A study published in the International Journal of Medical Engineering and Informatics has shown that artificial intelligence models based on the Transformer architecture can effectively identify signs of cardiovascular disease using electrocardiogram (ECG) data.

Cardiovascular diseases remain one of the leading causes of premature death worldwide, claiming approximately 18 million lives each year. One of the major challenges is early diagnosis: the sooner a condition is detected, the greater the chances of successful treatment. However, interpreting ECGs requires considerable expertise, can be time-consuming, and is subject to human error.

In the new study, researchers developed a one-dimensional (1D) Transformer model capable of analyzing ECG signals alongside additional clinical information. Originally designed for natural language processing, the Transformer architecture was adapted to process time-series medical data.

During testing, the model achieved an accuracy of up to 94.2% in detecting early-stage cardiovascular disease across several well-established medical datasets. According to the authors, this level of performance could significantly improve diagnostic reliability, particularly when used as a decision-support tool alongside cardiologists.

The researchers believe that AI-powered systems of this kind could become valuable aids in clinical practice by accelerating decision-making and reducing the risk of diagnostic errors. However, they emphasize that the approach still requires validation on independent clinical datasets and further testing before it can be implemented in real-world healthcare settings.

If future studies confirm these findings, AI-assisted ECG analysis could help identify cardiovascular disease earlier, enabling timely intervention and potentially improving patient outcomes.