New AI ECG Tool Detects Early Heart Disease

Early detection is crucial because timely intervention can change prognosis and conserve resources. Electrocardiography (ECG) is a frontline test, yet interpretation is time-consuming and prone to error. Researchers have now developed a language-inspired artificial intelligence (AI) approach to flag early signs of heart disease on ECGs.

A one-dimensional (1D) Transformer model, adapted from architectures originally designed for language processing, has been introduced for cardiovascular risk detection. The system was detailed in a study published in the International Journal of Medical Engineering and Informatics. It is designed to assist clinical teams by analyzing ECG data with a level of consistency that supports expert review and downstream decision-making.

The model processes ECG signals alongside other clinical data. By running these inputs in parallel, it aims to capture patterns that indicate the early stages of cardiac disease. The approach is intended to complement, not replace, clinician interpretation, providing an additional layer of screening support in settings where workload and variability can hinder timely assessment.

In tests using several well-known medical datasets, the model achieved up to 94.2% accuracy in detecting early-stage heart disease. The results indicate that the technique performed well across disparate sources of ECG data. Such performance could, when paired with expert clinical assessment, improve confidence in determining next diagnostic steps and potential treatment pathways.

The researchers emphasize that further development is needed before clinical deployment. They recommend validation with independent clinical datasets to confirm generalizability and reliability in real-world settings.