Po-05-003 determining the accuracy of virtual heart arrhythmia ablation targeting in predicting critical substrate in scar-dependent ventricular tachycardia

Substrate-based ablation for ventricular tachycardia (VT) is used in those with hemodynamically unstable VT. Personalised virtual heart arrhythmia ablation targeting (VAAT) has been shown to be feasible in predicting optimal ablation targets to control scar-dependent VT. The predictive accuracy of VAAT modelling in detecting substrate abnormalities is unknown. To determine the accuracy of VAAT in predicting critical substrate in patients with scar-dependent VT. Gadolinium-enhanced MR images were used to create patient-specific models. Simulated ventricular pacing was performed to demonstrate VT inducibility and to predict optimum lesion sets that terminate all likely VTs in the models. The 3D reconstructed MR image with predicted lesion sets was merged with the electroanatomical map obtained during catheter ablation. Mapping was performed during sinus rhythm (SR), right ventricular (RV) pacing and extrastimulus at 20ms above ventricular effective refractory period. Blinded to location, electrogram (EGM) duration, fractionation, double potentials, late potentials and other local abnormal ventricular activity (LAVA) were noted. Comparison was made between predicted and non-predicted sites within bipolar low voltage (<1.5mV) regions. Comparison was made between acquisition rhythms. 10 patients (100% male) with scar-dependent VT underwent VAAT modelling and VT ablation. 11766 EGMs were analysed. During RV pacing, mean bipolar voltage was similar between predicted and other sites (0.81 ± 0.37 vs 0.79 ± 0.37 mV, p = 0.13). EGM duration was longer in the predicted sites vs other sites (73.90 ± 27.06 vs 62.68 ±21.05 ms, p = < 0.001). EGM abnormalities were greater in the predicted sites vs other sites (574/1354 (42.4%) vs 512/2648 (19.3%), p < 0.001). Table 1. EGM duration was longest during extrastimulus mapping (75.63 ± 29.51ms), followed by RV pacing (66.47 ± 23.84 ms) and shortest during SR (62.88 ± 20.20 ms), p < 0.001. EGM abnormalities were most prevalent during extrastimulus mapping (present in 38.1% of all EGMs) followed by RV pacing (27.1%) and least during SR (21.8%, p < 0.001). EGM abnormalities were more frequent in predicted sites compared to other sites regardless of acquisition rhythm. Figure 1. VAAT predicted optimum ablation targets are associated with a greater prevalence of EGM abnormalities and longer EGM duration than non-predicted, voltage-matched areas. Computational modelling may aid in the detection of critical substrate in scar-dependent VT.Tabled 1Table 1: Comparison of EGM parameters and abnormalities during RV pacing within areas of bipolar low voltage (<1.5mV) between predicted and non-predicted areas.ParameterPredicted (n = 1354)Other (n = 2658)p-valueBipolar voltage (mV)0.81 (± 0.37)0.79 (± 0.37)< 0.001Fractionated EGM514/1354 (38.0%)441/2648 (16.6%)< 0.001Double potential62/1354 (4.6%)54/2648 (2.0%)< 0.001Late potential48/1354 (3.5%)18/2648 (0.7%)< 0.001Other LAVA104/1354 (7.7%)50/2648 (1.9%)< 0.001Any EGM abnormality574/1354 (42.4%)512/2648 (19.3%)< 0.001EGM duration (ms)73.90 (± 27.06)62.68 (± 21.05)< 0.001 Open table in a new tab