FaceAge: Decoding biological age from face photographs using deep learning
Because humans age at different rates, a person’s physical appearance may yield insights into their biological age and physiological health more reliably than their chronological age. In medicine, however, appearance is incorporated into medical judgments in a subjective and non-standardized fashion. We developed FaceAge, a deep learning system to estimate biological age from face photographs. FaceAge was trained on data from 58,851 healthy individuals, and clinical utility was evaluated on data from 6,196 patients with cancer diagnoses from two trans-Atlantic institutions. We found that, on average, cancer patients look older than their chronological age, and looking older is correlated with worse overall survival. FaceAge demonstrated significant independent prognostic performance in a range of cancer types and stages. We found that FaceAge can improve physicians’ survival predictions in incurable patients receiving palliative treatments, highlighting the clinical utility of the algorithm to support end-of-life decision-making. FaceAge was also found to be significantly associated with molecular mechanisms of senescence through gene analysis, while age was not. Our results demonstrate that deep learning can provide a means to estimate biological age from easily obtainable and low-cost face photographs, improving prognostication across a spectrum of cancer diagnoses. These findings may extend to diseases beyond cancer, motivating using deep learning algorithms to translate a patient’s visual appearance into objective, quantitative, and clinically useful measures.
Publication
Bontempi, D et al. Decoding biological age from face photographs using deep learning. Manuscript under review.
Data availability
All of the datasets used to develop the FaceAge model are publicly available and can be downloaded directly from the respective project web pages.
No clinical datasets used for validation can be shared owing to institutional REB protocols and privacy concerns regarding face photographs of patients.
Code availability
Our GitHub repository includes the code for replicating the technical validation of the FaceAge pipeline, as well as all the code for reproducing our statistical analyses and clinical. In addition to sharing reproducible code, we provide trained model weights, example outputs, and extensive documentation.
AIM Investigators
Acknowledgements
The authors acknowledge financial support from NIH (HA: NIH-USA U24CA194354, NIH-USA U01CA190234, NIH-USA U01CA209414, NIH-USA R35CA22052, and NIH-USA U54CA27451; BHK: NIH-USA K08DE030216-01), and the European Union - European Research Council (HA: 866504).