Deep learning tool analyzes biometric data from selfies to detect heart disease
New medical research published in the European Heart Journal discusses how four patient biometric “selfies” could be enough for a doctor to detect heart disease, thanks to a deep learning computer algorithm that analyzes facial features to expose coronary artery disease (CAD), writes Science Daily.
The algorithm based on facial recognition techniques is still a prototype and requires further testing on people from a variety of ethnic groups, so that it could ultimately be deployed as a screening method for heart diseases.
The biometric research was led by Professor Zhe Zheng, vice director of the National Center for Cardiovascular Diseases and vice president of Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, People’s Republic of China.
“To our knowledge, this is the first work demonstrating that artificial intelligence can be used to analyze faces to detect heart disease,” said the professor in a prepared statement. “It is a step towards the development of a deep learning-based tool that could be used to assess the risk of heart disease, either in outpatient clinics or by means of patients taking ‘selfies’ to perform their own screening. This could guide further diagnostic testing or a clinical visit.”
He continued: “Our ultimate goal is to develop a self-reported application for high risk communities to assess heart disease risk in advance of visiting a clinic. This could be a cheap, simple and effective of identifying patients who need further investigation. However, the algorithm requires further refinement and external validation in other populations and ethnicities.”
The research cites industry knowledge that specific facial features can be linked to an increased risk of heart disease, such as thinning hair, wrinkles, ear lobe crease, and cholesterol deposits under the skin and around eyelids.
The project involved the blood vessel study of 5,796 patients from eight hospitals in China between July 2017 and March 2019. Digital cameras were used to take four facial photos from different angles. Nurses interviewed patients about their socioeconomic status, lifestyle and medical history, while radiologists analyzed their angiograms. The information collected was then used to train the deep learning algorithm, which was next tested on 1,013 patients from nine hospitals.
The test found that the deep learning algorithm performed better than existing heart disease prediction methods. It accurately detected heart problems in 80 percent of cases, and found that there was no heart disease in 61 percent of cases. The sensitivity was 80 percent and specificity 54 percent.
“The algorithm had a moderate performance, and additional clinical information did not improve its performance, which means it could be used easily to predict potential heart disease based on facial photos alone,” explained Professor Xiang-Yang Ji, director of the Brain and Cognition Institute in the Department of Automation at Tsinghua University, Beijing, in a prepared statement. “The cheek, forehead and nose contributed more information to the algorithm than other facial areas. However, we need to improve the specificity as a false positive rate of as much as 46 percent may cause anxiety and inconvenience to patients, as well as potentially overloading clinics with patients requiring unnecessary tests.”
The deep learning tool for clinical evaluation requires further testing in multiple ethnic groups, because the study was mostly conducted on similar patients. The project’s innovation was applauded; however, its limitations were pointed out in an article by Charalambos Antoniades, Professor of Cardiovascular Medicine at the University of Oxford, UK, and Dr. Christos Kotanidis, a DPhil student working under Professor Antoniades at Oxford.
Some limitations include not validating the tool in a larger population and the ethical concerns about “unwanted dissemination of sensitive health record data, that can easily be extracted from a facial photo, renders technologies such as that discussed here a significant threat to personal data protection, potentially affecting insurance options.”
Professor Zheng acknowledged some limitations to the research and emphasizes that privacy and ethics are important to the team, as is the guarantee that the tool would not be used for any other purposes except medicine.
AI company Lapetus Solutions has been developing analytical solutions for the insurance market that leverage biometric facial recognition with selfies to treat the face as a biomarker of human aging.
artificial intelligence | biometric data | biometric research | biometrics | deep learning | face photo | healthcare