A novel transformer-based approach for cardiovascular disease detection

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Cardiology

Background:

• Cardiovascular diseases (CVDs) are a leading cause of global mortality, accounting for 17.9 million deaths annually.
• Early detection of CVDs, including arrhythmia and heart failure, is crucial for reducing mortality rates.
• Existing diagnostic methods require improvement in efficiency and cost-effectiveness for widespread application.

Purpose of the Study:

• To develop a novel, efficient, and low-cost transformer-based algorithm for the classification of cardiovascular diseases.
• To enhance the early prediction capabilities for various CVDs using electrocardiography (ECG) data.
• To evaluate the performance of the proposed algorithm against existing state-of-the-art methods.

Main Methods:

• Extraction of 56 features from 1,200 ECG records representing four distinct cardiovascular diseases.
• Application of Random Forest for feature selection, identifying the 13 most significant features.
• Development and implementation of a novel transformer-based algorithm for classifying the four CVD classes.

Main Results:

• The proposed transformer algorithm achieved exceptional classification performance.
• Maximum accuracy reached 0.9979, with precision, recall, and F1 scores all exceeding 0.9958.
• The algorithm demonstrated superior performance compared to all previously reported state-of-the-art CVD classification methods.

Conclusions:

• The developed transformer-based algorithm offers a highly accurate and efficient solution for cardiovascular disease classification.
• This novel approach holds significant potential for improving early CVD detection and patient outcomes.
• The low-cost and efficient nature of the algorithm makes it suitable for broader clinical implementation.