Convolutional Neural Network-Based ECG-Assisted Diagnosis for Coal Workers.

Area of Science:

• Artificial Intelligence in Medicine
• Cardiology
• Occupational Health

Background:

• Electrocardiograms (ECGs) are crucial for diagnosing cardiac conditions.
• Developing automated systems for ECG analysis can improve diagnostic efficiency, especially in occupational health settings.
• Coal workers face specific health risks that may be reflected in cardiac health.

Purpose of the Study:

• To develop and evaluate a convolutional neural network (CNN) model for processing and extracting features from electrocardiogram (ECG) images.
• To establish an ECG-assisted diagnostic model for identifying common cardiac abnormalities in coal workers.
• To assess the performance of the CNN model in classifying specific ECG abnormalities.

Main Methods:

• Selected coal workers from Gequan Mine Hospital and Dongpang Mine Hospital for ECG analysis.
• Preprocessed ECG images and utilized Python software with a CNN for image recognition and classification.
• Employed various metrics including accuracy, sensitivity, specificity, Brier score, and AUC to evaluate model performance.

Main Results:

• The CNN model demonstrated high accuracy in identifying specific ECG abnormalities: sinus bradycardia (97.66%), non-specific intraventricular conduction delay (96.49%), myocardial ischemia (93.62%), and sinus tachycardia (93.02%).
• High sensitivity and specificity were observed across the models for different conditions.
• The study identified sinus bradycardia, non-specific intraventricular conduction delay, myocardial ischemia, and sinus tachycardia as prevalent ECG abnormalities in the study population.

Conclusions:

• The developed CNN model accurately identifies key ECG abnormality types in coal workers.
• The model shows potential for assisting in the diagnosis of cardiac conditions within this occupational group.
• Common ECG abnormalities identified include sinus bradycardia, non-specific intraventricular conduction delay, myocardial ischemia, and sinus tachycardia.