New AI tool can predict cancer survival by a single patient selfie (www.theguardian.com)

Researchers at Mass General Brigham have unveiled an AI tool capable of predicting cancer survival rates using just a single facial photograph. This innovative system, detailed in the Lancet Digital Health journal, leverages an expansive dataset of over 40 million images to assess biological age and estimate survival outcomes for cancer patients. The tool employs two methods: FAHR-FaceAge, which predicts biological age and correlates it with survival chances, and FAHR-FaceSurvival, which directly estimates survival risk based on facial features. This dual approach allows for a nuanced understanding of patient health, making it a significant advancement in facial health recognition. This breakthrough holds considerable implications for the AI/ML community and healthcare at large. By demonstrating a foundation-model approach, which allows for scalable development of facial health biomarkers, the research paves the way for AI-driven assessments that could inform treatment decisions for aging patients requiring intensive therapies. While encouragingly effective across various conditions, researchers note that further studies are essential before the tool can be deployed clinically. The findings suggest a strong link between perceived biological age and survival, underscoring the potential for AI to enhance personalized medicine in oncology.