A novel convolutional neural network structure for differential diagnosis of wide QRS complex tachycardia. (March 2023)

• Record Type: Journal Article
• Authors: Fayyazifar, Najmeh; Dwivedi, Girish; Suter, David; Ahderom, Selam; Maiorana, Andrew; Clarkin, Owen; Balamane, Saad; Saha, Nishita; King, Benjamin; Green, Martin S.; Golian, Mehrdad; Chow, Benjamin J.W.
• Is Part Of: Biomedical signal processing and control. Volume 81(2023)
• Publication Date: 2023-03
• Subjects: Deep learning -- Convolutional neural network -- Wide QRS complex tachycardia -- Neural architecture search; AF Atrial Fibrillation -- AHA American Heart Association -- AI Artificial intelligence -- ANN Artificial neural network -- BN Batch normalization -- CNN Convolutional neural network -- CUDB Creighton University ventricular tachyarrhythmia database -- CVD Cardiovascular disease -- DAG Directed acyclic graph -- DARTS Differentiable architecture search -- DWT Discrete Wavelet Transform -- ECG Electrocardiograms -- EMD Empirical Mode Decomposition -- FFREWT Fixed Frequency Range Empirical Wavelet Transform -- GRAD-CAM Gradient-weighted class activation mapping -- HR Heart rate -- IF Instantaneous Frequency -- LSTM Long-short term memory -- MITDB MIT-BIH dataset -- ML Machine learning -- NAS Neural architecture search -- NPV Negative predictive value -- PPV positive predictive value -- RCNN Recurrent convolutional neural network -- SGD Stochastic gradient descent -- SVG Scalable vector graphic -- SVM Support vector machines -- SVT Supraventricular tachycardia with aberration -- VF Ventricular Fibrillation -- VFDB MIT-BIH- malignant ventricular arrhythmia database -- VMD Variational Mode Decomposition -- VT Ventricular tachycardia -- WCT Wide QRS complex tachycardia; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104506
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Accurate rhythm diagnosis on electrocardiograms (ECG) is critical in patients presenting with wide QRS complex tachycardia (WCT) arrhythmia. Real-time visual interpretation of ECG of complex arrhythmias is difficult and requires expertise. We designed a convolutional neural model through a neural architecture search that could accurately classify WCT into those that are ventricular in origin (87.5%) or supraventricular tachycardia (91.7%). Our model was also shown to be useful for arrhythmia diagnosis from ECG data generated by both 12 leads as well as single-lead devices. Our model can potentially be implemented in real clinical settings to assist physicians in more accurate and timely diagnosis of WCTs. Abstract: Background and objectives: Cardiac arrhythmias are a significant cause of morbidity and mortality in patients with cardiovascular disease. Accurate rhythm diagnosis is critical in patients presenting with wide QRS complex tachycardia (WCT). Real-time visual interpretation of electrocardiograms (ECG) of complex arrhythmias is difficult and requires expertise. We designed a convolutional neural network (CNN) that could accurately classify WCT into those that are ventricular in origin (ventricular tachycardia (VT)) or supraventricular tachycardia with aberrancy (SVT). Methods: A total of 3065 patients with wide complex ECGs were screened (415 with VT and 2650 with SVT). A CNN model was designed through a Neural Architecture Search (NAS) method. This CNN consisted of a stem convolution layer and five cells, each cell containing separable-convolution and dilated-separable-convolution layers. Results: Using 5-fold cross-validation and executing algorithm for five independent runs (with five different seeds), the proposed CNN model achieved a detection accuracy of 87.5 ± 0.0025 and 91.7 %±0.0004 for VT and SVT, respectively. The total sensitivity, specificity, positive predictive value, negative predictive value and F1-score of the CNN model were 88.50 %, 88.50 %, 88.54 %, 88.54 %, and 88.49 %, respectively. Conclusions: In a cohort of patients presenting with a WCT, our CNN model achieved an accuracy of 87.5% and 91.7% to correctly diagnose VT and SVT, respectively. This model has the potential of being used in real-time settings and to assist physicians with interpretation and decision making.