Spectral Analysis for Slow Pathway Characterization in Atrioventricular Nodal Reentry Tachycardia

Giorgia Bongiovanni1, Matteo Fioravanti2, Francesco Soliani3, Antonio Crocamo4, Francesca Notarangelo5, Cristiana Corsi1
1University of Bologna, 2Biosense-Webster, 3Biosense Webster, 4University Hospital Of Parma, 5Azienda Ospedaliero-Universitaria di Parma


Abstract

Aims: Despite significant technological advances in supraventricular arrhythmias mapping, the procedure for ablating atrioventricular nodal reentrant tachycardia (AVNRT) still relies on a qualitative and experience-based identification of the slow pathway (SP) potentials described by Jackman and Haissaguerre. This study aims to characterize the SP signal in the frequency domain, trying to identify one or more parameters to make the mapping more objective and effective than current techniques.

Methods: The study analyzed 355 atrial signals from 42 patients undergoing AVNRT ablation (using the CARTO 3 system, version 7.5). In particular, 123 signals were target signals recorded from the regions of Koch's triangle where junctional rhythm appeared during radiofrequency delivery, and 232 signals were from non-target regions. A multiparametric analysis was performed on these signals, utilizing 10 spectral and 1 temporal parameters in 2047 combinations to find the best set of parameters for distinguishing target signals from non-target ones, using ROC curve analysis.

Results: The findings indicated that the best performance in terms of discrimination was achieved by combining the amplitude of the second peak with the power spectrum density slope between the dominant and secondary peaks (AUC 0.63), achieving 80% sensitivity and 55% specificity. This represents a 50% reduction in false positives compared to operator experience-based evaluations.

Conclusions: Enhancing accuracy and automating the SP signal mapping is essential to avoid multiple non-target radiofrequency deliveries. Based on our results, to improve SP signal characterization, a shift from the time to the frequency domain seems necessary. Interestingly, the amplitude of the second peak might represent the fast component described by Jackman in the temporal domain. Overall, frequency domain signal analysis could be a valuable tool and, following a comprehensive validation on a larger dataset, its development and integration in current mapping systems could support ablation strategy planning.