Author + information
- Received October 19, 2017
- Revision received December 7, 2017
- Accepted December 11, 2017
- Published online March 19, 2018.
- Andreu Porta-Sánchez, MD, MSca,b,
- Nicholas Jackson, MBBSc,
- Peter Lukac, MD, PhDd,
- Steen Buus Kristiansen, MDd,
- Jan Moller Nielsen, MDd,
- Sigfus Gizurarson, MD, PhDa,
- Stéphane Massé, MASca,
- Christopher Labos, MD, MSce,
- Karthik Viswanathan, MBBSa,
- Benjamin King, MBBSa,
- Andrew C.T. Ha, MDa,
- Eugene Downar, MDa and
- Kumaraswamy Nanthakumar, MDa,∗ ()
- aPeter Munk Cardiac Centre, University Health Network, Toronto, Ontario, Canada
- bDepartment de Medicina, Universitat de Barcelona, Barcelona, Spain
- cJohn Hunter Hospital, Newcastle, Australia
- dÅrhus University Hospital, Skejby, Denmark
- eMcGill University, Montreal, Canada
- ↵∗Address for correspondence:
Dr. Kumaraswamy Nanthakumar, Division of Cardiology, University Health Network, Toronto General Hospital, 150 Gerrard Street West, GW3-522, Toronto, Ontario M5G 2C4, Canada.
Objectives The authors conducted a multicenter study of decrement-evoked potential (DEEP)–based functional ventricular tachycardia (VT) substrate modification to evaluate if such a mechanistic and physiological strategy is feasible and efficient in clinical practice and provides reduction in the VT burden.
Background Only a fraction of the myocardium targeted in current VT substrate modification procedures is involved in the initiation and perpetuation of VT. The physiological basis of the DEEP strategy for identification of areas of initiation and maintenance of VT was recently established.
Methods We included 20 consecutive patients with ischemic cardiomyopathy. During substrate mapping, fractionated and late potentials (LPs) were tagged, and an extra stimulus was performed to determine which LPs displayed decrement (DEEPs). All patients underwent DEEP-focused ablation: elimination of DEEP + further radiofrequency (RF) if VT was still inducible. Patients were followed during 6 months.
Results Patients were predominantly male (95%), and their mean age was 64.6 ± 17.1 years. Mean left ventricular ejection fraction was 33.4 ± 11.4%. Mean ablation time was 30.6 ± 20.4 min. Specificity of DEEPs to detect the isthmus of VT was better than that of LPs (0.97 [95% confidence interval [CI]: 0.95 to 0.98] vs. 0.82 [95% CI: 0.73 to 0.89]), without significant differences in terms of sensitivity (0.61 [95% CI: 0.52 to 0.69] vs. 0.60 [95% CI: 0.44 to 0.74], respectively). Fifteen of 20 (75%) patients were free of any VT after DEEP-RF at 6 months of follow-up and there was a strong reduction in VT burden compared to 6 months pre-ablation.
Conclusions In a multicenter prospective study, DEEP substrate mapping identified the functional substrate critical to the VT circuit with high specificity. DEEP-guided VT ablation, by its physiological nature, may enable greater access to focused ablation therapy for patients requiring VT treatment.
- catheter ablation
- myocardial infarction
- nonischemic cardiomyopathy
- substrate ablation
- ventricular tachycardia
Contemporary substrate-based approaches for ventricular tachycardia (VT) ablation targets abnormal potentials and late potentials, and can result in large areas of viable myocardium being ablated that may or may not be linked to the mechanisms of initiation and/or maintenance of VT (1,2). An alternative ablation strategy would be to identify which of those potentials actually participate in the initiation and maintenance of VT as targets for VT ablation. We have recently shown which of these many abnormal and late potentials actually initiate and/or maintain VT. In our previous study, regions that displayed decremental behavior evoked during right ventricular (RV) pacing with extra stimuli (decrement-evoked potential; DEEP), colocalized with the regions of the initiation and diastolic circuit of VT more accurately than those areas displaying nondecremental late potentials (LPs) (3).
This mechanistic physiological demonstration, where decrement precedes unidirectional block, allowed us to identify the region of diastolic path of VT by delivering extra stimulus and evoking the delay of the electrogram (EGM) so that the hidden mechanistic substrate could manifest itself. This obviates the need for VT induction for identifying the critical component of the circuit. This physiological substrate deduction comes from patients who underwent intraoperative global mapping of substrate and activation of VT including the moment of initiation of VT. That study provided the mechanistic basis of DEEP in substrate ablation of VT, allowing VT ablators to identify regions that are responsible for the mechanism of VT without inducing VT. However, that study did not test the practical use and feasibility in the catheterization laboratory. More importantly, it was not established whether DEEP-guided substrate ablation could be adapted to contemporary clinical practice and could be implemented by other operators and sequential mapping systems.
To address these issues, as a follow-up to our original mechanistic study, we designed a multicenter prospective observational study to: 1) establish if the methodology for DEEP mapping with extra stimulus to identify mechanistic substrate is implementable using contemporary 3-dimensional (3D) electroanatomic mapping systems in the catheterization laboratory; and 2) describe the initial results of a multicenter DEEP-guided ablation strategy for reducing VT burden using this focused ablation strategy in the ischemic substrate.
Consecutive patients with ischemic cardiomyopathy (ICM) and recurrent episodes of VT despite medical therapy listed for VT ablation at 4 different institutions (Toronto General Hospital, Ontario, Canada; John Hunter Hospital and Lake Macquarie Private Hospital, Newcastle, Australia; and Århus University Hospital, Skejby, Denmark) were evaluated for participation in the study. The protocol of the study was reviewed and approved by the research ethics boards at all institutions and complies with the Declaration of Helsinki; all patients provided informed consent. All patients underwent DEEP mapping (n = 20) and the ablation strategy focused on eliminating DEEP sites and perform further radiofrequency (RF) if VT was still inducible.
Substrate maps were created in sinus rhythm or during RV pacing. Access to the left ventricular endocardial surface was achieved with a retroaortic (n = 13) or transseptal approach (n = 7). In 16 patients, multielectrode catheters were used for substrate mapping (Decanav in 9 patients, Pentarray in 6 patients [Biosense Webster, Diamond Bar, California], and Orion in 1 patient [Boston Scientific, Marlborough, Massachusetts]). The rest of patients (n = 4) underwent mapping with a 3.5-mm irrigated tip ablation catheter (Thermocool SF catheter, Biosense Webster). LPs (potentials with complex high frequency or multicomponent after or at the QRS offset either present in sinus rhythm or seen during RV pacing) were identified and annotated either as location tags or as local activation time (LAT) maps by manually annotating onset of the delayed bipolar EGM.
For all the points showing LPs, a systematic assessment for local decrement was performed with a drive train (S1) from the RV at 600 ms with a single extra stimulus (S2, coupled at 20 ms above the ventricular effective refractory period [VERP]). If the difference in the time interval measured from surface ventricular far field signal onset to the local LP bipolar EGM during the S1 drive compared to the same interval measured immediately after the S2 was >10 ms, the LP was defined as a DEEP (Figure 1). The same strategy was used for multicomponent EGMs from which DEEP were identified if their components split by >10 ms after S2. All DEEP and non-DEEP-LPs (hereafter referred as LPs) were given a different annotation marker in the substrate map. Care was taken to map the areas of interest with the same density of points to avoid over-representation of the DEEP or LP areas. Analysis of areas of DEEPs and LPs could also be performed by creating an LAT map during RV pacing annotating the onset of the bipolar sharp delayed component for timing of LPs/DEEPs (Figure 2). LAT maps could only be created either in sinus rhythm or during RV pacing, but never together to allow for a correct interpretation of activation times. The threshold of 10 ms for local decrement was chosen based on the intraoperative mapping data and to reduce interobserver disagreement from small variations in measurements (3). Percentage area of LPs and DEEPs was calculated based on the number of LP tags and DEEP tags compared to the total number of points for patients who had >290 points in their maps (above the median of our study) (n = 14). Supplemental information on the procedure can be found in the Methods section of the Online Appendix.
Identification of the VT circuit
In those patients with hemodynamically tolerated VT with high density activation maps (arbitrarily defined as >100 points; n = 13 VTs) the diastolic path of the VT was identified via an LAT map and used to delineate the re-entrant circuit including exit sites as per the classical criteria (4,5). The channel containing all mid-diastolic and pre-systolic signals was identified as the path of VT and used for comparison to the LP and DEEP regions to derive their diagnostic accuracy measures (sensitivity and specificity); points located <5 mm from the mid-diastolic path were considered to be in the re-entrant circuit during VT (Figures 2, 3, and 4⇓⇓).
Ablation strategy and procedural endpoints
Ablation was performed with either an SF Navistar irrigated-tip catheter (Biosense Webster) or a Smart-Touch Navistar irrigated-tip catheter (Biosense Webster) or an IntellaNav open-irrigated catheter (Boston Scientific). Areas with DEEPs were ablated first, and inducibility of the clinical tachycardia was assessed after targeting DEEP areas only. Acute success of the ablation was defined as achieving noninducibility of all clinical tachycardias. If the patient was inducible and became noninducible after DEEP-RF, the procedure was terminated. If the patient was noninducible at the beginning of the case, no further reinductions were performed. If the tachycardia was still inducible, further RF was performed at the discretion of the operator. Details on the RF strategies used, procedural end-points, and follow-up after ablation can be found in the Online Appendix.
Statistical analysis of data was performed with the aid of a biostatistician (C.L.) using STATA (StataCorp, 2013, Stata Statistical Software: Release 13, StataCorp LP, College Station, Texas). For continuous variables with normal distribution, mean ± SD or ranges are reported. For those with non-normal distribution, median and interquartile range (IQR) is expressed. For comparisons of continuous variables, a Student t test was used if data were normally distributed and a sign-rank test was used for non-normally distributed variables. Categorical data was compared with the Fisher exact test or chi-square test as needed.
A bivariate mixed-effects regression model was used to calculate summary statistics for sensitivity and specificity, as well as to construct a hierarchical summary receiver operator curve (ROC) of DEEP and LP for the diagnosis of the isthmus of VT.
Twenty patients were prospectively enrolled. Their baseline characteristics are summarized in Table 1. The location of scar (bipolar voltage <1.5 mV) was inferior in 12 patients, anterior in 9 patients, lateral in 3 patients, and septal in 1 patient. Procedural details are available in the Online Appendix.
LP and DEEP maps
Points with LPs accounted for a median area of 16.8% (IQR: 8.9% to 73.7%) of the myocardium mapped. The median area with DEEP points was 4.8% (IQR: 2.2% to 25.7%) of the myocardium mapped (p < 0.001 compared to LP area). Examples of maps are shown in Figures 2, 3, and 4.
All patients had VT inducible at the beginning of the procedure; the majority (n = 16, 80%) became noninducible after DEEP-focused RF. Further RF on those patients who were still inducible did not achieve noninducibility in any of them. The RF time was 30.6 ± 21.4 min.
VT burden before and after DEEP-guided ablation
The median (IQR) number of VT episodes 6 months pre-procedure was 11 (IQR: 5 to 25) and the median number of shocks was 1.5 (IQR: 0 to 4.5); this decreased to 0 (IQR: 0 to 2) for VT episodes and to 0 (IQR: 0 to 0) shocks 6 months post-procedure (p = 0.02 and p = 0.03, respectively). Fifteen of 20 patients (75%) were free of any VT after DEEP-RF at 6 month of follow-up (Online Figure).
Relationship between LPs and DEEPs and diastolic signals in VT
The diagnostic performance of DEEP and LPs was analyzed in only those VTs mapped where a detailed LAT map was created (13 VTs in 9 patients) with a median number of activation points of 485.5 (IQR: 352 to 890 points). Areas with DEEPs performed better than LPs at colocalizing within the region of the diastole of VT with a sensitivity of 0.61 (95% confidence interval [CI]: 0.52 to 0.69) and a specificity of 0.97 (95% CI: 0.95 to 0.98) with an ROC area under the curve of 0.86 (95% CI: 0.82 to 0.88) compared to LPs sensitivity of 0.60 (95% CI: 0.44 to 0.74) and a specificity of 0.82 (95% CI: 0.73 to 0.89) with an ROC area under the curve of 0.79 (95% CI: 0.75 to 0.82) (Online Figure).
The primary findings from this multicenter prospective observational study of ischemic VT ablation are: 1) identification of critical myocardial regions for VT maintenance with decremental conduction properties (DEEP mapping) by extra stimulus is feasible using contemporary 3D mapping systems; 2) targeting the DEEP regions, which is a physiological assessment of the substrate, deems VT noninducible in the majority of cases; 3) areas of DEEPs are localized more frequently in the diastolic pathway of the VT than LPs; and finally, 4) mid-term outcomes of this limited focused mechanistic ablation strategy using extra stimulus are in keeping with the current ablation outcomes described in the published data.
The insight that decrement precedes unidirectional block is observable in atrial tissue sections in studies undertaken by Lammers et al. (6). This has also been shown later in modeling studies as well as in experimental models and in humans (7–9). We have shown previously the sequence of decrement preceding unidirectional block leading to re-entry in patients with VT, and the unique clinical finding of improved sensitivity, specificity, and accuracy of DEEP over LP mapping for identifying the initiating regions and diastolic pathway of VT using a tightly coupled S2 (VERP + 20 ms) (3). The results described in this paper prove clinical feasibility and act as a validation of the findings from the derivation cohort. With the inherent limitations of a catheter-based sequential mapping approach available currently, we have been able to identify regions with decremental conduction (DEEP) properties, and we found that they frequently colocalize to the areas that initiate and maintain VT, providing a functional and mechanistic target for ablation.
Conventional substrate mapping currently involves high-density mapping during sinus rhythm or during ventricular pacing to identify and target areas with fractionated EGMs, double potentials, LPs, or all abnormal EGMs in the case of substrate homogenization. Identification of those targets is subject to interobserver variability and will not always identify circuits that are involved in VT initiation or maintenance and can be linked to ablating a very large area of myocardium. Additionally, ablating abnormal tissue that is not actively participating in VT perpetuation could potentially create new areas of block/decrement that may predispose to new VT circuits.
Our work builds on our previous study in identifying which of the abnormal EGMs are involved in initiation and maintaining VT and provides a mechanistic ablation strategy for substrate modification. This is an alternative strategy to targeting all abnormal potentials. Even when LPs are easily identifiable beyond the end of the QRS, they could represent fixed delayed conduction through bundles of viable myocardium inside the scar (10). In contrast with those fixed-delayed LPs, DEEPs display decremental conduction properties, allowing the time for blocked regions to recover excitability and potentially initiate re-entry. It has also been shown that regions with the latest activation during sinus rhythm are infrequently linked to successful ablation sites, whereas the slowly conducting regions that actually propagate into the latest zone of activation are more likely to correlate with the critical isthmus of a VT (11). Additionally, with recent high-resolution mapping tools, there is growing evidence about the importance of functional block in the critical areas that sustain re-entry (12).
Current substrate modification procedures targeting an extensive substrate with scar homogenization are associated with long procedural times with broad ablation targets in some series and outcomes after VT ablation are influenced by procedural length (13). In our current study, although the methodology is to use a focused ablation strategy, for purposes of proving the concept it was important to identify the critical isthmus in VT with activation mapping, pacing maneuvers, and often multiple VT inductions and that leads to procedural times that are similar to other contemporary series (2). Furthermore, annotation of the DEEP points is not automated at present and must be performed manually. For these reasons, this study was unlikely to show shorter procedure times than contemporary substrate-based ablation studies. We postulate that automation of DEEP mapping could lead to a decrease in procedure times. When implemented, the combination of the focused DEEP mapping, automated annotation, and limited ablation would make VT ablation procedures less cumbersome and therefore accessible to more patients; this is our goal for future development of DEEP.
There are numerous VT substrate-based strategies with individual merit that include RF transection of the scar (14) or scar “dechanneling” with (15) or without the use of RV pacing and extra stimuli to identify the origin of channels of slow conduction near the low voltage areas (16). These methodologies are weighted towards the anatomic/structural channels and are essentially challenged by the physiological/functional block and channels that are claimed to be significant by Josephson (17) and others (18). Although some reports have suggested the lack of incremental value of mechanistic strategy compared to substrate homogenization (19), the fundamental problem of identifying whether the myocardium being ablated is part of initiation and maintenance of the VT circuit and ablating such regions of myocardium is the focus of this paper.
Recurrence rates after a DEEP-guided ablation were similar to what has been seen in the literature with a similar reduction in the VT burden and the follow-up time for our study was limited to 6 months to focus on the feasibility aspect of the mapping technique. Additionally it has been recently described that VT recurrences occurring after 6 months from the procedure have limited impact on outcomes (20).
This is a nonrandomized cohort of 20 patients. A randomized study with a comparator control is needed to further confirm its clinical value. The validity of this strategy in VT ablation in nonischemic or patchy scar patterns must also be established as the use of DEEP in this population remains to be proven. Identification of DEEP points was performed with RV apical pacing. Whether our findings would be reproducible and/or perform better with pacing from other regions of the myocardium is currently under investigation. Previous work from Baldinger et al. (21) has shown that RV pacing could also preclude the identification of a significant proportion of LPs by creating a wider QRS; for that reason our work has not restricted the identification of LPs to sinus rhythm or RV pacing, but has used both strategies to increase the sensitivity of mapping. In rare instances during DEEP mapping, VT was induced with S1/S2, usually requiring us to use a longer coupling interval for S2 to characterize LPs as DEEP. In patients with incessant VT and/or extremely easy inducibility of VT DEEP, mapping could become cumbersome and is not feasible. We also acknowledge the limitations of not creating an ultra high-density map and performing entrainment maneuvers systematically to define the re-entry circuits which has been reported to define smaller or larger sizes of the isthmus respectively (22) and we limited our analysis to proximity between LPs and DEEPs to the shortest path of diastolic EGMs during VT. Furthermore, whether the calculations of the diagnostic performance of DEEPs were influenced by VT stability and circuit characteristics is unknown, but our initial study by Jackson et al. (3) showed that when the whole VT circuit is mapped with a nonsequential mapping system, DEEPs perform better at identifying the isthmus of VT than nondecremental LPs.
In this prospective multicenter study, we have evaluated the feasibility of a mechanistic physiological approach to identify functional substrate modification for VT therapy by targeting limited regions of the diseased myocardium that are involved in the initiation and maintenance of VT. DEEP mapping with an extra stimulus can be implemented using contemporary 3D mapping systems with a meaningful reduction in VT burden. Initial results of catheter ablation guided solely by this technique are encouraging, with the majority of patients rendered noninducible for clinical VT. Automation of the DEEP technique will allow for rapid identification of DEEP potentials enabling further validation of such a strategy in a randomized study.
COMPETENCY IN MEDICAL KNOWLEDGE 1: Patients with unstable, nonmappable VT may have an extensive substrate in which only a small proportion participates in VT. It is by no means mechanistic to eliminate or practically strategize to eliminate all of the substrate by RF ablation. A mechanistic-focused strategy such as DEEP mapping can provide a physiologically relevant ablation target and yield the same results as extensive and tedious substrate modification.
COMPETENCY IN MEDICAL KNOWLEDGE 2: Electrophysiologists performing VT ablation could adopt this DEEP technique in cases where VT is noninducible or nonmappable due to hemodynamic instability. The diagnostic performance of areas of the substrate with DEEP properties provides higher specificity than late potentials and similar sensitivity for detection of critical isthmuses of VT as shown in our multicenter study.
TRANSLATIONAL OUTLOOK: This multicenter prospective experience is a validation of our previous mechanistic work that showed the concept that decrement in the region of the diastolic circuit and critical isthmus precedes unidirectional block, and that unidirectional block precedes VT using global endocardial mapping in patients. This current study translates those mechanistic findings and proves their applicability with the currently available electro-anatomic mapping technologies in the electrophysiology laboratory for the benefit of patients undergoing VT ablation.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Dr. Massé has received consulting fees from Abbott Laboratories. Dr. Nanthakumar has received consulting fees and grants from Abbott Laboratories and Biosense Webster. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- decrement-evoked potential
- implantable cardioverter defibrillator
- ischemic cardiomyopathy
- local activation time
- late potentials
- nonischemic cardiomyopathy
- stimulus 1
- stimulus 2
- ventricular effective refractory period
- ventricular tachycardia
- Received October 19, 2017.
- Revision received December 7, 2017.
- Accepted December 11, 2017.
- 2018 American College of Cardiology Foundation
- Di Biase L.,
- Santangeli P.,
- Burkhardt D.J.,
- et al.
- Tzou W.S.,
- Frankel D.S.,
- Hegeman T.,
- et al.
- Jackson N.,
- Gizurarson S.,
- Viswanathan K.,
- et al.
- Brunckhorst C.B.,
- Delacretaz E.,
- Soejima K.,
- Maisel W.H.,
- Friedman P.L.,
- Stevenson W.G.
- Stevenson W.G.,
- Friedman P.L.,
- Sager P.T.,
- et al.
- el-Sherif N.,
- Gough W.B.,
- Restivo M.
- Ciaccio E.J.,
- Coromilas J.,
- Wit A.L.,
- Peters N.S.,
- Garan H.
- de Bakker J.M.,
- van Capelle F.J.,
- Janse M.J.,
- et al.
- Irie T.,
- Yu R.,
- Bradfield J.S.,
- et al.
- Yu R.,
- Ma S.,
- Tung R.,
- et al.
- Soejima K.,
- Suzuki M.,
- Maisel W.H.,
- et al.
- Fernandez-Armenta J.,
- Penela D.,
- Acosta J.,
- et al.
- Acosta J.,
- Andreu D.,
- Penela D.,
- et al.
- Josephson M.E.,
- Anter E.
- Dillon S.M.,
- Allessie M.A.,
- Ursell P.C.,
- Wit A.L.
- Di Biase L.,
- Burkhardt J.D.,
- Lakkireddy D.,
- et al.
- Muser D.,
- Mendelson T.,
- Fahed J.,
- et al.
- Baldinger S.H.,
- Kumar S.,
- Fujii A.,
- et al.
- Anter E.,
- Tschabrunn C.M.,
- Buxton A.E.,
- Josephson M.E.