Innovative AI Using Selfies to Estimate Biological Age and Predict Cancer Outcomes

A groundbreaking AI technology is now enabling the estimation of biological age using just a selfie, potentially transforming how healthcare professionals assess patient health and disease prognosis. FaceAge, a deep learning algorithm introduced in The Lancet Digital Health, analyzes facial images to generate a biological age score that reflects a person’s true physiological aging process, which may differ from their chronological age.

Developed and trained on over 58,000 photographs of healthy adults aged 60 and above, FaceAge has demonstrated promising results in clinical settings. When applied to images of cancer patients in the U.S. and Netherlands, the model found that these individuals appeared nearly five years biologically older than their actual age on average. Importantly, higher FaceAge scores were linked to poorer survival rates, even after adjusting for traditional factors such as age, sex, and tumor type. This suggests the potential of the tool to help guide treatment decisions—identifying patients who may tolerate aggressive therapies and those who might need gentler approaches.

The technology could prove invaluable in various medical decisions—from cancer treatment to surgery and end-of-life care—by offering a more nuanced understanding of a patient’s health status. For example, two patients with similar chronological ages might have significantly different biological ages, influencing the course and intensity of treatment.

FaceAge’s ability to assess aging at an individual level stems from its training on nearly 59,000 portraits of presumed healthy older adults. When tested on cancer patients, it accurately predicted biological aging markers, aligning with clinical outcomes. Interestingly, the model weighs subtle facial features like muscle tone more heavily than obvious signs such as gray hair or baldness, providing a more detailed picture of aging.

Despite its promising potential, AI tools like FaceAge raise important ethical considerations. Early evaluations suggest minimal racial bias, but further testing is underway to ensure fairness across diverse populations. The developers are also investigating how elements like makeup, cosmetic surgery, and lighting might influence results.

The researchers plan to launch a public platform where individuals can upload selfies to participate in ongoing studies. While commercial versions for clinicians are anticipated, these will require extensive validation before widespread use. Overall, FaceAge represents a significant step forward in personalized medicine, offering a non-invasive, cost-effective method to better understand and predict health outcomes based on facial analysis.

Source: https://medicalxpress.com/news/2025-05-ai-tool-selfies-biological-age.html