Contrast-Enhanced MRI–Derived Scar Patterns and Associated Ventricular Tachycardias in Nonischemic CardiomyopathyClinical Perspective
Implications for the Ablation Strategy
Background—There are limited data on typical arrhythmogenic substrates and associated ventricular tachycardias (VT) in patients with nonischemic cardiomyopathy. The substrate location may have implications for the ablation strategy.
Methods and Results—Nineteen consecutive patients with nonischemic cardiomyopathy (age 58±14 years, 79% men, left ventricular ejection fraction 41±11%) who underwent contrast-enhanced MRI and VT ablation were included. On the basis of 3-dimensional contrast-enhanced MRI–derived scar reconstructions, 8 patients (42%) had predominant basal anteroseptal scar, 9 patients (47%) had predominant inferolateral scar, and 2 patients (11%) had other scar types. Three distinct VT morphologies (≥1 of 3 inducible in 16/19 patients) were associated with underlying scar type. In 9 patients with anteroseptal scar–related VT (8/9 predominant scar, 1/9 nonpredominant), ablation target sites (defined as sites with ≥11/12 pacemap, concealed entrainment or VT termination during ablation) were located in the aortic root and/or anteroseptal left ventricular endocardium in 8 patients (89%) and in the anterior cardiac vein in 1 patient (11%), with additional target sites at the right ventricular septum in 2 patients (22%) and at the epicardium in 1 patient (11%). In contrast, in 8 patients with predominant inferolateral scar–related VT, target sites were located at the epicardium in 5 patients (63%) and in the endocardial inferolateral left ventricle in 3 patients (37%).
Conclusions—Two typical scar patterns (anteroseptal and inferolateral) account for 89% of arrhythmogenic substrates in patients with nonischemic cardiomyopathy. Three distinct VT morphologies are highly suggestive of the presence of these scars. Anteroseptal scars were, in general, most effectively approached from the aortic root or anteroseptal left ventricular endocardium, whereas inferolateral scars frequently required an epicardial approach.
The substrate for ventricular tachycardia (VT) in patients with nonischemic cardiomyopathy (NICM) is typically located near the base of the left ventricle (LV) close to the valve annuli, with an intramural or subepicardial distribution pattern.1–3 Isolated septal scars have been reported in ≈12% of patients2 but may be difficult to detect by electroanatomical mapping (EAM),2,4 thereby underestimating its prevalence. Viable myocardium overlying the substrate may prevent scar detection by bipolar EAM.2,5 Unipolar voltage mapping has been suggested to have a larger field of view than bipolar voltage mapping, allowing detection of subepicardial scars during endocardial mapping.6 The exact size of this field of view is, however, unclear and has not been validated by contrast-enhanced MRI (CE-MRI), which is considered to be the gold standard to detect scar.
Clinical Perspective on p 883
CE-MRI–derived 3-dimensional (3D) scar reconstructions may provide insights into the geometry of the arrhythmogenic substrate4 and its spatial relation to anatomic structures, which may have implications for the procedural strategy, the required access, and potential limitations for ablation.
The 12-lead morphology of scar-related VT depends on 3D scar geometry and VT exit site, and it may help to identify the substrate in patients for whom CE-MRI is not available. Distinct VT morphologies have been described in selected patients with NICM and isolated septal scars2 but not in other scar types.
The aims of the present study were (1) to identify typical MRI-derived scar patterns and the associated 12-lead VT morphologies in consecutive patients with NICM who underwent CE-MRI and VT ablation, (2) to evaluate its implications for the ablation strategy, and (3) to analyze the value of bipolar and unipolar endocardial voltage mapping to detect the CE-MRI–derived VT substrate in patients with NICM.
The study sample consisted of 19 of 21 consecutive patients with NICM who underwent CE-MRI and VT ablation. Two patients with poor CE-MRI image quality were excluded. Patients with coronary artery disease (>50% stenosis, assessed by coronary angiography in all patients), congenital heart disease, LV noncompaction, (sub)acute myocarditis, cardiac sarcoidosis, tachycardia-induced cardiomyopathy, primary valvular abnormalities, or hypertrophic, arrhythmogenic right ventricular (RV), and restrictive cardiomyopathy were excluded. All patients were treated according to the routine clinical protocol and have provided informed consent.
CE-MRI Acquisition and Processing
CE-MRI was performed on a 1.5 T Gyroscan ACS-NT/Intera MR system (Philips Medical Systems, Best, The Netherlands). A standardized protocol was followed including cine MRI in long axis (2- and 4-chamber views) and short axis covering the LV. In addition, the proximal aorta was imaged using a black-blood turbo spin-echo sequence. Contrast-enhanced images were acquired 15 minutes after bolus injection of gadolinium (Magnevist; Schering, Berlin, Germany; 0.15 mmol/kg) using a 3D turbo-field echo sequence with parallel imaging. The heart was imaged in 1 or 2 breath-holds with 20 to 24 imaging levels (dependent on heart size) in short-axis views. Images were acquired 600 to 700 ms after the R-wave.
Using Mass research software, the centerline of the proximal coronary arteries was manually defined in the black-blood spin-echo image stack. The aortic, endocardial, and epicardial contours were semiautomatically detected on short-axis views and converted into 3D meshes. Myocardial tissue with 35% to 50% and ≥50% of maximal signal intensity were considered gray zone and core scar, respectively. The LV endocardial and epicardial meshes were color coded for scar transmurality in the inner and outer half of the wall, respectively. All meshes were imported into the CARTO system using IPE-software (Figure 1).
Computed Tomography Acquisition and Processing
Before epicardial ablation, ECG-gated cardiac computed tomographic (CT) imaging was performed with an intravenous iodinated contrast agent. The endocardial, epicardial, and pericardial contours were manually traced on short-axis reformatted CT slices to create 3D meshes color coded for fat thickness (distance between epicardial and pericardial contours).7,8 Subsequently, the original CT data and the 3D meshes were imported into the mapping system. These images were used during epicardial VT ablation procedures to avoid radiofrequency (RF) applications in the close vicinity of coronary arteries and to provide detailed information on epicardial fat thickness, which may facilitate the interpretation of electrograms and explain ineffective RF applications.7
Anti-arrhythmic drugs were discontinued for ≥5 half-lives with the exception of amiodarone (n=2). Programmed electric stimulation was performed before sedation and opioid administration. The simulation protocol consisted of 3 drive cycle lengths (600, 500, and 400 ms) with 1 to 3 ventricular extrastimuli (≥200 ms) from 2 RV sites and burst pacing (cycle length ≥200 ms), repeated with isoproterenol (2–10 µg/min) when necessary. Positive end point was induction of a sustained monomorphic VT (>30 s or requiring termination because of hemodynamic compromise).
All induced sustained monomorphic VTs were categorized as right bundle-branch block–like or left bundle-branch block (–like morphology (defined as predominant R or S V1), inferior or superior axis (predominant R or S in aVF), left or right axis (predominant R or S in I), and precordial transition (first lead with a predominant R or S for left bundle-branch block and right bundle-branch block VTs, respectively).
EAM, Real-Time Image Integration, and Ablation
Epicardial access was obtained through a subxyphoid puncture. Electroanatomical bipolar and unipolar voltage mapping was performed using a 3.5 mm irrigated-tip catheter (2-mm ring electrode, 1-mm interelectrode spacing; NaviStar ThermoCool, Biosense Webster Inc, Diamond Bar, CA) and the CARTO system. Electrograms were filtered at 30 to 400 Hz (bipolar) and 1 to 240 Hz (unipolar). Limited EAM of the aortic root was performed and the position of the left main coronary artery, confirmed by undiluted contrast injection through the mapping catheter, was tagged on the map. The LV was mapped retrogradely focusing on the MRI-derived scar areas. The RV was mapped if indicated. CE-MRI–derived images were aligned with the EAM using the left main landmark. Then, the LV surfaces were aligned with the translation tool until the lowest mean surface registration error was reached. The CT-derived coronary anatomy and epicardial fat meshes were integrated with the EAM in patients who underwent epicardial mapping, as previously described.8
Potential re-entry circuit sites targeted for ablation were identified on the basis of activation mapping and entrainment mapping for stable VT. For unstable VT, the area of interest was identified by substrate mapping and pace mapping. Then, VT was reinduced and briefly mapped in an attempt to identify diastolic activity and terminate the VT by ablation. For analyses purposes, ablation target sites were defined as sites with (1) ≥11/12 pace map; or (2) concealed entrainment, suggestive for a central isthmus or exit site; or (3) VT slowing and termination during ablation.
At the epicardium, ablation was usually withheld when the estimated distance to a coronary artery was <5 mm, as assessed by integrated CT–derived coronary anatomy and coronary angiography.7 High-output pacing (10 mA, 2 ms) was performed to determine the location of the phrenic nerve. RF energy was applied at 30 to 45 W (maximum temperature, 45°C; flow, 20–30 mL/min; 60 s) for endocardial sites and ≤50 W (flow 20 mL/min) for epicardial sites.
Acute and Long-Term Outcome
After ablation, the entire stimulation protocol was repeated. Isoprenaline was administered if required to induce VT before ablation. Complete procedural success was defined as noninducibility of any sustained monomorphic VT, partial success as elimination of the clinical VT but inducibility of ≥1 nonclinical VT, and failure as inducibility of the clinical VT. Patients were followed at 3, 6, and 12 months and at 6- to 12-monthly intervals thereafter.
CE-MRI–Derived 3D Scar Reconstructions
On the basis of the short-axis slices, 3D scar reconstructions were created (MATLAB software, version 2009b). For each patient, the MRI-derived scar transmurality (including core and gray zone) was calculated and displayed on a bull’s eye image containing the 17 LV segments. On the basis of the reconstructions, the locations of the predominant (ie, largest) scar and of additional scars (defined as scars involving ≥25% of a remote segment) were assessed. The scar distribution within the LV wall was determined by calculating the average endocardial and epicardial viable layers for each scar.
After the procedure, all mapping points were superimposed on the MRI-derived 3D scar reconstruction. For each point, a 5-mm diameter transmural cylinder was created, and the presence of late enhancement and the thickness of the endocardial viable layer were assessed.
Categorical variables are displayed as number (percentage) and continuous variables as mean±SD or median (first to third quartile [Q1–Q3]). Categorical variables were compared using the χ2 test or Fisher exact test. Continuous variables were compared using the Student t test when normally distributed or the Mann–Whitney U test when non-normally distributed. Mapping point-based receiver operating characteristic curve analysis was performed to determine the optimal cutoff values of bipolar and unipolar voltage for scar detection, defined as the values maximizing the sum of sensitivity and specificity. An additional cutoff value with a fixed specificity of 80% was calculated for unipolar voltage. The bipolar and unipolar voltage of mapping points at scars with an endocardial viable layer of 0 to 2 mm, 2 to 4 mm, 4 to 6 mm, 6 to 8 mm, >8 mm, and nonscar areas were averaged in each patient. Paired t tests were subsequently performed to analyze the difference between the average voltage at nonscar areas and the average voltage at scar areas with different endocardial viable layers in the same patient. All analyses were performed with SPSS version 20.0 (SPSS Inc, Chicago, IL). All P values are 2 sided, and P values <0.05 were considered significant.
Nineteen patients (age, 58±14 years; 15 [79%] men; LV ejection fraction, 41±11%) underwent VT ablation with CE-MRI integration. In 11 patients who underwent epicardial mapping, CT-derived coronary anatomy and epicardial fat thickness were also integrated. Baseline characteristics are summarized in Table. None of the patients had documented pericarditis or myocarditis. Three patients (16%) had a family history of NICM/sudden cardiac death.
Typical MRI-Derived Scar Patterns
As illustrated by the bull’s eye plot color coded for the average MRI-derived scar transmurality (Figure 2), 2 typical scar patterns could be identified. One was located in the basal anteroseptal segment (referred to as anteroseptal scar) and the other in the basal/midinferior, inferolateral, or lateral segments (referred to as inferolateral scar). A predominant anteroseptal scar was present in 8 of 19 patients (42%) and a predominant inferolateral scar in 9 (47%; Figure 3). Of the remaining 2 patients (11%), 1 had a subepicardial inferoapical scar and the other had a subepicardial scar along the anterior interventricular groove.
Two of 8 patients (25%) with a predominant anteroseptal scar also had a smaller inferolateral scar, and 6 of 9 patients (67%) with a predominant inferolateral scar also had a smaller anteroseptal scar.
Distinct ECG Patterns of VTs
A total of 60 VTs were induced (median, 2 VTs per patient; Q1–Q3, 1–3). Three distinct ECG patterns were observed as follows (Figure 4):
Right bundle-branch block, left/right inferior axis, and positive concordance (Figure 4A) in 12 induced VTs (20%) in 8 patients (42%).
Left bundle-branch block, left/right inferior axis, and early (≤V3) precordial transition (Figure 4B) in 7 induced VTs (12%) in 6 patients (32%).
Right bundle-branch block, right inferior/superior axis, and late (≥V5) precordial transition (Figure 4C) in 6 induced VTs (10%) in 6 patients (32%).
Together, 25 of 60 VTs (42%) had 1 of these 3 morphologies. All patients with type 1 or type 2 VTs had anteroseptal scar, and all patients with type 3 VTs had inferolateral scar. A less distinct ECG pattern, left bundle-branch block VTs with a superior axis and varying transition, was observed in patients with anteroseptal, inferolateral, or nontypical scar extending to the inferoseptum. Sixteen of 19 patients (84%) had ≥1 type 1, 2, or 3 VT. The remaining 3 patients (16%) had extension of the scar on CE-MRI consistent with the VT morphology.
Electroanatomical Substrate Mapping and Ablation Target Sites
Procedural data are summarized in Table. CE-MRI–derived data were successfully integrated in all patients (mean registration error, 3.7±0.8 mm). Substrate mapping of the LV (all patients; mean, 128 points) and the epicardium (11 patients [58%]; mean, 155 points) was performed focusing on the area of interest as identified by MRI (Table; Figure 5).
A ≥11/12 pacemap was obtained for 31 VTs (52%). Diastolic potentials were identified during 26 VTs (43%), and for 7 VTs (12%), concealed entrainment was suggestive for a central isthmus or exit site (stimulus-QRS, 43–254 ms). Eighteen VTs (30%) slowed and terminated during RF applications after a median of 9 s (Q1–Q3, 7–13 s) without premature ventricular contraction. Overall, ablation target sites were identified for 40 VTs (67%). The locations of these target sites are displayed in Figure 6. All target sites were located within 10 mm from the CE-MRI–derived scar area.
In 9 patients with basal anteroseptal scar–related VT (predominant in 8), target sites were located in the aortic root and/or basal anteroseptal LV endocardium in 8 patients (89%) and in the anterior cardiac vein in 1 patient (11%), with additional target sites at the RV septum in 2 patients (22%) and at the basal anterior epicardium in 1 patient (11%).
In contrast, in 8 patients with inferolateral scar–related VT (all predominant scars), target sites were located at the inferolateral epicardium in 5 patients (63%) and in the inferolateral endocardial LV in 3 patients (37%; both endocardial and epicardial target sites in 1 patient), with an additional target site at the epicardial RV outflow tract in 1 patient. No target sites could be identified in 1 patient.
Anatomic Limitations of Mapping and Ablation
In patients with a predominant anteroseptal scar, ablation was limited by the proximity of the proximal conduction system in 2 patients (25%) and by the left main ostium in another 2 patients (25%). On the basis of integrated CT–derived and MRI–derived information, the entire scar could not be reached from the epicardium because of the RV, pulmonary artery, overlying coronary arteries, and/or epicardial fat in 5 patients (63%; Figure 1), and only part of the scar could be reached in the remaining 3 patients (37%). In 3 patients with an anteroseptal substrate, epicardial mapping and ablation were performed after image integration. The expected epicardial accessibility of the scar was confirmed in these patients (only part of the scar reachable in 2 patients, the entire scar not reachable in 1 patient).
In patients with a predominant inferolateral scar and endocardial ablation target sites, endocardial ablation was not limited. On the basis of integrated CT–derived information, coronary angiography, and high-output pacing, RF applications had to be withheld at the epicardium because of the right coronary artery or the obtuse marginal branch in 3 of 7 patients (43%) and because of phrenic nerve capture in 1 patient.
CE-MRI–Derived 3D Scar Reconstructions
A total of 2444 points acquired in the LV were projected on 3D scar reconstructions. Scar was present at 1146 sites (47%). For endocardial bipolar voltage, the optimal cutoff value to differentiate between scar and viable myocardium was 2.04 mV (area under the curve, 0.67; sensitivity, 48%; specificity, 81%; Figure 5). For endocardial unipolar voltage, the optimal cutoff value was 9.84 mV (area under the curve, 0.75; sensitivity, 73%; specificity, 69%). The specificity could be increased to 80% using a cutoff value of 8.01 mV, which resulted in a sensitivity of 59%. Endocardial bipolar voltage was only affected by scar involving the endocardial 2 mm, whereas endocardial unipolar voltage was affected by scar involving the endocardial 4 mm and, less pronounced, by scar >4 mm from the endocardium (Figure 5). Two patients (1 patient with a basal anteroseptal scar and 1 patient with an inferolateral scar) had an epicardial scar identified by MRI that significantly exceeded small endocardial bipolar and unipolar low-voltage areas (Figure 7). Low bipolar and unipolar voltage areas, thickness of endocardial and epicardial viable layers, and average scar transmurality were similar for anteroseptal and inferolateral scars (all P>0.05; Figure 5D–5H).
Acute and Long-Term Procedural Outcome
Complete success was achieved in 10 patients (53%), partial success in 6 patients (32%), and ablation failed in 3 patients (16%). The acute procedural outcome did not differ among patients with a predominant anteroseptal or inferolateral scar (Table). Seventeen of 19 patients (90%) were discharged with an implantable cardioverter defibrillator. During a median 15 (Q1–Q3, 11–27) months follow-up, 7 of 19 patients (37%) had VT recurrence (4/8 with a predominant anteroseptal scar, 2/9 with a predominant inferolateral scar, 1/2 with other scar type). The 6-month VT burden was reduced by ≥75% in 16 of 19 patients (84%). One patient died because of a lung cancer.
In the present study, CE-MRI–derived scar was associated with VT in patients with NICM. Two typical scar patterns (ie, anteroseptal and inferolateral scars) were present in 89% of patients referred for VT ablation. Three distinct 12-lead ECG morphologies of VTs were common and diagnostic for a typical scar pattern. Ablation target sites in patients with a basal anteroseptal scar were mainly identified in the aortic root and the basal anteroseptal LV endocardium, whereas target sites in patients with an inferolateral scar were located at the epicardium in 63% of patients. The typical scar patterns may therefore have implications for the mapping and ablation strategy. Endocardial EAM is insufficient to delineate these scars because endocardial bipolar voltage was only affected by scar involving the endocardial 2 mm and endocardial unipolar voltage by scar involving the endocardial 4 mm and, less pronounced, by scar >4 mm from the endocardium.
Typical Arrhythmogenic Substrates
Previous studies have described arrhythmogenic substrates in NICM but have not systematically used CE-MRI as a gold standard for scar detection.2,9 In the current study, the EAM and CE-MRI–derived 3D scar reconstructions allowed a detailed evaluation of the arrhythmogenic substrate.
Importantly, all ablation target sites were located within 10 mm of the CE-MRI–derived scar area. A predominant basal anteroseptal scar was present in 42% of patients but often associated with only a small endocardial low bipolar voltage area, difficult to distinguish from a normal low-voltage area at the membranous part of the interventricular septum. In this group, endocardial target sites could be identified in all but 1 patient. The ablation was withheld because of the location of the conduction system in 1 patient and caused expected atrioventricular block in another. Epicardial mapping in the presence of this scar pattern was considered impossible because of or hampered by overlying coronary arteries, fat, and the left atrial appendage.
Although basal anteroseptal scars have been described in previous studies and 2 case reports,2–4,9,10 they are reported to occur in only 11.3% of patients with NICM.2 On the basis of the current data and 2 previous studies, these scars may, however, be difficult to detect by substrate mapping,2,4 and as a result, scar-related VTs may even be mistaken for benign idiopathic outflow tract VTs, which may have important prognostic and therapeutic consequences.4
A predominant inferolateral scar was present in 47% of patients but also difficult to detect by endocardial bipolar voltage mapping. Endocardial unipolar voltage mapping allowed improved but still limited scar detection. Epicardial mapping was not limited, and epicardial target sites were identified in 63% of these patients. Epicardial ablation was, however, partly withheld because of coronary arteries and the phrenic nerve.
In the current study, CE-MRI served as a gold standard for LV scar. For both scar patterns, endocardial bipolar voltage was only affected by scar involving the endocardial 2 mm, whereas endocardial unipolar voltage was affected by scar involving the endocardial 4 mm and, less pronounced, by scar >4 mm from the endocardium. Accordingly, the sensitivity for bipolar and unipolar voltage mapping to detect scar in the NICM population is poor with an overall sensitivity of 48% and 59%, respectively. Therefore CE-MRI should be considered in all patients with suspected NICM before VT ablation and implantable cardioverter defibrillator implantation.
In only 1 patient in our study, a causative mutation (phospholamban) for NICM has been identified (yet), and none had a history of myocarditis. However, typical septal and inferolateral scars have been described in various diseases, such as myocarditis,11 arrhythmogenic RV cardiomyopathy/dysplasia with LV involvement,12 laminopathy,13 and Anderson–Fabry disease.14 Rather than disease specific, these areas may be more vulnerable to various insults resulting in fibrosis and ultimately VT.
Typical 12-Lead ECG VT Morphologies
At least 1 of 3 distinct VT morphologies was observed in 84% of patients, and these morphologies were related to the 2 scar patterns. Documentation of these distinct morphologies may be helpful if CE-MRI is not available. Two typical VT morphologies have been previously described in patients with NICM but only with isolated septal scars, including inferior limb discordance (dominant R in II and dominant S in III, or the opposite) in 16% of VTs and precordial transition break in V2 (qR/Rs morphology in V1 and V3 but reversal of this in V2) in 24% of VTs.2 In the current study, each criterion was observed in 5 of 37 induced VTs (14%) in 2 of 8 patients (25%) with a predominant anteroseptal scar.
Implications for the Procedural Strategy
VTs related to basal anteroseptal scar were effectively approached from the aortic root or basal anteroseptal LV endocardium, whereas VTs related to an inferolateral scar often required an epicardial approach. This finding may influence the procedural strategy in patients with NICM and a typical substrate based on imaging or a distinct VT morphology. An initial endocardial approach may be appropriate in patients with a presumed basal anteroseptal substrate. A retrograde approach to the LV may be preferred instead of an antegrade approach, allowing mapping in the aortic sinus cusps. In patients with an inferolateral scars, epicardial mapping and ablation are likely to be necessary and an initial combined endocardial and epicardial approach may be reasonable.
Although the present study demonstrates a clear predilection for scars in the basal anteroseptal and inferolateral segments, there may be nonischemic scar distributions (and associated VT morphologies) that are not identified in this study. The number of patients was limited because only patients without an implantable device underwent CE-MRI. No ablation target sites could be identified for 33% of VTs, which may be because of intramural circuits or subepicardial circuits in areas covered with fat. Epicardial mapping was not performed in 4 patients with noncomplete procedural success because the septal substrate was likely not accessible from the epicardium based on CE-MRI.
Basal anteroseptal and inferolateral scars account for 89% of arrhythmogenic substrates in patients with NICM. Typical VT morphologies, observed in 84% patients, may help to identify the underlying substrate. Basal anteroseptal scars were, in general, most effectively approached from the aortic root or basal anteroseptal LV endocardium, whereas inferolateral scars frequently required an epicardial approach. These findings are likely to have implications for the procedural strategy in patients with NICM who are referred for VT ablation.
Sources of Funding
Carine F.B. van Huls van Taxis is supported by The Netherlands Heart Society (grant no: 2008B074).
The department of cardiology receives unrestricted research grants from Boston Scientific, Medtronic, and Biotronik.
- Received April 14, 2013.
- Accepted August 30, 2013.
- © 2013 American Heart Association, Inc.
- Nakahara S,
- Tung R,
- Ramirez RJ,
- Michowitz Y,
- Vaseghi M,
- Buch E,
- Gima J,
- Wiener I,
- Mahajan A,
- Boyle NG,
- Shivkumar K
- Haqqani HM,
- Tschabrunn CM,
- Tzou WS,
- Dixit S,
- Cooper JM,
- Riley MP,
- Lin D,
- Hutchinson MD,
- Garcia FC,
- Bala R,
- Verdino RJ,
- Callans DJ,
- Gerstenfeld EP,
- Zado ES,
- Marchlinski FE
- Tian J,
- Ahmad G,
- Mesubi O,
- Jeudy J,
- Dickfeld T
- Hutchinson MD,
- Gerstenfeld EP,
- Desjardins B,
- Bala R,
- Riley MP,
- Garcia FC,
- Dixit S,
- Lin D,
- Tzou WS,
- Cooper JM,
- Verdino RJ,
- Callans DJ,
- Marchlinski FE
- van Huls van Taxis CF,
- Wijnmaalen AP,
- Piers SR,
- van der Geest RJ,
- Schalij MJ,
- Zeppenfeld K
- Piers SR,
- van Huls van Taxis CF,
- Tao Q,
- van der Geest RJ,
- Askar SF,
- Siebelink HM,
- Schalij MJ,
- Zeppenfeld K
- Mavrogeni S,
- Spargias C,
- Bratis C,
- Kolovou G,
- Markussis V,
- Papadopoulou E,
- Constadoulakis P,
- Papadimitropoulos M,
- Douskou M,
- Pavlides G,
- Cokkinos D
- Corrado D,
- Basso C,
- Thiene G,
- McKenna WJ,
- Davies MJ,
- Fontaliran F,
- Nava A,
- Silvestri F,
- Blomstrom-Lundqvist C,
- Wlodarska EK,
- Fontaine G,
- Camerini F
- Holmström M,
- Kivistö S,
- Heliö T,
- Jurkko R,
- Kaartinen M,
- Antila M,
- Reissell E,
- Kuusisto J,
- Kärkkäinen S,
- Peuhkurinen K,
- Koikkalainen J,
- Lötjönen J,
- Lauerma K
- Moon JC,
- Sachdev B,
- Elkington AG,
- McKenna WJ,
- Mehta A,
- Pennell DJ,
- Leed PJ,
- Elliott PM
Ventricular tachycardia ablation in patients with nonischemic left ventricular cardiomyopathy is considered to be challenging, which may partly because of the frequent intramural and subepicardial location of the arrhythmogenic substrate. Relatively little is known about typical scar patterns and its implications for the ablation strategy. In the present study, 19 consecutive patients with nonischemic cardiomyopathy who underwent contrast-enhanced MRI and ventricular tachycardia ablation are described. Two contrast-enhanced MRI–derived scar patterns (ie, basal anteroseptal and inferolateral) accounted for 89% of arrhythmogenic substrates, and 3 distinct 12-lead ECG morphologies of ventricular tachycardias were common and diagnostic for an underlying scar pattern. Basal anteroseptal scars were most effectively approached from the aortic root or basal anteroseptal left ventricular endocardium, and they were usually not well accessible from the epicardium, suggesting that an initial endocardial approach may be reasonable in these patients. In patients with inferolateral scars, however, an initial epicardial approach may be considered because epicardial ablation target sites were frequently identified in these patients. Endocardial electroanatomical voltage mapping could identify these scars only to a limited extent, with endocardial bipolar voltage only being affected when the scar involved the endocardial 2 mm and endocardial unipolar voltage when the scar involved the endocardial 4 mm and, less pronounced, by scar >4 mm from the endocardium. These findings provide insights into the arrhythmogenic substrate in patients with nonischemic cardiomyopathy and may have important practical implications for preprocedural planning and intraprocedural substrate identification in patients undergoing ventricular tachycardia ablation.