The principal findings of our study are:

1.

DEEP-assisted ablation of scar-related ventricular tachycardia is a feasible strategy with comparable acute success rates to fixed-substrate-based techniques.

2.

Although VT recurrence rates in our study after an average follow-up period of 12 months were as high as 65%, the rates are equal in the DEEP-guided and LP-guided groups, denoting the absence of significant difference between the 2 strategies in terms of long-term efficacy.

3.

DEEP-guided ablation of scar-related VT is associated with rates of all-cause mortality and cardiovascular mortality as well as time to first VT, which are similar to a fixed substrate-guided strategy. However, the procedural mortality was higher in the DEEP group (10% vs. 0%), a difference which, despite being non-significant, warrants further evaluation in large RCTs. Longer procedural times in this arm might have been a contributor.

4.

In addition, there is a significant reduction in the number of ablation points, which is at the expense of a longer procedural time. The difference in mean procedural time between the DEEP and the non-DEEP groups was 65 min, which is a considerable figure despite missing statistical significance. Further experience with the technique, in addition to utilizing novel mapping features are expected to remarkably shorter procedural time.

To our knowledge, this is the first randomized double-armed prospective trial addressing the feasibility of DEEP-guided ablation of scar-related ventricular tachycardia in a head-to-head comparison with a fixed-substrate-centered strategy. Traditionally, conventional mapping techniques, namely activation and entrainment mapping, have been the gold standard for VT ablation [20]. With 70% of VTs being non-inducible or hemodynamically not tolerated, several substrate-based techniques have been devised as an alternative strategy, targeting EGM abnormalities during stable rhythm (sinus rhythm or paced rhythm).

Jaïs et al. demonstrated the feasibility and safety of targeting local abnormal ventricular activity (LAVA) elimination as a procedural endpoint with a significant reduction in the combined endpoint of VT recurrence or death [21]. Elimination of abnormal potentials has been compared to conventional mapping techniques and has proven superiority in terms of long-term recurrence (15.5% vs. 48.3%, respectively; log-rank p < 0.001) [22]. A recurrence rate of 41.4% was reported in another trial after a mean follow-up duration of 3.14 years postablation, but with a significant reduction in the burden of ICD shocks and VTs even in the patient cohort that experienced VT recurrence [23]. Another single-armed prospective trial of LP abolition strategy has shown an acute success rate of 71.4% and a long-term VT recurrence rate of 20% [24]. Of note, the recurrence rate reached 75% in patients with incomplete LP abolition. Incomplete LP abolition is an independent predictor of VT recurrence and is attributed to a septal location of the substrate with proximity to the conductive system, higher LV mass, and the use of conventional—as compared to high-density—mapping catheters [25]. In our study, the LP elimination arm had an acute success rate of 85%, which is midway between that reported by Vergara et al. (71.4%) [24] and Roca-Luque et al. (71%) on one side [25] and that reported by Luigi et al. (100%). [22].

Although effective, substrate mapping protocols are not without limitations. The identification of abnormal EGMs is prone to inter-observer variability. Besides, targeting all abnormal electrograms can lead to extensive ablation that often involves myocardial areas not incriminated in the VT circuit. The deleterious effect of such a strategy on the global ventricular function has not been ruled out, although in our study there was no significant reduction in the mean EF after 12 months in either group. Over and above, unnecessary ablation lesions can theoretically create areas of partial myocardial viability and slow conduction, calling forth new substrates for reentry. Functional substrate mapping techniques have been introduced with the aim of achieving optimal short- and long-term outcomes with the least number of ablation lesions.

In this paper, we compared a DEEP-guided functional mapping strategy to conventional static substrate mapping. The theory behind DEEP mapping is based on the observation that decrement precedes unidirectional functional block leading to reentry, as has been demonstrated in atrial tissue specimens [26]. The mechanism was reproduced in infarcted human hearts where unidirectional block preceded the initiation of sustained monomorphic VT [27].

The electro-physiologic implication of this notion has been demonstrated in the mechanistic study by Jackson et al., where DEEPs were found to colocalize with the diastolic corridor of VT with higher specificity than late potentials with no significant difference in sensitivity [28]. Such findings paved the way for testing such a strategy in the clinical setting. Porta-Sanchez et al. performed DEEP-guided ablation of VT in 20 patients with ischemic cardiomyopathy [8]. The percentage of LPs and DEEPs were 16.8% and 4.8%, respectively, which are comparable to ours. Likewise, the rate of VT non-inducibility was 80%. At a 6-month follow-up, 75% of patients were free from VT compared to 60% in the DEEP arm of our study.

Although a DEEP-guided ablation strategy is plausible, it inherently has several downsides. First of all, repeated pacing in patients with a depressed ventricular function that is aggravated by recurrent VT episodes is not without risk. Furthermore, electrophysiologically speaking, the magnitude of delay is subject to several variables, including the pacing location, the characteristics of the conducting channel as well as the coupling interval of the extrastimulus [29, 30]. Stimulation in the vicinity of the scar tissue from the side opposite to the entrance site produces longer delays. Shorter delays occur with pacing at farther sites from the scar and sites close to conduction barriers [29]. Potentials in unprotected conducting channels with multiple side branching have shorter DEEP delays compared to protected channels with fewer additional side branches [29].

A recently published study has shown that delivering an S2 at 400 ms and using a 20-ms decrement threshold for defining DEEP resulted in improved specificity for the identification of VT isthmuses in 13 patients with ischemic VT without compromising sensitivity, leading to a 59% reduction in the area targeted for ablation. At the 6-month-follow-up, freedom from device-detected VTs was 92% [30]. The paced electrogram feature analysis (PEFA) technique accounts for the electrogram duration as well as latency in response to RV apical stimulation at variable coupling intervals to identify VT isthmuses and has shown promising VT-freedom rates [13].

Another argument against the DEEP strategy is that drive train pacing before extra-stimulation allows for the adaptation of sodium channels [31], a phenomenon that may mask the abrupt changes in conduction properties that occur in the real world when single extrasystoles initiate VTs. In 2020, Srinivasan et al. introduced the Barts Sense Protocol, which entails high-density mapping during sinus rhythm and with single-sensed extrastimuli [12]. Interestingly, the Barts Sense Protocol could identify an area of ablation that is larger than that identified during intrinsic rhythm, with a highly specific correlation to the critical isthmus (96%). The authors attributed such contradiction with DEEP mapping results to the advantage of the absence of tissue adaptation to extrastimulation with single-sensed beats [12]. With employing such a protocol, the rate of VT non-inducibility was 96%, and 90% of patients were free from VT at 12 months of follow-up.

Limitations

The mean age in our study was 54 years, which is younger than the usual mean age in prior scar-related VT studies. A possible explanation is that our cohort included patients with congenital heart diseases, ARVC, and dilated non-ischemic cardiomyopathies. Such patient categories are generally younger when they present with VTs than the ischemic population represented in most of the published literature.

Besides, 45% of our patients had a presentation of VT storm. This is a remarkably high percentage, given that most of the time, VT ablation is performed on an elective basis for recurrent VT. This is correct for regions where the availability of resources is not problematic. However, in our country, where the resources for VT ablation are limited, patients presenting with VT storm have the highest priority, while patients with 1st or recurrent VT episodes are given every trial of medical management before referral for elective ablation.

Our study has shown a 1-year VT recurrence rate of 65% after DEEP-guided ablation. Several factors are proposed to have contributed to such high rates. In our current study, the multi-electrode Pentaray catheter was only used in 15% of patients in each group. The remaining cases underwent point-by-point mapping using conventional bipolar catheters. In the study by Porta-Sanchez et al., mapping was performed by multipolar catheters in 80% of cases [8]. This factor might have contributed to the high recurrence rates in our study, given the published data that suggest a correlation between the density of mapping and ICD-therapy-free survival [32]. The ability of the catheter to detect abnormal electrograms depends on the size of electrodes, spacing between electrodes and wavefront direction in relation to bipolar pairs [33, 34].

Another possible cause of higher recurrence rates is that, unlike the study by Porta-Sanchez et al., our study included patients with non-ischemic etiology of VT (45% of the study sample and 40% of the DEEP arm). Long-term outcomes of VT ablation in such a cohort are generally less favorable [35]. Moreover, our patient population had a pre-procedural burden of shocks and ICD therapies that is to some extent higher than that in the study by Porta-Sanchez et al. This might highlight the possibility that our patients had a more aggressive disease that was subsequently translated to poorer outcomes. We believe that the relatively high procedural complication rate in our study (15%) is at least in part explained by the instability of our patients’ condition prior to and during the procedure.

Another crucial limitation of our study is that although randomized, the trial arms showed an imbalance in terms of endocardial versus epicardial mapping. More patients in the non-DEEP group underwent epicardial mapping, while the number of mapped points was similar between the 2 groups, which carries the possibility that the non-DEEP group had less dense mapping. This issue should be considered while drawing conclusions from our study.

Last but not least, our trial was single-centered with limited sample size. We recommend further large-scale multi-center randomized trials that compare the several available functional substrate mapping techniques, integrating the ever-developing mapping technology with our growing understanding of VT electrophysiology to obtain the optimum patient outcomes.