-
Views
-
Cite
Cite
B. Imielski, C. Kowaleski, M. Tamboli, M. Alhusseini, T. Baykaner, G. Meckler, F. Shenasa, D. Krummen, M. Viswanathan, P. Wang, J. Brachmann, J. Miller, D. Vidmar, W.J. Rappel, S. Narayan, P777
Interobserver accuracy in identifying sources for atrial fibrillation by different mapping algorithms at proven locations of termination, European Heart Journal, Volume 38, Issue suppl_1, August 2017, ehx501.P777, https://doi.org/10.1093/eurheartj/ehx501.P777Close - Share Icon Share
Extract
Background: It has been suggested that variations in outcome from driver/rotor ablation for atrial fibrillation (AF) may reflect human interpretation in reading complex maps.
Purpose: To study interobserver variability in identifying sites where ablation terminated persistent AF before pulmonary vein isolation (PVI). We compared the sensitivities of 3 observers for such sites, mapped by 2 published methods in the same patient data.
Methods: We identified 33 patients (26 M, 62±9 yrs) with persistent AF in whom ablation terminated AF to sinus rhythm (20) or atrial tachycardia before PVI. Ablation was targeted targeted by method (1) Activation/Phase (FIRM, Narayan JACC 2012) on 64 pole basket data. 3 reviewers read datasets in a blinded fashion, which were then unblinded to actual termination sites by (1) Index method (without automated source detection) and (2) Phase (Kuklik, IEEE Trans Biomed Eng. 2015). Statistical analysis was performed using Fleiss' Kappa.
Results: Table 1a shows reviewer agreement between visually identified sources by FIRM/Kuklik methods (table 1b) compared to known procedural termination sites. Individual reviewer sensitivity was 39–67% and 55–73% for localizing sources within 1 and 1.5 electrode accuracy. There was concordance between reviewer identified sites and actual termination sites by methods 1 (p=0.03), which was non-significant using method 2 (p=0.24).
