relationship between transmural extent of necrosis and quantitative recovery of regional strains...

J A C C : C A R D I O V A S C U L A R I M A G I N G V O L . 3 , N O . 7 , 2 0 1 0

P U B L I S H E D B Y E L S E V I E R I N C . D O I : 1 0 . 1 0 1 6 / j . j c m g . 2 0 1 0 . 0 3 . 0 0 8

Relationship Between Transmural Extent ofNecrosis and Quantitative Recovery of RegionalStrains After RevascularizationBernhard L. Gerber, MD, Julie Darchis, MD, Jean-Benoît le Polain de Waroux, MD,Gabin Legros, MD, Anne-Catherine Pouleur, MD, David Vancraeynest, MD,Agnès Pasquet, MD, Jean-Louis Vanoverschelde, MD

Brussels, Belgium

O B J E C T I V E S To better understand the quantitative relationship of recovery of regional and global

dysfunction after revascularization in chronic infarcts with variable transmural extent of necrosis by

delayed enhanced cardiac magnetic resonance.

B A C K G R O U N D Studies relating transmurality of delayed enhanced magnetic resonance to

functional recovery in dysfunctional myocardium using semiquantitative Likert scales have demon-

strated the intermediate likelihood (50% probability) of recovery of dysfunction in subendocardial scars.

M E T H O D S Forty-two patients with chronic left ventricular dysfunction due to coronary artery

disease underwent tagged and delayed enhanced magnetic resonance before and 10 � 7 months after

revascularization (coronary artery bypass graft: 35, percutaneous transluminal coronary angioplasty: 7).

Left ventricular ejection fraction and regional mid-myocardial Eulerian radial thickening strain (Err) and

mid-myocardial, subendocardial, and subepicardial Eulerian circumferential shortening strain (Ecc)

strains were quantified in 16 segments per patient before and after revascularization and related to

pre-operatively measured transmurality of necrosis.

R E S U L T S At baseline, 256 of 672 segments were dysfunctional, having �2 SD (i.e., ��10%)

mid-myocardial Ecc. The magnitude of recovery of mid-myocardial Ecc (r � �0.33, p � 0.01) was

inversely correlated with transmurality of necrosis before revascularization. Segments with �25%

necrosis improved mid-myocardial Ecc and Err. No significant improvement of mid-myocardial Ecc or Err

occurred when transmurality was �25%. However, subendocardial Ecc improved up to 75% transmural

necrosis. Receiver-operator characteristic analysis determined optimal sensitivity (54%) and specificity

(82%) for normalization of mid-myocardial Ecc (to ��10% Ecc) at a cutoff value of �18% transmural

necrosis. Improvement of left ventricular ejection fraction (from 35 � 15% to 40 � 7%, p � 0.001) was

best predicted (67% sensitivity, 58% specificity) by the presence of �4.5 dysfunctional segments with

�75% transmural necrosis.

C O N C L U S I O N S The quantitative relationship between necrosis transmurality and improvement

of regional and global dysfunction after revascularization is complex. Although improvement of recovery

of regional mid-myocardial dysfunction after revascularization was observed only for scarring not

exceeding 25% transmurality, global dysfunction significantly improved even when more extensive

College of Cardiology Foundation

From the Division of Cardiology, Cliniques Universitaires St. Luc, Brussels, Belgium. Grant support by the FondationNationale de la Recherche Scientifique of the Belgian Government (FRSM 3.4557.02). Contrast media for cardiac magneticresonance was donated by Amersham-Nycomed (now GE Healthcare).

Manuscript received January 26, 2010; revised manuscript received March 29, 2010; accepted March 31, 2010.

1uwwcrtbhi

taapnuwatsst

Pecvcpch3gaii

accpSpf4spfftBpaA(owCpgulawLgaoqmHdt(Mt1dgrDsrlRvmci

� multidetector CT

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 7 , 2 0 1 0

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

Quantitative Recovery of Strain

everal studies have evaluated the value ofdelayed enhancement (DE) cardiac mag-netic resonance (CMR) for assessment ofmyocardial viability in chronic ischemic dys-

unction (1,2). They described an inverse relation-hip between the transmural extent of necrosis andikelihood of improvement of regional contractilityfter revascularization. Segments with high (75% to

See page 731

00%) transmurality of necrosis were shown to benable to recover function after revascularization,hereas segments with no or very little necrosisere found to likely improve function after revas-

ularization. In segments with subendocardial scar-ing of 25% to 75% transmurality of DE, however,he likelihood of recovery of function was found toe intermediate, around 50%, and thus DE-CMRas a low accuracy in predicting functional recovery

n such segments (3).However, a major limitation of earlier studies was

hat both the transmural extent of DE and themount of functional recovery were assessed visu-lly, using semiquantitative Likert scales. We hy-othesized that the binary nature of assessment ofecrosis and functional recovery might explain thencertainty of predicting viability in such segmentsith subendocardial necrosis, and therefore we

ttempted to examine the relationship of the quan-itatively transmural extent of DE using automatedoftware to the magnitude of quantitatively mea-ured regional myocardial strains obtained fromagged CMR images.

E T H O D S

atients and study protocol. This prospective studynrolled consecutive patients with chronic regionalontractile dysfunction in areas supplied by a se-erely stenotic or occluded coronary artery withlinical indication for revascularization by eitherercutaneous transluminal coronary angioplasty ororonary artery bypass graft (CABG). Patients whoad an acute myocardial infarction within the past0 days, hemodynamic instability, higher thanrade II mitral or aortic insufficiency, significantortic stenosis, chronic atrial fibrillation, or contra-ndications to CMR were not considered fornclusion.

Patients underwent tagged and DE-CMR beforend �4 months after revascularization. Verification

ged using either recatheterization or multidetectoromputed tomography (MDCT), if no contraindi-ation (i.e., serum creatinine level �1.4 mg/dl) wasresent. Study enrollment is shown in Figure 1.ixty-one patients underwent baseline CMR. Fiveatients died during follow-up, and 14 refusedollow-up. The final study population consisted of2 consecutive patients (Table 1). There were noignificant differences in the patients who com-leted the protocol and those who were lost toollow-up. All participants gave informed consentor the study protocol, which had been approved byhe institutional review board of our university.aseline coronary angiography. Coronary angiogra-hy was evaluated visually and semiquantitatively byblinded reviewer using the standard 16-segmentmerican Heart Association classification system

4). Luminal narrowing exceeding 50%f the mean reference luminal diameteras considered significant.MR protocol. CMR was performed asreviously described (5). To compute re-ional strains, we acquired 6 to 8 contig-ous short-axis images covering the entireeft ventricle (LV) from base to apex with

prospectively triggered cine sequenceith grid tagging. Subsequently, to assessV function, we acquired retrospectivelyated steady-state free-precession cine im-ges. To evaluate myocardial viability, 2-r 3-dimensional DE images were ac-uired 15 min after administration of 0.20mol/kg�1 gadodiamide (Omniscan; GEealthcare, Oslo, Norway) during mid-

iastole. All images had the same orienta-ion, slice thickness (8 mm), and spacing2 mm) as the tagged short-axis images.DCT for verification of completeness of revasculariza-ion. To verify completeness of revascularization,6- or 40-slice MDCT was performed as previouslyescribed (6). Patency of arterial and venous bypassrafts and stents was assessed visually by a blindedeviewer (B.G.).ata analysis. REGISTRATION OF DATASETS. Theets of tagged and DE images acquired beforeevascularization were acquired using the same sliceocations and thus did not need to be registered.egistration of CMR images before and after re-

ascularization was performed using standard land-arks and facilitated by standardized acquisition

overing the entire left ventricle. A standard Amer-can Heart Association 17-segment model omitting

AUC �

CI � c

CMR �

DE �

Ecc �

shorte

EF � e

Err �

strain

LV �

he apex (i.e., 16 segments per patient) was u

R E V I A T I O N S

A C R O N YM S

area under the receiver-

tor characteristic curve

� coronary artery bypass

onfidence interval

cardiac magnetic

delayed enhancement

Eulerian circumferential

ning strain

jection fraction

Eulerian radial thickening

left ventricle/ventricular

sed to

ircbTliaisncanwttaicts

unmlm(

fefirarpsptaracmw12dhwCfdEmm2c0iTff

mpmedouGsooimooSp

ous transluminal coron

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

nterpret segmental strain and transmural DE on aegional basis. To obtain these 16 segments, 2 to 3onsecutive short-axis slices were assigned to theasal, mid-ventricular, and apical levels of the heart.he same number of slices was assigned to each

evel before and after revascularization. Using thenferoseptal insertion of the right ventricular wall as

landmark, each short-axis slice was subdividednto 4 (apex) or 6 (mid- and basal levels) radiallypaced segments. Segments were assigned to coro-ary artery territories (7) according to standardriteria, taking into account the individual coronarynatomy of the patient (8). When the right coro-ary artery supplied at least 1 posterolateral branch,e assumed right coronary dominance and assigned

he basal and inferoseptal and inferior segments tohe right coronary artery. When the right coronaryrtery did not reach the crux cordis and when thenferior septal branches were supplied by the leftoronary artery, we assumed left dominant circula-ion and assigned basal and inferoseptal and inferioregments to the left circumflex coronary artery.

EGIONAL STRAIN. Tagged images were analyzedsing HARP (Diagnosoft, San Francisco, Califor-ia). Within each of the 16 segments, mid-yocardial, subendocardial, and subepicardial Eu-

erian circumferential shortening strain (Ecc) andid-myocardial Eulerian radial thickening strain

RevascularizationCABGPTCA

Baseline CMRn=61

FU CMR (10±8 months)n=42

Died:n=5

tive n=3, sudden n=2)

Refused FU: n=14

of the Study Cohort

erwent baseline cardiac magnetic resonance (CMR) before revas-nts died during follow-up (FU) and 14 refused FU. The final studyf 42 consecutive patients who underwent a 10 � 8 month FUation. CABG � coronary artery bypass graft; PTCA � percutane-ary angioplasty.

Err) were computed. Strains were expressed as the m

ractional change in myocardial length betweennd-diastole and end-systole. End-systole was de-ned as nadir of Ecc in normally contractingegions. The minimal (Ecc)/maximal (Err) strain in

range of 2 to 3 frames from end-systole waseported, excluding diastole to avoid measuringost-systolic strain. By convention, shorteningtrains had a negative sign and elongation strainsositive. Normal values and interobserver and in-raobserver reproducibility of mid-myocardial Eccnd Err were assessed in a previous study in 10andomly selected patients (5). Both intraobservernd interobserver reproducibility had an intraclassorrelation coefficient of 0.81. Mean values ofid-myocardial strain in normal healthy volunteersere Ecc �8 � 4 (95% confidence interval [CI]:0% to 26%) and Err 16 � 5 (95% CI: 7% to6.0%). Therefore, we defined regional contractileysfunction as Ecc ��10% (i.e., lower 95% CIs inealthy volunteers) (5). Recovery of dysfunctionas defined as return of Ecc or Err to lower 95%Is of healthy volunteers (i.e., ��10% and �7%

or Ecc and Err, respectively). Improvement ofysfunction was defined as improvement of Ecc orrr �95% CI of test-rest values of these measure-ents assessed in 5 healthy volunteers after 4onths. For Ecc, test-to-test variation was 0.5 �

.5 (95% CI: �4.3 to 5.5), intraclass correlationoefficient � 0.61. For Err, test-retest variation was.45 � 4.1 (95% CI: �7.6 to 8.5), and reproduc-bility intraclass correlation coefficient � 0.73.herefore, we defined significant improvement of

unction as improvement ��5% for Ecc or �8%or Err.

RANSMURAL EXTENT OF INFARCTION. Assess-ent of the transmural extent of DE-CMR was

erformed using the freely available software Seg-ent (9). After manually tracing endocardial and

picardial contours, the program semiautomaticallyetected and reported the mean transmural extentf DE in each of the 16 same segments that weresed for computation of strain.lobal cardiac function. LV end-diastolic and end-ystolic volumes and ejection fraction (EF) werebtained by manual tracing of endocardial contoursn anonymized steady-state free-precession CMRmages before and after revascularization. Measure-

ents were performed in duplicate by 2 blindedbservers, and the average measurement of bothbservers was reported.tatistical analysis. All continuous values are re-orted as mean � 1 SD. Only dysfunctional seg-

(periopera

Figure 1. Illustration

Sixty-one patients undcularization. Five patiepopulation consisted ostudy after revasculariz

ents from territories with significant coronary

amrisMsamdt(tfirscpPtRumOsso

Sp(t1pprlvflohhapF(Nsb

SBnatwsbdqoCdaapr(a

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

rtery stenosis were considered for analysis. Seg-ents belonging to nonrevascularized coronary ter-

itories and segments in which revascularization wasncomplete due to documented graft or stent occlu-ion or restenosis by either recatheterization or

DCT were excluded from analysis. To correct fortatistical independence of segments with differentmounts of necrosis within patients, a mixed linearodel was used to compare strains in remote and

ysfunctional segments stratified according to quar-iles of transmural infarction. The segment groupnormo- or dysfunctional stratified by quartile ofransmural extent of necrosis) was considered axed factor, whereas the patient was considered aandom factor. Individual comparisons betweenector groups were performed post hoc using Wil-oxon tests with a Bonferroni correction for multi-le comparisons. Strain was also compared usingearson regression analysis against transmural ex-ent of necrosis in dysfunctional segments.eceiver-operator characteristic (ROC) curves weresed to evaluate diagnostic accuracy of mean trans-urality of necrosis to regional and global function.ptimal cutoff points were considered the greatest

um of sensitivity and specificity. All tests were 2ided and a p value �0.05 was considered indicativef statistical significance.

E S U L T S

tudy population. The clinical characteristics of theatient population are shown in Table 1. Thirty-six76%) patients had a history of myocardial infarc-ion, occurring on average 54 � 76 months (range

month to 20 years) before the study. Fourteenatients had infarcts �1 year ago. The majority ofatients had multivessel disease and underwentevascularization with CABG. Two patients hadeft-main disease. At the time of the study, 37essels (11 left anterior descending, 9 left circum-ex, and 17 right coronary arteries) were completelyccluded, 47 had �75% diameter stenosis and 22ad 50% to 75% diameter stenosis.. Three patientsad undergone previous revascularization (1 CABGnd 2 percutaneous transluminal coronary angio-lasty). Twenty patients had an EF �35%.ollow-up CMR was performed 10 � 7 months

range 5 months to 4 years) after revascularization.o patient experienced a clinically significant event

uch as a new infarct, or revascularization between

trains and transmural extent of necrosis at baseline.ased on the 95% lower CI of Ecc obtained fromormal volunteers (�10% mid-myocardial Ecc) (5)t baseline, 395 of all 672 segments were consideredo have a normal function, whereas 277 segmentsere considered to be dysfunctional. We only con-

idered the 256 dysfunctional segments subtendedy significant (�50%) coronary stenosis. Theseysfunctional segments were further separated intouartiles according to their mean transmural extentf necrosis (Fig. 2).oronary revascularization. Thirty-five patients un-erwent revascularization with CABG. The leftnterior mammary artery was grafted to the leftnterior descending coronary artery in all but 1atient (n � 34) (who was a redo CABG). Theight mammary artery (n � 11), free radial arteryn � 2), and gastroepiploic arteries (n � 3) werelso used. Twenty-one patients received saphenous

Table 1. Baseline Characteristics of the Study Population

Completed Follow-Up(n � 42)

Lost to Fo(n �

Age (yrs) 62 � 9 67 �

Sex (M/F) 38/4 13/

Coronary risk factors

Smoking 18 (42%) 7 (37

Hypertension 26 (62%) 13 (68

Diabetes mellitus 9 (21%) 8 (42

Hypercholesterolemia 26 (62%) 9 (47

Family history of CAD 13 (31%) 6 (31

No. of diseased vessels

1 3 (7%) 2 (10

2 14 (33%) 5 (26

3 25 (60%) 12 (63

Baseline CMR

EDVI (ml/m2) 116 � 39 95 �

ESVI (ml/m2) 79 � 39 67 �

EF (%) 35 � 13 33 �

Symptoms

Chest pain 22 (53%) 5 (26

Heart failure symptoms 18 (42%) 7 (37

Previous MI 32 (76%) 12 (63

Previous revascularization 3 (7%) 2 (10

Q waves 24 (57%) 13 (68

LBBB 5 (11%) 0 (0%

Treatment

PTCA 7 (17%) 0 (0%

CABG 35 (83%) 19 (10

CABG � coronary artery bypass graft; CAD � coronary artery disease; CMR � carECG � electrocardiogram; EDVI � end-diastolic volume index; EF � ejection fravolume index; F � female; LBBB � left bundle branch block; M � male; MIPTCA � percutaneous transluminal coronary angioplasty.

llow-Up19) p Value

10 0.09 (NS)

6 0.06 (NS)

%) 0.65 (NS)

%) 0.62 (NS)

%) 0.09 (NS)

%) 0.28 (NS)

) 0.96 (NS)

%) 0.75 (NS)

35 0.65 (NS)

34 0.76 (NS)

16 0.56 (NS)

%) 0.06 (NS)

%) 0.65 (NS)

%) 0.29 (NS)

%) 0.64 (NS)

%) 0.40 (NS)

) 0.31 (NS)

0%) 0.09 (NS)

diac magnetic resonance;ction; ESVI � end-systolic� myocardial infarction;

ein grafts. CABG revascularization was complete

(wgdp4

tumitwpNugfsaa2cmeccfipItB

ebmbdmmnsusmtotnmpn�0obo

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

i.e., of all stenotic vessels) in all but 3 patients inhom the native vessels were too small to receiverafts. Seven patients underwent single (n � 6) orouble angioplasty. Revascularization using angio-lasty was complete in 3 patients and incomplete inpatients.Completeness of revascularization was ascer-

ained in 29 patients by MDCT and by 6 patientssing repeat cardiac catheterization within 2onths of CMR. In the remaining 7 patients, renal

nsufficiency contraindicated MDCT for verifica-ion of vessel patency. Revascularization patencyas verified in 6 of 7 patients undergoing angio-lasty and in 28 of 35 patients undergoing CABG.o restenosis was found in the 6 patients who

nderwent angioplasty. Furthermore, all bypassrafts were found to be patent at the time ofollow-up; however, in 1 patient with CABG,tenosis was found to be present angiographically innative vessel situated beyond the graft. Overall,

mong the 256 dysfunctional segments at baseline,4 segments were considered to not have beenorrectly revascularized and 2 revascularized seg-ents were occluded (Fig. 2). These segments were

xcluded from analysis. Forty dysfunctional revas-ularized segments that lacked verification of theirompleteness were considered for analysis. Atollow-up CMR, no patient presented with newnfarcts, as demonstrated by regions of late DE notresent at baseline study.mprovement of strains after revascularization. Twoypical patient studies are shown in Figure 3.

Figure 2. Distribution of Segments

The 256 dysfunctional segments having mid-myocardial circumferencoronary stenosis were stratified according to their mean transmuraments with incomplete or failed were excluded. Only fully revascula

ecause 26 segments were considered to not have p

een correctly revascularized, only 230 dysfunc-ional segments were compared between baselinend follow-up. The distribution of mean transmu-ality of necrosis in these segments is shown inigure 2.Mean mid-myocardial, subendocardial, and sub-

picardial Ecc and mid-myocardial Err strain ataseline and follow-up in different types of seg-ents are depicted in Figure 4 and Table 2. At

aseline, mid-myocardial Ecc was not significantlyifferent in dysfunctional segments with a differentean transmural extent of delayed hyperenhance-ent (p � 0.07 by analysis of variance). In the 395

ondysfunctional segments, Ecc or Err did notignificantly change between baseline and follow-p. In the 230 correctly revascularized dysfunctionalegments, mid-myocardial and epicardial Ecc andid-myocardial Err improved after revasculariza-

ion when necrosis at baseline was absent or when itccupied �25%, but not beyond �25% meanransmurality of necrosis. Subendocardial Ecc sig-ificantly improved in segments with up to 50%ean transmurality of necrosis. Significant, albeit

oor correlations between mean transmurality ofecrosis and recovery of mid-myocardial Ecc (r �0.33, p � 0.01) but not with recovery of Err (r �

.09, p � NS) in dysfunctional segments werebserved. Results were similar when only segmentselonging to patients with infarcts �1 year previ-usly were considered.ROC analysis (Fig. 5) demonstrated similar ac-

uracy of the transmural extent of necrosis to

shortening strain ��10% at baseline subtended by a significanttent of necrosis (percentage of delayed enhancement [DE]). Seg-d segments were considered for follow-up (FU).

tiall ex

redict improvement of mid-myocardial Ecc

�i�tip�8ed

ommIl1Tv0u0H0

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

�5% (area under the receiver-operator character-stic curve [AUC] � 0.67) normalization of Ecc to

�10% (AUC � 0.67), and normalization of Erro �7% (AUC � 0.64). Accuracy for predictingmprovement of Err �8% was lower (AUC � 0.56,

� 0.05). Using an optimized cutoff value of18%, mean transmurality had 54% sensitivity and

2% specificity for predicting recovery or nonrecov-ry of function to mid-myocardial ECC ��10% inysfunctional segments at baseline.The probability of improving Ecc by ��5% and

hat of normalizing to Ecc ��10% in dysfunctionalegments at baseline after revascularization is shown inigure 6. We observed that the probability of normal-

Figure 3. Examples of DE-CMR and End-Systolic Shortening StrainWith Different Transmurality of Necrosis

Strain is color coded: green represents absence of contraction, bluenecrosis and inferior septal dysfunction, with recovery after revascumural extent. Inferior and lateral dysfunction at baseline recovereddial scarring with �50% transmurality of necrosis. Dysfunction persdysfunction at follow-up. Ecc � Eulerian circumferential shortening

zing dysfunction steadily declined with each quartile i

f increasing mean transmurality from 62% for seg-ents with 0% transmurality to only 10% for seg-ents with �75% mean transmurality.

mprovement of global cardiac function after revascu-arization. LVEF improved significantly (from 35 �3% to 40 � 7%, p � 0.001) after revascularization.his was due to a reduction of LV end-systolic

olume (from 152 � 78 ml to 133 � 79 ml, p �.002), while LV end-diastolic volume remainednchanged (from 226 � 79 to 213 � 80 ml, p �.12). Patient symptoms improved from New Yorkeart Association functional class average 2.8 �

.7 to 1.6 � 0.9 (p � 0.001).EF before revascularization was significantly but

fore and After Revascularization in 4 Representative Patients

mal shortening, and red abnormal elongation. (A) Absence ofation. (B) Subendocardial scarring of lateral wall of �25% trans-pletely at follow-up. (C) Anterior, septal, and inferior subendocar-at follow-up. (D) Transmural inferior scarring with persisting

in; BL � baseline; other abbreviations as in Figures 1 and 2.

norlarizcomisted

nversely correlated with the number of dysfunc-

tScrpadiacNsnhptdsnp

iiEstdmapb�fi

OrsedtstTssti

dtapwcnpaoptafotdp(mne

0.01 after versus befor

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

ional segments at baseline (r � �0.86, p � 0.001).imilarly, EF after revascularization was signifi-antly correlated with the number of segmentsemaining dysfunctional at follow-up (r � �0.62,

� 0.001). The change in EF between baselinend follow-up was correlated with the number ofysfunctional segments improving after revascular-

zation (r � 0.36, p � 0.02). The change in EFfter revascularization was, however, not signifi-antly correlated with infarct size (r � �0.11, p �S) with the average transmurality in dysfunctional

egments (r � �0.11, p � NS) nor with theumber of dysfunctional revascularized segmentsaving �25% (r � 0.26, p � NS), �50% (r � 0.26,� NS), and 75% (r � 0.29, p � NS) mean

ransmurality of necrosis at baseline. ROC analysisemonstrated that presence of �4.5 dysfunctionalegments with �75% mean transmural extent ofecrosis had 67% sensitivity and 58% specificity to

re and After Revascularization According to Transmurality

d Eulerian radial thickening strain (Err) in segments in normofunc-larized dysfunctional segments of patients with variable amountsbefore (red bars) and after (green bars) revascularization. *p �

e revascularization. Abbreviation as in Figure 3.

redict improvement of EF �5% after revascular- W

zation (AUC � 0.64). Diagnostic accuracy tomprove EF �5% was similar in patients with lowerF (EF �35%, n � 22, AUC � 0.61, optimal

ensitivity and specificity 73% and 61%, respec-ively) and in those with larger ventricles (end-iastolic volume indexed to body surface �100l/m2, n � 20, AUC � 0.66 optimal sensitivity

nd specificity 71%/67%. Diagnostic accuracy toredict improvement of EF �5% was somewhatetter in patients with infarcts older than 1 year (n

29, AUC � 0.77, optimal sensitivity and speci-city, 82% and 62%, respectively).

I S C U S S I O N

ur study demonstrated a quantitative yet complexelationship between improvement of regionaltrains in dysfunctional segments and transmuralxtent of necrosis. At mid-myocardial and epicar-ial level, Ecc and Err improved only up to 25% ofhe mean transmurality of necrosis. In contrast, inubendocardial levels, Ecc improved even whenransmurality of necrosis was higher, up to 75%.ransmission of contractile forces and cross-fiber

hortening from viable myocardium in noninfarctedubepicardial layers likely contribute to this effect inhe subendocardium even when this layer is entirelynfarcted.

The cutoff value of 25% for recovery of regionalysfunction at the mid-myocardial level was lowerhan that observed in earlier work (1,2,10,11) andlso lower than what we previously observed inatients with acute myocardial infarction (12), inhom recovery of circumferential shortening oc-

urred in segments with up to 75% transmurality ofecrosis. This is likely explained by the fact thatatients in our study were sicker, having lower EFsnd higher LV volumes than those described byther groups in the literature. It is possible that inatients with more severely depressed cardiac func-ion, more complex mechanisms such as tetheringnd remodeling are involved in dysfunction. There-ore, in such patients, recovery of dysfunction maynly be achieved for lower transmurality of necrosishan in patients with less severe dysfunction and LVilation. We already observed similar findings inrevious work using positron emission tomography13). Another explanation could be that our auto-ated method computed the mean transmurality of

ecrosis, whereas others using visual assessmentstimated the maximal transmurality of necrosis.

Figure 4. Strains Befoof Necrosis

Mid-myocardial Ecc antional and fully revascuof myocardial necrosis

hen we computed maximal rather than mean

ngandT

ing s bepicardial level.

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

ransmurality of necrosis, mid-myocardial Ecc andrr increased up to 75% transmurality of necrosis.The present study also confirmed earlier findings;

hat is, the high specificity (82%) of DE-CMR toredict the absence of functional recovery in trans-ural scars in relation to the presence of irreversible

Table 2. Strains in Segments and Dysfunctional Segments With

Normofunctional(n � 395)

0% Necrosis(n � 89)

Ecc mid

Baseline �15.8 � 3.7 �5.2 � 3.0

Follow-up �15.3 � 4.7 �10.4 � 5.2

p Value 0.06 (NS) �0.001

Ecc subendo

Baseline �16.2 � 4.4 �6.5 � 3.4

Follow-up �15.7 � 5.1 �10.8 � 5.1 �

p Value 0.06 (NS) �0.001

Ecc subepi

Baseline �13.9 � 4.0 �5.1 � 3.0

Follow-up �13.3 � 5.0 �9.2 � 4.8

p Value 0.03 (NS)* �0.001

Err mid

Baseline �11.9 � 6.4 �7.4 � 6.6

Follow-up �12.6 � 6.4 �10.6 � 5.6 �

p Value 0.03 (NS)* �0.001

*Considered nonsignificant given multiple comparisons.Ecc � Eulerian circumferential shortening strain; Err � Eulerian radial thicken

0.0 0.2 0.4 0.6

1-Sensitivity (%

Mean Transmurality of

Figure 5. ROC Curves Evaluating Diagnostic Accuracy to Recove

Receiver-operator characteristic (ROC) curve evaluating diagnostic amid-myocardial Ecc (to ��10%), improvement of Ecc by ��5%, refully revascularized segments at baseline. AUC � area under the re

ecrosis. Sensitivity for predicting recovery of re-ional dysfunction in segments with no or a smallmount of necrosis was low (54%). In particular,ot all segments and patients with viable myocar-ium had improved function after revascularization.here may be different explanations for this lack of

iable Amount of Myocardial Necrosis at Baseline and After Revas

Dysfunctional

Necrosis71)

26%–50% Necrosis(n � 36)

51%–75% Necrosis(n � 24)

� 3.1 �5.4 � 3.1 �3.1 � 4.8

� 5.0 �7.3 � 4.4 �4.4 � 5.4

.001 0.03 (NS)* 0.20 (NS)

� 3.6 �6.0 � 3.8 �3.1 � 5.3

� 5.4 �8.1 � 4.8 �6.5 � 6.6

.001 0.004 0.01

� 3.3 �4.6 � 2.8 �3.3 � 5.1

� 5.3 �6.2 � 4.6 �4.7 � 7.0

.001 0.05 (NS)* 0.24 (NS)

� 6.2 �7.6 � 6.6 �4.0 � 5.8

� 6.2 �9.6 � 6.1 �5.6 � 5.9

.01 0.12 0.47 (NS)

train; Mid � mid-myocardial level; subendo � subendocardial level; subepi � su

1.00.8

crosis

Normalization of Ecc to <-10%, AUC=0.67

Improvement of Ecc by >5%, AUC=0.67

Normalization of Err to >8%, AUC=0.64

Improvement of Err by >7%, AUC=0.56

Normalize Strains

acy of mean segmental transmurality for predicting recovery ofery of Err to �7% or improvement of Err by �8% in dysfunctionaler-operator characteristic curve; other abbreviations as in Figures

Var cularization

25%(n �

76%–100% Necrosis(n � 10)

�5.2 �4.7 � 2.3

�9.5 �5.2 � 3.9

�0 0.64 (NS)

�6.3 �3.8 � 2.4

10.4 �5.0 � 4.5

�0 0.45 (NS)

�4.9 �3.8 � 2.6

�8.1 �4.1 � 2.3

�0 0.80 (NS)

�7.7 �2.9 � 4.7

10.1 �2.0 � 6.4

�0 0.68 (NS)

ccurcovceiv

faffdltwcltstHn(hrta

piDfecdcntcvaamdcafbo

ppfoafsCsgd�oeemwiscspefeiflarz

amounts of mean tran

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

unctional recovery of segments with no or a smallmount of necrosis after revascularization. Success-ul revascularization is an important prerequisite tounctional recovery of chronically ischemic myocar-ium. We ascertained the completeness of revascu-

arization by MDCT or repeat cardiac catheteriza-ion in 83% of patients and excluded segments thatere not completely revascularized or were reoc-

luded at follow-up. Therefore, this factor is un-ikely to account for the low proportion of func-ional recovery in noninfarcted segments in ourtudy. Periprocedural necrosis may also account forhe absence of functional recovery at follow-up.

owever, we did not observe such periproceduralecrosis in the present study. Several studies10,14,15) found that recovery of dysfunction inibernating myocardium is not instantaneous, butather a time-dependent phenomenon that mayake as long as 1 year to complete. Because the

of Normalizing or Recovering Dysfunction According

ing regional function to mid-myocardial Ecc ��10% (A) or of% (B) in dysfunctional fully revascularized segments with variablesmurality. Abbreviation as in Figure 3.

verage follow-up after revascularization in the p

resent study was 10 months, insufficient follow-ups also unlikely to explain the low sensitivity of

E-CMR to predict functional recovery in nonin-arcted segments. Therefore, we believe that otherxplanations might account for low functional re-overy of some noninfarcted or small subendocar-ial scarring in our work and that of others. Me-hanical factors such as passive tethering fromeighboring transmurally necrotic segments or ex-reme LV remodeling may prevent functional re-overy of noninfarcted segments, especially in se-erely dilated left ventricles with poor LV function,s in our study. Despite the presence of coronaryrtery disease, dysfunction in noninfarcted seg-ents could also have resulted from other myocar-

ial diseases (16), such as coexisting structuralardiomyopathy, which are not likely to improvefter revascularization, or from as yet unidentifiedactors, accounting for the variable relationshipetween transmurality of necrosis and improvementf regional strain.It is interesting to note that limited sensitivity to

redict recovery of dysfunction has also been re-orted for nuclear imaging techniques (17). There-ore, some authors (3) suggested that demonstrationf preserved inotropic reserve during stress dobut-mine CMR might have better value for predictingunctional recovery than demonstration of the ab-ence of myocardial necrosis.linical implications. The findings of the presenttudy illustrate the complex relationship between re-ional viability and recovery of regional and globalysfunction. Although in our study, segments with25% transmurality of necrosis had limited recovery

f contractile function at the mid-myocardial andpicardial levels, EF and functional status improvedven when necrosis was �25% transmural. Improve-ent of subendocardial shortening in those patientsith more severe necrosis might have contributed to

mprovement of global function. Also, isometric ten-ion development in subepicardial myocardium canontribute to tension development without any actualtrain. This suggests also that revascularization ofartially infarcted myocardium might have other ben-ficial effects beyond improvement of regional dys-unction. Some studies indicated that revascularizationven in the absence of residual inotropic reserve maymprove survival (18), presumably by restoring bloodow in subepicardial layers of nontransmural infarctsnd reducing the potential to induce ventricular ar-hythmias originating from the peri-infarct borderones (19,20). However, this has not been definitely

Figure 6. Probabilityto Transmurality

Probability of normalizimproving Ecc by ��5

roven. Thus, the presence of subendocardial scarring

oopiSEt2bmssiotttbcfstasdmGwstr

wfrtssot

Oqsrlsrogooftf

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

n DE-CMR, which might indicate a low probabilityf recovery of regional contractile dysfunction, shouldrobably not be a reason to withhold revascularizationn patients with severe global dysfunction.tudy limitations. In this study, we only evaluatedcc and Err but not longitudinal shortening and

wist. Similar to other studies performing CMR atdifferent time points, misalignment of segments

etween the pre- and post-revascularization studiesight have occurred. We attempted to minimize

uch misalignment error by careful registration ofegments to fixed anatomic landmarks. Also, DEmages (mid-diastole) and tagged images (through-ut systole) were not acquired in the same part ofhe cardiac cycle. This could theoretically causeemporal misalignment due to through-plane mo-ion. However, this likely had few repercussionsecause 1) our patients had a quite low EF andonsequently low through-plane motion and 2) datarom 2 to 3 adjacent slices were averaged to obtaintrain and DE in 16 segments per patient. Al-hough we took individual variations of coronarynatomy into account, the exact assignment ofegments to coronary artery territories may beifficult, and some misregistration between seg-ents and coronary territories might have occurred.iven that most patients had 3-vessel disease andere fully revascularized, we do not believe that this

ignificantly influenced our study findings. Al-hough we excluded segments from incompletely

Force on practice guidelines (Com-mittee on Coronary Angiography). segmentation and

hether revascularized segments had improved per-usion after revascularization. That the lack ofeperfusion at the tissue level may have contributedo reduced recovery of dysfunction in the presenttudy cannot be excluded. Finally, the number ofegments with �50% scarring was quite low, andur study might have been underpowered for de-ecting small increases in strain in these segments.

O N C L U S I O N S

ur study evaluated the relationship between theuantitatively measured transmural extent of necro-is and improvement of regional function afterevascularization. We demonstrated an inverse re-ationship between the transmural extent of DE andtrain, yet with limited sensitivity for predictingecovery of regional dysfunction in segments with-ut or with very little necrosis. Nevertheless, in thisroup of patients with severe LV dysfunction, webserved more beneficial effects of revascularizationn improvement of global than on regional dys-unction. This illustrates the complex relationshiphat may exist between metabolic viability andunctional recovery after revascularization.

eprint requests and correspondence: Dr. Bernhard L.erber, Division of Cardiology, Cliniques St. Luc, Uni-

ersité Catholique de Louvain, Av. Hippocrate 10/2803,-1200 Brussels, Belgium. E-mail: bernhard.gerber@

evascularized segments, we did not evaluate uclouvain.be.

E F E R E N C E S

1. Kim RJ, Wu E, Rafael A, et al. Theuse of contrast-enhanced magneticresonance imaging to identify revers-ible myocardial dysfunction. N EnglJ Med 2000;343:1445–53.

2. Selvanayagam JB, Kardos A, FrancisJM, et al. Value of delayed-enhancement cardiovascular magneticresonance imaging in predicting myo-cardial viability after surgical revascu-larization. Circulation 2004;1535–41.

3. Wellnhofer E, Olariu A, Klein C, etal. Magnetic resonance low-dose do-butamine test is superior to scar quan-tification for the prediction of func-tional recovery. Circulation 2004;109:2172–4.

4. Scanlon PJ, Faxon DP, Audet AM, etal. ACC/AHA guidelines for coro-nary angiography. A report of theAmerican College of Cardiology/American Heart Association Task

Developed in collaboration with theSociety for Cardiac Angiography andInterventions. J Am Coll Cardiol1999;33:1756–824.

5. Gerber BL, Thanh HT, Roelants V,Pasquet A, Vancraeynest D, Vanover-schelde JL. Flow-function relation-ships in chronic left-ventricular isch-emic dysfunction: impact of thetransmurality of infarction. J NuclCardiol 2008;15:363–74.

6. Pouleur AC, le Polain de Waroux JB,Kefer J, Pasquet A, VanoverscheldeJL, Gerber BL. Direct comparison ofwhole-heart navigator-gated magneticresonance coronary angiography and40- and 64-slice multidetector rowcomputed tomography to detect thecoronary artery stenosis in patientsscheduled for conventional coronaryangiography. Circ Cardiovasc Imag-ing 2008;1:114–21.

7. Cerqueira MD, Weissman NJ, Dilsi-zian V, et al. Standardized myocardial

nomenclature for

tomographic imaging of the heart: astatement for healthcare professionalsfrom the Cardiac Imaging Committeeof the Council on Clinical Cardiologyof the American Heart Association.Circulation 2002;105:539–42.

8. Pereztol-Valdes O, Candell-Riera J,Santana-Boado C, et al. Correspon-dence between left ventricular 17myocardial segments and coronary ar-teries. Eur Heart J 2005;26:2637–43.

9. Heiberg E, Sjögren J, Ugander M,Carlsson M, Engblom H, Arheden H.Design and validation of segment—freely available software for cardiovas-cular image analysis. BMC Med Im-aging 2010;10:1.

0. Bondarenko O, Beek AM, Twisk JW,Visser CA, van Rossum AC. Timecourse of functional recovery after re-vascularization of hibernating myocar-dium: a contrast-enhanced cardiovas-cular magnetic resonance study. Eur

J U L Y 2 0 1 0 : 7 2 0 – 3 0

Gerber et al.

1. Knuesel PR, Nanz D, Wyss C, et al.Characterization of dysfunctionalmyocardium by positron emission to-mography and magnetic resonance:relation to functional outcome afterrevascularization. Circulation 2003;108:1095–100.

2. Gerber BL, Garot J, Bluemke DA,Wu KC, Lima JA. Accuracy ofcontrast-enhanced magnetic reso-nance imaging in predicting improve-ment of regional myocardial functionin patients after acute myocardialinfarction. Circulation 2002;106:1083–9.

3. Gerber BL, Ordoubadi FF, Wijns W,et al. Positron emission tomographyusing (18)F-fluoro-deoxyglucose andeuglycaemic hyperinsulinaemic glucoseclamp: optimal criteria for the predic-tion of recovery of post-ischaemic leftventricular dysfunction. Results fromthe European Community ConcertedAction Multicenter study on use of(18)F-fluoro-deoxyglucose positronemission tomography for the detectionof myocardial viability. Eur Heart J

14. Vanoverschelde JJ, Depre C, GerberBL, et al. Time course of functionalrecovery after coronary artery by-pass graft surgery in patients withchronic left ventricular ischemic dys-function. Am J Cardiol 2000;85:1432–9.

15. Bax JJ, Visser FC, Poldermans D, etal. Time course of functional recoveryof stunned and hibernating segmentsafter surgical revascularization. Circu-lation 2001;104:I314–8.

16. le Polain de Waroux JB, Pouleur AC,Goffinet C, Pasquet A, VanoverscheldeJL, Gerber BL. Combined coronaryand late-enhanced multidetector-computed tomography for delineationof the etiology of left ventricular dys-function: comparison with coronary an-giography and contrast-enhanced car-diac magnetic resonance imaging. EurHeart J 2008;29:2544–51.

17. Vanoverschelde JL, D’Hondt AM,Marwick T et al. Head-to-head com-parison of exercise-redistribution-reinjection thallium single-photonemission computed tomography andlow dose dobutamine echocardiogra-phy for prediction of reversibility of

function. J Am Coll Cardiol 1996;28:432–42.

8. Pasquet A, Robert A, D’Hondt AM,Dion R, Melin JA, VanoverscheldeJL. Prognostic value of myocardialischemia and viability in patientswith chronic left ventricular isch-emic dysfunction. Circulation 1999;100:141– 8.

9. Yan AT, Shayne AJ, Brown KA, et al.Characterization of the peri-infarctzone by contrast-enhanced cardiacmagnetic resonance imaging is a pow-erful predictor of post-myocardial in-farction mortality. Circulation 2006;114:32–9.

0. Schmidt A, Azevedo CF, Cheng A, etal. Infarct tissue heterogeneity bymagnetic resonance imaging identifiesenhanced cardiac arrhythmia suscepti-bility in patients with left ventriculardysfunction. Circulation 2007;115:2006–14.

ey Words: cardiac magneticesonance y myocardialnfarction y myocardial viability

strain.

relationship between transmural extent of necrosis and quantitative recovery of regional strains...

Documents

revascularization in the rabbit hindlimb: dissociation...

transmyocardial revascularization with co2 laser in patients...

ischemic mitral regurgitation: revascularization alone...

what is the best perfusion temperature for coronary...

transmural extent of acute myocardial infarction predicts...

a general linear model for sea ice extent

assessment of myocardial oxidative stress in patients after...

native coronary disease progression exceeds failed...

land-salinisation-upland-nsw-murray-darling-basin-extent

the occluded renal artery: durability of revascularization

reperfusion injury and inflammatory responses following...

prospective randomized trial comparing eus and egd for...

the extent and chronology of cenozoic global glaciation

the role of transmural ventricular heterogeneities in...

2014 esc/eacts guidelines on myocardial revascularization:...

off-pump coronary revascularization: a potential benefit for...

extent & impact of land degradation and rehabilitation

vascular endothelial growth factor gene delivery for...

aprotinin therapy for reoperative myocardial...

anterograde techniques for percutaneous revascularization of...