Could AI Use a Simple Photo of Your Face to Reveal Clues About Your Health?

Researchers at Mass General Brigham in Somerville, Massachusetts, have unveiled an innovative artificial intelligence (AI) model capable of assessing an individual’s health based solely on a photograph of their face. This groundbreaking technology, named Foundation Artificial Intelligence for Health Recognition – Face (FAHR-Face), leverages a vast database of over 40 million facial images to inform its predictions.

The team behind the research developed two specific applications of this model: FAHR-FaceAge, which estimates an individual’s biological age, and FAHR-FaceSurvival, designed to predict survival outcomes for cancer patients. A study detailing these findings was published on 19 October in the journal *Lancet Digital Health*, showcasing the potential implications of this technology in the medical field.

One of the most compelling aspects of FAHR-Face is its correlation between the estimated FAHR-FaceAge and a patient’s actual age, revealing critical links to their survival prognosis. The FaceSurvival tool focuses particularly on the areas beneath the eyes and the bridge of the nose when making its assessments, demonstrating significant predictive capabilities across various cancer types and stages, including cases where the disease has metastasised to the brain or bones.

Participants in the study included cancer patients from both the United States and the Netherlands, ranging from their 50s and 60s to children and some nonagenarians. Dr. Hugo Aerts, co-senior author of the study and director of the Artificial Intelligence in Medicine (AIM) programme at Mass General Brigham, emphasised the model’s unique ability to learn and predict survival risks directly from images of patients. “This is the first demonstration that an AI model can directly predict cancer survival from a simple photograph,” he stated.

Further analysis revealed that the two tools capture complementary information about patients’ health, enhancing the accuracy of predictions when used together. Notably, the FaceAge tool identified specific lifestyle factors that correlate with facial ageing. For example, current smokers appeared older than those who had quit, and the latter looked older than individuals who had never smoked. Heavy alcohol consumption and drug use were also associated with an increase in perceived age.

Dr. Raymond Mak, a radiation oncologist and corresponding author of the study, highlighted the surprising consistency of FAHR-FaceAge across various common variables, including different poses, lighting conditions, facial expressions, and the application of makeup. He noted that its reliability extended to the majority of tested cosmetic procedures, with only minor changes observed post-rhinoplasty.

While the model has shown promising results, further research and validation are essential before it can be implemented in clinical settings. The researchers are optimistic that this technology could ultimately assist doctors in making informed treatment recommendations for their patients. “Our goal is to make face photograph-based health assessment more quantitative, consistent, and scalable,” Dr. Mak remarked, stressing the intention to enhance—rather than replace—physician judgement in healthcare.

To facilitate ongoing research, the team has launched the FaceAge Research Study, which invites members of the public to participate voluntarily. Participants can upload their own photographs to receive insights about their facial features, thereby contributing additional data to this pioneering research effort.

As the landscape of medical technology continues to evolve, the development of tools like FAHR-Face represents a significant step towards more personalised and accessible healthcare assessments. By transforming facial photographs into valuable health insights, researchers hope to support medical professionals in delivering tailored treatments that consider not only clinical data but also the visual cues provided by patients’ appearances.