AI-powered Deep-learning Model Predicts Risk of Low LVEF using ECG Images

JJohn August 2, 2023 9:57 PM

Researchers have developed a deep-learning model which uses artificial intelligence to identify patients at risk of developing low left ventricular ejection fraction (LVEF) by analyzing ordinary electrocardiogram (ECG) images. The approach appears to predict not only current patients with low LVEF but also those who may develop this condition in the future.

AI model predicts future cardiac risks

A deep-learning model developed by researchers has shown promising results in leveraging ordinary ECG images to identify patients with a left ventricular (LV) ejection fraction below 40%. Notably, the model can also predict which patients, who currently do not exhibit signs of LV systolic dysfunction, are at an increased risk of developing it in the future. This predictive capability of the model could become a game changer in preemptively diagnosing and potentially preventing severe cardiac conditions.

Large dataset ensures model reliability

The development and testing of the model involved nearly 400,000 12-lead ECGs and corresponding echocardiographic data. This extensive data set allowed the researchers to fine-tune the model for high accuracy and sensitivity. Given the large amount of data used, the findings are expected to be reliable and applicable to a broad spectrum of patients, potentially leading to a significant breakthrough in the detection of low LVEF.

Model validated in various healthcare settings

The model's versatility lies in its independency - it can be applied in various institutions without the need for specific adjustments. It has been validated in outpatient Yale New Haven Hospital clinics, and externally through ECG images taken at various other hospitals and health centers. This independent application and broad validation process increases the robustness and reliability of the model, making it a highly useful tool in detecting and predicting low LVEF.

Although the model shows promise in predicting the risk of low LVEF, it is still in the investigational stage and has not yet received approval from the US Food and Drug Administration. The researchers are hopeful of developing AI-based ECG screening approaches for a variety of cardiovascular conditions, demonstrating the potential of AI in revolutionizing cardiovascular healthcare.

More articles

Also read

Here are some interesting articles on other sites from our network.