Abstract 2514: Automatic Biventricular Capture Verification Based On Morphological Analysis of the Intracadiac Electrogram Signal

Abstract

Objective: We test a new biventricular capture verification algorithm, which can be used to adjust the pacing output in right ventricle (RV) and left ventricle (LV), to ensure the safety and efficacy of the cardiac resynchronization therapy (CRT). Methods: The algorithm calculates a novel index called ASCI, which measures the morphological similarity between the intracardiac electrogram (IEGM) signal and a template waveform representing captured paces. Ventricular capture (CAP) is verified if ASCI is greater than a predefined threshold, whereas non-capture (NC) is declared otherwise. CRT pacemakers were implanted in 6 dogs. During a mean follow-up of 23 days, 175 biventricular threshold tests were conducted by running a customized program, with various combinations of pace/sense polarities. Wideband RV and LV IEGM were recorded and downloaded for off-line analysis by the algorithm. Template signals for each pace/sense configuration in each chamber were created from the average of 4 captured beats. Each pace was annotated for CAP or NC by visual examination of the IEGM and based on stored data log of the pacing parameters. Results: A total of 9991 CAP paces (5104 RV, 4887 LV) and 4474 NC paces (2136 RV, 2338 LV) were eligible for morphology analysis. In both RV and LV, the calculated ASCIs were well separated for CAP and NC paces (Figure ): 0.83 ± 0.12 (RV CAP), 0.85 ± 0.11 (LV CAP), -0.06 ± 0.13 (RV NC), -0.07 ± 0.14 (LV NC). The classification accuracy of the algorithm remained >= 99% for any ASCI threshold value choosing between 0.18 and 0.52. Conclusion: This study demonstrated the feasibility to perform automatic biventricular capture verification based on morphology analysis of the IEGM.