Abstract 19049: Feasibility of Simulation-Based Prediction of Optimal Ventricular Tachycardia (VT) Ablation Targets in Patients With ICD Artifact Burden
Introduction: We recently developed methodology to predict infarct-related VTs and the optimal ablation targets using personalized ventricular models reconstructed from LGE MRI scans. However, most patients receiving VT ablation have ICDs, which imposes image artifact burden and hinders model generation by reducing the amount of available structural information. This study aimed to develop an enhanced methodology for ventricular model generation and to demonstrate, in a retrospective study, the feasibility of using simulation-based prediction of optimal VT ablation targets in patients with ICDs despite the artifact burden.
Methods: Four patients with ICDs and infarct-related VT underwent LGE MRI prior to clinical ablation; artifact burden was up to 62%. Myocardial regions covered by artifact were delineated based on the 3D radial distance from the ICD. 3D models of patients’ hearts were generated from the scans by extrapolating ventricular geometry over the artifact region (Fig columns 1, 2); the latter was assumed to not contain infarcted tissue. Outside the artifact, tissues were classified as normal, scar, and gray zone based on pixel intensity. In silico pacing from 26 sites induced VTs, VT analysis identified optimal ablation targets, which were then compared to clinical ablation lesions.
Results: With the new image processing steps, generated models successfully predicted infarct-related VTs localized outside the artifact (Fig column 3). In silico ablation terminated the induced sustained VTs. Predicted ablation targets corresponded with clinical lesions but were smaller in size (Fig column 4). In patient 3, ablation in anterior region was not predicted by the model due to the occlusion of this region (containing part of the infarct) by artifact.
Conclusion: Personalized ventricular models could accurately predict non-invasively the optimal VT ablation sites where scar was not obscured by ICD artifact, and thus may be used to guide clinical ablation.
Author Disclosures: A. Prakosa: None. P. Nikolov: None. S. Zahid: None. E.G. Ipek: None. S. Nazarian: Research Grant; Significant; Biosense Webster, Inc.. Consultant/Advisory Board; Modest; Medtronic, Inc., CardioSolv, LLC. N.A. Trayanova: Ownership Interest; Modest; CardioSolv, LLC..
- © 2016 by American Heart Association, Inc.