AI trained to identify valvular heart disease

News

Chest radiograph data collected from patients has been used to develop an AI system that classifies cardiac functions and pinpoints valvular heart disease.

Left: Chest radiograph Right: Visualisation of the grounds for the AI's judgment
Left: Chest radiograph Right: Visualisation of the grounds for the AI's judgment - Daiju Ueda, OMU

The development from Osaka Metropolitan University, Japan, is claimed to identify valvular heart disease with unprecedented accuracy. The results have been published in The Lancet Digital Health.

Valvular heart disease is a cause of heart failure that is often diagnosed using echocardiography. This technique requires specialised skills and there is a corresponding shortage of qualified technicians.

Chest radiography is one of the most common tests to identify diseases, primarily of the lungs. The heart is also visible in chest radiographs little was known about the ability of chest radiographs to detect cardiac function or disease until now.

Chest radiographs (chest X-Rays), are performed in many hospitals and very little time is required to conduct them, making them highly accessible and reproducible. The research team led by Dr. Daiju Ueda, from the Department of Diagnostic and Interventional Radiology at the Graduate School of Medicine, calculated that if cardiac function and disease could be determined from chest radiographs, this test could serve as a supplement to echocardiography.

Since AI trained on a single dataset faces potential bias, leading to low accuracy, the team aimed for multi-institutional data. Accordingly, a total of 22,551 chest radiographs associated with 22,551 echocardiograms were collected from 16,946 patients at four facilities between 2013 and 2021. With the chest radiographs set as input data and the echocardiograms set as output data, the AI model was trained to learn features connecting both datasets.

MORE FROM MEDICAL & HEALTHCARE

The AI model was able to precisely categorise six selected types of valvular heart disease, with the Area Under the Curve (AUC, a rating index for AI models using a value range from 0 to 1. The closer to 1, the better the model). The AUC was 0.92 at a 40 per cent cut-off for detecting left ventricular ejection fraction, an important measure for monitoring cardiac function.

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

AI medical diagnosis