AI finds its way to people’s hearts (literally!)

Body

Osaka, Japan – AI (artificial intelligence) may sound like a cold robotic system, but Osaka Metropolitan University scientists have shown that AI (artificial intelligence) can provide heartwarming support—or, more accurately, “heartwarming.” They unveiled the innovative use of AI to classify heart function and determine valvular disease with unprecedented accuracy, demonstrating continued progress in combining medicine and technology to advance patient care. The results will be published in Digital Health Lancet.

Valvular heart disease, a cause of heart failure, is often diagnosed using echocardiography. This technique, however, requires specialized skills, so there is a shortage of qualified technicians. Meanwhile, a chest X-ray is one of the most common tests to identify diseases, especially the lungs. Although the heart is also visible on a chest X-ray, little was previously known about the ability of a chest X-ray to detect heart function or disease. A chest X-ray, or chest X-ray, is performed in many hospitals and requires very little time to perform, making it very accessible and easy to reproduce. Therefore, the research team led by Dr. Daiju Ueda, of the Department of Diagnostic and Interventional Radiology at Osaka Metropolitan University Graduate School of Medicine, thinks that if heart function and disease can be determined from chest radiographs, this test can serve as a complement to echocardiography.

Team Dr. Ueda successfully developed a model that utilizes AI to accurately classify heart function and valvular disease from chest radiographs. Because AI trained on a single data set faces the potential for bias, leading to low accuracy, the team targeted multi-agency data. Thus, a total of 22,551 chest radiographs associated with 22,551 echocardiograms were collected from 16,946 patients at the four facilities between 2013 and 2021. With chest radiographs set as input data and echocardiograms set as output data, an AI model is trained to learn features that relate the two datasets. .

The AI ​​model was able to correctly categorize six selected types of heart valve disease, with an Area Under the Curve, or AUC, ranging from 0.83 to 0.92. (AUC is a rating index indicating the ability of an AI model and uses a range of values ​​from 0 to 1, with the closer to 1, the better.) The AUC is 0.92 with a 40% limit for detecting left ventricular ejection fraction — an important measure for monitoring cardiac function .

“It took us a very long time to get these results, but I believe this is significant research,” said Dr. Ueda. “In addition to increasing the diagnostic efficiency of doctors, the system can also be used in areas where there are no specialists, in night emergencies, and for patients who have difficulty undergoing echocardiography.”

###

About OMU

Osaka Metropolitan University is the third largest public university in Japan, formed by the merger of Osaka City University and Osaka Prefectural University in 2022. OMU upholds “Knowledge Convergence” through 11 undergraduate schools, colleges, and 15 graduate schools. For more research news, visit https://www.omu.ac.jp/en/ or follow us on Twitter: @OsakaMetUniv_enor Facebook.