The era of AI. Artificial intelligence is constantly showing its 'magic'. Now the doctors said, they have created such an AI tool, which can tell the probability of survival of a cancer patient just by looking at the selfie taken by him. And that's why this AI system has been trained using more than 40 million facial images.
A report published in 'Lancet Digital Health' reported this new research. This AI system was specially developed by Mass General Brigham, a non-profit academic health services organization in Massachusetts. This tool is able to determine the biological age of the patient and directly predict how long they can live in patients diagnosed with cancer.
The study is one of the most comprehensive examples of how AI analysis of selfies can assess a patient's health and guide their medical management. The AI system was trained using more than 40 million facial images of people of different demographics and ages. As reported, two complementary tools are used for this. In this 'FAHR-FaceAge' determines the biological age of the patient. It also predicts the chances of surviving cancer. On the other hand, 'FAHR-FaceSurvival' estimates survival probability directly from face images. And in this case there is no need to rely on biological age.
How exactly does AI work? Each algorithm observes different parts of the face to predict age and risk of death. 'FAHR-FaceAge' mainly focuses on the folds that extend from the sides of the nose to the corners of the mouth, the forehead just above the eye sockets and the areas next to the ears. On the other hand, 'FAHR-FaceSurvival' analyzes the area under the eyes and around the bridge of the nose. Then predictions are made from that analysis.
Among cancer patients in the US undergoing or having completed radiotherapy, those assessed by FAHR-FaceAge as looking five years or older than their actual age had a 21 percent higher risk of death during an average of four years of follow-up. On the other hand, the risk was 17 percent lower for those who appeared to be five years or less younger than their actual age.