An Innovative Approach for the Prediction of Future Arrhythmia through T-wave Alternans on Surface Electrocardiogram (ECG)

Objective: To explore the techniques for predicting risk-causing arrhythmia in cardiac patients. Study Design: Prospective longitudinal Study Place and Duration of Study: Armed Forces Institute of Cardiology, Rawalpindi Pakistan, and Government College University, Faisalabad from Jul to Oct 2017. Methodology: The Electrocardiograms of 24-hour Holter monitoring were collected from the Electrophysiology Department. Electrocardiogram data was collected in the portable document format that was further transformed into Image format for computational analysis. Administrative data were analysed in multiple episodes of cardiac arrhythmogenesis. Data were classified by using a Convolutional Neural Network (CNN) based on computing the results of selected T-waves in three consecutive peaks within each cardiac cycle of patients. Results: One hundred twenty-six patients diagnosed with arrhythmia were selected. The mean episode of premature ventricular contractions in participants was 21.5±30. The mean duration of significant ECG episodes was 3.33±9.65 (seconds). The accuracy and precision rate of the classifier was about 81% for the overall significance of data that exhibited the risk of causing future life-threatening arrhythmia. Conclusion: This study introduces an innovative approach based on clinical paradigms that may help prevent the upcoming cardiac arrhythmogenesis events.