A novel automated peak frequency annotation algorithm for identifying deceleration zones and ventricular tachycardia ablation sites

BACKGROUND Current annotation of local fractionated signals during ventricular electroanatomic mapping (EAM) requires manual input subject to variability and error. OBJECTIVES The purpose of this study was to evaluate a novel peak frequency (PF) annotation software for its ability automatically detect late potentials (LPs) and local abnormal ventricular activity (LAVA), determine an optimal range for display, and assess its impact on isochronal late activation mapping (ILAM). METHODS EAM data from 25 patients who underwent ventricular tachycardia (VT) ablation were retrospectively analyzed. Sam-plings of electrogram PFs from areas of normal bipolar voltage, areas of low voltage, and areas of low voltage with fractioned signals were performed. An optimal range of frequency display was identified from these patients and applied to a validation cohort of 10 prospective patients to assess high PF within scar as a predictor of VT ablation target sites, in particular deceleration zones (DZs) identified by ILAM, LP, and LAVA. RESULTS Voltage and PF ranges of normal endocardial tissue varied widely. Using 220 Hz as a frequency cutoff value in areas low bipolar voltage, areas of high fractionation were identified with sensitivity of 91% and specificity of 85% There was no nificant reduction in targeted DZ surface areas, and colocalization with DZs was observed in all cases. Applied to the prospective cohort, PF predicted fractionated areas and DZ in 9 of 10 patients. CONCLUSION A PF annotation algorithm with a cutoff of 220 Hz accurately identifies areas of fractioned signals and accurately predicts DZs during ILAM.