Multiparametric Electrogram Feature Analysis for Ventricular Tachycardia Functional Extra-Stimulus Substrate Mapping

BACKGROUND:Functional extra-stimulus (ES) substrate mapping has been shown to reduce the recurrence of ventricular tachycardia (VT) compared with intrinsic rhythm mapping. However, there is no standardized approach to functional ES mapping. OBJECTIVES:This study investigated the accuracy of a range of functional ES mapping methods using S1 and S2 ES, evaluating multiple electrogram (EGM) components. METHODS:A multicenter, international VT ablation cohort was investigated. Offline annotation of stimulus artifact, first deflection, last deflection, and near-field (NF) EGM components enabled construction of 6 functional ES maps. Comprehensive accuracy testing was performed in identifying the critical arrhythmogenic substrate. RESULTS:Nineteen patients (mean age 61.1 ± 14.2 years; mean ejection fraction 31.7% ± 11.3%) were included. The EGM duration-based decremental evoked potential (DeEP) map had the highest area under the receiver-operating characteristic curve (0.804 ± 0.107; P < 0.001), with an optimal decrement threshold of 20 milliseconds (F1 score, P = 0.014) or 30 milliseconds (F1 score, P = 0.029). The inclusion of latency to EGM duration, assessed with a stimulus to the last deflection-based DeEP map, did not improve diagnostic performance (area under the receiver-operating characteristic curve 0.625 ± 0.173). Automated NF EGM annotation did not improve accuracy with or without the inclusion of latency. The EGM duration DeEP map exhibited superior accuracy across almost all left ventricular locations, providing higher precision in lateral/basal regions (Δ average precision, 0.24 ± 0.04; P < 0.001). CONCLUSIONS:Functional ES mapping using EGM duration prolongation DeEP maps yielded the highest accuracy in defining the arrhythmogenic substrate. The inclusion of latency or annotation of NF EGM components does not improve functional substrate mapping accuracy.