Automated Ventricular Substrate Mapping—Evaluation in an Ovine Chronic Myocardial Infarction Model

Introduction: We hypothesized that automated electrogram analysis might enable rapid localization of ventricular scar. This would allow the delivery of interventions such as radiofrequency ablation or therapeutic agents to critical areas within the scar and scar periphery. Methods: Substrate mapping was performed on seven sheep 36.5 ± 32.9 weeks after a left anterior descending artery myocardial infarction had been induced. Contact electrograms and the mapping catheter three-dimensional (3D) location were recorded simultaneously. A computer program was written in-house to automatically identify sinus beats, analyze electrogram characteristics (e.g., electrogram amplitude and minimum slope), and integrate the analysis results into a 3D scar map. Results: The total time required to produce the scar maps was a mean of 8.3 ± 2.0 minutes. The automated substrate mapping (ASM) system beat detection algorithm had a high sensitivity (i.e., detected 87.4% of the recorded beats) and excellent specificity (only one false activation over 58.2 minutes of total recorded data). The system was able to classify the detected beats (‘sinus’ or ‘ectopic’) with high specificity (specificity = 97.3% confidence interval (CI): 96.9–97.7) and moderate sensitivity (sensitivity = 78.3% CI: 77.3%–79.5%). The scar area identified by the ASM system correlated well with the pathologically defined scar area (R2= 0.87 p < 0.001). Conclusions: ASM enables accurate scar maps to be produced rapidly. This strategy may play an important role for both clinical and research applications, allowing therapeutic agents and radiofrequency ablation to be delivered to critical locations in and around ventricular scar.