Identification of Coronary Culprit Lesion in ST Elevation Myocardial Infarction by Using Deep Learning

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Imaging

Background:

• ST-elevation myocardial infarction (STEMI) requires prompt identification of the infarct-related coronary artery.
• Current methods rely on electrocardiogram (ECG) interpretation, which can be challenging.
• Accurate culprit artery identification is crucial for timely and effective revascularization, reducing mortality and morbidity.

Purpose of the Study:

• To develop and evaluate a deep learning model for more accurate prediction of culprit coronary arteries in STEMI patients.
• To compare the efficacy of different ECG signal preprocessing techniques, specifically short-time Fourier transform (STFT) and continuous wavelet transform (CWT), within a deep learning framework.

Main Methods:

• A convolutional neural network (CNN) deep learning model was trained using ECG signals from 384 STEMI patients who underwent primary percutaneous coronary intervention (PCI).
• ECG signals were preprocessed using STFT and CWT (Morlet wavelet) with varying input lengths.
• Model performance was assessed by evaluating sensitivity, specificity, and overall accuracy for predicting the infarct-related artery (LAD, LCX, RCA).

Main Results:

• STFT preprocessing yielded an overall accuracy of 79.3%.
• CWT preprocessing achieved a higher overall accuracy of 83.7%.
• CWT demonstrated superior sensitivity and specificity for predicting all three major coronary arteries (LAD, LCX, RCA) compared to STFT.

Conclusions:

• Deep learning, utilizing CNNs, can significantly aid in identifying the culprit coronary artery in STEMI patients.
• Continuous wavelet transform (CWT) is a more effective ECG signal preprocessing method than STFT for this deep learning application.
• This approach holds promise for improving the diagnostic accuracy and guiding treatment decisions in STEMI management.