Single photo of the face used to predict biological age and mortality risk

Mass General Brigham/Lancet Digital Health

By Olivia Bowthorpe

A single photograph of the face may be able to identify features linked to biological ageing and mortality risk among patients with cancer, a study suggests.

Researchers found the facial analysis AI tool may add useful information to risk assessments when planning cancer treatment.

The tool was trained on more than 40 million facial photographs, then fine-tuned to separately estimate age and survival. The models were evaluated in a US cohort of more than 38,000 radiotherapy patients and validated on nearly 5,000 patients with cancer in the Netherlands.

This approach is “capable of transforming subtle facial feature variations into quantifiable imaging biomarkers," reported the authors in The Lancet Digital Health.

The image above shows yellow areas of the face which are most important for determining biological age. Other areas of the face are used for predicting mortality risk.

Hugo Aerts, professor at Harvard University and director of the AI in Medicine Program at Mass General Brigham in Massachusetts, and colleagues wrote: "Influenced by various physiological factors, disease processes, and lifestyle choices, the human face might provide valuable insights into an individual’s health status.