Studies have shown that artificial intelligence (AI) analyzes the face of cancer patients to predict the opportunity of survival, and in some cases, more accurate results are produced than short -term survival predictions of clinicians.
The Financial Times reported on August 8 that a researcher on the Massachusetts Hospital Network Mass Network Mass Brigham announced the AI tool ‘FACEAGE’, which analyzes cancer patients’ facial photos and predicts the opportunity of survival.
In keeping with the study, the biological age of the patient through the face ace showed that the face of the cancer patients was about five years older than the actual age.
This technology is certainly one of the biomarker’s research to predict the chance of disease by utilizing the degree of aging of the body, and the event of AI technology has made it possible to learn vast health data and quantitatively predict risks.
Dr. Hugo Aertz, co -director of the study, said, “One selfie accommodates essential information that might help to determine clinical decisions and treatment plans.”
The researchers first collected 58,8851 photos of healthy people from public data, learned Face Age, and tested algorithms with 6196 cancer patients taken at first of radiation therapy.
Consequently, the upper the face age among the many cancer patients, the lower the survival rate. This was maintained significantly after correcting aspects similar to actual age, gender, and cancer type. Specifically, the survival results of patients who were estimated to be over 85 years old were the poorest.
The researchers also provided 10 clinical and researchers with only photos of terminal cancer patients and predicted whether or not they survived inside six months. The predictive accuracy of only easy photographs was 61%, however the accuracy was improved to 80%when the face age evaluation information was provided together.
After all, there are limitations in face age. The researchers also mentioned the chance of distorting the outcomes because of model errors, moderately than the opportunity of bias and actual biological differences in the info learned by the algorithm.
Currently, the researchers are reviewing the Face Age to numerous patients’ groups to make use of them for general health, disease predictions, and life expectancy evaluation.
Meanwhile, biomarker studies related to body aging are actively underway. In February, a blood test method was published to detect the chance of greater than 30 diseases, including long -term aging and detect the chance of 30 diseases, including lung cancer.
Interest in health prediction technology using ‘visible age’ can also be increasing. Amongst them, studies have continued that ‘perceived ages’ judged by expert eyes can predict actual mortality and various aging diseases. Nevertheless, there’s a limit that this human commentary -based approach takes money and time.
AI expert Jaume Bacardit, a professor of Recent Castle University, said, “The evaluation was quite thorough,” but added that it’s needed to elucidate more about what AI judges based on the face.
By Park Chan, reporter cpark@aitimes.com
