AI will estimate biological age based on the photo and create a survival prediction for cancer patients

The system was called FAHR-Face - from the English Foundation Artificial Intelligence for Health Recognition. It was trained on the basis of over 40 million facial photos.

Researchers created two separate tools based on it. The first, FAHR-FaceAge, estimates biological age, i.e. age related to the condition of the body, and not only the date of birth. The second one - FAHR-FaceSurvival - is intended to directly assess the risk of death in oncology patients based on a single photograph.

The research results were published in the journal "The Lancet Digital Health".

Both tools analyzed different facial features and provided complementary information. The model assessing biological age focused, among others, on: on the nasolabial folds, forehead and temples. The survival prediction algorithm paid more attention to the area under the eyes and around the bridge of the nose.

Combining the results of both models improved the fit of survival predictions in an external validation cohort of cancer patients.

Support in choosing treatment

According to the authors, the technology may in the future help assess the so-called the patient's physiological reserve and fragility - especially before intensive treatment such as surgery, radiotherapy or systemic therapy.

Our goal is to make facial photo-based health assessment more quantitative, consistent and scalable so that it can support personalized therapeutic decisions rather than replace a physician's judgment, said Dr. Raymond Mak of Mass General Brigham Cancer Institute, quoted by News Medical.

Scientists declare that they checked the stability of the model in typical photo-taking conditions: with different lighting, facial positioning, facial expressions and make-up. According to their data, in most cases the algorithm was also resistant to the effects of plastic surgery procedures. Little change was noted with rhinoplasty.

The analysis also showed a relationship between lifestyle and the biological age of the face. Biologically older appearance was associated with, among others: heavy alcohol consumption, drug use and cigarette smoking.

The strongest association was with smoking. Current smokers were judged by the system to be biologically older than former smokers, and people who had quit smoking were judged to be biologically older than people who had never smoked.

More research needed

The technology is at the research and validation stage. Although the results are promising, they do not mean that a photo of a face can independently determine the health condition or prognosis of a specific patient. The implementation of such systems into clinical practice requires further testing, including: in diverse populations, and address privacy issues and possible algorithmic biases.