AI tool uses face photos to predict health and cancer outcomes
Key Takeaways:
- FaceAge, an AI tool developed at Mass General Brigham, uses facial photographs to predict biological age and survival outcomes in people with cancer.
- The algorithm outperformed clinicians in predicting short-term life expectancy in individuals undergoing palliative radiotherapy.
- FaceAge may become a novel, non-invasive biomarker to support personalised care decisions across cancer and potentially broader healthcare contexts.
Introduction: A New Frontier in Biomarkers
A team of researchers at Mass General Brigham has developed a deep learning tool known as FaceAge that estimates a person’s biological age and predicts survival outcomes using only a photograph of their face. The innovation demonstrates that even a simple facial image—such as a selfie—can reveal clinically significant insights about an individual’s health status and future prognosis, particularly in the context of cancer care.
The findings, published in The Lancet Digital Health, reveal that cancer patients generally appear biologically older than their actual chronological age, and that this discrepancy is strongly associated with poorer clinical outcomes.
“We can use artificial intelligence (AI) to estimate a person’s biological age from face pictures, and our study shows that information can be clinically meaningful,” said Dr Hugo Aerts, PhD, co-senior author and director of the Artificial Intelligence in Medicine (AIM) programme at Mass General Brigham.
The Link Between Appearance and Survival
The study found that individuals with cancer had an average FaceAge approximately five years older than their chronological age. Those whose predicted biological age appeared particularly elevated—especially beyond the age of 85—tended to have worse survival outcomes, even after adjusting for known factors such as sex, cancer type, and actual age.
“How old someone looks compared to their chronological age really matters—individuals with FaceAges that are younger than their chronological ages do significantly better after cancer therapy,” added Dr Aerts.
This observation supports longstanding clinical intuition: a patient’s appearance can often offer subtle indicators of their underlying health. However, human assessment is fallible and subject to unconscious bias. FaceAge provides a more objective, scalable, and potentially more accurate measure of biological ageing.
Testing the Algorithm: Cancer and Radiotherapy Cohorts
To build the FaceAge tool, researchers trained the algorithm on 58,851 images of presumed healthy individuals drawn from public datasets. They then tested the algorithm on 6,196 patients with cancer at two medical centres, using routine photographs taken at the beginning of radiotherapy treatment.
These evaluations showed a consistent pattern: cancer patients not only had higher FaceAges, but those with greater discrepancies between biological and chronological age had markedly poorer survival rates.
Outperforming Clinicians in Prognostic Judgement
One of the study’s most compelling insights emerged from a sub-study involving 100 patients receiving palliative radiotherapy. The researchers invited 10 clinicians and researchers to estimate short-term life expectancy using facial photographs and clinical context, including chronological age and cancer status.
Despite their experience, clinicians’ predictions were only marginally better than chance. However, when provided with the FaceAge data, their accuracy improved significantly.
“This work demonstrates that a photo like a simple selfie contains important information that could help to inform clinical decision-making and care plans for patients and clinicians,” said Dr Aerts.
A Glimpse Into the Future of AI-Driven Health Tools
While the results are promising, the research team cautions that FaceAge is not yet ready for clinical deployment. Ongoing and future research aims to:
- Validate FaceAge across different hospitals and populations
- Assess its utility in various stages of cancer
- Monitor how FaceAge changes over time
- Investigate its robustness against potential confounders, such as cosmetic procedures
“This opens the door to a whole new realm of biomarker discovery from photographs, and its potential goes far beyond cancer care or predicting age,” said Dr Ray Mak, co-senior author and faculty member in the AIM programme.
“As we increasingly think of different chronic diseases as diseases of ageing, it becomes even more important to be able to accurately predict an individual’s ageing trajectory. I hope we can ultimately use this technology as an early detection system in a variety of applications, within a strong regulatory and ethical framework, to help save lives.”
Contributors, Conflicts of Interest, and Funding
Study Authors:
The research was conducted by a multidisciplinary team from Mass General Brigham including Dennis Bontempi, Osbert Zalay, Danielle S. Bitterman, Fridolin Haugg, Jack M. Qian, Hannah Roberts, Subha Perni, Vasco Prudente, Suraj Pai, Christian Guthier, Tracy Balboni, Laura Warren, Monica Krishan, and Benjamin H. Kann.
Intellectual Property:
Mass General Brigham has filed provisional patents on two next-generation facial health algorithms developed in conjunction with this research.
Funding:
This study received funding from the following sources:
- National Institutes of Health (NIH-USA grants: U24CA194354, U01CA190234, U01CA209414, R35CA22052, and K08DE030216-01)
- European Union – European Research Council (Grant 866504)
Conclusion
FaceAge marks a significant advance in the intersection of AI and personalised medicine. By turning a simple image into a clinically valuable biomarker, this technology has the potential to support decision-making, improve prognostic accuracy, and ultimately enhance outcomes for people living with cancer and other ageing-related conditions.
