Applying ablation index to catheter ablation of ventricular tachycardia: The search for the more perfect ablation lesion continues

Catheter ablation has emerged as a mainstay therapy for the treatment of scar-related ventricular tachycardia (VT). However, despite advances in ablation technology and techniques to identify and prioritize critical ablation targets, VT ablation with radiofrequency energy remains associated with prolonged procedure times and significant risks of complications. Achieving the appropriate balance between ensuring a transmural lesion that can abolish conduction through critical isthmuses of VT reentry while avoiding unnecessary ablation and the risks of collateral injury remains a critical goal of point-by-point radiofrequency ablation of VT. Mathematical formulas such as the ablation index (AI) and lesion size index (LSI), that incorporate variables such as power, current, contact force, and duration during radiofrequency ablation energy delivery, have been developed in an effort to provide real-time estimation of lesion size.1, 2 The clinical use of AI and LSI to date has mainly been applied to the ablation of atrial tissue during pulmonary vein (PV) isolation. Several studies have shown that the use of specific AI-goals during PV isolation can lead to increased rates of first-pass isolation as well as reduced PV reconnection and atrial fibrillation recurrence.3, 4 Furthermore, the use of an AI-guided high power-short duration ablation strategy for PV isolation has been associated with decreased procedure times. However, the applicability of AI and LSI guidance for delivery of radiofrequency ablation in ventricular tissue has not been well-studied. In a single retrospective study of patients undergoing catheter ablation of idiopathic outflow tract premature ventricular complexes, higher maximum and mean AI values achieved during the procedure was associated with increased rates of acute as well as 6-month success.5 To date, there are little to no clinical data on the clinical utility of AI to guide radiofrequency ablation of VT associated with structural heart disease. In this issue of the Journal of Cardiovascular Electrophysiology, Huang et al.6 report the results from a retrospective, single-center cohort study of patients undergoing catheter ablation of VT associated with structural heart disease. Of 103 consecutive patients undergoing VT ablation, 74 patients were included in the final study analysis which comprised 37 patients in an AI-guided ablation group and 37 patients in a non-AI-guided ablation group who were propensity-matched for baseline characteristics. In this study, AI-guided ablation was defined by adherence to an ablation index (CARTO 3; Biosense Webster) cut-off of 550 for stopping energy delivery. Non-AI-guided ablation was performed as per conventional operator practice. There were no protocols for specific power delivery in either treatment group. The two patient cohorts were reasonably well-matched using propensity scores and were predominantly male and had ischemic cardiomyopathy. Left ventricular ejection fractions were not reported in the study groups and there was a numerically higher proportion of patients in the non-AI-guided cohort who required inotropic and mechanical circulatory support during the VT ablation procedure. A variety of VT ablation strategies including scar homogenization, core isolation, scar dechanneling, and late potential/local abnormal ventricular activity ablation was used. The authors found that compared to the non-AI-guided group, patients in the AI-guided group had shorter procedure times, lower average radiofrequency ablation duration per lesion and decreased total intraprocedural fluids administered. There were no differences in acute outcomes between the two groups with respect to clinical VT inducibility and procedural complications. During follow-up, the rates of VT recurrence and appropriate device therapy were similar between the two treatment groups. This study provides needed insights into the clinical feasibility and impact of using AI to guide VT associated with structural heart disease. By using AI to limit ablation lesion duration, the authors found shorter procedure times and no apparent differences in VT inducibility and clinical recurrence. However, several caveats should be discussed. First, it is important to note that for the AI-guided treatment group in this study, AI was not used as a target but rather as a limit to ablation. The median AI measurements among all ablation lesions in the AI-guided cohort was 504, which was well short of 550. Notably, the median AI per lesion in the non-AI-guided group was higher at 517 but was still <550. Therefore, the AI-guided cohort may be more aptly described as an “AI-limited cohort.” Second, power delivery was not balanced between the two treatment groups. The maximal power delivery was significantly higher in the AI-guided group compared to non-AI group (46 W vs. 42 W; p < .001). This was likely the key driver of the shorter ablation lesion durations in the AI group as there were no significant differences in mean contact force between the two groups. The differences in power delivery may reflect fact that the AI and non-AI group patients were likely not contemporaneously treated and raises the possibility that changes in operator practice involving higher power delivery occurred independently of the use of AI guidance. Third, in addition to the likely presence of unmeasured confounders, there was heterogeneity in patient characteristics and VT ablation strategies between treatment groups, even with propensity score matching. Finally, the patient sample sizes were not powered sufficiently and the clinical follow-up was too limited to permit conclusions to be made about the lack of differences in VT recurrence or readmissions. There may be fundamental reasons why the application of ablation index to guide radiofrequency ablation of ventricular tissue, particularly scarred myocardium, may have limitations. A recent study by Younis et al.7 found that varying power delivery to achieve the same AI goal can result in significantly different lesion sizes. Using ex vivo experiments with healthy swine ventricles, they found that for any fixed target AI ≥ 600, the delivery of 40 W led to shorter ablation lesions compared to the delivery of 30 W, but at the expense of significantly smaller lesions. In the current study by Huang et al.6 higher power delivery, together with an AI limit of 550, was a likely key driver of shorter ablation lesions. While a high-power short duration strategy with AI guidance may be effective with PV isolation, employing a similar strategy with ablating ventricular tissue may lead to suboptimal ablation lesions. Thicker ventricular tissue may require lower power delivery for longer duration to ensure sufficient conductive tissue heating to permit deeper penetration. Furthermore, Younis and colleagues found that with in vivo beating heart models, AI values did not correlate well with lesion depth, particularly in infarcted ventricular myocardium. This underscores how factors such as tissue architecture and scar, local tissue impedance, and surrounding blood flow are all determinants of radiofrequency ablation lesion size that are not accounted for in the ablation index. In summary, optimization of energy delivery during catheter-based radiofrequency ablation represents a major pillar for improving efficacy and safety in our treatment of patients with VT and structural heart disease. Huang et al. should be commended for studying the feasibility of introducing the concept of ablation index into the ventricular space. Improvements in our understanding of the biophysics of radiofrequency ablation of ventricular tissue of varying thickness and tissue characteristics should drive continued efforts to improve real-time modeling of lesion adequacy. Whether the incorporation of other variables such as temperature and current output into mathematical models or the introduction of novel imaging techniques will hold the key to this remains to be seen. Studies such as the those by Huang and colleagues are important first steps for highlighting the gaps that need to be addressed in our search for the more perfect ablation lesion.