AI tool to predict biological age and cancer outcomes using selfies

Scientists have developed an AI tool called FaceAge that analyzes selfies to estimate biological age, predict cancer survival rates, and assist in healthcare decisions. The tool has demonstrated remarkable accuracy, outperforming clinicians in some cases, and could revolutionize health diagnostics.

Published in The Lancet Digital Health, the study involved training the AI on nearly 59,000 facial images to estimate biological age—an indicator of cellular health distinct from chronological age.

FaceAge was tested on 6,200 cancer patients, revealing that on average, they appeared biologically five years older than their chronological age. This data suggested a link between accelerated aging and cancer. Notably, the AI was also tested on 100 palliative care patients to predict life expectancy. Its predictions were more accurate than those of 10 experienced clinicians.

Dr. Ray Mak, one of the study’s authors, emphasized the tool’s potential as an early detection system for health issues, particularly as aging-related diseases become more prevalent. 

However, the tool’s accuracy may vary due to factors like limited training diversity and external influences such as makeup and lighting. Researchers are working to refine the model by expanding datasets and conducting further testing.

Though its clinical implementation may be years away, FaceAge is opening new avenues in health diagnostics and biomarker discovery, promising advancements in early detection and personalized care.

Copied