Algorithmic detection of the beginning and end of bipolar electrograms: Implications for novel methods to assess local activation time during atrial tachycardia

Activation mapping is required to effectively ablate atrial tachycardia (AT). Conventional tools to assess local activation time (LAT) are based upon the peak of the bipolar electrogram (B-EGM, LAT_Peak) and the maximal negative slope of the unipolar electrogram (U-EGM, LAT_Slope). Bipolar electrograms are influenced by wavefront direction, bipole orientation, and inter-electrode spacing causing ambiguity in peak detection, whereas unipolar electrograms are disturbed by the presence of far-field signals. We developed a new algorithm to detect the beginning and end of bipolar electrograms (t_begin and t_end). Then, we introduced new LAT methods related to the onset of B-EGMs (LAT_Onset), the center of mass of B-EGMs (LAT_CoM), and the slope of U-EGMs within a pre-defined window (LAT_Slope-hybrid).In total 3752 recordings from 31 AT patients were retrospectively analyzed. The signal-to-noise ratio (SNR) for B-EGMs was calculated to differentiate algorithmically high from low quality electrograms (HQ and LQ). In a subset of 328 B-EGMs, five experts validated the t_begin as determined by the algorithm by visual rating. The newly developed LAT methods were compared to the conventional LAT methods and to one another (Bland–Altman plots) in both HQ (n = 3003) and LQ EGMs (n = 749).The t_begin algorithm was accurate (deviation < ±10 ms) in 96 ± 4% of HQ and 91 ± 8% of LQ B-EGMs. BA plots revealed the following difference (bias) and variation in HQ and LQ EGMs respectively: (1) LAT_Onset vs. LAT_Peak: 27 ± 30 ms and 24 ± 62 ms; (2) LAT_CoM vs. LAT_Peak: 0 ± 16 ms and 2 ± 38 ms; (3) LAT_Slope-hybrid vs. LAT_Slope: 1 ± 32 ms and 15 ± 110 ms; (4) LAT_Onset vs. LAT_CoM: 22 ± 24 ms and 18 ± 22 ms; (5) LAT_Onset vs. LAT_Slope-hybrid: 16 ± 18 ms and 13 ± 22 ms; and (6) LAT_CoM vs. LAT_Slope-hybrid: 5 ± 20 ms and 4 ± 18 ms.In the present study, we introduced three new methods to assess local activation time in AT, based upon an algorithm detecting accurately the beginning and end of the B-EGM complex. BA analysis of the new methods showed similar variation in high and low quality EGMs, suggesting that they introduce less ambiguity than the conventional peak method. LAT_Onset consistently yielded an earlier activation moment. LAT_Slope-hybrid – by blanking far-field potentials – seems to be the optimal method for detection of the maximal negative slope in U-EGMs. Interestingly, LAT_CoM in B-EGMs coincided with the maximal negative slope in U-EGMs, suggesting its physiological sense and future use. The new LAT methods can be implemented in real-time mapping applications.