Specific electrogram characteristics impact substrate ablation target area in patients with scar‐related ventricular tachycardia—insights from automated ultrahigh‐density mapping

Substrate‐based catheter ablation approaches to ventricular tachycardia (VT) focus on low‐voltage areas and abnormal electrograms. However, specific electrogram characteristics in sinus rhythm are not clearly defined and can be subject to variable interpretation. We analyzed the potential ablation target size using automatic abnormal electrogram detection and studied findings during substrate mapping in the VT isthmus area.

Disclosures: C. Meyer, S. Willems, P. Jaïs, F. Sacher, H. Estner report honoraria as consultants or speaker's fees relevant to this topic. The remaining authors declare no conflict of interests.
Methods and Results: Electrogram characteristics in 61 patients undergoing scarrelated VT ablation using ultrahigh-density 3D-mapping with a 64-electrode minibasket catheter were analyzed retrospectively. Forty-four complete substrate maps with a mean number of 10319 ± 889 points were acquired. Fractionated potentials detected by automated annotation and manual review were present in 43 ± 21% of the entire low-voltage area (<1.0 mV), highly fractionated potentials in 7 ± 8%, late potentials in 13 ± 15%, fractionated late potentials in 7 ± 9% and isolated late potentials in 2 ± 4%, respectively. Highly fractionated potentials (>10 ± 1 fractionations) were found in all isthmus areas of identified VT during substrate mapping, while isolated late potentials were distant from the critical isthmus area in 29%.

Conclusion:
The ablation target area varies enormously in size, depending on the definition of abnormal electrograms. Clear linking of abnormal electrograms with critical VT isthmus areas during substrate mapping remains difficult due to a lack of specificity rather than sensitivity. However, highly fractionated, low-voltage electrograms were found to be present in all critical VT isthmus sites. Substrate-based ablation approaches, targeting low-voltage areas and abnormal electrograms are an established technique, since mapping and ablation during VT is often hemodynamically not tolerated. However, specific electrogram characteristics during substrate mapping have not been linked precisely to the critical isthmus area of VT, which is only a small part of the entire scar. Whenever VT mapping (activation or entrainment) and direct characterization of the circuit is not possible, ablation covering the entire substrate appears to achieve better outcomes than incomplete substrate modification. 2 The scar area after myocardial infarction can be large, leading to extensive ablation. 3 Furthermore subtle, non-transmural myocardial scars might not be delineated using 3.5 mm tip catheters only. 4 Still, modifying the substrate without eliminating it completely may promote arrhythmia recurrences. Therefore, beyond the anatomical low-voltage substrate, thorough analysis of the functional substrate in regards of activation pattern and timing in sinus rhythm (SR) is of high importance. 5 Of particular interest in this context are fractionated and late electrograms (late potentials [LP]). [5][6][7] However, definitions are not homogeneous and as for low-voltage areas, not all abnormal electrograms are associated with the critical reentry isthmus as they can be found over wide areas.
With advances in catheter technology, especially decrease in electrode size, spacing and electrode design, novel insights into VT mechanisms and arrhythmia substrate have been described. 8 Multielectrode mapping of abnormal electrograms has been found to increase the sensitivity with which areas of scar are identified. 4 Small and closely spaced electrodes allow identification of distinct diastolic activity (including diastolic pathways) that may not be seen with standard linear catheters. Moreover, automatic detection of abnormal electrograms may help overcoming subjective judgment and accelerate revision of ultrahigh-density 3D-maps with a huge number of acquired points. Whether and how this might enable a further indept understanding of the arrhythmia's mechanism remains to be demonstrated.
In the present study, we aimed to investigate the size of potential ablation target areas depending on automatic detection of abnormal electrograms and furthermore analyzed the specific electrograms at the critical VT isthmus site using ultrahigh-density 3D-mapping.

| Study design
Postprocedural analysis of 65 consecutive procedures in 61 patients who were referred to our tertiary-care center for radiofrequency catheter ablation of scar-related VT was performed (see Figure 1). Data collection and analysis were performed under a protocol approved by the institutional ethics committee. All patients gave written informed consent. SCHWARZL ET AL. | 377

| Electrophysiological evaluation and instrumentation
All patients underwent the procedure in the fasting state under conscious sedation. Hyperthyroidism or other reversible causes of VT were excluded before the procedure. Detailed procedure methods have been described before. 3 Briefly, the catheter setting consisted of a 6 French (F) quadripolar diagnostic catheter, placed in the right ventricular (RV) apex to induce VT by programmed stimulation with a fixed protocol.
The induced VT was defined as the clinical VT, when cycle length (CL) and morphology matched previous recordings (CL within 20 ms; 12-lead-ECG and/or device recordings). A catheter in the coronary sinus (CS) served as reference for the 3D-electroanatomical mapping system in most cases. Unfractionated Heparin was administered intravenously to maintain an activated clotting time greater than 300 s during the procedure.

| Ultrahigh-density 3D-mapping with multisize electrode catheters
Ultrahigh-density 3D-mapping was performed using the Rhythmia (Boston Scientific) mapping system as previously described. 3  The basket catheter was used to create an ultrahigh-density electroanatomical map of the LV or RV for substrate and, if possible, activation mapping during VT. Electrogram annotation was performed automatically by the mapping system as previously described. 3 F I G U R E 1 Flow chart of the study population. 65 consecutive scar-related ventricular tachycardia (VT) ablation procedures resulting in 44 complete endocardial substrate maps and 22 complete VT maps F I G U R E 2 Specific electrogram characteristics during substrate mapping. Bipolar recordings with minimum abnormal characteristics of the different electrogram groups from a multipolar mapping catheter with 0.4 mm 2 electrode area and 2.5 mm 2 interelectrode spacing are shown. The green line represents the mapping systems trigger (maximum QRS amplitude during substrate mapping). Note that the group "fractionated electrograms" was defined by a minimum of five fractionations and thus also contained electrograms with higher fractionationlikewise, the group "late potentials" contained all electrograms with activation after the QRS offset. "Highly fractionated" was defined by a minimum of ten fractionations and fractionated late potentials had a minimum number of five fractionations. Isolated late potentials needed a minimum of 20 ms isoelectric interval between a first electrogram component Substrate maps were considered complete when the entire chamber anatomy was reconstructed with the best achievable electrode-tissue contact and scar borders were clearly defined. Activation maps were considered complete when ≥90% of the VT CL was mapped. Maps were acquired with the basket catheter and subsequently completed with the single-tip ablation catheter in areas less accessible.
In line with previous studies for scar demarcation a bipolar endocardial voltage of 0.1/1 mV (dense scar <0.1 mV, border zone 0.1-1.0 mV, healthy tissue greater than 1.0 mV) with individual adaptation was chosen. 3,10 All voltage maps were generated during SR, CS, or ventricular pacing. The extent of areas with fractionated and late electrograms, LV-and low-voltage area were measured using the integrated measuring tool in all patients with complete substrate maps after the procedure. Two areas were considered continuous if the distance inbetween was less than 0.5 cm. Slow conduction was defined by crowding in the isochronal map.
Whenever VT was inducible and hemodynamically tolerated, activation mapping was performed. The critical isthmus was defined as region between conduction barriers and between inward curvature (entrance) and outward curvature (exit of the common pathway). 8 Hemodynamic instability was defined as mean arterial blood pressure below 50 mmHg. If activation mapping was possible and the reentrant circuit was completely identified, detailed review of electrograms during sinus or paced rhythm in the critical VT isthmus zone was performed.

| 379
In hemodynamically not tolerated VT we aimed to map repeatedly for short lasting episodes. Additional entrainment or pace mapping was performed at the operator's discretion. 11

| Analysis of abnormal electrograms
Electrogram review was performed by two independent electrophysiologists. Minimum characteristics for electrogram classification are shown in Figure 2.
During ultrahigh-density mapping, areas with fractionated potentials were identified using the novel integrated software tool LUMIPOINT, which has been developed aiming at automatic detec-

| Study population
A total of 65 consecutive procedures in 61 patients were analyzed (see Figure 1). Baseline patient characteristics are shown in Table 1.
All patients had a cardiomyopathy (ischemic in 37 patients [60.7%]) with a mean left ventricular ejection fraction of 36 ± 12%.

| Procedural data
The

| Incidence, distribution and characteristics of abnormal electrograms
Mean low-voltage area (border zone and dense scar, <1.0 mV) was 52 ± 24% and mean dense scar (<0.1 mV) was 9 ± 10% of total mapped LV surface. In all patients with identified VT circuit, the  Table 2. FrP (29 ± 18% of total mapped LV area) had the greatest prevalence and variability, iLP (0.8 ± 1.3% of total mapped LV area, Figure 4) the smallest. The percentage of abnormal electrogram area on total low-voltage area is shown in Table 2 and was distributed correspondent to the percentage on total mapped LV area. Voltage maps indicated areas with abnormal electrograms most often in border zone scar ( Figure 4). Only in two patients FrP were found in areas with predominantly normal voltage. In the isochronal map, areas of slow conduction during sinus or paced rhythm were found in 64% of substrate maps, most often in border zone scar (86%, 7% in dense scar, 7% normal voltage area; Figure S1).
A subgroup analysis of patients with ischemic and nonischemic cardiomyopathy is shown in  Figure 5).

| Abnormal electrogram characteristics in critical VT areas
A postinterventional analysis of electrograms during sinus or paced rhythm associated to VT circuit showed low-voltage hFrP in all isthmus areas (Table 3, Figure 6, Figure S2). The mean number of fractionations showed that there were no significant differences regarding the type (e.g. FrP/LP) of abnormal signals in the critical VT area. In 2 out of 14 patients a focal VT mechanism was identified (Table 3, Figure S3).

| DISCUSSION
The main findings of the present study are: (1). Depending on the definition of abnormal electrograms (low-voltage, fractionated potentials, late potentials) the potential ablation target area varies enormously in size. (2). Clear linking between substrate map electrograms and the critical VT isthmus is limited due to a lack of specificity rather than sensitivity.
(3). Highly fractionated potentials were present in all identified VT isthmus sites during substrate mapping and can be rapidly identified by automated annotation. Furthermore, slow conduction on isochronal maps was often found to be associated with these sites.
(4). The latest and isolated late potentials were often found distant from the identified critical VT isthmus.

| Substrate ablation area
The definition of myocardial scar is well known to be a key step of any substrate-based ablation approach. 16 The importance of electrogram analysis in this context has been pioneered by Josephson and colleagues, while in recent years novel procedural endpoints involving a combination of low amplitude and several abnormal electrogram characteristics (local abnormal ventricular activities) have become widely used. 7 We show, that within the low-voltage zone, the size of the ablation target area varies enormously, depending on the defi- Note that the local activation timing during SR is during and not late after QRS. The local amplitude at isthmus and exit is higher during VT than during SR. VT, ventricular tachycardia using conventional ablation catheters, we saw more comparable results with the novel mini-electrode ablation catheter ( Figure 9).
Moreover, complete manual review of all points, at least during the procedure, is challenging but often necessary regardless of the approach. In our experience, identification of areas of interest (e.g., with fractionated or late potentials) with the system's integrated annotation algorithm enabled subsequent focused manual review in practicable time.

| Specific electrogram characteristics-impact of fractionation and timing
Abnormal electrograms originate in consequence of scarred myocardium where fibrous tissue separates myocardial fibers. to only a small sample size and may represent certain VT, as many other were not completely mappable.
We do not provide a prospective validation of specific electrogram characteristics as an ablation target and thus can only generate hypotheses for further studies aiming at a clear identification of substrate map electrograms linking to VT origin.
Comparison of different activation wavefronts might additionally improve critical VT isthmus identification and novel approaches like the "one acquisition-two maps" technique might give additional insights. 28 Of note, the exact discrimination between different parts of reentry circuit (entrance or isthmus) was often not possible due to comparably high interpoint distance (>0.25 mm) and a slightly different anatomic shell between maps.
However, detailed division is maybe not necessary as recent studies have shown that the dimensions are relatively small. 8

| CONCLUSION
The target area of substrate-based ablation of scar related VT varies enormously in size, depending on the definition of abnormal electrograms. Clear linking between SR electrograms and critical VT reentrant circuit is difficult due to lack of specificity rather than sensitivity. However, highly fractionated, low-voltage electrograms in ultrahigh-density maps can be easily identified by automated annotation with manual revision and were found to be present in all

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.