부산대학교 도서관

Background: During ventricular tachycardia (VT) ablation, clinical VT is often non-inducible or hemodynamically unstable. Hence, a substrate-based approach is often necessary. Although many strategies have been proposed, none have been automated nor incorporated into current day electro-anatomical mapping (EAM) systems. One strategy, recommended in 2019 HRS guideline, is Decrement Evoked Potential (DeEP) mapping for identification of critical VT substrate. We have developed an automated method to identify critical DeEP VT substrate in an EAM system (CARTO 3).Objective: The performance of this novel automated algorithm was retrospectively evaluated in patients who underwent substrate-based VT ablation.Methods: Entire VT ablation using DeEP mapping was performed and recorded in 12 consecutive patients. CARTO 3 case data was downloaded and analyzed by a novel automated algorithm for DeEP detection. A blinded electrophysiologist verified the automated electrogram DeEP annotations and assessed its performance in a dichotomous fashion. The effect of bipolar voltage threshold (0.01, 0.02, 0.05, 0.075, 0.1, 0.20, 0.25 and 0.30 mV) on algorithm performance was evaluated. Sensitivity, specificity and ROC curve of the algorithms were calculated. Finally, results from the analysis were merged with voltage map to produce DeEP maps.Results: Results from the ROC curve shows that a bipolar threshold of 0.075mV optimizes specificity and sensitivity. At that threshold the algorithm found 848 true positive and 626 true negatives out of a total of 1975 EGMS being analyzed. Sensitivity and specificity were found to be 65.8% and 82.3% respectively.Conclusions: We have developed an automatic detection algorithm that identifies and locates critical VT substrates. The performance characteristics have been evaluated for a multimodal VT substrate map to provide clinicians mechanistic substrate map in CARTO 3 for VT ablation.