Heart-rate adaptive match filter based procedure to detect and quantify T-wave alternans

Aim of the present study was to test our heart-rate adaptive match filter procedure for TWA analysis, which includes a) a heart-rate adaptive match filter, that filters out every ECG component but the TWA and quantifies TWA in terms of the duration (TWAD), amplitude (TWAA), and magnitude (TWAM, defined as TWAD × TWAA); and b) a statistical threshold criterion (STC) to identify, among all TWA detected cases, the ones (TWA+) most likely associated to an increased risk of sudden cardiac death. Our test, performed on the database assembled for the PhysioNet/Computers in Cardiology Challenge 2008, yielded the following results. On average, the data set was characterized by TWAD= 63±31 beats, TWAA= 48±77 μV, and TWAM= 3249±5107 beats μV. Based on our STC, TWA+ cases were 21, yielding a ranking score of 0.684. Use of STC, in the perspective of discriminating risky cases, among all detected TWA cases, might have limited our score