rather than chronological age.
Raymond Mak, associate professor at Harvard Medical School and co-senior author of the studies, said, “One of the first numbers they put in is your chronological age. It’s done by every single primary care doctor, the same with a pre-op evaluation, the same with a lot of our risk calculators and cancer care. What we’re arguing is why use chronological age when we’re seeing these massive deflections between biological age and chronological age? Why not use something that might be more precise for an individual?”
The first study showed that face age is older than actual age for 65 percent of over 24,000 cancer patients examined. Those who looked five years or more younger had significantly better outcomes; those who appeared ten years or more older fared worse. The second study introduced an updated FaceAge algorithm trained on millions of images worldwide. Hugo Aerts, professor at Dana-Farber Cancer Institute and director at Mass General Brigham’s Artificial Intelligence in Medicine Program as well as co-senior author said: “Those first two layers, the 40 million and the 700,000 [images], are a potential resource… If you decide to go in a different direction, you don’t have to reinvent the wheel… Now do this other thing.”
Researchers noted ongoing efforts to refine FaceAge’s accuracy across diverse skin types or cosmetic alterations. They also suggested future work could develop similar tools for estimating ‘heart age’ or ‘liver age.’ Interest from other specialties has led to clinical trials using an online portal where people can upload their own images for assessment.
Aerts concluded: “CT and MRI will generate much, much more information. But you cannot take an MRI every day of every individual in the world… The beauty of this is you can get rougher but more frequent health assessments using a very simple picture.”
