Scar Characteristics for Prediction of VentricularArrhythmia in Ischemic CardiomyopathySHERIF GOUDA, M.SC.,* AMIR ABDEL WAHAB, M.D.,* MOHAMED SALEM, M.D.,†and MAGDY ABDEL HAMID, M.D.*From the *Department of Cardiovascular Medicine, Cairo University, Cairo, Egypt; and †Department ofRadio-Diagnosis, Cairo University, Cairo, Egypt

Background: Better risk-stratification tools are needed to identify the best candidates for implantablecardioverter defibrillator implantation. Infarct characterization by cardiac magnetic resonance (CMR) hasbecome an evolving potential tool for risk stratification.

Objective: We assessed the ability of scar characteristics by CMR in patients with postinfarction leftventricular (LV) dysfunction to predict sustained monomorphic ventricular tachycardia (SMVT).

Methods: Forty-eight patients with postinfarction LV dysfunction underwent CMR study. Twenty-fourpatients had history of SMVT and the other 24 were control group and underwent electrophysiologicalstudy to assess SMVT inducibilty. Various scar characteristics were assessed in the spontaneous SMVTgroup and were compared with the inducible and noninducible SMVT groups.

Results: Only six patients in the control group had inducible SMVT. In univariable analysis, total scar(absolute and as percent of LV), scar core (absolute and as percent of LV), peri-infarct zone (absolute andas percent of LV), mean infarct transmurality, and number of segments with late gadolinium enhancement(LGE) were statistically significant predictors of spontaneous SMVT experience and SMVT inducibility. Inmultivariable analysis, total infarct as percent of LV mass was the only significant independent predictor ofspontaneous SMVT experience (odds ratio [OR] 1.33 per% change, 95% confidence interval [CI] 1.12–1.6,P = 0.001) and SMVT inducibility (OR 1.3 per% change, 95% CI 1.1–1.6, P = 0.004).

Conclusion: Characterization of myocardial infarct by LGE-CMR, specifically total infarct size, is betterpredictor of spontaneous SMVT experience and SMVT inducibility than LV ejection fraction. (PACE 2014;00:1–8)

sudden cardiac death, ventricular tachycardia, ischemic cardiomyopathy, cardiac magneticresonance

IntroductionMost patients with coronary artery disease

(CAD) die of sudden cardiac death (SCD) orcongestive heart failure.1 Currently, risk stratifi-cation for SCD relies on left ventricular ejectionfraction (LVEF) for eligibility for implantablecardioverter defibrillator (ICD) implantation.2–6

However, LVEF has its limitations as risk iden-tifier. In population studies, up to 70% of patientssuffering SCD have a preserved LVEF and arenot identified for prophylactic ICD implantation.7SCD in patients with CAD is predominantlycaused by ventricular tachycardia (VT) or ventric-ular fibrillation (VF).2,8 Given that approximately

Conflict of Interest: none.

Address for reprints: Sherif Gouda, M.Sc., Department of Car-diovascular Medicine, Cairo University, 12451 Sayeda KhadijaStreet, 6th October City, Cairo, Egypt. Fax: +2038350933;e-mail: sherifgoda@yahoo.com

Received July 3, 2014; revised September 9, 2014; acceptedOctober 19, 2014.

doi: 10.1111/pace.12536

14–18 patients with left ventricular (LV) dysfunc-tion need to have an ICD implanted to preventone death3,6 and considering the substantialcost,9 better risk-stratification tools are needed toidentify the best candidates for ICD implantation.Late gadolinium enhancement cardiac magneticresonance (LGE-CMR) delineates, with high spa-tial resolution, regions of scar tissue potentiallyforming part of the arrhythmia substrate inpatients with ischemic cardiomyopathy.10 Infarctcharacterization by CMR has become an evolvingpotential tool for risk stratification.11–14 In thisstudy, we sought to assess scar characteristics byLGE-CMR in patients with LV dysfunction dueto prior myocardial infarction (MI) and history ofspontaneous sustained monomorphic VT (SMVT)and compare them with a control group subjectedto electrophysiological study (EPS).

MethodsStudy Population

Study population included 48 subjects. Werecruited 24 consecutive patients with LV dys-function due to prior MI (LVEF < 50%) and

© 2014 Wiley Periodicals, Inc.

PACE, Vol. 00 2014 1

GOUDA, ET AL.

history of spontaneous SMVT at least 1 monthafter the most recent MI (according to definitionof healed MI).15 We also recruited a control groupof 24 consecutive patients with stable CAD andLV dysfunction (LVEF < 50%) but without historyof spontaneous SMVT, syncope, or cardiac arrestwho had their most recent MI at least 1 monthbefore enrollment.

Patients with contraindication for magneticresonance imaging (MRI) due to implantationof medical devices and those with creatinineclearance �30 mL/min were excluded from thestudy.

Electrophysiological Study

Patients in the control group had undergoneEPS. The stimulation protocol included pacingoutput at twice the diastolic threshold currentand a pulse width of 2 ms. Single ventricularextrastimuli (VESs) at pacing drive cycle lengthsof 600 ms and 400 ms were delivered, first fromthe right ventricular apex and then from theright ventricular outflow tract. The prematurityof extra-stimuli was increased until refractorinessor induction of SMVT is achieved. If that fails toinduce VT, double and then triple VESs were usedin the same manner. We limited prematurity of theVESs to a minimum of 200 ms. The EPS end pointsincluded the induction of SMVT or completion ofthe protocol. All patients gave written informedconsent for the EPS.

MRI Protocol

After recruitment, all patients underwentCMR using a 1.5-T scanner (Philips Achieva,Philips Medical Systems, Eindhoven, the Nether-lands) with phased-array receiver coil placedon the chest. Cine images were acquired witha steady-state free-precession pulse sequence inlong-axis planes and 10–12 contiguous short-axis slices covering the left ventricle. Imagingparameters were as follows: repetition time (TR)2.9 ms, echo time (TE) 1.4 ms, average in-plane resolution 1.25 × 1.25 mm, flip angle 60°,8-mm-slice thickness, 2-mm gap, and temporalresolution of 40 ms.

Fifteen to twenty minutes after bolusintravenous administration of 0.2 mmol/kg ofgadolinium-based contrast agent (Magnevist;Schering, Berlin, Germany), delayed contrast-enhanced images were acquired using segmentedinversion-recovery fast gradient-echo pulsesequences16 in the same short-axis locations as thecine images. Look-Locker sequence was obtainedfirst to choose the optimal inversion time (TI) forsubsequent infarct imaging. Imaging parameterswere as follows: TR 6.1 ms, TE 2.9, averagein-plane spatial resolution 1.25 × 1.25 mm,

8-mm-slice thickness, 2-mm gap, TI 175–300 ms(adjusted as needed to optimally null the signalof normal myocardium), and flip angle of 25°.

MRI Analysis

The CMR Digital Imaging and Communica-tions in Medicine (DICOM) images were analyzedusing the freely available software Segmentversion 1.9 R3248 (http://segment.heiberg.se).17

LVEF, LV volumes, and LV mass were com-puted from the short-axis cine images aftersemiautomated contour tracing of endocardial andepicardial borders of the left ventricle.

Subsequently, LGE-CMR images were ana-lyzed using the same software. The location ofhyperenhanced segments on LGE-CMR imageswas determined by visual inspection with theAmerican Heart Association 17-segment model.18

We also determined the mean transmural extentof the infarct. Each short-axis image was dividedinto 80 radial sectors then for each sector,transmurality was calculated and the mean infarcttransmurality was derived. After the endocardialand epicardial borders were traced, the hyperen-hanced areas were outlined as the scar region.A region of interest was traced in the remoteviable myocardium. Using 3-SD threshold abovereference myocardium to define total scar willinclude the gray zone.19 Accordingly we usedthreshold of 5 SD above the mean signal intensity(SI) of remote myocardium to define the infarctcore and the peri-infarct zone was defined as SIbetween 3 SD and 5 SD greater than the remote.The total scar was defined as the sum of the scarcore, and the peri-infarct zone. Total scar, scarcore, and peri-infarct zone were also expressed aspercentages of the total LV mass. Figure 1 showsshort-axis LGE-CMR images in two of the studysubjects.

Statistical Analysis

Continuous variables are presented as mean ±standard deviation (SD) for normally distributeddata and as median and interquartile range (IQR)for nonnormally distributed variables. Categor-ical data are summarized as frequencies andpercentages. Fisher’s exact was performed fornoncontinuous variables. Comparison of multipledata sets was performed using analysis of variancefor normally distributed continuous variables,and specific differences were identified by aStudent’s t-test. Kruskall-Wallis test was used forcomparison of nonparametric parameters.

A multivariable logistic regression was per-formed to identify independent predictors of spon-taneous SMVT experience and SMVT inducibility.Interdependent variables were not included inthe same regression model. Receiver operating

2 2014 PACE, Vol. 00

PREDICTION OF VENTRICULAR ARRHYTHMIA IN ISCHEMIC CARDIOMYOPATHY

Figure 1. LGE-CMR short-axis images of two of the study population, one (A) with large anterior MI with overlyingthrombus (arrow) and the other (B) with inferolateral MI. In both (A) and (B), green and red contours correspondto epicardial and endocardial borders, respectively; yellow contour encircles the scar. CMR = cardiac magneticresonance; LGE = late gadolinium enhancement; MI = myocardial infarction.

characteristic (ROC) curve analysis was performedto define cut-off points for total infarct size aspercent of LV for prediction of spontaneous SMVTand inducibility of SMVT. Areas under the curve(AUC) were calculated.

All tests were two-sided, and a P value lessthan 0.05 was considered statistically significant.All analyses were performed using the SPSSsoftware package 22.0 (IBM Corp., Armonk, NY,USA).

ResultsStudy Population

Baseline characteristics of the study pop-ulation are summarized in Table I. In brief,48 patients were enrolled, 24 with history ofspontaneous SMVT. Out of the other 24 patients,six developed SMVT by programmed electricalstimulation (PES). Patients who had noninducibleVT during the EPS were less likely to be smokerscompared to the other two study groups. Most ofthe spontaneous VT group was taking amiodaroneat enrollment. None of the other two groups wason amiodarone before the EPS. No other statis-tically significant differences in baseline charac-teristics were observed among the three studygroups.

In the 24 patients who underwent EPS, sixpatients had eight different SMVT with cyclelength of 265 ± 54 ms and QRS duration of

175 ± 31 ms. In five patients, only single VTmorphology was induced in one patient; threemorphologies were captured due to a changeinduced by ventricular overdrive pacing.

MRI Variables

MRI findings are listed in Table II. LVEF in theentire study population was 23% (IQR 17–32%)and it was not statistically different among thestudy groups (P = 0.25). LV end-diastolic volume(EDV), LV end-systolic volume (ESV), and LV masswere also comparable among the three groups withP values of 0.2, 0.3, and 0.9, respectively. Allpatients had LGE by CMR indicating the presenceof prior MI.

Total scar mass, absolute (54, IQR 41–70 g vs19, IQR 12–28 g, P <0.001) and as percent of LVmass (26, IQR 23–40% vs 9, IQR 5–14%, P < 0.001)was larger in spontaneous SMVT group than thenoninducible group. Scar core mass, absolute (45,IQR 35–63 g vs 16, IQR 9–24 g, P < 0.001) andas percent of LV mass (22, IQR 19–35% vs 8, IQR4–13%, P < 0.001) was larger in spontaneousSMVT group than the noninducible group. Peri-infarct zone mass, absolute (8, IQR 5–12 g vs2.3, IQR 1.2–5 g, P < 0.001) and as percent ofthe LV mass (4.5, IQR 2.6–6% vs 1, IQR 0.7–2%,P < 0.001) was greater in spontaneous SMVTgroup than the noninducible group. Additionally,the number of segments with LGE (8, IQR 6–9

PACE, Vol. 00 2014 3

GOUDA, ET AL.

Table I.

Clinical Baseline Characteristics

Total Spontaneous Inducible Noninducible P Value*Population SMVT SMVT SMVT (Difference

Variable (n = 48) (n = 24) (n = 6) (n = 18) among Groups)

Male 48 (100%) 24 (100%) 6 (100%) 18 (100%)Age (years) 60 ± 10 59 ± 11 52 ± 5 63 ± 9 0.07Location of prior MI 0.05

Anterior ± other territory 33 (69%) 18 (75%) 6 (100%) 9 (50%)Inferior and/or lateral only 15 (31%) 6 (25%) 0 (0%) 9 (50%)

Duration since last MI (months) 8.5 (2–117) 5.5 (1.5–129) 7.5 (4.8–72) 11 (3.8–84) 0.7Family history of SCD 1 (2%) 1 (4.2%) 0 (0%) 0 (0%) 1.0Diabetes 14 (29%) 5 (20.8%) 1 (16.7%) 8 (44.4%) 0.2Hypertension 26 (54%) 13 (54.2%) 3 (50%) 10 (55.6%) 1.0Dyslipidemia 25 (52%) 16 (66.7%) 2 (33.3%) 7 (38.9%) 0.1Smoking 38 (79%) 23 (95.8%) 5 (83.3%) 10 (55.6%) 0.003Family history of IHD 6 (13%) 3 (12.5%) 1 (16.7%) 2 (11.1%) 1.0Prior PCI 24 (50%) 12 (50%) 3 (50%) 9 (50%) 1.0Prior coronary bypass surgery 12 (25%) 7 (29.2%) 1 (16.7%) 4 (22.2%) 0.9BMI 29 ± 4 29.3 ± 3.7 27.8 ± 3.7 28.5 ± 4.6 0.7NYHA 1.0

Class II 43 (90%) 21 (87.5%) 6 (100%) 16 (89%)Class III 2 (4%) 1 (4.2%) 0 1 (5.6%)Class IV 3 (6%) 2 (8.3%) 0 1 (5.6%)

Baseline ECGRate (beats/min) 74 ± 15 73 ± 16 74 ± 19 76 ± 12 0.8Rhythm 0.37

Sinus 46 (96%) 24 (100%) 6 (100%) 16 (88.9%)Non-sinus 2 (4%) 0 0 2 (11.1%)

QRS duration (ms) 109 ± 24 105 ± 22 107 ± 19 115 ± 27 0.4Creatinine (mg/dL) 1.2 ± 0.5 1.2 ± 0.4 1.2 ± 0.2 1.3 ± 0.7 0.7Medications

β-Blockers 40 (83%) 22 (91.7%) 5 (83.3%) 13 (72.2%) 0.2Amiodarone 19 (39.6%) 19 (79.2%) 0 0 <0.001Digoxin 2 (4%) 0 1 (16.7%) 1 (5.6%) 0.1

Continuous data as mean ± SD and median (IQR) for normally distributed and nonnormally distributed data, respectively. Categoricaldata are expressed as n (%).*Fisher’s exact test for categorical data, analysis of variance test, or Kruskall-Wallis test (duration since last MI) for continuous data.BMI = body mass index; ECD = electrocardiogram; IHD = ischemic heart disease; IQR = interquartile range; MI = myocardial infarction;NYHA = New York Heart Association; PCI = percutaneous coronary intervention; SCD = sudden cardiac death; SD = standard deviation;SMVT = sustained monomorphic ventricular tachycardia.

vs 4, IQR 3–6, P = 0.001) and mean infarcttransmurality (78, IQR 67–83% vs 48, IQR36–68%, P < 0.001) were higher in the sponta-neous SMVT group than the noninducible group.

Similar findings were observed when compar-ing inducible SMVT group to noninducible group.Total infarct mass, absolute (61, IQR 42–72 g vs19, IQR 12–28 g, P = 0.001) and as percent of LV(32, IQR 25–41% vs 9, IQR 5–14%, P = 0.002)was greater in the inducible SMVT group thanthe noninducible one. Scar core mass, absolute(54, IQR 36–64 g vs 16, IQR 9–24 g, P = 0.001) and

as percent of LV (27, IQR 21–34% vs 8, IQR 4–13%,P = 0.002) was higher in the inducible SMVTgroup than the noninducible one. Peri-infarct zonemass, absolute (6, IQR 4–14 g vs 2.3, IQR 1.2–5 g,P = 0.042) and as percent of LV (4, IQR 2–6%vs 1, IQR 0.7–2%, P = 0.028) was larger in theinducible SMVT group than the noninducible one.Moreover, the number of segments with LGE (8,IQR 6–10 vs 4, IQR 3–6, P = 0.02) and meaninfarct transmurality (80, IQR 73–84% vs 48, IQR36–68%, P = 0.003) were higher in the inducibleSMVT group than the noninducible group.

4 2014 PACE, Vol. 00

PREDICTION OF VENTRICULAR ARRHYTHMIA IN ISCHEMIC CARDIOMYOPATHY

Table II.

MRI Variables among the Three Study Groups

Total Spontaneous Inducible Noninducible P Value*Population SMVT SMVT SMVT (Difference

Variable (n = 48) (n = 24) (n = 6) (n = 18) among Groups)

LV EDV (mL) 220 ± 64 231 ± 62 235 ± 63 199 ± 65 0.2LV ESV (mL) 166 ± 57 169 ± 51 194 ± 61 153 ± 63 0.3LVEF (%) 23 (17–32) 24 (18–34) 18 (15–23) 23 (17–34) 0.25LV mass (g) 200 ± 45 197 ± 45 199 ± 46 204 ± 48 0.9Total scar (g) 45 (22–64) 54 (41–70) 61 (42–72) 19 (12–28) <0.001Scar core (g) 36 (20–55) 45 (35–63) 54 (36–64) 16 (9–24) <0.001Peri-infarct zone (g) 6 (2.5–9) 8 (5–12) 6 (4–14) 2.3 (1.2–5) <0.001Total scar (%LV) 23 (12.5–34) 26 (23–40) 32 (25–41) 9 (5–14) <0.001Scar core (%LV) 20 (10–31) 22 (19–35) 27 (21–34) 8 (4–13) <0.001Peri-infarct zone (%LV) 3 (1–5) 4.5 (2.6–6) 4 (2–6) 1 (0.7–2) <0.001Mean infarct transmurality (%) 72.5 (51–81) 78 (67–83) 80 (73–84) 48 (36–68) <0.001No. of segments with LGE 7 (4–9) 8 (6–9) 8 (6–10) 4 (3–6) 0.002

Continuous data as mean ± SD and median (IQR) for normally distributed and nonnormally distributed data, respectively.*Analysis of variance test or Kruskall-Wallis test (LVEF; total scar, scar core, and peri-infarct zone as absolute mass or as percent of LV;mean infarct transmurality; no. of segments with LGE).EDV = end-diastolic volume; LGE = late gadolinium enhancement; LV = left ventricular; LVEF = left ventricular ejection fraction.Other abbreviations as in Table I.

There was no statistically significant differ-ence of any of the scar characteristics betweenpatients with spontaneous SMVT and those withinducible SMVT during the EPS.

Predictors of Inducible or Spontaneous VT

Tables III and IV show univariable andmultivariable analysis for variables postulated tobe associated with spontaneous SMVT experienceand SMVT inducibility during EPS.

In univariable analysis, total scar (absoluteand as percent of LV), scar core (absolute and aspercent of LV), peri-infarct zone (absolute and aspercent of LV), mean infarct transmurality, andnumber of segments with LGE were statisticallysignificant predictors of spontaneous SMVT expe-rience and SMVT inducibility by PES.

We designed a multinomial logistic regressionmodel with noninducible SMVT group as thereference group and in which LVEF, total scarmass (as percent of LV), and peri-infarct zonemass (as percent of LV) were included. Inthis model, total infarct as percent of LV wasthe only significant independent predictor ofspontaneous SMVT experience (odds ratio [OR]1.33 per% change, 95% confidence interval [CI]1.12–1.6, P = 0.001) and SMVT inducibility (OR1.3 per% change, 95% CI 1.1–1.6, P = 0.004).Peri-infarct zone as percent of LV, which wasa significant predictor in univariable analysis,became a statistically nonsignificant predictor for

spontaneous SMVT (OR 0.9 per% change, 95% CI0.6–1.4, P = 0.7) and SMVT inducibility (OR 0.8per% change, 95% CI 0.48–1.47, P = 0.5).

Using ROC curve analysis, a cut-off value oftotal scar percent of LV at 19% had sensitivityof 96% and specificity of 89% for predictionof spontaneous VT experience (AUC: 0.95, P <0.001). Similarly, total scar percent of LV at 14%had sensitivity of 100% and specificity of 78%for prediction of SMVT inducibility during EPS(AUC: 0.93, P = 0.002).

DiscussionThis study demonstrates that extent of my-

ocardial scarring is clearly associated with spon-taneous SMVT experience and SMVT inducibilityby PES. In multivariable analysis, total scar size (aspercent of LV) is the only independent predictorassociated with spontaneous SMVT experienceand SMVT inducibility by PES.

Several clinical outcome-based studies havedemonstrated the prognostic value of infarct sizeand/or infarct tissue heterogeneity by LGE-CMR inpatients with ischemic cardiomyopathy.12–14,20,21

There are potential mechanisms linking in-creased mortality risk to myocardial scarring.Prior MI may serve as a substrate for reentranttachyarrhythmias that can lead to VF and SCD.22,23

Bello et al.11 showed in a group of 48 CAD patientsthat infarct mass and infarct surface area arebetter predictors of the inducibility of VT than

PACE, Vol. 00 2014 5

GOUDA, ET AL.

Table III.

Univariable and Multivariable Analyses for Spontaneous VT Group with Noninducible VT Group as Reference

Univariable Analysis Multivariable Analysis

Variable P OR (95% CI) P OR (95% CI)

LV EDV (mL) 0.1 1.01 (0.99 –1.02)LVEF (%) 0.6 1.02 (0.96–1.07) 0.38 1.04 (0.95–1.13)Total scar (g) 0.001 1.16 (1.07–1.26)Scar core (g) 0.001 1.2 (1.08–1.33)Peri-infarct zone (g) 0.01 1.3 (1.07–1.6)Total scar (%LV) 0.001 1.3 (1.12–1.5) 0.001 1.33 (1.12–1.6)Scar core (%LV) 0.001 1.35 (1.14–1.6)Peri-infarct zone (%LV) 0.005 1.8 (1.2–2.6) 0.7 0.9 (0.6–1.4)Mean infarct transmurality (%) <0.001 1.12 (1.05–1.2)No. of segments with LGE 0.002 1.6 (1.2–2.2)

CI = confidence interval; OR = odds ratio; VT = ventricular tachycardia. Other abbreviations as in Tables I and II.

Table IV.

Univariable and Multivariable Analyses for Inducible VT Group with Noninducible VT Group as Reference

Univariable Analysis Multivariable Analysis

Variable P OR (95% CI) P OR (95% CI)

LV EDV (mL) 0.2 1.01 (0.99–1.03)LVEF (%) 0.12 0.9 (0.8–1.03) 0.24 0.9 (0.7–1.1)Total scar (g) 0.001 1.16 (1.06–1.27)Scar core (g) 0.001 1.2 (1.07–1.34)Peri-infarct zone (g) 0.02 1.3 (1.04–1.7)Total scar (%LV) 0.002 1.3 (1.11–1.56) 0.004 1.3 (1.1–1.6)Scar core (%LV) 0.001 1.36 (1.13–1.64)Peri-infarct zone (%LV) 0.02 1.7 (1.1–2.7) 0.5 0.8 (0.48–1.47)Mean infarct transmurality (%) 0.018 1.17 (1.03–1.3)No. of segments with LGE 0.016 1.67 (1.1–2.5)

Abbreviations as in Tables I-III.

LVEF. They identified a value of 10% of LV massto be the critical substrate necessary to developsustained ventricular arrhythmias. Additionally,Schmidt et al.24 demonstrated that peri-infarctzone is associated with VT inducibility in agroup of patients with ischemic cardiomyopathy(LVEF � 35%) referred for ICD placement forprimary prevention of SCD. These studies providea potential pathophysiological explanation forthe increased mortality, that is, larger scars andincreased peri-infarct zone may provide the poten-tial substrate for reentrant ventricular arrhythmiasthat could lead to sudden death. The results of ourstudy corroborate those of the earlier two reportsinvestigating the association between LGE and VTinducibility. In fact, our study extends beyond that

and tries to find differences in scar characteristicsbetween patients with spontaneous SMVT andthose with inducible SMVT during PES.

In addition to serving as a substrate forarrhythmias, larger infarcts are associated withLV remodeling, which may lead to depressed LVfunction and heart failure.25 Thus, in additionto being at high risk for SCD, patients withlarge infarcts may experience death secondary toprogressive heart failure.

Infarct Tissue Heterogeneity

In our study, the peri-infarct zone is not asso-ciated with spontaneous SMVT and inducibilityof SMVT in multivariable analysis. While some

6 2014 PACE, Vol. 00

PREDICTION OF VENTRICULAR ARRHYTHMIA IN ISCHEMIC CARDIOMYOPATHY

studies14,24 have demonstrated the importance ofthe peri-infarct zone as an independent predictorof inducibility of SMVT and adverse cardiacevents, other studies like ours have not.20,21

In addition to different study populationsand outcomes among these studies, several otherfactors could explain this discrepancy. Thereis no consensus on the optimal thresholds fordefining the peri-infarct zone that best correlatewith various clinical outcomes. In a recent study,26

Rubenstein et al. performed LGE in 47 patientswith CAD referred for EPS to assess for VT.26

They used three previously published methods toquantify the peri-infarct zone in the LGE images.Only using one method, peri-infarct zone was abetter predictor of inducible SMVT during EPSthan LVEF and infarct size, while by using theother two methods, the peri-infarct zone was notsignificantly different between the inducible andnoninducible groups.

Partial volume effects, where normal my-ocardium, blood pool, or epicardial fat mayexist within the same voxel volume along withthe infarct, can produce intermediate-intensityvalues. This may produce the appearance of grayzone on LGE imaging, even if the actual tissueis not an admixture of viable and nonviable my-ocardium and presumably not arrhythmogenic.27

In an animal study by Schelbert et al.28 they usedex vivo high-resolution LGE imaging to charac-terize the peri-infarct zone. The percentage ofapparent intermediate SI myocardium increasedsignificantly when image resolution was degradedto resemble clinical resolution consistent withsignificant partial volume averaging.

Additionally, the distribution of the grayzone in the infarct and not only its mass mayhave played a role. Gray zone is not necessarilyconfined to the border of the infarct, and it canbe seen at locations that are more central.24 Perez-David et al.29 showed that heterogeneous channelswere more frequent in the SMVT group comparedto control group and these channels were relatedto VT critical isthmuses.

Spontaneous Versus Inducible VT Groups

The differences in scar characteristics be-tween patients with spontaneous SMVT and thosewith inducible SMVT by PES have not beenexplored in previous studies. Our study does

not show any differences in the various scarcharacteristics assessed by LGE-CMR betweenpatients who presented with spontaneous SMVTand those who had SMVT induced by PES. It iswell known that the presence of infarcted tissueor scar forms the substrate for malignant reentrantarrhythmias.30 Therefore, these two patient groupsshould have larger substrate for VT than thenoninducible group. Other factors could havetriggered VT in the spontaneous SMVT group.These triggers may include subclinical myocardialischemia and/or abnormal cardiac sympatheticinnervation.

Study Limitations

This is an observational study and has all thelimitations inherent to this kind of study.31 Asnone of the patients had an ICD prior to enroll-ment, we may have introduced survival bias byoverlooking the patients who might have died oftheir ventricular arrhythmia. Spontaneous SMVTand inducible SMVT groups were comparable,but because of the small sample size no definiteconclusion could be made. Our study populationhas a relatively low ejection fraction 23% (IQR17–32%) and therefore these results could not beextrapolated to patients with preserved LV systolicfunction. Negative EP study is not reassuring anddoes not indicate low likelihood for arrhythmicdeath, especially in patients with low LVEF.32

Therefore, we cannot assume that patients withnoninducible SMVT in this study are at lower riskof SCD. Additionally, spontaneous VT occurrenceoverestimates the risk of SCD as not every VTepisode equates with mortal event.33

ConclusionsCharacterization of myocardial infarct by

LGE-CMR, specifically total infarct size, is a betterpredictor of spontaneous SMVT experience andSMVT inducibility than LVEF. This highlightsthe potential importance of myocardial scarringassessment in risk stratification of patients withischemic cardiomyopathy for selection of patientswho will benefit from ICDs. The total scar percentof LV mass at 14% could be included andvalidated prospectively in a larger study forselection of ICD candidates for primary preventionof SCD.

References1. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ,

Dai S, et al. Heart disease and stroke statistics–2014 update: A reportfrom the American Heart Association. Circulation 2014; 129:e28–e292.

2. Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, KleinH, Levine JH, et al. Improved survival with an implanted

defibrillator in patients with coronary disease at high riskfor ventricular arrhythmia. Multicenter Automatic DefibrillatorImplantation Trial Investigators. N Engl J Med 1996; 335:1933–1940.

3. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS,Daubert JP, et al. Prophylactic implantation of a defibrillator in

PACE, Vol. 00 2014 7

GOUDA, ET AL.

patients with myocardial infarction and reduced ejection fraction.N Engl J Med 2002; 346:877–883.

4. Buxton AE, Lee KL, Fisher JD, Josephson ME, Prystowsky EN,Hafley G. A randomized study of the prevention of sudden deathin patients with coronary artery disease. Multicenter UnsustainedTachycardia Trial Investigators. N Engl J Med 1999; 341:1882–1890.

5. Kadish A, Dyer A, Daubert JP, Quigg R, Estes NA, Anderson KP,Calkins H, et al. Prophylactic defibrillator implantation in patientswith nonischemic dilated cardiomyopathy. N Engl J Med 2004;350:2151–2158.

6. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R,Domanski M, et al. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med 2005;352:225–237.

7. Stecker EC, Vickers C, Waltz J, Socoteanu C, John BT, Mariani R,McAnulty JH, et al. Population-based analysis of sudden cardiacdeath with and without left ventricular systolic dysfunction: Two-year findings from the Oregon Sudden Unexpected Death Study. JAm Coll Cardiol 2006; 47:1161–1166.

8. The Antiarrhythmics versus Implantable Defibrillators (AVID)Investigators. A comparison of antiarrhythmic-drug therapy withimplantable defibrillators in patients resuscitated from near-fatalventricular arrhythmias. N Engl J Med 1997; 337:1576–1583.

9. Groeneveld PW, Matta MA, Suh JJ, Heidenreich PA, Shea JA. Costsand quality-of-life effects of implantable cardioverter-defibrillators.Am J Cardiol 2006; 98:1409–1415.

10. Mahrholdt H, Wagner A, Holly TA, Elliott MD, Bonow RO, Kim RJ,Judd RM. Reproducibility of chronic infarct size measurement bycontrast-enhanced magnetic resonance imaging. Circulation 2002;106:2322–2327.

11. Bello D, Fieno DS, Kim RJ, Pereles FS, Passman R, Song G, KadishAH, et al. Infarct morphology identifies patients with substrate forsustained ventricular tachycardia. J Am Coll Cardiol 2005; 45:1104–1108.

12. Kwong RY, Chan AK, Brown KA, Chan CW, Reynolds HG, TsangS, Davis RB. Impact of unrecognized myocardial scar detectedby cardiac magnetic resonance imaging on event-free survival inpatients presenting with signs or symptoms of coronary arterydisease. Circulation 2006; 113:2733–2743.

13. Wu E, Ortiz JT, Tejedor P, Lee DC, Bucciarelli-Ducci C, Kansal P,Carr JC, et al. Infarct size by contrast enhanced cardiac magneticresonance is a stronger predictor of outcomes than left ventricularejection fraction or end-systolic volume index: Prospective cohortstudy. Heart 2008; 94:730–736.

14. Yan AT, Shayne AJ, Brown KA, Gupta SN, Chan CW, Luu TM, DiCarli MF, et al. Characterization of the peri-infarct zone by contrast-enhanced cardiac magnetic resonance imaging is a powerfulpredictor of post-myocardial infarction mortality. Circulation 2006;114:32–39.

15. Thygesen K, Alpert JS, White HD, Jaffe AS, Apple FS, GalvaniM, Katus HA, et al. Universal definition of myocardial infarction.Circulation 2007; 116:2634–2653.

16. Simonetti OP, Kim RJ, Fieno DS, Hillenbrand HB, Wu E, BundyJM, Finn JP, et al. An improved MR imaging technique for thevisualization of myocardial infarction. Radiology 2001; 218:215–223.

17. Heiberg E, Sjogren J, Ugander M, Carlsson M, Engblom H, ArhedenH. Design and validation of Segment–freely available software forcardiovascular image analysis. BMC Med Imaging 2010; 10:1–13.

18. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, LaskeyWK, Pennell DJ, et al. Standardized myocardial segmentation andnomenclature for tomographic imaging of the heart. A statement forhealthcare professionals from the Cardiac Imaging Committee of theCouncil on Clinical Cardiology of the American Heart Association.Int J Cardiovasc Imaging 2002; 18:539–542.

19. Kim HW, Farzaneh-Far A, Kim RJ. Cardiovascular magneticresonance in patients with myocardial infarction: Current andemerging applications. J Am Coll Cardiol 2009; 55:1–16.

20. Gao P, Yee R, Gula L, Krahn AD, Skanes A, Leong-Sit P,Klein GJ, et al. Prediction of arrhythmic events in ischemic anddilated cardiomyopathy patients referred for implantable cardiacdefibrillator: Evaluation of multiple scar quantification measuresfor late gadolinium enhancement magnetic resonance imaging. CircCardiovasc Imaging 2012; 5:448–456.

21. Klem I, Weinsaft JW, Bahnson TD, Hegland D, Kim HW, HayesB, Parker MA, et al. Assessment of myocardial scarring improvesrisk stratification in patients evaluated for cardiac defibrillatorimplantation. J Am Coll Cardiol 2012; 60:408–420.

22. Geltman EM, Ehsani AA, Campbell MK, Schechtman K, RobertsR, Sobel BE. The influence of location and extent of myocardialinfarction on long-term ventricular dysrhythmia and mortality.Circulation 1979; 60:805–814.

23. Bayes de Luna A, Coumel P, Leclercq JF. Ambulatory suddencardiac death: Mechanisms of production of fatal arrhythmia onthe basis of data from 157 cases. Am Heart J 1989; 117:151–159.

24. Schmidt A, Azevedo CF, Cheng A, Gupta SN, Bluemke DA, FooTK, Gerstenblith G, et al. Infarct tissue heterogeneity by magneticresonance imaging identifies enhanced cardiac arrhythmia suscep-tibility in patients with left ventricular dysfunction. Circulation2007; 115:2006–2014.

25. Bello D, Shah DJ, Farah GM, Di Luzio S, Parker M, Johnson MR, CottsWG, et al. Gadolinium cardiovascular magnetic resonance predictsreversible myocardial dysfunction and remodeling in patients withheart failure undergoing beta-blocker therapy. Circulation 2003;108:1945–1953.

26. Rubenstein JC, Lee DC, Wu E, Kadish AH, Passman R, Bello D,Goldberger JJ. A comparison of cardiac magnetic resonance imagingperi-infarct border zone quantification strategies for the predictionof ventricular tachyarrhythmia inducibility. Cardiol J 2013; 20:68–77.

27. Klocke FJ, Wu E, Lee DC. “Shades of gray” in cardiac magneticresonance images of infarcted myocardium: Can they tell us whatwe’d like them to? Circulation 2006; 114:8–10.

28. Schelbert EB, Hsu LY, Anderson SA, Mohanty BD, Karim SM,Kellman P, Aletras AH, et al. Late gadolinium-enhancement cardiacmagnetic resonance identifies postinfarction myocardial fibrosisand the border zone at the near cellular level in ex vivo rat heart.Circ Cardiovasc Imaging 2010; 3:743–752.

29. Perez-David E, Arenal A, Rubio-Guivernau JL, del Castillo R,Atea L, Arbelo E, Caballero E, et al. Noninvasive identification ofventricular tachycardia-related conducting channels using contrast-enhanced magnetic resonance imaging in patients with chronicmyocardial infarction: Comparison of signal intensity scar mappingand endocardial voltage mapping. J Am Coll Cardiol 2011; 57:184–194.

30. Zipes DP, Wellens HJ. Sudden cardiac death. Circulation 1998;98:2334–2351.

31. Jepsen P, Johnsen SP, Gillman MW, Sorensen HT. Interpretation ofobservational studies. Heart 2004; 90:956–960.

32. Daubert JP, Zareba W, Hall WJ, Schuger C, Corsello A, Leon AR,Andrews ML, et al. Predictive value of ventricular arrhythmiainducibility for subsequent ventricular tachycardia or ventricularfibrillation in Multicenter Automatic Defibrillator ImplantationTrial (MADIT) II patients. J Am Coll Cardiol 2006; 47:98–107.

33. Moss AJ, Greenberg H, Case RB, Zareba W, Hall WJ, BrownMW, Daubert JP, et al. Long-term clinical course of patientsafter termination of ventricular tachyarrhythmia by an implanteddefibrillator. Circulation 2004; 110:3760–3765.

8 2014 PACE, Vol. 00