Development and validation of a new ECG algorithm based on the analysis of lead V3 to determine the origin of outflow tract ventricular arrhythmias

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Introduction (Aim) To preoperatively differentiate the site of origin outflow tract ventricular arrhythmias (OT-VA) it is of utmost importance for procedure planning yet challenging for those with a precordial transition (PT) at lead V3: close-proximity anatomical structures produce similar VA morphologies, thus leading to possible misdiagnosis. We sought to create an ECG algorithm that could be accurate and useful in overcoming this problem. Methods and Results 74 consecutive patients (Pts) with OT-VA who underwent successful radio-frequency catheter ablation (V3 PT 60.8%) were enrolled. The ECG characteristics of the first 30 Pts were analyzed (retrospective cohort); those with a PT at lead V3 underwent activation-mapping of both OT to ensure a correct diagnosis. LV-OT and RV-OT groups shared similar characteristics, including BMI and BSA. Above all ECG measurements, the V3 duration index (DI) and amplitude percentage, both calculated with the formula [R\(R+S)]*100 (computing duration and amplitude measurements respectively), showed the greatest AUC. A V3 DI < 50% established a certain diagnosis of origin at RV-OT (sensitivity 86.7%, specificity 100%, PPV 100%, NPV 88.2%; AUC 0.931). If the V3 DI was ≥ 50%, a V3 R wave percentage ≥ 50% established a certain diagnosis of origin at LV-OT (sensitivity 86.7%, specificity 100%, PPV 100%, NPV 88.2%, AUC 0.951). While a direct study of the RV-OT or LV-OT is suggested in these first two cases, few classification errors occurred only in the rare event of a V3 DI ≥ 50% with a V3 R wave percentage < 50% (Figure 1). In the prospective cohort, the two indices were confirmed to have the two best AUCs (0.992 and 0.986, respectively), and the algorithm showed an accuracy of 95.45%. Since the aortic cusps are structures posterior to RV-OT, therefore further away from V2, these foci more frequently show a QRS onset at ≥ V3; a characteristic which in our series was associated with an odds ratio for LV-OT origin of 4.1 [95% CI 1.47 – 11.39], p = 0.007. Using the transition ratio, we found a statistical significance only in lead V3, with a cut-off of ≥ 1 for predicting an LV-OT origin (LV-OT vs. RV-OT: 7.70 ± 12.75 vs. 0.70 ± 0.55, p < 0.001; AUC 0.898). Analyzing the overall case series, we can generalize the following: duration and amplitude indices showed an increasing AUC passing from lead V1 to V3 (Figure 2). The indices based on the complete measurement of the wave R compared to the QRS showed better AUCs than those on partial measures in V2 and V3: the duration index was better than the deflection index, as well as the amplitude percentage was better than the amplitude ratio. Conclusions The usefulness of the algorithm lies in providing preoperatively two assured outcomes in most cases (86.36%), allowing to limit the procedure to one OT directly - even in the event of a V3 PT - while it can select a small subgroup of complex Pts in which a study of both OT is recommended.