Po-05-037 efficacy and clinical characteristics of avert-vt: ablation at virtual-heart predicted vt

We have previously demonstrated the feasibility of virtual-heart simulations in determining noninvasively the optimal infarct-related ventricular tachycardia (VT) ablation targets and guiding the clinical ablation. We are now proceeding with AVERT-VT, an FDA-approved study for ten prospective patients. To develop the end-to-end personalized computational modeling prediction pipeline enabling AVERT-VT and to assess the corresponding clinical characteristics of the predicted virtual-heart ablation targets. We developed a comprehensive end-to-end computational pipeline enabling AVERT-VT (Fig). 3D LGE-MRI scans with 1.5x1.5mm in-plane spatial resolution and slice thickness <2mm were used to reconstruct the 3D left ventricular (LV) model incorporating deep scar, border (gray) zone, and non-infarcted tissue. In silico rapid pacing from 7 LV regions were performed to induce VTs. The optimal ablation targets were determined by thorough analyses of the wavefront propagation of the induced in silico VTs. A linear ablation connecting a VT ablation target to a transmural deep scar region, the basal region, or the other VT target was performed to eliminate the potential VT pathway formed by the target ablation. Potential emergent VTs in the post-ablation substrate were investigated by repeating the VT induction pacing protocol. This process was iterated until complete VT non-inducibility was achieved. The initial and repeat ablation targets, linear ablations, as well as the LV, right ventricle, infarct area, aortic arch, and left and right coronary cusp geometries, were exported as an ablation strategy plan compatible with the clinical mapping system. The figure presents results from one patient in the clinical study. The ablation plan was successfully imported and co-registered to the patient’s heart. The extent of the model infarct corresponded to the bipolar low-voltage regions. Abnormal electrograms were observed at the proposed targets. 2 VTs were observed during the procedure. Areas of latest activation in sinus rhythm and best pace maps, with a long stimulus to QRS, were found at 2 of the proposed targets. One of the seven proposed targets was not ablated due to the proximity to the left sided His potential. The patient was not inducible of any VT at the end of the procedure. AVERT-VT proposed targets corresponded to abnormal electrograms, and ablating at the targets led to VT non-inducibility.