A Novel Algorithm Combining Continuous Wavelet Transform and Philips Method For QT Interval Analysis

Abstract

QT surveillance is the most important tool to diagnose the possibility of sudden death and torsade de pointes caused by usage of pro-arrhythmic drugs using the treatment of abnormalities in the heart. However, T wave, which represents the repolarization of heart ventricles, changes make QT interval measurement difficult in terms of subject-to-subject difference, morphological variation and noise effect. On the other hand, automated QT algorithm is necessitated because of manual QT measurement with some difficulties such as reading long records, fatigue in observation and individual evaluation by experts leading to different results. In this study, a novel fully automated method based on continuous wavelet transform and Philips method was developed to determine QT interval measurement. The ECG recordings were obtained from Physionet QT database marked by manual and standard automated method, and then the presented algorithm had scores of 15.46 and 11.87 msec mean error with 10.85 and 9.91 msec standard deviation in terms of gold and silver standard, respectively. The proposed algorithm was more stable and robust with high performance than the previous studies compared. As a result, the presented algorithm is a useful and novel method to perform QT interval analysis.