AI-generated from publicly available materials.Researchers have unveiled a groundbreaking AI model capable of predicting cancer survival risk using just facial photographs. Developed by a team from Mass General Brigham, this model employs two algorithms: one estimating biological age and the other directly forecasting survival outcomes.
The first algorithm, known as FAHR-FaceAge, demonstrated an impressive accuracy in identifying biological age, with a margin of error around five years. This accuracy surpasses that of existing tools, and notably, the difference between a patient's biological and chronological age was linked to their survival chances. The second algorithm, FAHR-FaceSurvival, takes a more innovative approach by predicting survival risk from a single facial image, marking a significant advancement in the field.
Trained on over 40 million facial images from diverse ethnic backgrounds, the algorithms utilize common facial features to make their predictions. Unlike traditional models that require specialized imaging, this AI system relies on consumer-grade photographs, enhancing its potential for widespread clinical application. The researchers emphasize the importance of such tools as the aging population necessitates better methods for assessing physical health before deciding on intensive treatments.
This research not only aims to refine health assessments through facial analysis but also invites public participation via the FaceAge Research Study. While experts acknowledge the study's promise, they caution that further validation is needed to ensure these AI predictions offer clinically valuable insights beyond traditional medical assessments.