A morphology-based spatial consistency algorithm to improve EGM delineation in ventricular electroanatomical mapping

Activation mapping using electroanatomical mapping (EAM) systems helps to guide catheter ablation treatment of common arrhythmias. In focal tachycardias, the earliest activation area becomes the ablation target. Recently, we proposed a single-point wavelet-based algorithm to automatically identify electrogram (EGM) activation onsets for activation mapping. In this work, we propose an EGM morphology-based spatially-consistent algorithm for improving activation mapping in areas with a high-density of mapping points. The algorithm aligns those EGMs spatially close and morphologically similar and checks if the detected bipolar EGM activation onset is determined within a tolerance of ± 5 ms. If not, a weighted average bipolar EGM activation signal is computed and delineated. Then, the new activation onset is used to compute the local activation time (LAT). Automatically detected onsets are compared with manual annotations obtained during ablation procedure by an expert technician in a total of 15 electroanatomical maps (1763 mapping points). The presented algorithm modifies 31% of the studied mapping points and in those cases reduces the difference with manual annotations from 5.1 ± 13 ms to 4.3 ± 11.6 ms.