A collaborative approach to improve TWA detection

摘要

T wave Alternans (TWA) is an important risk identification marker of ventricular arrhythmia and Sudden Cardiac Death (SCD). Number of methods is proposed to detect this micro level phenomenon. The correlation (CM) and modified moving average (MMA) are two time domain, popular and well studied methods. The MMA does not have detection stage but quantification results are good and robust. While the CM has both detection and quantification stages but later is underestimated. Thus considering the shortcomings of each, a new collaborative approach in which detection stage of CM and quantification stage of MMA is proposed. The algorithm is tested on TWA challenge database record. Concordance Correlation Coefficient of 0.41 and 0.70 while RMSE values of 18 and 13 are obtained respectively for CM and MMA. A simple correlation coefficient is almost same for two. Thus the understanding limitations of the existing methods and efforts to overcome those shows value addition and improvement in TWA analysis. This will help to enhance overall prognostic and diagnostic performance of TWA test.