Developed by U.S. Harvard Medical School researchers
AI trained on 40 million facial photos

Can facial photos reveal the survival prospects of cancer patients? Researchers overseas have made headlines by developing an artificial intelligence (AI) system to predict cancer patients' survival rates.


According to the British newspaper The Guardian on October 19 local time, researchers at MGB, a group of hospitals affiliated with Harvard Medical School in the United States, developed an AI model that estimates biological age from facial photos and calculates cancer patients' chances of survival.


An artificial intelligence (AI) system has been developed to predict cancer patients' chances of survival using only facial photos taken with smartphones and other devices. Pixabay

An artificial intelligence (AI) system has been developed to predict cancer patients' chances of survival using only facial photos taken with smartphones and other devices. Pixabay

View original image

The researchers' findings were also published in the prestigious international journal The Lancet. The AI model they developed is called FaceAge. It was trained on 40 million facial photos of people of various ethnicities and age groups. Using these photos, the AI estimates cancer patients' age from their faces and calculates their risk of death.


According to data from actual experiments, among cancer patients who received radiation therapy in the United States, those whom FaceAge judged to look at least five years older than their actual age had an approximately 21% higher risk of death than the other patients over an average follow-up period of four years. By contrast, patients who appeared at least five years younger than their actual age had a 17% lower risk of death.


Hugo Aerts, a co-author of the study, told the newspaper, "FaceAge takes facial health recognition AI a step further by directly training it on photos of cancer patients' faces." He added, "This demonstrates that an AI model can predict cancer survival from a simple photo."


Raymond Mak, a radiation oncologist at MGB, also stressed, "The results were consistent despite everyday variations in posture, lighting, facial expressions, and makeup in the photos." He added, "The AI's performance remained stable even after plastic surgery, with only a slight change after nose surgery."



AI Predicts Cancer Survival Chances from Facial Photos... U.S. System Makes Headlines View original image

However, further research is reportedly needed before the tool can be introduced into actual medical services. MGB has filed a U.S. patent application for the AI.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing