AI Predicts Cancer Survival Using Facial Photos
A groundbreaking deep learning system named FaceAge has been developed and validated to estimate a person's biological age using simple photographs of their faces. A recent study, featured in The Lancet Digital Health, highlights FaceAge's remarkable ability to predict short-term outcomes for patients diagnosed with cancer.
Understanding FaceAge: AI That Reads Your Biological Clock
The study demonstrated that FaceAge could effectively determine biological age and compare it with a patient's chronological age. A significant observation was that the biological age in patients with cancer tended to be more advanced than in those without the disease.
The Link Between Biological Age and Cancer Survival
Furthermore, the model revealed a critical connection: when a cancer patient's FaceAge (the biological age estimated by the AI) was older than their actual chronological age, it was associated with worse survival outcomes. This finding held true even after researchers adjusted for factors such as the specific type of cancer, the patient's sex, and their chronological age. The association was particularly strong for patients whose FaceAge suggested they looked older than 85.
Clinical Significance: How FaceAge Could Reshape Patient Care
"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," stated co-senior and corresponding author Dr. Hugo Aerts, PhD, Director of the Artificial Intelligence in Medicine (AIM) program at Mass General Brigham. "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. 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."
Inside the Research: Training FaceAge and Its Striking Results
Investigators trained FaceAge on an extensive dataset from 58,851 presumed-healthy individuals aged 60 or older. This included 56,304 individuals from the IMDb-Wiki data set for training and 2,547 from the UTKFace data set for validation. Additionally, the AI tool was trained with data from 6,196 patients with cancer from the Netherlands and the United States who were undergoing radiotherapy. This group was compared with 535 patients without cancer.
The study's results showed that an older biological age than the patient’s actual chronological age correlated with worse overall survival. Specifically, this was indicated by a hazard ratio (HR) of 1.151 (P = .013) for a pan-cancer cohort and an HR of 1.117 (P = .021) in a palliative cohort. Most patients with cancer also presented with a higher biological age than their chronological age, with a mean increase of 4.79 years (P < .0001). Among patients with incurable cancer receiving palliative radiotherapy, FaceAge demonstrated the ability to improve physicians’ survival predictions. The model increased the Area Under the Curve (AUC), a measure of predictive accuracy, from 0.74 to 0.8 (P < .0001). The study authors suggest this improvement may significantly help with end-of-life decision-making processes.
The Future is Visual: Expanding AI's Role in Healthcare
"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 co-senior author Dr. Ray Mak, MD, a faculty member in the AIM program at Mass General Brigham. "As we increasingly think of different chronic diseases as diseases of aging, it becomes even more important to be able to accurately predict an individual’s aging 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."
Next Steps and the Promise of Visual Biomarkers
"Subject to further testing and validation, approaches such as FaceAge could be used to translate a patient’s visual appearance into objective, quantitative, and clinically valuable measures," the study authors concluded.
Disclosure: For full disclosures of the study authors, please visit thelancet.com.
The content in this post is based on a study published in The Lancet Digital Health and has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®). It does not necessarily reflect the ideas and opinions of ASCO®.
