Background: Virtual models of cardiac electrophysiology which are capable of replicating like-for-like clinical observations for a given patient are referred to as cardiac digital twins. Due to their immanent predictive potential, especially in electrophysiology there are high expectations concerning this tool as assistance for clinical decision making and safe/ ethical testing for new electrophysiology techniques and ablation lines. Complex procedures such as VT ablation might benefit from pre-interventional testing using a digital twin of the heart.

Methods and Results: Clinical data of one patient with ischemic cardiomyopathy due to posterior AMI and recurrent VTs was retrospectively collected. ECG showed pathological Q-waves in the inferior leads. Pre-interventional LGE-MRI showed posterior transmural scar. We constructed a personalized 3-D model of the heart using the pre-ablation LGE MRI and 12-lead ECG. Automatic biventricular chamber segmentation with visualization of the transmural scar and border zones was performed. The activation times of the ventricular myocardium during sinus rhythm were estimated by the digital heart. Latest activation times were within the scar. Virtual ECG traces produced by the model had good qualitative agreement with clinical measurements (1ms error in QRS duration, 2ms error in QT interval, 17 degrees error in QRS axis). In the model, we performed virtual ramp pacing with cycle length at 600ms from different locations in the LV endocardium to induce VT. VT was defined as being inducible when it persists for >30s. In the model, a VT was inducible. Visual inspection of the VT circuit predicted by the model showed a re-entry path through scar tissue and allowed the definition of candidate locations for RF ablation. In a retrospective analysis, ablation lesion geometry was reconstructed from post-intervention LGE MRI and incorporated in a post-ablation model of the heart as non-conductive tissue. Ramp pacing was repeated in the post-ablation model resulting in unsuccessful VT inducibility of the clinical VT. Non-invasive programmed stimulation in the real patient was performed 3 days after VT ablation with no inducible VT. Furthermore, during follow-up of 26 months the patient had no further VTs.

Conclusions: Our results suggest that digital twin of the heart technologies have the potential to virtually predict VT inducibility and therefore could be useful to guide optimal ablation targets in patients with ventricular tachycardias. The concepts and information presented in this abstract are based on research results that are not commercially available. Future availability cannot be guaranteed.