Clinical accuracy of a 3-dimensional multielectrode non-invasive ECG mapping vest in premature ventricular contraction ablation procedures

Abstract

Abstract Background Invasive 3D mapping is the current gold standard in determining premature ventricular contraction (PVC) localization. However, given its invasive nature, unawareness of PVC origin beforehand may lead to prolonged procedure times and increased complications risk. Purpose In this study, we compared the diagnostic accuracy of a non-invasive 3-dimensional mapping ECG vest (MECGV) with invasive 3D mapping and ablation successful points as a reference. We also evaluated 12-lead ECG based pre-procedural PVC localization accuracy of 2 expert operators blinded to clinical findings. Methods In 15 consecutive adult patients between the ages of 35 to 69 (male n: 9, median age: 48), MECGV and subsequently a 3D mapping with uneventful successful PVC RF catheter ablation were performed and included in the study. PVC localization estimations of MECGV method and the EP experts' locatization predictions were compared separately with invasive 3D ablation results using the Cohen's kappa method and with the help of a 16-segmentation model we created for the ease of the analyses . Results Results were presented on Table-1. MECGV was found to be highly compatible (Cohen's kappa 0.968, p<0.001) with the RF ablation results. However operators assessments were found to be less coherent with invasive 3D mapping (Operator 1 Cohen's kappa 0.571, p 0.033, Operator 2 Cohen's kappa 0.618, p: 0.019, respectively). MECGV accuracy was %87 in pinpointing the PVC focus. The only 2 localizations that MECGV failed were on the septal aspects of the heart (Picture-2 Failure of defining the focus on LVOT septal instead of RVOT septal) Conclusion MECGV method seems to be highly compatible with invasive 3D mapping techniques in detection of PVC localization and bests EP expert estimates in this regard. However, it also seem to have some pitfalls especially on septal aspects of the heart.Table-1 and segmentation schemePicture-1