AI Tool Uses Face Photos to Estimate Biological Age—and Predict Cancer Survival
A simple selfie could soon provide critical insights into a cancer patient’s prognosis, thanks to an artificial intelligence (AI) tool developed by researchers at Mass General Brigham. The tool, called FaceAge, analyzes facial characteristics from photographs to estimate a person’s biological age—a measure of how quickly the body is aging at a cellular level—and links it to survival outcomes in cancer patients. According to a study published in The Lancet Digital Health, patients with cancer whose FaceAge predicted them to be biologically older than their chronological age faced significantly worse survival rates across multiple cancer types.
The findings suggest that facial features—long considered a window into overall health—may hold objective data that could help clinicians tailor treatment plans, particularly for patients with advanced disease. “You can use artificial intelligence to estimate a person’s biological age from face pictures, and our study shows that information can be clinically meaningful,” said Hugo Aerts, PhD, director of the Artificial Intelligence in Medicine (AIM) program at Mass General Brigham and co-senior author of the study. “This work demonstrates that a photo like a simple selfie contains significant information that could help to inform clinical decision-making and care plans for patients, and clinicians.”
For patients and families grappling with a cancer diagnosis, the implications are profound. Biological age, unlike chronological age, reflects the cumulative impact of genetics, lifestyle, and disease on the body. A 50-year-old with a biological age of 60, for example, may have a weaker immune system, reduced organ function, and a higher risk of complications from aggressive treatments. The FaceAge tool could help oncologists identify patients who may benefit from modified therapies, closer monitoring, or palliative care earlier in their treatment journey.
How FaceAge Works—and What It Reveals About Cancer Survival
The FaceAge algorithm was trained on thousands of facial photographs paired with clinical data from patients with and without cancer. Using deep learning techniques, the AI identifies subtle patterns in facial features—such as skin texture, wrinkles, and facial structure—that correlate with biological aging. In the study, researchers found that patients with cancer, on average, had a FaceAge that was five years older than their actual age. More critically, a higher FaceAge was associated with poorer overall survival across multiple cancer types, including lung, breast, and colorectal cancers.
One of the most striking findings was FaceAge’s ability to outperform clinicians in predicting short-term life expectancies for patients receiving palliative radiotherapy. While doctors often rely on a combination of lab results, imaging, and clinical judgment to estimate prognosis, the AI tool provided a more objective and consistent measure. This could be particularly valuable in settings where access to advanced diagnostic tools is limited, offering a low-cost, non-invasive way to assess a patient’s likely trajectory.
The study also explored whether analyzing multiple photographs taken over time could provide even deeper insights. Preliminary results suggest that tracking changes in FaceAge—such as a rapid increase in biological age—may help clinicians identify patients who are responding poorly to treatment or experiencing accelerated disease progression. While this approach is still under investigation, it highlights the potential for AI to move beyond static snapshots and toward dynamic, personalized health monitoring.
Why Biological Age Matters in Cancer Care
Chronological age—the number of years a person has lived—has long been a factor in cancer treatment decisions. Older patients, for example, may be less likely to tolerate aggressive chemotherapy or surgery due to age-related declines in organ function. However, chronological age is an imperfect proxy for a patient’s true physiological state. Two 70-year-olds can have vastly different health profiles: one may be biologically “younger” due to a healthy lifestyle, while the other may be biologically “older” due to chronic illness, smoking, or poor nutrition.
Biological age, by contrast, is a more nuanced measure that accounts for the cumulative effects of disease, genetics, and environmental factors. It is often estimated using biomarkers such as telomere length, epigenetic changes, or inflammatory markers. However, these methods typically require blood tests or tissue samples, which can be invasive, expensive, or unavailable in resource-limited settings. FaceAge offers a non-invasive, accessible alternative that could be deployed in clinics, hospitals, or even telemedicine platforms with nothing more than a smartphone camera.
Dr. Aerts emphasized that the tool is not intended to replace clinical judgment but to augment it. “This is about providing an additional layer of data that clinicians can use to make more informed decisions,” he said. “It’s not about replacing the human element of medicine but about giving doctors another tool in their toolkit.”
Beyond Cancer: The Broader Potential of FaceAge
While the current study focused on cancer, the researchers believe FaceAge could have applications across a range of chronic diseases. Biological aging is a key driver of conditions such as cardiovascular disease, diabetes, and neurodegenerative disorders. A tool that can quickly and accurately estimate biological age from a photograph could help identify individuals at higher risk of these diseases before symptoms appear, enabling earlier interventions.
The AIM Lab at Brigham and Women’s Hospital, where FaceAge was developed, has already demonstrated the tool’s potential in predicting survival outcomes for patients with other chronic illnesses. On its official website, the lab notes that understanding one’s biological age can help individuals make more informed decisions about their health and lifestyle. For example, a person with a higher-than-expected FaceAge might be motivated to adopt healthier habits, such as quitting smoking, improving their diet, or increasing physical activity, to slow the aging process.
Critics, however, caution that the use of facial analysis in medicine raises ethical and privacy concerns. AI tools that rely on facial data must navigate issues such as bias in training datasets, the potential for misuse (e.g., by insurers or employers), and the need for transparency in how the technology works. The Mass General Brigham team has acknowledged these challenges and is working to ensure that FaceAge is developed and deployed responsibly. “We are committed to addressing these concerns through rigorous validation, diverse datasets, and clear communication about the tool’s limitations,” said Aerts.
What’s Next for FaceAge?
The study published in The Lancet Digital Health represents a significant step forward, but the researchers are already looking ahead. Future work will focus on validating FaceAge in larger, more diverse patient populations to ensure its accuracy across different ethnicities, ages, and geographic regions. The team is also exploring how the tool can be integrated into electronic health records (EHRs) to provide real-time insights for clinicians.
For patients, the most immediate question is whether FaceAge will become a standard part of cancer care. While the tool is not yet widely available, the researchers hope to make it accessible to clinicians within the next few years. In the meantime, the study underscores the growing role of AI in medicine—not as a replacement for doctors, but as a powerful ally in the fight against disease.
“This is just the beginning,” said Aerts. “As AI continues to evolve, we’re going to see more tools like FaceAge that can extract meaningful health insights from data we already have—like a simple photograph. The goal is to make medicine more precise, more personalized, and more effective.”
Key Takeaways
- FaceAge is an AI tool developed by Mass General Brigham that estimates biological age from facial photographs, offering insights into a patient’s overall health and disease trajectory.
- Patients with cancer in the study had a FaceAge that was, on average, five years older than their chronological age, and a higher FaceAge was linked to poorer survival outcomes.
- The tool outperformed clinicians in predicting short-term life expectancies for patients receiving palliative radiotherapy, suggesting it could be a valuable addition to clinical decision-making.
- Biological age, unlike chronological age, reflects the cumulative impact of genetics, lifestyle, and disease, making it a more accurate predictor of health risks and treatment tolerance.
- While the current focus is on cancer, FaceAge could have applications in other chronic diseases, such as cardiovascular disease and diabetes, by identifying individuals at higher risk before symptoms appear.
- Ethical and privacy concerns, including bias in training datasets and potential misuse, must be addressed as the technology is developed and deployed.
What Readers Can Do
For those interested in learning more about biological age and its implications for health, the AIM Lab’s FaceAge website provides additional information about the tool and its development. Patients with cancer or other chronic diseases should discuss any concerns about biological age or treatment options with their healthcare providers. As AI tools like FaceAge become more integrated into medicine, staying informed about their benefits and limitations will be key to making empowered health decisions.
The next major milestone for FaceAge will be its validation in larger, more diverse patient populations. Researchers are also working on integrating the tool into electronic health records to provide real-time insights for clinicians. For updates on the tool’s progress, readers can follow Mass General Brigham’s newsroom or the The Lancet Digital Health journal for future studies.
Have you or a loved one been affected by cancer? How do you perceive about the potential of AI tools like FaceAge to improve treatment outcomes? Share your thoughts in the comments below, and don’t forget to share this article with others who may be interested in the future of personalized medicine.