A groundbreaking study published in The Lancet Digital Health reveals that artificial intelligence (AI) can predict cancer survival outcomes by analyzing a simple selfie. Developed by researchers at Mass General Brigham, the AI tool, named FaceAge, estimates a patient’s biological age from facial features, providing insights into their overall health and potential response to cancer treatment.
How FaceAge Works
FaceAge utilizes a deep learning algorithm trained on nearly 59,000 photographs of healthy individuals to assess biological age based on facial characteristics. The tool was subsequently tested on over 6,000 cancer patients, using images taken at the onset of their radiotherapy treatment. Findings indicated that patients whose FaceAge appeared older than their chronological age had poorer survival rates, particularly among those perceived as older than 85, regardless of cancer type or gender . In a comparative study, clinicians were asked to predict six-month survival outcomes for patients undergoing palliative radiotherapy. Without FaceAge, their accuracy stood at 61%. However, when supplemented with FaceAge analysis, accuracy improved to 80%, demonstrating the tool’s potential to augment clinical assessments.
Broader Implications and Future Research
The concept of “perceived ageing,” or how old a person appears, has emerged as a significant biomarker for disease prognosis. FaceAge offers an objective method to quantify this, potentially aiding in personalized treatment planning. Researchers are exploring the tool’s applicability beyond cancer, assessing its effectiveness in predicting general health status, disease risk, and lifespan.
Ethical Considerations
While FaceAge presents promising advancements, concerns about data bias and the need for transparency in AI decision-making have been raised. Experts emphasize the importance of understanding which facial features the AI prioritizes to ensure accurate and unbiased predictions . This study underscores the potential of integrating AI-based age estimation into clinical decision-making, paving the way for more individualized patient care.
