Isolated septal substrate for ventricular tachycardia innonischemic dilated cardiomyopathy: Incidence,characterization, and implications

Haris M. Haqqani, MBBS, PhD, Cory M. Tschabrunn, CEPS, Wendy S. Tzou, MD, Sanjay Dixit, MD, FHRS,Joshua M. Cooper, MD, Michael P. Riley, MD, PhD, David Lin, MD, Mathew D. Hutchinson, MD, FHRS,Fermin C. Garcia, MD, Rupa Bala, MD, Ralph J. Verdino, MD, David J. Callans, MD, FHRS,Edward P. Gerstenfeld, MD, Erica S. Zado, PA-C, FHRS, Francis E. Marchlinski, MD, FHRS

From the Section of Cardiac Electrophysiology, Cardiovascular Division, Hospital of the University of Pennsylvania,Philadelphia, Pennsylvania.

p

CBaeupm

Km

A

BACKGROUND The substrate for ventricular tachycardia (VT) innonischemic cardiomyopathy (NICM) has a predilection for thebasolateral left ventricle with right bundle branch block VT mor-phology.

OBJECTIVE The purpose of this study was to describe a uniquegroup of NICM patients with septal VT substrate.

METHODS Between 1999 and 2010, 31 (11.6%) of 266 patientswith NICM undergoing VT ablation had septal substrate and nolateral involvement. Mean age was 59 � 12 years, and ejectionfraction was 30% � 14%. Eight patients had heart block.

RESULTS Cardiac magnetic resonance showed septal delayed en-hancement in 8 of 9 patients. Electroanatomic mapping demon-strated bipolar low voltage (�1.5 mV) extending from the basalseptum in 22 of 31 patients. The remaining 9 patients had normalendocardial bipolar voltage but abnormal unipolar septal voltage(�8.3 mV) consistent with intramural abnormalities. Epicardialmapping in 14 patients showed no scar in 9 and patchy basal leftventricular summit scar in 5. VTs were mapped to the septalsubstrate, with 62% having right bundle branch block morphologyand V2 precordial transition pattern break in 17% suggesting

m

tbttsuphs.upenn.edu. (Received January 4, 2011; accepted March, 2, 2011.)

1547-5271/$ -see front matter © 2011 Heart Rhythm Society. All rights reserved

inducible in 66% and no “clinical targeted” VT in 86%. Over amean follow-up of 20 � 28 months, VT recurred in 10 (32%)atients.

ONCLUSION Isolated septal VT substrate is uncommon in NICM.iventricular low-voltage zones extending from the basal septumre characteristic, but septal scarring can be entirely intramural asvidenced by unipolar/bipolar electrograms and imaging. Multiplenmappable morphologies are the rule, often requiring severalrocedures aggressively targeting the septal substrate to achieveoderate long-term VT control.

EYWORDS Cardiomyopathy; Catheter ablation; Electroanatomicapping; Heart failure; Ventricular tachycardia

BBREVIATIONS CMR � cardiac magnetic resonance; DGE � de-layed gadolinium enhancement; EAM � electroanatomic mapping;ICD � implantable cardioverter-defibrillator; ICE � intracardiacechocardiography; LBBB � left bundle branch block; LV � left ven-tricle; NICM � nonischemic cardiomyopathy; RBBB � right bundlebranch block; RV � right ventricle; VT � ventricular tachycardia

(Heart Rhythm 2011;8:1169–1176) © 2011 Heart Rhythm Society. All

periseptal exit. After substrate and targeted VT ablation, no VT was rights reserved.

IntroductionMonomorphic ventricular tachycardia (VT) in the context ofnonischemic cardiomyopathy (NICM) can present significantmanagement challenges, and catheter ablation has emerged as

Dr. Haqqani is the recipient of an Overseas Training Fellowship(544309) from the National Health and Medical Research Council ofAustralia and the Bayer Fellowship from the Royal Australasian College ofPhysicians. Dr. Marchlinski has received honoraria and research fundingfrom Biosense Webster unrelated to this study. Address reprint requestsand correspondence: Dr. Francis E. Marchlinski, Cardiovascular Division,Hospital of the University of Pennsylvania, 3400 Spruce Street, Founders 9,Philadelphia, Pennsylvania 19104. E-mail address: francis.marchlinski@

an important therapy.1–4 Unmappable VT accounts for theajority of VT morphologies induced at these procedures,5

and the development of substrate-based ablation strategies hasallowed for the targeting and ablation of these arrhythmiasduring sinus rhythm.1 We previously showed that the substratefor VT in NICM characteristically affects the basal periannularregion of the left ventricle (LV), on both the endocardium andepicardium.6–8 Although the majority of these patients appearo have isolated basolateral substrate or both basolateral andasal septal involvement,7 this is not invariable. The purpose ofhis study was to characterize the electrophysiologic and elec-roanatomic characteristics of NICM patients with isolated

eptal substrate for VT.

. doi:10.1016/j.hrthm.2011.03.008

dc

ed“vin

sl

l�t

s

1170 Heart Rhythm, Vol 8, No 8, August 2011

MethodsBetween 1999 and 2010, 266 consecutive NICM patientswith recurrent sustained monomorphic VT underwent cath-eter ablation at the University of Pennsylvania. Of thesepatients, 30 (11.3%) identified as having isolated septalsubstrate for VT with minimal extension beyond the septumon electroanatomic mapping (EAM) or cardiac magneticresonance (CMR) imaging formed the study population.The diagnosis of NICM was established in all patients byvirtue of LV systolic dysfunction in the absence of priormyocardial infarction, occlusive coronary artery disease,valvular heart disease, or congenital heart disease. Othercauses of NICM, including sarcoidosis, arrhythmogenicright ventricular (RV) dysplasia, tachycardia-mediated car-diomyopathy, and alcoholic cardiomyopathy, were ex-cluded. Episodes of spontaneous recurrent VT were docu-mented in all patients by either 12-lead ECGs or from storedelectrograms recorded by an implantable cardioverter-defi-brillator (ICD). All patients underwent detailed electro-physiologic evaluation after providing written informedconsent in accordance with the University of PennsylvaniaHealth System’s institutional guidelines.

Preprocedural imagingAll patients underwent transthoracic echocardiography toexclude LV thrombus. CMR imaging was performed tofurther define the potential arrhythmogenic substrate in se-lected patients. CMR was performed on a clinical 1.5-Tscanner (Avanto, Siemens, Erlangen, Germany) using astandard protocol that included assessment of delayed gad-olinium enhancement (DGE). To reduce the risk of radiof-requency-related heating of the intracardiac lead in ICDpatients, the specific absorption rate was limited to 2 W/kgusing a TGRAPPA cine sequence (TR 14–43 ms, TE 1–2ms, 1.9 � 1.9-mm resolution, 8-mm slice thickness).9 Allevices were interrogated before and after scanning, with nohange in pacing threshold, lead, or battery data observed.

Endocardial electroanatomic substrate mappingPatients were studied in the postabsorptive state under con-scious sedation or general anesthesia. After systemic anti-coagulation, endocardial LV mapping was performed. En-docardial RV was mapped where necessary. An intracardiacechocardiography (ICE) catheter (Acuson AcuNav, Sie-mens, Mountain View, CA, USA) was deployed in the RVin all patients. EAM was performed in the basal rhythmusing the CARTO system (Biosense Webster, DiamondBar, CA, USA) using a 3.5-mm-tip open irrigated catheter,which contains a 2-mm ring electrode and a 1-mm inter-electrode distance (NaviStar ThermoCool, Biosense Web-ster) or a solid-tip 4-mm catheter (NaviStar, Biosense Web-ster). Bipolar electrograms were filtered at 30 to 500 Hz,displayed at 200 mm/s sweep speed and stored for offlineanalysis. Unipolar electrograms were filtered at 1 to 240 Hz.Acquired electrograms were reconstructed into three-di-mensional voltage maps and area measurements made with

the incorporated CARTO software. Mapping density was m

sufficient to allow complete surface representation with afill threshold of 15 mm. Normal bipolar endocardial elec-trogram voltage was defined as a peak-to-peak amplitude�1.5 mV; “dense scar” was defined as �0.5 mV.1 Whenndocardial bipolar voltage was preserved overall, we en-eavored to look for intramural substrate with the widerfield of view” of unipolar electrograms. As we previouslyalidated, we defined normal RV free-wall endocardial un-polar electrograms as having an amplitude �5.5 mV10 andormal LV and septal unipolar electrograms as �8.3 mV.11

Epicardial electroanatomic substrate mappingWhen required, pericardial access was obtained percutane-ously using a subxiphoid approach as described by Sosa etal.12 High-density EAM was performed, and epicardial sub-trate was defined as previously described to minimize theikelihood of epicardial fat being confused with fibrosis.8,13

Briefly, epicardial scar zones include those areas displayingbipolar signal amplitude �1.0 mV and containing pro-onged (�80 ms), split (with �2 components separated by

20 ms isoelectric interval), or isolated late (occurring afterhe surface QRS complex) electrograms.

VT mapping and catheter ablationBiventricular programmed stimulation was performed withup to three extrastimuli at �2 drive cycle lengths. InducedVTs were displayed at 100 mm/s and analyzed for cyclelength, morphology (including clues to an epicardial exitsite14,15), and hemodynamic tolerability. The diagnosis ofbundle branch reentrant VT was assessed with a His record-ing during tachycardia. Mappable VTs were then targetedusing entrainment techniques and ablated using an open-irrigated ablation catheter (NaviStar ThermoCool, BiosenseWebster) with power of 30 to 50 W titrated to an impedancedrop of 12 to 18 � (constrained by an electrode tip temper-ature limit of 45°C) or with a 4-mm-tip nonirrigated catheter(NaviStar, Biosense Webster) in the earlier patients. Un-mappable tachycardias were ablated using substrate-basedablation strategies with the deployment of linear or clus-tered lesions targeting putative exit sites with good pace-maps, sites of long stimulus to QRS delay, higher voltagechannels between islands of dense scar, and isolated latepotentials within dense scar. Deep intramural septal sub-strate was targeted with biventricular ablation from bothsides of the septum. The procedural endpoint was the non-inducibility of all targeted VTs.

Follow-up assessmentWherever possible, antiarrhythmic drugs were stopped fol-lowing ablation or, in the case of amiodarone, the dose wasreduced to �200 mg/day. Recurrence of clinically docu-mented VT was generally managed with repeat catheterablation. Patients were followed at regular intervals follow-ing discharge. An ICD monitor zone was used to detect slowVTs. Sustained VT episodes were defined as lasting �15econds or requiring clinical or device intervention for ter-

ination.

wcpv

cmtbd

1111111111222222222233

dine; S

1171Haqqani et al Septal Substrate for VT in Dilated Cardiomyopathy

Statistical analysisAll continuous data are expressed as mean � 1 SD or asmedian with range where a normal distribution was ex-cluded with a one-sample Kolmogorov-Smirnov test.

ResultsBaseline characteristicsBaseline characteristics of the 31 patients in this study arelisted in Table 1. Mean age was 57 � 12 years, and 26(84%) were men. Patients underwent 50 procedures in total(range 1–4). Mean LV ejection fraction was 30% � 14%,and mean LV end-diastolic dimension was 62 � 11 mm.Preexisting complete AV block was noted in 8 patients, withleft bundle branch block (LBBB) in another 10. Six patientswere paced in the RV, and 6 others had biventricular pacing.A total of 26 (84%) patients had an ICD. All patients weretreated with beta-blockers, and the majority were treatedwith antiarrhythmic drugs prior to ablation (Table 1). Tenpatients had undergone a mean of 1.7 prior ablation proce-dures (range 1–3) at other centers.

Endocardial substrateAll patients underwent detailed endocardial LV EAM (305 �

Table 1 Baseline characteristics

Patientno.

Age(years)

Left ventricular ejectionfraction (%) ICD

1 58 10 Y2 70 35 Y3 74 39 Y4 68 35 Y5 58 20 Y6 42 10 Y7 46 35 Y8 54 40 Y9 48 45 Y0 52 10 Y1 70 15 Y2 59 35 N3 53 45 N4 51 35 Y5 69 15 Y6 64 10 Y7 50 45 N8 78 35 Y9 60 10 Y0 70 30 N1 61 20 Y2 73 35 Y3 48 45 Y4 44 25 N5 43 45 Y6 65 30 Y7 60 18 Y8 34 30 Y9 56 30 Y0 59 30 Y1 33 40 Y

AAD � antiarrhythmic drug; Amio � amiodarone; Azim � azimilide; BBLido � lidocaine; Mex � mexiletine; Proc � procainamide; Quin � quini

121 points sampled). Endocardial LV access was via the

retrograde approach in 40 (80%) of 50 patients and trans-septal in the remainder. Lateral LV wall voltages and elec-trograms were normal in all patients. The majority ofpatients (22/31 [71%]) displayed bipolar low-voltage ar-eas on the basal midseptum extending to the inferosep-tum and anteroseptum, often involving the basal anteriorwall. No zone of periannular sparing was seen (Figure 1).The dense scar area in these patients averaged 28 � 23cm2 (or 11% � 8% of LV endocardial surface area),

hereas the total area of low voltage (�1.5 mV) was 63 � 41m2 (or 26% � 13% of LV surface area). Nine (29%)atients had completely normal endocardial LV bipolaroltage maps.

A total of 17 (57%) patients also underwent RV endo-ardial EAM (188 � 130 points) when suggested by VTorphology and by failure of LV ablation to eliminate all

argeted VTs. Of these patients, 9 (53%) displayed normalipolar voltage. Mean bipolar dense scar area in the remain-er was 10 � 14 cm2 (or 6% � 8% of RV surface area),

whereas the total low-voltage zone averaged 23 � 27 cm2

(or 14% � 16%). The low-voltage zones were found in theanterior or midseptal regions of the RV, or the septal aspect

AAD prior to procedurePreexistingheart block

No. of priorprocedures

Amio/Mex/Lido N 0Amio/Mex/Lido N 1Amio/Sot N 0Sot/Mex Y 2Amio/Lido N 2BB N 0Amio N 1Amio Y 1Amio/Mex N 0Amio/Lido N 0Amio/Mex/Lido N 0BB N 0BB N 0Sot/Mex Y 0Amio/Quin/Lido N 1Amio/Lido N 0Lido/Proc N 0Amio/Sot/Mex N 0Amio/Sot/Lido N 0Amio/Lido N 0Amio/Lido N 1Amio N 3Amio/Sot/Azim N 0Amio N 2Amio/Mex N 0Sot N 0Amio/Proc N 3BB N 0Amio N 0Amio N 0Amio/Flec/Sot N 0

a-blocker; Flec � flecainide; ICD � implantable cardioverter-defibrillator;ot � Sotalol.

� bet

of the infundibulum.

evul8

it

s4pep

1172 Heart Rhythm, Vol 8, No 8, August 2011

Unipolar low-voltage zones (amplitude �8.3 mV) werexamined in the 8 patients with normal endocardial bipolaroltage. In all of these patients, large confluent areas ofnipolar low voltage were identified on the septum, withateral sparing seen. Mean LV unipolar substrate area was0 � 36 cm2, or 40% � 16% of LV surface area. Although

the septum was always included in this area, there wassignificant extension to the anterior and inferior walls (Fig-ure 2C). Mean RV unipolar substrate area was 72 � 51 cm2

(or 43% � 34%).

Epicardial substrateA total of 14 (45%) patients underwent epicardial mappingfor one or more of the following reasons: (1) presence ofspecific 12-lead ECG features suggestive of an epicardialVT exit; (2) presence of significant suspected epicardialsubstrate on the basis of imaging or unipolar electrograms;and (3) failure of endocardial ablation. Eight of these hadunremarkable epicardial electroanatomic maps, with onlythe expected pericoronary or periannular low-voltage zonesdue to presumed epicardial fat; no abnormal electrogramsseen (Figure 2B). The remaining 5 patients had a circum-scribed zone (mean area 12 � 6 cm2) of fractionated orsolated late potentials in the anterobasal preaortic region of

Figure 1 A1: Left ventricular endocardial electroanatomic bipolar vocontiguously from the basal midseptum and involving the basal anterior preand black, respectively. The anterior, lateral, and inferior walls show preseeptal echogenic stripe (arrows) corresponding to the low-voltage zone2-year-old woman with extensive confluent septal scar. She had multiprogressive heart failure afterward and underwent cardiac transplantationdocardial fibrosis extending from the septal periannular region and coosteroanterior; RAO � right anterior oblique.

he LV summit.

Periprocedural imaging resultsPreprocedural CMR was performed in 9 (29%) of the31 patients, 7 of whom had an ICD in situ that variablycompromised image quality but did not preclude satisfac-tory assessment of the septum in any patient. Eight (89%) ofthe 9 patients had septal DGE seen on CMR. Intramuralseptal DGE was seen in 7 patients, whereas transmural DGEextending to the endocardium of the RV and LV (withcontiguous involvement of the basal preaortic area) wasseen in one. Minor degrees of additional inferolateral, pre-dominantly midmural, DGE occurred in 3 patients (Figure2D). No complications were noted by CMR imaging in ICDpatients.

Imaging of the septum with ICE showed intramural mid-septal echogenic abnormalities in 5 of the 31 patients (Fig-ure 1A). Two of them had also undergone prior CMRshowing DGE in a similar region. These patients all hadconfluent septal bipolar low-voltage zones recorded on LVendocardial EAM.

VT characteristicsVT characteristics are summarized in Table 2. Inductionwas by programmed stimulation for 50% of the VTs, burstpacing for 12%, catheter manipulation for 24%, and spon-taneously for the remainder. Termination occurred by rapid

aps from a 69-year-old man showing dense periannular scar extendingegion. Numerous fractionated and isolated potentials are annotated in pinkoltage. A2: Intracardiac echocardiogram from the same patient showing aateral sparing is again demonstrated in the electroanatomic maps of a

ricular tachycardias mapped and ablated on the septal substrate but hadThe explanted heart of the patient shown in B1 shows confluent thickding to the low-voltage zone seen on electroanatomic mapping. PA �

ltage maortic rrved v

. B1: Lle ventn. B2:rrespon

pacing in 40% of tachycardias, cardioversion in 16%, ra-

r

cntramur

QMLLSDVP

1173Haqqani et al Septal Substrate for VT in Dilated Cardiomyopathy

diofrequency ablation in 15%, and spontaneously in the rest.Because most VTs were not mappable by entrainment,putative VT exits were determined predominantly by pace-mapping (Figure 3). Patients with intact AV conduction hada mean HV interval of 58 � 12 ms, and bundle branchreentry was not established as the mechanism of tachycardiain any of the VTs.

VT morphologies were variable but predominantly rightbundle branch block (RBBB), with both superior and infe-rior axis VTs seen. Two particular configurations werenoted. First, of the 151 VTs, 24 (16%) displayed an inferiorlimb lead discordance, usually LBBB configuration with anet positive vector in II, negative in III, and positive in aVL.

Figure 2 Intramural septal scar in a 33-year-old man with nonischemic cventricular tachycardia (one of four VTs). A: Normal left ventricular anpatient’s epicardial substrate map was unremarkable, with no abnormal eldistribution (B). C: Unipolar endocardial voltage maps showing septal lowonfirmed by cardiac magnetic resonance imaging (D), which showed exten

tachycardia was mapped to the septal infundibular region overlying this i

Table 2 Ventricular tachycardia characteristics

No. of ventricular tachycardias 151Cycle length (ms) 396 � 90RS duration (ms) 186 � 34appable 54 (36%)eft bundle branch block configuration 57 (38%)eft/right axis (%) 48/52uperior/inferior axis (%) 30/60iscordant inferior leads 15 (10%)2 precordial pattern break 26 (17%)

recordial limb lead amplitude (mV) 0.92 � 0.12

This morphology was ablated from the basal midseptalparahisian region, from either the RV or LV side. Three ofthese VTs displayed RBBB configuration with reverseddiscordance (negative vector in II and positive in III) andwere ablated from the LV side of the basal midseptal region.The second morphology accounted for 36 (24%) of 151 VTsand was characterized by a precordial transition patternbreak in V2 (with qR/Rs morphology in V1 and V3 buteversal of this in V2) usually in the presence of an inferior

axis RBBB configuration (Figure 3). This morphology wasmapped to the superior LV septum with exit in the preaorticLV summit region and was ablated from a number oflocations in this vicinity, including the superior LV septum,basal preaortic epicardium, anterior interventricular vein,and left coronary cusp . A different pattern of V2 precordialtransition pattern break was seen in three cases with asuperior axis and broad R/S, atypical LBBB configurationin V1. This morphology was mapped/ablated in the infero-septal endocardium or in the opposite epicardial region.

Procedural resultsAblation targeted all induced VTs and required a total of 38 �21 radiofrequency applications per procedure, with a meanablation duration of 2,455 � 1,692 seconds. Mean proce-

yopathy (ejection fraction 40%) and left bundle branch block configurationventricular endocardial bipolar voltage in the posteroanterior view. Theams and preserved bipolar voltages apart from the expected pericoronarye (�8.3 mV), raising suspicion of the deeper intramural process. This wastramural septal delayed gadolinium enhancement (arrows). The ventricularal fibrosis. LAO � left anterior oblique.

ardiomd rightectrogrvoltagsive in

dural time was 361 � 134 minutes. One procedure was

ah

Vo

1174 Heart Rhythm, Vol 8, No 8, August 2011

aborted due to pericardial bleeding late after routine epicar-dial access. The bleeding stopped with conservative man-agement. Another procedure was aborted due to the earlydevelopment of cardiogenic shock. In the remaining 48ablations, acute procedural success (as defined by nonin-ducibility of all targeted VTs) was achieved in 33 (66%)procedures. Partial success, in which the clinical targetedVT was no longer inducible, was observed in an additional10 (20%) procedures. Clinical VTs remained inducible after7 (14%) procedures.

Five patients developed complete AV block during ab-lation (including septal alcohol ablation in one) and newLBBB developed in two other cases. Conduction resumedby the end of the procedure in two of these patients. Onepatient developed angiographic stenosis of the left anteriordescending artery after endocardial ablation in the antero-septal aspect of the RV infundibulum. This resolved onrepeat angiography 2 days later and was presumed to be dueto spasm.

Long-term outcomeTwelve patients required more than one procedure (mean1.6 per patient, range 2–4) to achieve VT control. Overmean follow-up of 20 � 28 months (median 12 months)fter the last procedure, 8 (26%) patients died, 2 of whom

Figure 3 Endocardial left ventricular (LV) and infundibular right ventri(LV ejection fraction 35%) and isolated septal substrate. The patient’s induconfiguration VT with inferior axis (VT 1) was pace-mapped to the septaV2 precordial transition pattern break was pace-mapped to the anterior inter

T was ablated from the endocardial preaortic region underneath the leftblique; PA � posteroanterior.

ad recurrent VT. Five (16%) patients underwent cardiac b

transplantation for heart failure, none of whom had recur-rent VT. Of the 18 remaining patients, 9 have had norecurrent VT, with 6 on antiarrhythmic drugs, includingamiodarone in 4. In total, VT recurred in 10 (32%) patients.No clinical, procedural, or VT characteristics predicted ar-rhythmia recurrence.

DiscussionThis is the first report to characterize a group of patientswith NICM who have an isolated septal substrate for VT asidentified by EAM or CMR imaging. They represent 11.6%of NICM patients undergoing catheter ablation for VT froma large consecutive series. The location of the VT substratein the anatomically complex region of the ventricular sep-tum results in a number of salient characteristics, includingthe following:

A. Preponderance of endocardial LV or RV low-voltagezones with relative sparing of the periseptal epicardium

B. Possibility of an exclusively intramural substrate deepwithin the interventricular septum that may be evident onlyon imaging or with unipolar voltage mapping

C. Presence of multiple left and right bundle branchblock VTs with variable axis and early breakthrough in theperiseptal region leading to precordial V2 transition pattern

V) voltage maps of a 68-year-old man with nonischemic cardiomyopathytricular tachycardia (VTs) were not mappable. A left bundle branch block

fundibular region. A right bundle branch block VT with inferior axis andlar vein, where proximity to the circumflex artery precluded ablation. Thisry cusp. AoV � aortic valve; MV � mitral valve; LAO � left anterior

cular (Rced ven

l RV inventricucorona

reak

ciif

achftaisoctu

btaopfse

iciparrisiwbtthptfrpmt

1175Haqqani et al Septal Substrate for VT in Dilated Cardiomyopathy

D. Incomplete response to ablation of all inducible VTsand risk of damage to the conduction system during ablationin this area.

The presence of septal scar in ischemic cardiomyopathyhas been well described, usually in association with exten-sive anterior or inferior infarction. Postinfarction VT arisingfrom septal scar has been studied in both animal16,17 andlinical17,18 models. The difficulties of precisely establish-ng septal breakout sites of VT and the challenges in ablat-ng these circuits (often requiring that lesions be deliveredrom both sides of the septum) are well recognized.16,18,19

In the postinfarction situation, the scarring progresses intra-murally to a variable depth from the subendocardium andcharacteristically spares the basal periannular region. InNICM, however, the fibrotic process is more heterogeneousand geometrically more complex. We previously describedthe stereotypical basolateral perivalvular low-voltage zoneswith contiguous extension back to the annular plane,7 aswell as the propensity for extensive transmural and epicar-dial involvement.8 Although some NICM patients may haveseptal involvement in addition to the basolateral process, thepatients in the current series displayed isolated scarring ofthe septum and adjacent areas with no basolateral involve-ment on EAM or CMR imaging.

Given its potential intramural location within the septum,the full extent of fibrosis may be difficult to determine.While the majority of patients have large contiguous endo-cardial LV or RV septal low-voltage zones, a significantnumber of patients have preserved bipolar endocardial volt-age. Mapping the epicardium would be expected to helpwhen no endocardial substrate is revealed in the usualNICM patient in whom a large basolateral epicardial scarmay be present. However, when epicardial mapping wasperformed in these patients with a septal substrate, either noscar was seen or only a small abnormal area was present inthe basal preaortic region. The large intramural septal scarcould remain undetected in such cases without the use ofimaging20 and unipolar voltage mapping. The presence ofn ICD remains a contraindication to CMR imaging in mostenters; however, as part of an investigational protocol, weave imaged increasing numbers of ICD patients. Artifactrom the pulse generator or lead remains a limitation, but inhe 9 patients from this series who were imaged, the locationnd extent of the septal substrate were identified by CMRmaging in 8. Real-time ICE imaging was able to identifyeptal echogenic zones in a proportionately smaller numberf patients but was useful because it allowed imaging of theatheter tip and lesion formation relative to the location ofhe septal substrate. We previously identified the utility ofnipolar voltage mapping in both the LV11 and RV10 for

recognition of deeper layers of fibrosis when bipolar voltageis normal. In these NICM patients with intramural septalsubstrate, the unipolar voltage maps were useful for identi-fying the true extent of the substrate abnormality.

The site of origin of the VTs induced in these patients

was mapped to the region of the septal substrate by pace-

mapping and by entrainment mapping when possible. Whenarising from the RV or midinterventricular septum, VTmostly had an LBBB configuration. Several of these VTsshowed discordance in the inferior limb lead vector andwere ablated from either the LV or RV side of the basalparahisian midseptal region. The high proportion of VTswith RBBB morphology was accounted for by VTs exitingfrom sites adjacent to the true interventricular septum, in-cluding the basal anteroseptal and preaortic region, the an-terior interventricular vein, and the epicardial LV summit. Acommon ECG signature seen with the RBBB morphologyVTs was a precordial transition pattern break in lead V2.The unexpected negative vector in V2 in these cases woulde expected to be the result of an earliest breakthrough onhe epicardial aspect of the LV summit, but in many casesblation was not possible here using either a percutaneousr transvenous (via the anterior interventricular vein) ap-roach due to the proximity of the coronary arteries. There-ore, these VTs were often targeted from contiguous regionsuch as the left coronary cusp and, when demonstrated byntrainment mapping, superior septal mid-isthmus sites.

Catheter ablation of VT in this group of NICM patientss challenging for several reasons, including (1) the diffi-ulty in mapping multiple nontolerated VTs, (2) the oftenncomplete delineation of arrhythmogenic substrate, (3) theossibility of collateral damage during epicardial ablation,nd (4) the inability to deliver transmural lesions with cur-ently available ablation technology. These difficulties areeflected in the 66% acute success rate in eliminating allnducible VTs, although the rate rose to 86% when defininguccess by elimination of the clinical VT. The significantncidence of heart block or LBBB caused by ablation isorthy of note, as is the proximity of the coronary arteriesoth with superior or inferior endocardial septal RV abla-ion, or with any consideration of epicardial ablation. For-unately, 5 of 7 patients with heart block or LBBB alreadyad a biventricular device, and aggressive ablation was inart pursued because of the presence of this treatment op-ion. Moreover, the long-term outcome of these patientsollowing ablation is mixed, particularly because they are atisk for progressive heart failure requiring mechanical sup-ort or transplantation. Despite the good outcome in theajority of patients, recurrent VT was seen in one third of

his cohort during long-term follow-up.

Study limitationsThis study was a nonrandomized analysis of a sequentialcase series of NICM patients undergoing VT ablation. Exitsites of VT were defined using pace-mapping in the major-ity of cases (as only a minority were mappable with entrain-ment), and this may have led to inaccurate localization.Registration of periprocedural imaging into the electroana-tomic map was not performed in this study. An evolution inthe routine use of CMR imaging in ICD patients, epicardialmapping, and irrigated-tip ablation occurred during the

study period and may have affected the procedural results.

1

1

1

1

1

1

1

1

1

1

2

1176 Heart Rhythm, Vol 8, No 8, August 2011

ConclusionIsolated or predominant septal substrate for scar-related VTin NICM represents an uncommon but challenging problemfor successful catheter ablation. Biventricular endocardiallow-voltage zones extending from the basal septum (with orwithout patchy epicardial involvement) are characteristic,but septal scar may be intramural and seen best on CMRimaging. Variable VT morphologies (both left and rightbundle branch block) are the rule, and a characteristic pre-cordial transition pattern break may be seen in lead V2.Multiple procedures may be required for successful VTablation, with the potential to affect the conducting system.

References1. Marchlinski FE, Callans DJ, Gottlieb CD, Zado E. Linear ablation lesions for

control of unmappable ventricular tachycardia in patients with ischemic andnonischemic cardiomyopathy. Circulation 2000;101:1288–1296.

2. Soejima K, Stevenson WG, Sapp JL, Selwyn AP, Couper G, Epstein LM.Endocardial and epicardial radiofrequency ablation of ventricular tachycardiaassociated with dilated cardiomyopathy: the importance of low-voltage scars.J Am Coll Cardiol 2004;43:1834–1842.

3. Sacher F, Tedrow UB, Field ME, et al. Ventricular tachycardia ablation: evo-lution of patients and procedures over 8 years. Circ Arrhythm Electrophysiol2008;1:153–161.

4. Nakahara S, Tung R, Ramirez RJ, et al. Characterization of the arrhythmogenicsubstrate in ischemic and nonischemic cardiomyopathy implications for catheterablation of hemodynamically unstable ventricular tachycardia. J Am Coll Car-diol 2010;55:2355–2365.

5. Cesario DA, Vaseghi M, Boyle NG, et al. Value of high-density endocardial andepicardial mapping for catheter ablation of hemodynamically unstable ventric-ular tachycardia. Heart Rhythm 2006;3:1–10.

6. Hsia HH, Marchlinski FE. Characterization of the electroanatomic substrate formonomorphic ventricular tachycardia in patients with nonischemic cardiomy-opathy. Pacing Clin Electrophysiol 2002;25:1114–1127.

7. Hsia HH, Callans DJ, Marchlinski FE. Characterization of endocardial electro-

physiological substrate in patients with nonischemic cardiomyopathy and mono-morphic ventricular tachycardia. Circulation 2003;108:704–710.

8. Cano O, Hutchinson M, Lin D, et al. Electroanatomic substrate and ablationoutcome for suspected epicardial ventricular tachycardia in left ventricularnonischemic cardiomyopathy. J Am Coll Cardiol 2009;54:799–808.

9. Naehle CP, Strach K, Thomas D, et al. Magnetic resonance imaging at 1.5-T inpatients with implantable cardioverter-defibrillators. J Am Coll Cardiol 2009;54:549–555.

0. Polin GM, Haqqani HM, Tzou WS, et al. Endocardial unipolar voltage mappingto identify epicardial substrate in arrhythmogenic right ventricular dysplasia/cardiomyopathy. Heart Rhythm 2011;8:76–83.

1. Hutchinson MD, Gerstenfeld EP, Desjardins B, et al. Endocardial unipolarvoltage mapping to detect epicardial VT substrate in patients with nonischemicleft ventricular cardiomyopathy. Circ Arrhythm Electrophysiol 2011;4:49–55.

2. Sosa E, Scanavacca M, d’Avila A, Pilleggi F. A new technique to performepicardial mapping in the electrophysiology laboratory. J Cardiovasc Electro-physiol 1996;7:531–536.

3. Garcia FC, Bazan V, Zado ES, Ren JF, Marchlinski FE. Epicardial substrate andoutcome with epicardial ablation of ventricular tachycardia in arrhythmogenicright ventricular cardiomyopathy/dysplasia. Circulation 2009;120:366–375.

4. Valles E, Bazan V, Marchlinski FE. ECG criteria to identify epicardial ventric-ular tachycardia in nonischemic cardiomyopathy. Circ Arrhythm Electrophysiol2010;3:63–71.

5. Bazan V, Gerstenfeld EP, Garcia FC, et al. Site-specific twelve-lead ECGfeatures to identify an epicardial origin for left ventricular tachycardia in theabsence of myocardial infarction. Heart Rhythm 2007;4:1403–1410.

6. Sivagangabalan G, Pouliopoulos J, Huang K, et al. Simultaneous biventricularnoncontact mapping and ablation of septal ventricular tachycardia in a chronicovine infarct model. Circ Arrhythm Electrophysiol 2009;2:441–449.

7. Kawamura Y, Page PL, Cardinal R, Savard P, Nadeau R. Mapping of septalventricular tachycardia: clinical and experimental correlations. J Thorac Cardio-vasc Surg 1996;112:914–925.

8. Patel VV, Rho RW, Gerstenfeld EP, Hsia HH, Callans DJ, Marchlinski FE.Right bundle-branch block ventricular tachycardias: septal versus lateral ven-tricular origin based on activation time to the right ventricular apex. Circulation2004;110:2582–2587.

9. Menz V, Duthinh V, Callans DJ, Schwartzman D, Gottlieb CD, Marchlinski FE.Right ventricular radiofrequency ablation of ventricular tachycardia after myo-cardial infarction. Pacing Clin Electrophysiol 1997;20:1727–1731.

0. Assomull RG, Prasad SK, Lyne J, et al. Cardiovascular magnetic resonance,fibrosis, and prognosis in dilated cardiomyopathy. J Am Coll Cardiol 2006;48:

1977–1985.