Researchers from Mass General Brigham have developed a groundbreaking deep learning algorithm called “FaceAge,” which can estimate a person’s biological age simply by analyzing a photograph of their face.
This innovative tool has shown promise not only in predicting biological age but also in identifying survival outcomes for patients with cancer, potentially offering new insights for clinical care.
The researchers found that cancer patients tend to have a higher biological age as predicted by FaceAge compared to their actual chronological age. On average, cancer patients appeared about five years older than their real age when analyzed by the AI system.
This biological age gap, measured through facial features, was linked to poorer survival rates across various types of cancer. The study, which was published in The Lancet Digital Health, suggests that FaceAge could be more reliable than traditional clinical assessments in predicting short-term survival for patients undergoing palliative radiotherapy.
According to Hugo Aerts, Ph.D., director of the Artificial Intelligence in Medicine (AIM) program at Mass General Brigham and co-senior author of the study, the algorithm could revolutionize how doctors evaluate patients.
He explained that a simple photo could reveal critical health information that doctors often estimate based on intuition and visual cues. He noted that patients who look younger than their actual age tend to recover better after cancer treatments, while those who appear older have worse outcomes.
When doctors see patients, they often make quick judgments about their overall health based on appearance, combined with their age and other medical information. But these visual assessments can sometimes be biased.
For example, a person who looks older might be assumed to be sicker, even if their medical records don’t suggest it. This is where FaceAge could provide a more objective and accurate measure, reducing bias in medical decision-making.
To develop FaceAge, the researchers used deep learning and facial recognition technology. They trained the tool on 58,851 photos of healthy individuals from public datasets. The system learned to identify facial characteristics associated with aging and health status.
After training, FaceAge was tested on 6,196 cancer patients from two different medical centers. These patients had their photos taken before starting radiotherapy, providing a perfect opportunity to compare biological age predictions with real health outcomes.
The results showed that cancer patients, on average, had FaceAge predictions about five years older than their actual age. Those who appeared much older than their chronological age were more likely to have poor survival rates.
This difference remained significant even after adjusting for factors like age, gender, and cancer type. Patients over the age of 85, according to FaceAge, had particularly poor survival prospects, highlighting the algorithm’s potential in identifying those at the highest risk.
In addition to predicting survival, the researchers tested FaceAge’s effectiveness in clinical settings. They asked ten clinicians and researchers to estimate the short-term life expectancy of 100 cancer patients who were receiving palliative radiotherapy. Even with full medical histories and age information, the clinicians were only slightly better than random guesses.
However, when they were given FaceAge estimates, their predictions improved considerably. This suggests that FaceAge provides valuable information that doctors might not always pick up on just by looking at patients or reviewing their medical records.
The research team believes this technology could change the way doctors think about age and health. Traditionally, age has been viewed as a number, but FaceAge shows that biological aging—how old someone looks—might be more important for understanding their health risks.
Ray Mak, MD, co-senior author of the study, highlighted that this technology could open doors to new ways of detecting chronic diseases early, not just cancer. He envisions using FaceAge to track aging over time, potentially catching signs of health problems before they become serious.
Although the results are promising, the researchers caution that more studies are needed before FaceAge can be widely used in hospitals. They plan to expand their research to different medical centers, test FaceAge on patients at various cancer stages, and evaluate how it performs against data from people who have had plastic surgery or use makeup that can alter facial features.
The team is also exploring the potential of FaceAge beyond cancer care, hoping to predict not just age but also the risk of other chronic conditions. Their goal is to use this facial analysis technology to better understand a person’s health trajectory, helping doctors make more informed treatment decisions.
This innovative research signals a shift towards more personalized medicine, where a simple photo could one day become a powerful diagnostic tool, helping doctors predict health outcomes with greater accuracy and reduce biases in patient care.
If FaceAge continues to prove effective, it could become a standard part of medical evaluations, giving doctors a new lens through which to view patient health and longevity.
If you care about cancer, please read studies that low-carb diet could increase overall cancer risk, and new way to increase the longevity of cancer survivors.
For more health information, please see recent studies about how to fight cancer with these anti-cancer superfoods, and results showing daily vitamin D3 supplementation may reduce cancer death risk.
The research findings can be found in The Lancet Digital Health.
Copyright © 2025 Knowridge Science Report. All rights reserved.
