Integrating multi-domain deep features of electrocardiogram and phonocardiogram for coronary artery disease detection

• Published in: Computers in biology and medicine Vol. 138; p. 104914
• Main Authors: Li, Han, Wang, Xinpei, Liu, Changchun, Li, Peng, Jiao, Yu
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.11.2021; Elsevier Limited
• Subjects: Abnormalities; Accuracy; Artificial neural networks; Cardiovascular disease; Classification; Coronary artery; Coronary artery disease; Coronary vessels; Deep learning; ECG; EKG; Electrocardiography; Feature extraction; Heart; Heart diseases; Human error; Internal Medicine; Machine learning; Medical imaging; Morphology; Multi-domain features; Neural networks; Other; PCG; Phonocardiography; Sensitivity; Time series; Two dimensional models; Vein & artery diseases; Wavelet transforms; China; Deep learning; ECG; Multi-domain features; PCG; Coronary artery disease
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2021.104914

Electrocardiogram (ECG) and phonocardiogram (PCG) are both noninvasive and convenient tools that can capture abnormal heart states caused by coronary artery disease (CAD). However, it is very challenging to detect CAD relying on ECG or PCG alone due to low diagnostic sensitivity. Recently, several studies have attempted to combine ECG and PCG signals for diagnosing heart abnormalities, but only conventional manual features have been used. Considering the strong feature extraction capabilities of deep learning, this paper develops a multi-input convolutional neural network (CNN) framework that integrates time, frequency, and time-frequency domain deep features of ECG and PCG for CAD detection. Simultaneously recorded ECG and PCG signals from 195 subjects are used. The proposed framework consists of 1-D and 2-D CNN models and uses signals, spectrum images, and time-frequency images of ECG and PCG as inputs. The framework combining multi-domain deep features of two-modal signals is very effective in classifying non-CAD and CAD subjects, achieving an accuracy, sensitivity, and specificity of 96.51%, 99.37%, and 90.08%, respectively. The comparison with existing studies demonstrates that our method is very competitive in CAD detection. The proposed approach is very promising in assisting the real-world CAD diagnosis, especially under general medical conditions. •Multi-domain deep features are integrated for coronary artery disease detection.•Simultaneously recorded ECG and PCG signals are utilized.•The proposed framework eliminates the need for feature engineering.•The proposed framework obtains better results than exiting studies.