Substrate Mapping and Ablation for Ventricular Tachycardia in Patients with Structural Heart Disease: How to Identify Ventricular Tachycardia Substrate

Catheter ablation for ventricular tachycardia (VT) has been increasingly used over the past two decades in patients with structural heart disease (SHD). In these individuals, a substrate mapping strategy is being more commonly applied to identify targets for VT ablation, which has been shown to be more effective versus targeting mappable VTs alone. There are a number of substrate mapping methods in existence that aim to explore potential VT isthmuses, although their success rates vary. Most of the reported electrogram-based mapping studies have been performed with ablation catheters; meanwhile, the use of multipolar mapping catheters with smaller electrodes and closer interelectrode spacing has emerged, which allows for an assessment of detailed near-field abnormal electrograms at a higher resolution. Another recent advancement has occurred in the use of imaging techniques in VT ablation, particularly in refining the substrate. The goal of this paper is to review the key developments and limitations of current mapping strategies of substrate-based VT ablation and their outcomes. In addition, we briefly summarize the role of cardiac imaging in delineating VT substrate.


Introduction
Over the last decade, catheter ablation for ventricular tachycardia (VT) has been increasingly performed as an adjunctive therapy to antiarrhythmic drugs. 1,2 With this came an increased understanding of the mechanisms of VT; notably, it commonly results due to scar-related reentry in patients with structural heart disease (SHD). Activation mapping and entrainment mapping are reasonable approaches to identify and target critical sites of the reentrant VT circuit for ablation in patients with tolerated reentrant VT. [3][4][5][6] However, the majority of patients presenting for catheter ablation have unstable VT that hampers the accurate definition of the critical areas of the reentrant circuit with activation or entrainment mapping. 7,8 Thus, substrate ablation is increasingly favored as a VT ablation strategy, with or without entrainment/activation mapping methods. Substrate-based approaches involve the identification of local abnormal ventricular electrograms that represent diseased areas consistent with potential isthmuses capable of supporting reentrant VT and may be followed even when VTs are not inducible or not hemodynamically tolerated. 9,10 Although a combination of several approaches is commonly employed during VT ablation, a number of studies have examined with variable results whether a substrate-based ablation strategy may be superior or comparable to one guided predominantly by the activation/entrainment mapping of inducible and hemodynamically tolerated VTs. [11][12][13][14][15][16][17][18] Critical in achieving more successful results with a substrate-based approach is an accurate representation of the pathologic substrate; various strategies to delineate this have been proposed to date. 4,10,[13][14][15][16][17][18][19][20][21][22] With the advent of three-dimensional (3D) electroanatomical mapping (EAM) systems in the 1990s, 23 there has been a significant improvement in our ability to represent both anatomical and functional electrical information in a realtime model of the ventricle. Furthermore, the development of multipolar catheters facilitates precise scar detection, [24][25][26] with potentially more successful results. 27 In addition, cardiac imaging in the form of cardiac magnetic resonance (CMR) imaging or multidetector computed tomography (MDCT) may play a potentially important role in the preprocedural assessment of cardiac anatomy and myocardial substrate and the intraprocedural integration of the structural and electrophysiological VT substrate. [28][29][30][31] In this paper, we review substrate mapping and ablation strategies and clinical outcomes for VTs in the setting of SHD and also discuss the importance of modalities for substrate detection in addition to those based on electrograms.

Electrogram-based substrate detection
The identification and modification of the arrhythmogenic substrate in the endocardium and/or epicardium are increasingly considered as composing a primary ablation strategy in patients with SHD. This technique was originally developed in the absence of 3D mapping 9,32-34 ; however, the development of 3D mapping systems in the late 1990s 23 accelerated the use of electrogram-based techniques in detecting and localizing substrate reproducibly and feasibly. Furthermore, the recent use of ultra-highdensity mapping with catheters with multiple small electrodes and closer interelectrode spacing has enhanced the speed, density, resolution, and detailed near-field signal assessment of mapping acquisition, reducing interpolation and possibly improving clinical outcomes [25][26][27][35][36][37][38] (Figure 1).
Bipolar voltage mapping to evaluate the electrogram peak-to-peak voltage is a widely accepted and frequently used technique to characterize substrate. Endocardially, a bipolar voltage amplitude of 1.5 mV or more normally identifies healthy tissue. Additionally, normal ventricular myocardial bipolar electrograms are defined as sharp, biphasic or triphasic signals with a duration of 70 ms or less and/or with an amplitude-to-duration ratio of more than 0.046. 9,39 Areas with voltages of 0.5 mV to 1.5 mV are often considered as border zones in the setting of a 3.5-mm to 4-mm-tip, 1-mm ring electrode, and 2-mm interelectrode spacing mapping catheter (Table 1), 10,32,34 even in the right ventricle. 40 The definitions above have been validated by human pathologic data and radiologic studies. [41][42][43] Areas with voltages of less than 0.5 mV are generally considered as "dense scar"; however, low-amplitude abnormal electrograms are frequently observed in these areas. 44 Therefore, in order to define unexcitable scar, the area should contain no visible electrograms (ideally using mapping catheters with smaller and narrower-spaced bipolar electrodes) and have no local pacing capture. 45 In the epicardium, a bipolar voltage cutoff of 1 mV or more 46 or 1.5 mV 47 is considered normal. With regard to the right ventricle, epicardial 1.5 mV can also be a reasonable bipolar voltage cutoff. 40 The majority of VTs have critical circuits located in the scar border zone, 48 which harbors abnormal electrograms 9,32,33 [ie, fractionated potentials, 49 double potentials, and late potentials (LPs), discrete and separated from the QRS by 40 ms 50 ], which can be targeted by catheter ablation. However, it has been reported that a certain proportion of abnormal potentials are also located in regions with bipolar voltages of more than 1.5 mV, 19,21 with abnormal electrograms occasionally unmasked by extrastimuli. 51,52 We have recorded at least 3% of substrate defined as local abnormal ventricular activity (LAVA) in voltage zones of more than 1.5 mV (because of far-field signal annotation) 53 (Figures 2A-2D). Moreover, Tung et al. 54 found that 18% of critical VT isthmuses were within low-voltage areas during pacing from the site but Comparison of bipolar voltage maps (endocardial: scar < 0.5 mV, border zone 0.5-1.5 mV, and healthy tissue > 1.5 mV; epicardial: border zone 0.5-1 mV and healthy tissue > 1 mV) using Navistar ® (Biosense Webster, Diamond Bar, CA, USA) (NAV) mapping versus PentaRay ® (Biosense Webster, Diamond Bar, CA, USA) (PR) mapping in the endocardium (A) and epicardium (B) of a sheep model with an iatrogenic-created anteroseptal scar and in humans (C) using the CARTO ® 3 system (Biosense Webster, Diamond Bar, CA, USA). All images are shown in an anteroposterior view. A: LAVA is represented in pink, the proximal conduction (left-sided His) system is in blue, and the Purkinje system is in yellow. Border zone areas and LAVA channels are visible within the scar with PR mapping, but none are visible with NAV mapping. B: Three LAVA channels are visible with PR mapping but none with NAV mapping. The border zone is smaller and demonstrates increased detail using PR mapping. C: Larger scar area and improved border zone definition using PR versus NAV mapping. Right: Point pair analysis (≤ 3 mm of distance between a PR and NAV point) from two examples by manual signal analysis in two different patients. Substrate maps are shown at 100% transparence. LAVA using PR (in purple) and NAV (in white) are tagged. The distance between tags is measured using the distance measurement tool in the mapping system. A red arrow indicates a clear LAVA visible with PR mapping but one that is barely or not visible with NAV mapping. Reproduced with permission from: Berte B, Relan J, Sacher  within normal amplitude (> 1.5 mV) areas with pacing from another site, indicating that voltage is affected by the activation wavefront. [54][55][56] Therefore, a voltage map based on standard (0.5-1.5 mV) voltage criteria is not necessarily capable of delineating the entire possible VT substrate.
Furthermore, with the use of high-density mapping with small electrodes and narrower interelectrode spacing ( Figure 1), traditional definitions of substrate voltage need to be adjusted depending on the mapping catheter. 24,36 The electric field recorded by a pair of electrodes on a novel mapping catheter is relatively small, recording precise local signals located just underneath the electrodes. 57,58 Meanwhile, endocardial unipolar voltage mapping has a large field of view and is useful to identify septal, intramural, and/or epicardial substrate, [58][59][60][61][62][63][64][65][66][67][68][69] with different amplitude thresholds present depending on the etiology of the disease. The use of bipolar mapping with small electrodes and closer interelectrode spacing in combination with unipolar mapping may constitute an optimal strategy.
In summary, there are several limitations 55 of conventional voltage mapping for substrate detection, with variations occurring due to the wavefront of activation, 54-56 catheter interelectrode spacing, 24,25 interpolation, 63 far-field peak annotation of multicomponent electrograms, catheter orientation 64 and contact, 65,66 and surrounding insulating tissue (eg, fat, edema). Conversely, nonelectrogram techniques of substrate detection such as cardiac imaging 29,30,[67][68][69] are unaffected by directions of wavefront activation or techniques based on specific electrogram characterizations (see later).

Substrate Mapping and Ablation for VT in SHD
The Journal of Innovations in Cardiac Rhythm Management, March 2019  Lastly, omnipolar mapping is a new development that may resolve some of these limitations by providing instantaneous catheter-tip wavefront direction and speed. 70,71 With this mapping technology, local electrical field signals are determined and used to assess the traveling wavefront on a multielectrode catheter, rather than activation-based data acquisition, which may allow for a beat-to-beat determination of wave propagation information that is independent of electrode orientation or activation time. Further clinical validation of this technology is ongoing.

Specific electrogram-based substrate ablation strategies
Substrate-based ablation approaches may differ between VT ablation centers. A number of methods are implemented during substrate modification, with a large variation in the amount of ablation energy delivered according to the preset mapping and endpoints of the procedure. The major strategies are summarized in Table 1. These studies must be cautiously interpreted because the mapping details or endpoints are often heterogeneous, even for the same strategy. In addition, although the advantage of multipolar mapping catheters has now been recognized, many previous studies have used ablation catheters for mapping.

Local abnormal ventricular activity-guided ablation.
In a seminal study by our group, we reported a mapping and ablation strategy to homogenize substrate defined as LAVAs (Figure 2). [19][20][21] Elimination of all LAVAs is associated with improved midterm and long-term arrhythmia-free survival. [19][20][21] LAVAs are defined as sharp, high-frequency ventricular potentials, possibly of low amplitude, that are distinct from the far-field ventricular electrogram that occurs at any time during or after the far-field ventricular electrogram in sinus rhythm or before the far-field ventricular electrogram during VT, which sometimes display fractionation or double or multiple components separated by very-low-amplitude signals or an isoelectric interval and which are poorly coupled to the rest of the myocardium. 19,21 Importantly, this strategy also targets abnormal substrate in so-called normal-voltage areas, although most LAVAs are generally observed in low-voltage areas. 53 Clinical outcomes have been reported as including a 55% (88/159) VT freedom rate during a median follow-up of 47 months (range: 33-82 months) without antiarrhythmic drug therapy except for β-blockers. 20

Linear ablation with cross-section of the scar and border-zone.
Marchlinski et al. 10 first described the use of linear ablation lesions to target multiple unmappable VTs. The technique involved the creation of contiguous lesions from the dense infarct area through the infarct border-zone and anchored to anatomic barriers or healthy myocardium. In their study, 10 77 reported that, after a mean follow-up of 13 months ± four months, VT recurred in 10 patients (20%).
Scar homogenization. Di Biase et al. 17 reported on a scar homogenization approach targeting all abnormal electrograms within low-voltage areas defined with conventional bipolar voltage criteria when mapping in sinus or paced rhythm. With this approach, abnormal electrograms are defined as any electrograms that have more than three deflections, an amplitude of less than 1.5 mV, and a duration of more than 70 ms. The acute ablation endpoint was either the elimination of abnormal electrograms or the loss of local capture at high-output pacing (20 mA output at a 10-ms pulse width). This approach can potentially eliminate more possible critical sites than more focused mapping approaches, but has limitations in patients with massive substrate, particularly under unstable conditions. During a mean follow-up of 25 months ± 10 months, the freedom from VT recurrence rate was 81% (35/43) in patients with ICM who showed scar homogenization. 17 More recently, the results of a multicenter randomized study comparing scar homogenization with standard limited substrate ablation in patients with ICM were reported. 11 At one year of follow-up, freedom from VT recurrence was achieved in 52% (31/60) of patients who underwent clinical VT ablation only versus in 85% (49/58) of patients who underwent scar homogenization.
Border-zone ablation/core isolation. The core isolation approach was recently developed by Tzou et al. 79 in an effort to limit the number of lesions required to eliminate all of the areas critical for VT maintenance within the dense scar. This is a stepwise approach that starts with the identification of the potential critical isthmus within the dense scar that is related to the patient's clinical and/ or induced VTs based on conventional criteria including voltage channels; sites with LPs; sites with good pacemaps; and the existence of long stimulus to QRS intervals, isthmus sites defined by entrainment mapping, and sites of VT termination with ablation. Therefore, this approach acts as a combined approach between conventional and substrate-based approaches. These areas are typically within areas of dense scar (< 0.5 mV). Once identified, the critical area is targeted with contiguous ablation lesions either completely surrounding the region of interest or by using anatomic anchors to minimize the amount of ablation necessary. The authors demonstrated that, after a mean follow-up of 18 months ± nine months, no VT recurrence was observed in 38 (86%) of 44 patients. 79 Scar dechanneling. This substrate ablation approach, which targets channels within the abnormal substrate, was originally described by Soejima et al. 45 and Arenal et al. 80 Although, in all studies, the concept of scar dechanneling encompasses targeting the VT channels, the identification of the channels differs in terms of technique. In the study by Soejima et al., 45 channels were identified within the low-voltage area using high-output pacing (10 mA, with pulse width of 2 ms). Electrically unexcitable scar was defined as a loss of capture at high-output pacing and marked on the voltage maps. Conversely, Arenal et al. 80 were able to visualize channels after adjusting voltage cutoffs on EAM. More recently, Tung et al. 81 and Berruezo et al. 82 described an approach that targets interconnected activation channels within the abnormal substrate, adopting clear endpoints with clinical follow-up. The method involved high-density mapping of the channels of activation of LPs. Once a specific sequence of LP

Frequency analysis mapping
High-frequency electrogram components are more often associated with critical sites of reentry as compared with low-frequency, large-amplitude components. Several studies have demonstrated that the frequency analysis of electrograms may aid with substrate identification 84-86 ; however, this analysis is still only available as an offline tool and the feasibility of an automated real-time tool needs to be further investigated.

Use of imaging to identify substrate
Cardiac imaging may play an important role in the preprocedural assessment of cardiac anatomy and myocardial scar as well as in the intraprocedural integration of the structural VT substrate. 28 Cardiac imaging has been mainly used as an adjunct either offline or online (realtime image integration) 27,83,87,88 to support the localization of scar in 3D mapping systems during substrate mapping and ablation (Figure 3). Cardiac imaging has several advantages, as follows: (1) it may provide precise anatomical information including endocardial/ intramural/epicardial scar location, while EAM systems can only provide derived 3D reconstructions from  catheter-electrode contact at the myocardial surface; (2) there is no possibility of inaccuracy due to extrapolation, lack of catheter contact, confounding effects of far-field electrograms, or the influence of wavefront activations; (3) it provides information about adjacent anatomical structures, which may affect mapping and ablation, such as papillary muscles, coronary arteries, phrenic nerves, and epicardial fat. However, there are also limitations in imaging techniques in terms of feasibility (eg, magnetic resonance imaging in some patients with old ICDs, MDCT in patients with severe chronic kidney failure), the availability of images with 3D mapping systems, and registration issues.
Recently, in an attempt to refine targeted VT ablation strategies further, several studies have focused on identifying specific scar regions that harbor critical VT isthmuses. 29,67,89 At this time, scar regions with increased transmurality, scar border zones, and regions at the scar-core-border-zone transition point have been identified as potential targets. 29 Other modalities such as intracardiac echocardiography 96,97 and nuclear imaging 98,99 may also help to identify substrate. In addition, electrocardiographic imaging incorporated with CMR or MDCT has the potential to identify VT isthmuses noninvasively. 100 Furthermore, an entirely noninvasive approach that combines anatomical imaging with electrocardiographic imaging and noninvasive cardiac radiotherapy for ablation has been reported, which represents another intriguing strategy that employs cardiac imaging. 101

Clinical outcomes of substrate-based ablation
Overall, the clinical outcome of substrate-based ablation is a VT recurrence-free rate of approximately 54% to 91% in mid-to long-term follow-up (Table 1). As described above, the success rates, however, vary widely with different strategies and across studies. We have also compared clinical outcomes between substrateand nonsubstrate-based ablation (  102 also conducted a randomized study comparing substrate-based ablation to conventional ablation and demonstrated comparable VT-free survival rates between the two strategies. A recent meta-analysis has also revealed similar acute procedural efficacy, complication, VT recurrence, and mortality rates while comparing a predominantly substrate-based ablation strategy to a strategy guided by activation and entrainment mapping of inducible and hemodynamically tolerated VTs. 103 A separate meta-analysis 104 also showed a significantly lower risk of the composite primary outcome of longterm VA recurrence and all-cause mortality among those undergoing substrate modification in comparison with standard ablation in a cohort of mostly patients with ICM. Furthermore, in this study, long-term success was improved when performing complete substrate modification. 104 Hence, substrate ablation may be superior to a conventional strategy in terms of VT recurrence when extensive substrate ablation is performed. 11 However, despite the substantial progress that has been made in the use of cardiac imaging to guide VT ablation, there is insufficient evidence at present to suggest that the use of imaging can add value to clinical outcomes. Although observational, nonrandomized studies suggest that image integration may have an impact on procedural outcomes, 27,83,87,88 well-designed, prospective randomized studies are needed to assess the true impact of image integration as well as to evaluate the potential mechanism(s) of any benefit.

Conclusion
Several substrate-based ablation strategies have been developed, which include extensive or less-extensive ablation lesions according to the preset mapping and endpoints of the procedure. Although multipolar mapping catheters with smaller and more-narrowly-spaced bipolar electrodes are now widely used, most currently available studies use data acquired by way of ablation catheters. Advances in cardiac imaging may be helpful in providing refined anatomical substrate details.
At this time, clinical outcomes of substrate-based ablation are at least comparable with and possibly superior to conventional VT ablation. The further development of mapping technologies, cardiac imaging, and novel modalities and the incorporation of these modalities in delineating VT substrate may additionally improve the clinical outcomes of substrate-based ablation.