Basic/Translational Science -> Computer Modeling/Simulation D-PO02 - Poster Session II (ID 47) Poster

D-PO02-030 - Integrating Endocardial Mapping And Electrocardiographic Imaging (ECGI) For Improving PVC Localization: A Feasibility Study (ID 168)

Disclosure
 W.W. Good: Nothing relevant to disclose.

Abstract

Background: Accurate localization of premature ventricular contractions (PVCs) deep within the myocardium is a significant challenge with current techniques reporting large and unpredictable errors. Such techniques are based on endocardial mapping alone, however combining mapping with non-invasive electrocardiographic imaging (ECGI) could provide valuable missing information. We propose and evaluate a novel integrated approach to improve PVC detection within the myocardium using both endocardial mapping and ECGI.
Objective: Determine the feasibility and improvements in accuracy possible by integrating non-invasive ECGI activation maps with endocardial mapping to detect the depth and location of PVCs within the myocardium.
Methods: We used the Cardiac Arrythmia Research Package (CARP) software to simulate activation from 7 PVC sites at varying depths throughout the myocardium. To replicate endocardial mapping, we sampled activation times over the ventricular chamber surface of the simulated heart at 1-2 mm resolution, from which we reconstructed activation times throughout the myocardium using radial basis interpolation. We then added epicardial activation times, to emulate the information provided by ECGI, and repeated the reconstruction of three-dimensional activation times and determined the PVC locations. Taking the position of the earliest activated node in the reconstructed volume and measuring the Euclidean distance to the ground truth location determined the PVC localization error.
Results: PVC localization using endocardial mapping alone produced errors of 0.5 - 17 mm, with the largest occurring from PVCs originating from the epicardium. Incorporating ECGI improved the reconstruction, decreasing the error range to 0.5 - 4.5 mm with the largest error occurring with the mid-myocardial PVCs. The integrated approach reduced localization error of mid-myocardial PVCs by 3.4 -8.7 mm.
Conclusion: Combining endocardial mapping with epicardial activation times as estimated from ECGI have the potential to improve localization of PVCs within the myocardium. Feasibility studies are now necessary to confirm the practical utility of this approach as well as its sensitivity to inevitable errors in either mapping or ECGI.
Collapse