A new image-based algorithm uses artificial intelligence (AI) to analyze a person's facial photos to detect and determine the possibility of having a heart-related illness. Initial consultations and check-ups could be simplified for heart-disease patients because the AI can give a preliminary assessment to determine the next step.
A group of Chinese researchers discovered a way to create a deep learning algorithm aimed at analyzing a person's facial photograph and detecting whether there are signs of progressing heart disease.
The European Society of Cardiology recently published a press release that elaborates more on the Chinese researchers' work towards determining heart diseases. 'Selfies' could now be used as a means and ways for a cheap and simple analysis of a possible ailment of the human heart.
The screening tool developed for heart diseases requires four photographs from a patient to analyze and have initial assessments. This minimizes the need to go on further steps to know an illness or disease from within.
"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." said Professor Zhe Zhang, lead researcher of the AI, vice director of the National Center for Cardiovascular Diseases, and vice president of Fuwai Hospital, Chinese Academy of Medical Sciences, Beijing, People's Republic of China.
The researchers note that the algorithm is currently in its development stage, with upgrades and tests needing to be done to larger groups of different ethnicities.
ALSO READ: Satellite Images Show California Wildfires at Night, Triggered Extended Heat Waves
'Selfie' Algorithm's Development
Scientists took advantage of specific factors that relate facial features to present signs and indications that a person is experiencing an illness within his or her heart. The symptoms they are looking at include, gray or thinning hair, wrinkles, ear lobe crease, xanthelasmata (small, yellow deposits of cholesterol that is shown underneath the skin usually found in the eyelids), and arcus cornea (fat and cholesterol deposits that appear white-ish, grayish, or blue opaque rings in the outer edges of a person's cornea).
The study was conceptualized way back in 2017, where scientists gathered 5,796 patients from eight hospitals in China divided into Training (5,216 patients or 90%) and Validation (580 patients or 10%). Their demographics and health records were also taken for the algorithm's data processing, x-ray angiogram images of the patients, and even the four photos that nurses took for reference.
Professor Xiang-Yang Ji stated that the patients' records did not contribute to the algorithm's performance, saying that the photos were enough.
The algorithm surmounted the presently used heart disease detection methods such as the Diamond-Forrester model and the CAD consortium clinical score. The researchers were able to detect 80 percent of the heart disease accurately with the validation control group, and 61 percent among those have no present conditions.
The algorithm's test group boasts of 80% sensitivity and 54% specificity.
'Selfie' Algorithm for CAD
The algorithm is primarily focused on detecting coronary artery disease or CAD, the most common form of heart disease known to man. CAD is currently the leading cause of death amongst American citizens in both men and women.
This disease results from the hardening of arteries that prevent blood circulation from flowing to the heart muscle and the rest of the body. Cholesterol and plaque buildup mainly causes this disease inside the arteries and stop the right flow of blood from within.
CAD is a disease that can weaken the heart's muscle and even cause it to fail.
ALSO READ: WHO Chief Hopes COVID-19 Gone in 2 Years Like Spanish Flu; Vaccine Not Enough to Stem Coronavirus
This article is owned by Tech Times
Written by Isaiah Alonzo
