In Silico Pace Mapping To Optimise Ventricular Tachycardia Ablation

Overview

People who suffer from incessant cardiac arrhythmias receive a small electrical device implanted into their chest that automatically senses when the heart beats arrhythmically and applies electrical pulse to re-establish normal activity. However, if problems persist, people can have an operation called catheter ablation therapy, which involves 'burning' small areas of the heart tissue in order to permanently disrupt the problematic electrical pathways driving these arrhythmias. However, procedure times and complication rates are high, whist success rates are punitively low (\~50% success), largely due to the significant challenge clinicians face in identifying the ideal 'target' to ablate within the patient's heart. In this project, the investigators aim to develop, and clinically validate, an in silico tool that reconstructs a personalised computational model of a patient's heart using advanced MRI data, upon which a virtual 'mapping' procedure is then performed in order to identify (in the model) the optimal ablation target. This pre-procedural planning tool utilises stored information about the patient's specific arrhythmia from their implanted device, ensuring optimal targets are selected. The approach aims to reduce procedure times whilst increasing their safety, and ensure significantly increased long-term effectiveness of these invasive ablation procedures, increasing survival rates and quality-of-life. This study is concerned with the clinical arm of the study, specifically, in the collection of data from patients in order to (retrospectively) validate the computational model. The model itself will not be applied or used to treat these patients.

Catheter ablation of ventricular tachycardia (VT) most frequently requires the identification of exit sites of slow-conducting diastolic isthmuses associated with infarcted tissue which are critical to sustain the reentrant arrhythmia. Pace mapping is a technique that identifies exit sites by matching the recorded 12-lead electrocardiogram (ECG) QRS of catheter-paced beats at different myocardial locations with the QRS of the clinical VT. When a high correlation between QRS morphologies is found, it is believed that the paced-beat lies at the VT exit site, helping guide the ablation. Procedures thus involve methodologically moving the catheter to multiple (accessible) locations to produce a pace map. Unfortunately the procedure is inherently limited due to the following reasons: i) Surface maps - Most often, pacing sites are limited to the endocardium (where catheter access is easiest). Consequently, identification of VTs with an intramural or epicardial substrate can be challenging, with mis-leading and/or difficult to interpret pace maps being obtained. ii) Low Resolution - Due to time-pressures and practical difficulties in catheter manipulation, pace map sites are often sparsely located (typically 40-60 separate sites). In complex structural VTs, with multiple possible anatomical circuits through the scar, it is often challenging to accurately identify from the map the relevant isthmus exit site responsible for the presenting clinical VT. iii) VT Induction - Pace mapping usually necessitates VT induction during the procedure, which is time-consuming and carries inherent risk (\~80% of induced VTs in ischemic heart disease (IHD) are either non-sustained or not haemodynamically tolerated), but induction of the exact clinical VT itself is often challenging. A non-invasive, pre-procedural approach that is capable of generating accurate, high-resolution 3-dimensional personalised pace maps that correlate with the presenting clinical VT without the need for VT induction would thus revolutionise VT ablation planning, guidance, safety and efficacy. Two recent studies have demonstrated the utility of performing a pace map based on comparison to stored implanted cardioverter defibrillator (ICD) electrograms (EGMs) of the clinical VT. The first study initially demonstrated that stored ICD EGMs allow differentiation of the clinical VT from other VTs, and their subsequent use in pace mapping may be useful for identifying the VT exit site. More recently, it has been prospectively shown that pace mapping of non-inducible clinical VTs based on ICD EGMs is feasible and resulted in higher freedom from recurrent VT, compared to targeting only inducible VTs. The investigators have recently published a pipeline for creating a patient-specific image-based whole torso-cardiac model to perform virtual ('in silico') pace mapping. The approach creates a high-resolution 3D pace map that successfully identifies the exit site of an induced (simulated) VT within the same model, using both simulated 12-lead ECG and ICD EGM data derived from the simulated VT episode. Here, it was demonstrated: 1) the importance of creating a dense, fully transmural (3D) pace map to allow differentiation of epicardial vs endocardial substrates; 2) that accuracy of exit site identification could be enhanced by considering multiple EGM sensing vectors, e.g. from multipolar cardiac resynchronisation devices (CRT-D) with additional epicardial electrodes. Developed on simulated data, this workflow now requires full validation with clinical data from ablation patients, as proposed here, and refinement of methodological algorithms.