oninvasive quantification of regional myocardialoxygen consumption has been an elusive and longstanding objective ofcardiovascular research. Its utilityis analogous to that of delineation of ventricular function, now readily accomplished by ventriculography orechocardiography among other methods; or of assessment of myocardial blood flow. Its potential value isunderscored by clinical issues such as optimization oftreatment of patients with ischemic heart disease requiring titration of myocardial oxygen requirementsand ventricularperformance,and by the importance ofdelineating viable myocardium after interventions suchas coronary thrombolysis, essential for identification ofpatients who may benefit from subsequent additional

Received Feb. 20, 1989; revision accepted June 12, 1989.For reprints contact: Mary Norine Walsh, MD, Cardiovascular

Division,Washington University School of Medicine, Box 8086,660 S. Euclid Ave., St. Louis, MO 63110.

measures such as angioplasty or coronary artery bypassgrafting. In addition, elucidation ofthe pathophysiologyof syndromes such as angina with angiographicallynormal coronary arteries (syndrome X) or of cardiomyopathies ofdiverse etiologies requires quantificationof regional myocardial oxygen consumption and callmation of oxidative metabolic reserve.

Despite the quantitative power of positron emissiontomography and its sensitivity for detection of labeled,physiologic substrates, quantification of carbohydrateor fattyacid metabolism alone does not provide a directmeasure of regional myocardial oxygen consumption(1,2). We have shown that admixture of diverse substrates obscures changes in myocardial oxygen consumption estimated on the basis of metabolism ofindividual substrates such as glucose or fatty acid (2).In addition, the myocardial handling of either labeledpalmitate or glucose (or the glucose analogue, fluorodeoxyglucose) is critically dependent upon the level of

The Journal of Nudear Medicine1798 Walsh,Geltman,Brownetal

Noninvasive Estimation of RegionalMyocardial Oxygen Consumption by PositronEmission Tomography with Carbon-11 Acetatein Patients with Myocardial InfarctionMary Norine Walsh, Edward M. Geltman, Michael A. Brown, C. Gregory Henes,Carla J. Weinheimer, Burton E. Sobel, and Steven R. Bergmann

Cardiovascular Division, Department oflnternal Medicine, Washington University School ofMedicine, St. Louis, Missouri

Wepreviouslydemonstratedin experimentalstudiesthat myocardialoxygenconsumption(MVO2)can be estimated noninvasively with positron emission tomography (PET) fromanalysisof the myocardialturnoverrateconstant(k)afteradministrationof carbon-I1 (11C)acetate.To determineregionalk in healthyhumansubjectsandto estimatealterationsinMVO2accompanyingmyocardialischemia,we administered[11C]acetateto five healthyhumanvolunteersand to six patientswith myocardialinfarction.Extractionof [11Cjacetatebythe myocardiumwas avidandclearancefromthe blood-poolrapidyieldingmyocardialimagesof excellentquality.Regionalk was homogeneousin myocardiumof healthyvolunteers(coefficient variation = 11%). In patients, k in regions remote from the area of infarction wasnot differentfromvaluesin myocardiumof healthyhumanvolunteers(0.061±0.025compared with 0.057 ±0.008 min1). In contrast, MVO2in the center of the infarct regionwasonly6% of that in remoteregions(p < 0.01).In four patientsstudiedwithin48 hr ofinfarction and again more than seven days after the acute event, regional k and MVO2did notchange.Theapproachdevelopedshouldfacilitateevaluationof the efficacyof interventionsdesigned to enhance recovery of jeopardized myocardium and permit estimation of regionalMVO2andmetabolicreserveundetlyingcardiacdiseaseof diverseetiologies.

J Nucl Med 30:1798—1808,1989

arterial substrate content as well as on the hormonalmilieu (3—6).

We recently hypothesized that measurement of oxidation of acetate would provide an indirect measure ofmyocardial oxygen consumption (7-9). In contradistinction to the case with glucose or fatty acid, metabolism of acetate is confined virtually exclusively to mitochondrial oxidation. We have demonstrated in isolated perfused hearts (7) and canine hearts in vivo (8)that the extraction of carbon-i 1- (‘‘C)labeled acetateby myocardium is avid and that the clearance from thetissue is very closely correlated with both oxidation oflabeled acetate to ‘‘CO2and with myocardial oxygenconsumption over a wide range of cardiac workloads.Furthermore, we have demonstrated that the quantitative relationshipbetween myocardial oxygen consumption and turnover of [“Clacetateis persistent despitemarked changes in the proportion of substrate presented to the heart in vivo (9).

The presentstudy was undertakento characterizetheturnover rateconstant of [‘‘C]acetateby positron emission tomography in normal human subjects and inpatients with acute myocardial infarction and to usethe ‘‘Cmyocardial turnover rate constant to estimateregional myocardial oxygen consumption.

METhODS

SubjectsThe protocol used was approved by the Washington Uni

versityInstitutionalReviewBoard(Human StudiesCommittee). After the nature ofthe study had been explained to eachsubject, written informed consent was obtained.

Five male volunteers with mean age of 28 yr (range 25 to36 yr) comprised the control group. All had normal electrocardiograms,no history of cardiac disease, and no knowncardiac risk factors. Six patients (five male and one female)with transmuralmyocardialinfarction(documented by analysis of plasma creatine kinase activity and by Q waves on theelectrocardiogram)werestudiedas well.None had been subjected to acute interventions such as coronary thrombolysisor balloon angioplasty.Their mean age was 48 yr (range= 38to 61 yr).One patienthadsustainedinferiorinfarction,andfivehad sustainedanterior infarctions.Four (onewithinferiorandthreewithanteriorinfarction)studiedinitiallywithin48hr ofinfarction were studied again at least 7 days after infarction to define sequential changes in [‘‘Cjacetatemetabolismin normal and infarct zones.

ProtocolEach subject was positioned in Super PElT I, a whole

bodypositronemissiontomographthat providessimultaneousacquisitionofdata sufficientfor reconstructionofseven transaxial slices with a center-to-centerslice separationof 1.5 cm,anda slicethicknessof 1.14cm (10). A transmissionscanofthe chest was obtained with a ring of germanium-68/gallium68 to correct for photon attenuation ofthe emitted radiation.The transmission scan was viewed prior to the collection of

emissiondata to verifyproper positioningof the patient sothat four to six of the seven cross-sectionaltransaxialimagingplanestransectedthe heart.Correctpositioningwas maintamed throughout the study with the use of a lightbeam andindelible marks on the subject's torso. Polyurethane moldswere made individually for each subject prior to each studyand used to stabilizethe head, neck, and upper torso tominimize movement and maintain identical positions formultipletomograms.Emissiondata werecollectedin the highresolutionmode witha reconstructedresolution(fullwidthathalfmaximum) of 13.5mm.

To characterizemyocardialperfusion,we usedthe diffusible traceroxygen-15- (‘SO,t½= 2.1 mm) labeled water and amethod previously developed in our laboratory and validatedin animals and patients (11—14).After collection of attenuationdata,0.4 mCi/kgof H2'5Owereinjectedas a bolusviaalarge bore catheter inserted into an antecubital vein. Datawerecollectedbeginningat the time ofadministration of tracerand continued for 150 sec in list mode with time-of-flightcorrection. After a 5-mm interval for decay ofactivity of tracerto baselinelevels,40 to 50 mCi of oxygen-15-labeledcarbonmonoxide (C'5O)were administeredby inhalation to label theblood pool. After a subsequent interval of 30 to 60 sec forclearanceof C'@Ofrom the lungs, data were collected for 5mm. After an additional 5-mm interval for decay of radioactivity to baseline levels, 0.4 mCi/kg of[' ‘C]acetate(t@= 20.3mm) were injected as a bolus. In general, ‘‘Cdata werecollectedfor 30 mm. In some controls, collectionswerecontinued for as long as 60 mm afteradministration of tracer.

Analysisof TomographicDataPerfusion. To display the distribution of nutritional perfu

sion after administration of â€@̃Owater, radioactivity emanatingfromthebloodpoolwascorrectedwiththeuseofC'5Oandamethodvalidatedpreviouslyin our laboratory(11—14).Ineach pixel of the tomographic reconstructionobtained afteradministration of H2'5O,radioactivityemanating from theblood pool was subtracted to provide an image of blood flowin the myocardium. Data reconstructedinto cumulative 120-sec images were used to identify regions of diminished perfusion in heartsof patients with myocardial infarction.

Metabolism. To characterize qualitatively the distributionof radiolabeled acetate in the heart and obtain a myocardialimage, data were reconstructed into a single static 5-mmcomposite scan. The reconstruction encompasseddata collected from 3 to 8 mm after administration of tracer, anintervalwhen blood-poolradioactivityhad declinedsubstantially. Images were used subsequently for placement of regionsof interestbut not for quantitativeanalysis.

For quantitative analysis, a single midventricular tomographic slice was selected from the reconstructed tomogramsin normal control subjects. In patients, the reconstructioncontaining the infarct region, as delineated from the H2150perfusionscan, was used for analysis.In the case of anteriorinfarction,midventriculartomogramswere used. In the caseof inferior infarction a more apical tomographic plane wasused.

Regions of interest with a volume of 1 cm3 were placedcircumferentiallyon each tomographic reconstruction. Formidventriculartomograms,nine regionsof interestwereemployed. For more apical tomograms, three posterior regionswere identified as well. An additional region of interest was

Volume30 •Number11 •November1989 1799

assigned to the center of the left atrium or the left ventricularcavityfordeterminationofactivityoftracer in the bloodpool.

Regionsof interest in tomograms obtained after administration of [‘‘C]acetateto patients were subdivided to encompass regions of interest in normal, infarcted, and pen-infarcttissue.Normal regionswereidentifiedas those fullycontainedwithin regions of myocardium with normal perfusion (> 50%of maximum flow) judging from the H2'5Otomograms. “Infarct―zoneswere identifiedas regionsof hypoperfusionwith< 50% of maximum perfusion. In addition, the 1 cm3 regionof interest in the center of this zone of hypoperfusionwasidentified as being the most ischemic zone and designated“centralinfarct―.Pen-infarct regions on either side of theinfarct zone weredelineatedas the regionsof interest immediately adjacent to the infarct but within a zone of perfusionwith counts exceeding 50% of maximum activity of H2'5O.

The [‘‘C]acetatemyocardialturnover rate constant (k),wasobtained from the equation:

q = e@t

where,q = counts/voxel/minute corrected for physical decay,k = myocardial turnover rate constant, andt = time

It wascalculatedfrom sequential90-120-secframes@A multiexponential, least squares, curve fitting procedurewas usedto fit myocardialtime-activitycurvesbeginningat the time ofoccurrence of maximal myocardial ‘‘Cactivity. Data wereexpressed also as the myocardial clearance half-time (t@1,=ln2/k).

Myocardialoxygen consumption (MVO2)was estimatedfor each defined region of interest based on the relationshipbetween the myocardial turnover rate constant, k, and directlymeasuredMVO2in 33 individualstudies in intact dogs(Fig.1) (8,9). Although this relationship has not been establisheddirectlyin humans, becauseof ethicalconsiderations,we didnot feelthatcoronarysinuscatheterizationwasjustifiableinnormal subjects.Coronary sinus catheterization in patientsanddeterminationofarterial-venousdifferencescouldprovideestimates of myocardial oxygen consumption, but would provide global, ratherthan regional, values.

Because clearance of radioactivity from the arterial bloodwasso rapid(seeResults)spilloverfrombloodto myocardiumdid not requirecorrection.Wehavepreviouslydemonstratedin experimental animals that spillover correction of data obtamed after administration of [‘‘Cjacetateis unnecessarybecauseofthe rapid clearancefrom arterial blood (8).

Preparation of TracersOxygen-15 water, C'5O and [‘‘Clacetatewere prepared as

previouslydescribedin detail (7,15,16).PUrityof[' ‘C]acetatewas typicallygreater than 99.5%. Preparations of 200—300mCi,witha specificactivityofgreaterthan 1Ci/mmol,wereachievedroutinely.

StatisticsDataarereportedas means±standarddeviations.Signifi

cance ofdifferences between groups were assessed with analysis of varianceand post-hoc unpairedanalysis. Within groupsdifferenceswerecomparedwith the useofpaired tests.Differences with a p < 0.05 were considered to be significant.

EI-z

(I)z0C-)

L@J

U)>

a:I-

0S

0.

S

0.;

0. n â€3̃3r@ 0.88ys 0.036X + 0.021

0 I 2 3 4 5 6 7 8 9 0MYOCARDIAL OXYGEN CONSUMPTION

(pmol/g 1mm)FIGURE 1Correlation between the myocardial turnover rate constantmeasurednoninvasivelywith PETafter i.v. administrationof [11C]acetateand MVO2measuredby direct arterialcoronary venous sampling Obtained from 33 individualmeasurements in 18 anesthetized, intact dogs studied atrest or after myocardial work had been altered wfth sympathetic stimulation or blockade, or after the pattern ofmyocardialsubstrateuse had beenchangedby alteringarterial substrate content after infusion of either glucoseor lipid. The data presented here Is a summary of theresults from previously published studies from our laborstory (8,9).Thisrelationshipwas usedto estimateregionalMVO2in humansfrom the myocardialturnoverrate conslant.

RESULTS

HemodynamicsNo adverse effects of administration of any tracer

were observed. All subjects tolerated the tomographicprocedures well. The rate-pressure product (heart ratex systolic blood pressure), an index of global myocardial work, was calculated for all subjects. The ratepressure product was lower in control subjects undergoing tomography than it was in patients with infarction(7051 ±1131 bpm x mmHg compared with9517 ±3023, p < 0.05), presumably because of differences inautonomic nervous system activity. The rate-pressureproduct was also lower at the time of follow-up 1 wkafter infarction in the four patients studied than it wasin studies within 48 hr of onset of infarction (7082 ±2654 at follow-up compared with 9879 ±3332 acutely,p < 0.05).

Assessment of Myocardial PerfusionNutritional perfusion, assessed with H2'50, was ho

mogeneous in all subjects in the control group (Fig. 2).In contrast, all patients with myocardial infarction displayed diminished perfusion in a tomographic regionthat correspondedto the locus ofthe region of infarction

TheJournalof NuclearMedicine1800 Walsh,Geitman,BrownetaI

•285

!14251756 1878

.@

, .@ :@@

@ @‘& ;@

8

FIGURE 2Mid-ventricular, transaxial, tomographic reconstructions from a representative control subject (P360). On the left is thereconstruction obtained after administration of 150 water, corrected for vascular radioactivity. On the right is thecorresponding tomogram obtained after administration of [11C]acetate.In these mid-ventricular, transaxial reconstructions,the septumis on the left, the lateralfreewallon the right,the anteriormyocardiumtowardthe top, andthe mitralvalveplane,devoldof significanttraceruptake,towardthebottom.Concordanceof distributionof tracerinthe perfusionand metabolicimagesis evident.The slightdiscontinuityof tracer uptakeseen in the septumin the H2150imageisrelated to statistical noise induced by correcting the image for radioactivity emanating from the blood pool. Uptake of[11C]acetatewas avid and resulted in myocardial images of high quality in all subjects.

delineated electrocardiographically (Fig. 3). The regionof diminished perfusion defined with H215O was usedto distinguish infarct, normal, and pen-infarction regions.

Analysis of Data after Administration of t―CjAcetateAn example of a midventriculartomographic recon

struction obtained after intravenous administration of[“Clacetateon a normal subject is shown in Figure 2.The homogeneity of myocardial ‘‘Cactivity observedand the quality ofthe image are representativeof resultsin each control.

A reconstruction from a patient with anterior myocardial infarction is displayed in Figure 3. Both theH215O and the F'‘C]acetatetomograms show markedlydecreased tracer in the anterior myocardium and relatively normal distribution ofperfusion and metabolismin the septal and lateralwalls.

Quantitative AnalysisAs shown directly in studies in experimental animals

(8,9), and confirmed in the human subjects studiedhere, [‘‘Cjacetateradioactivityin the blood pool dimin

ished by 90% to 95% from peak radioactivity, usuallywithin 2 mm (Fig. 4).

For each of the 9—12regions of interest placed ontomographicreconstructions,a multi-exponential curvefit was used to characterizedecay-correctedmyocardialtime-activity curves from the time of occurrence ofmaximal myocardial ‘‘Cradioactivity. In contrast tothe biexponential clearance of ‘‘Cradioactivity observed in isolated rabbit hearts (7) and in hearts of dogsin vivo (8,9), clearance of ‘‘Cwas monoexponential inboth controls and patients (Fig. 5). This observation isconsistent with the lower workload seen at rest innonanesthetized, noninstrumented human subjectscompared with that in anesthetized dogs and with preliminary observations in humans reportedby Pike et al(17) and by Henes et al. (18). No second exponentialcomponent could be identified even in two subjectsstudied for as long as 1 hr afteradministration of tracer.

Clearance ‘‘Cradioactivity from the myocardiumwas homogeneous in control subjects (Table 1, Fig. 5).The variation ofthe turnover rate constant within eachtomogram, reflectedby the coefficient of variation, was

FIGURE 3Mid-ventriculartomographicreconstructionsof perfusion(left)andoxidativemetabolism(right)in a patientwith acute, anterior, myocardial infarction (P354). A large anterior flowdeficit is evidentin the perfusiontomogram and a concordant decreasein myocardial oxygen consumptionreflected by decreased [11C]acetateuptake is evident in the tomogramobtained after administration of [11C]acetate.

1801Volume30@ Number11@ November1989

from k (Fig. 1), in normal human subjects, calculatedMVO2averaged0.97 ±0.23 @imol/g/min(Table 1).

In contrast to the homogeneous clearance observedin hearts of control subjects, clearance of radioactivityfrom the myocardium of patients with infarction washeterogeneous (Table 2, Fig. 5). Uptake oftracer in thezones ofinfarction was decreasedand the turnover rateconstant within the infarct zone was markedly diminished, indicative of a profound diminution of regionalMVO2 (Fig. 6). Regions immediately adjacent to theinfarct zone also had diminished MVO2 in comparisonto those regions more remote from the infarct (Fig. 6).

To define changes in oxidative metabolism over timein infarct and pen-infarct zones, four patients studiedinitially within 48 hr of the onset of symptoms werestudied again at least 7 days after the infarction. In eachofthese patients, uptakeof[' ‘C]acetateremained markedly diminished in the infarct zone. The turnover rateconstant, clearance half-time, and estimated myocardialoxygen consumption did not change with time (Table3, Fig. 7).

DISCUSSION

Myocardial Imaging with t―CiAcetateIn this study, uptake of [‘‘C]acetatewas found to be

avid in normal myocardium permitting acquisition ofmyocardial tomographic images of high quality. Therapid clearance of tracer from the blood pool results ingood contrast between blood and myocardium and

0 Remote£ Pen-infarct0 Infarct

2 6 20 24 28TIME(minutes)

N0‘C

-JLU><

FIGURE 4Clearance of 11Cradioactivity from the blood was rapid.Radioactivity cleared to < 10% of maximum within 2 mmafteradministrationof tracer.

11 ±3% in control subjects. Mean values ofk, clearancehalf-time and estimates of MVO2 averaged from allregions, were similar to values obtained from a regionof interest drawn to encompass the entire left ventricle(Table 1). On the basis of extrapolation from analysesin experimental animals in which the myocardial turnover rateconstant was compared with MVO2measureddirectly to yield a regression equation to estimate MVO2

4

l0@8

6

4

2

86

4

2

02

0 Septum0 AnteriorA Lateral

-JU)x0.C/,I—z3C.)I—U)0@

-JLU

0.

C/)I-

0

—-.-.-- 0a@ 0

0

86

4

2

86

4,

2

0' 5 10 5 20 5 @@iö 15 20TIME(min) TIME (mm)

FIGURE 5Representative myocardial time-activity curves obtained from one control subject(P336, A) and one patient with anteriormyocardial infarction (P362, B). Three of the 9—12 regions of interest comprising 1 cm@each are displayed. In hearts ofindividuals without cardiac disease, clearance of tracer is uniform, monoexponential, and homogeneous. In contrast, inpatients with myocardial infarction, initial uptake of [11Cjacetate in the infarct zone is diminished, and clearance isprolonged, consistent with a decrease in myocardial oxygen consumption. These examples from a control subject andfrom a patient with acute myocardial infarction are typiCalof those seen in all subjects studied.

1802 Walsh,Geltman,BrownetaI TheJournalofNuclearMedicine

0

;; .@Q0u—ç@CO' .@

a:COIor...F.@r@.,_@

0

I

I.I@II@ U

I@II@ .i.@

I@II;0 0 ‘j.

‘-0000.26:@

a.Cv)Cs1i-Cs4CsJ@

1-00,-0 @0

?U@@fl@@-do@@ II

2..@ :@

r

U;

@ @1

I.@.0

01

0>

C0.@.

E

.@

@1

.@

I

I

0) CD CsJ0)

,- 0 0 i- 0

10@00,@@CD

I....10 11)0 11),- I- I- C@JC')CO0CsJ

c@.@-:@-c44Cv)c@J0000004141414141411@ C%J 0 I@) CD F@C') CD 00)0),- 0 0 @-00

I

.@

.@

.@

I

>S

Il)

IC')

cs1

C')

c'J

C')

c@J

C')

c@J

C4J'-IDO)CD

I.- 0 0 0 ‘-

.c

I

II

CD ‘-0) CD F'-0 CD CD C's)0),- 0 0 ‘-0

§

0000ar;;00

“1W

0) CD CO ‘-Co

0

000000414141414141

000000000000

I@II@

I@II@

l@II@

00000

CDI@@.01O@-CD F'-

0000000000

It)CsJCO F-.

00000

C@J0COC')P-.

00000

00000

@—00000.: 00000

0C'J0@-1t)t'-.

000000

@ 0 It)0) CD CsJ

@ 00000

:@V Cl)

F-.C@Ji- CD@,-C'-.CO0C')1@ 0 0 ‘-@

,-000@

@00@@

CD F'-CD LI)C@)0CO@-.F'@1@ 0 0 0@

1803Volume30 5Number11@ November1989

0)

tuc,Ju@0)@.+1I-@ I-@

ciLqcsJc%J,-@@

CsJ

Cs1COC') CD 0 0)1@ @- I-@ C@J i

+1+111+1+1+1C')CO i- 0 CO CO0@c%J'-c@J

I@I@ It) It)

ICSJI@CD@

CDI IC')It) cvi

CqCqc').0@.0)C')F-.C')C@J

@-C@0@F-@CD

0)@ C')00)

C@@C')IO@0 It) C') 00

0) C')@ C')0

CO 0) i- ‘@‘- C') C') 0 CsJ

0 CDC) COCs1

CO CDCD C')Cs4C@J0

‘@.c@J 0) CO F-.@

@ Cvicvici o

j

@iti@oi

g:@ .@.IC,)

a:@

0) 00

Cd L V

Cl)@ @?8@8@I; a.C@ ociooacil.H II4.'.@ F'.

0@ 01

)( lCDaj @C,)@@‘5 j @:@@iI;.C lcD

w.@ ii

-@ @co8cv8ol3 .@

0 @o8o@lo> 00000aI+I

@, @cv) E

C@ 82 0 It)@E. 0E 01@ @Z8@(I)@ c:@c@@ddIo VC@ 41410414141141 a.0@ coF-.tt)1p@ IIC.) 0) F-.i-C@C@ c@c@od0@0) CS) CSICSJ @0 C.)>-@ c@i @cóCl)Ic,I

0..@ 41+I•@x0 (‘I,-. @“@i@―

C) r-.,.@ ‘@CDIO@0@ Itic;@ :@:@ i-CS)O)It)C@J@

C')

@ c'Jo (Oh) I@'

@51>-@E@@ .@ dd@doI0@ 4141d4141'HI@H

—1C>@ LO@ @-o)II)I0) II@@ C'I'-C')1C'J4 o,_, 000,0@@ 0000 a.

:t:,@ g@ “- 0I@ a:Eci@ d@@;IaWE

@@ +104100+1141

@@@@ >

=.cu)@ d olo@

n@

@ t@

Cl).c@?‘@C

to

.2 .@ C) CO CD@C')C 0 C@ F@I'@0 .,. ,-@

- .C

@:@ @±§@

4.' “@ 10@

@ .@ 0o0o@0I0

Cl) @1@@1I41

C @2 °@@@ 0@@ !@

(3 0 0 010@0 c-I @-O@.-0I0

.@ Cl C') @@i-@'-1tO

a: 0 dcslddääld .@@L@ > 41@:4141@141I41@

@@ C') c',oc@i'-I@ >EE0)@

E 2 ‘@@-ddd@0

I—@ ,.:,.:cvi,@cSJ,.:IIti28c;@ 0 ,- CDCS)0) CD@ 0) c@ 0

@ 4141.H.H4I4IFH.@@ @Ll)'@@c')It)c')ICO.@ 0@ @a.

E Cs) (Dollg@ Cs)(Or•-C0@C')II)

>5 sG8Go8I@ o@i

@ @@I0@

41414141•H'H141 D@C@ @‘•-?@ o@_

äooociälo @@(De

@ @“

4.' V

@ U). +1 >@ca

@a_Q.a.Q.Q.@I

1804 Walsh,Geltman,Brownetal TheJournalofNuclearMedicine

C

E

I-z4I-U)z0C-)

wI-4

w>9

20

ii -I-

E

I0

0wI-4-J

U-J4U

-I-

@-@bi*InfarctInfarct--i.::

‘-‘-I . -.

Subjects r@ -I Subjects ________________________I I

PATIENTS PATIENTS

FIGURE 6(A)Histogramdepictingthe regionalmyocardialturnoverrateconstantfromall regionsin normalsubjectsandin infarct,central infarct, pen-infarct, and remote regions in patients with myocardial infarction. (B) Estimates of regional myocardialoxygenconsumptionin all regionsfrom normalvolunteersand from infarct,central infarct, pen-infarct,and remoteregions in patients with myocardial infarction. There was no statistically significant difference between regions remotefrom infarction in patients and regions from normal human volunteers. Regional myocardial oxygen consumptiondiminishedin pen-infarctregionsin patients,andwas severelydepressedin the zoneof infarction(p < 0.05 comparedwith remoteregions).

facilitates quantitative analysis of myocardial turnoverrate without the necessity for correctingfor myocardialspillover from blood into myocardium (8,18).

In normal subjects, uptake and clearance of ‘‘Cwashomogeneous and monoexponential. Use of the relationship between k and MVO2 established previouslyin canine hearts in vivo and the data obtained in thepresent study in human subjects yielded a calculatedhuman myocardial oxygen consumption value of 0.97

@mol/g/min,somewhat lower but close to that preyously obtained by direct analysis with invasive procedures (19—21).

In patients with infarction, uptake of [‘‘C]acetateinthe center of the infarct zone, delineated by perfusionimaging with H2'5O, was markedly depressed and clearance was prolonged. Both are a result of diminishedperfusion to the region. The diminished clearance directly reflects the reduction of myocardial oxidativemetabolism. Overall oxygen consumption within zonesof hypoperfusion delineated by H2'5Oin humans was—30%of that in normal zones, a finding consistentwith the recognized magnitude of myocardial collateralperfusion in zones ofinfarction (22). Estimated oxygenconsumption in the center of the hypoperfused zonewas @—6%of normal. Regions remote from the centerof the infarction exhibited normal myocardial uptakeand turnover of tracerindicative of normal myocardialoxidative metabolism. Pen-infarction zones manifestedprolongation of clearance indicative of diminished oxidative metabolism. These regionsmay be metabolically“stunned―.Thus, the diminution ofblood flow to myocardium within these regions, although not sufficient toinduce infarction, is sufficient to diminish regionalwork and regional MVO2.Alternatively,pen-infarctionzones may contain a mixture of viable and infarcted

cells constituting a heterogeneous population manifesting phenomena integrated in the tomograms withinselected regions ofinterest because ofthe limited spatialresolution of PET. This latter interpretation is supported by the finding that MVO2 in peri-infarct regionsin patients with infarction studied acutely and again atleast 7 days after the acute event, at a time by whichfunction and metabolism in myocardium not destinedto undergo infarction have recovered, did not change.

Studies in experimental animals have demonstratedthat with reperfusion, oxidative metabolism assessedwith [‘‘C]acetaterecovers within 1wk in zones destinedto recover and is predictive of recovery of ventricularfunction (23). Studies by others have shown that restoration of ventricularfunction is predicated on recovery of myocardial oxidative metabolism (24,25). Although enhanced glycolytic flux is seen in reperfusedmyocardium (2) and can be demonstrated with fluorine-l8 (‘8f@)fluorodeoxyglucose (25), the prolongedimaging intervals requiredwith deoxyglucose as well asthe dependency of myocardial uptake of this tracer onplasma substrate and hormonal concentrations precludes its use to quantify regional myocardial oxidativemetabolism.

Technical ConsiderationsResults of this study indicate that the turnover rate

constant after i.v. administration of [‘‘Cjacetatecan bequantified by PET. We developed [“Cjacetateas atracer suitable for assessment of regional myocardialoxygen consumption because it is metabolized virtuallyexclusively via mitochondrial oxidation (7—9).Oxidalion ofacetate to CO2in mitochondria is coupled tightlyto oxidative phosphorylation. We have shown preyously that the turnover rate in isolated perfused hearts

1805Volume30@ Number11@ November1989

IIfi1r@L1rt1r@—1r@i

-a:?E@

-@C@ WV

XC0 @Q

V to

EV

j0E

.@

4-,

Cto

4.'

IE

4-,

I

a.

I

I

CD

3§F-.C000

8@41odo@

00

1@ C')LI)C')

3 88C,J0ooooc;418414141CO 0 ‘-C')It)‘0- ‘-0-CS)0 0000 @00

CDCsJw-CDCO

83883ctc@dod

414141414100COCDF'-It)C')―-0-C')ooc:@c@o00000,-

;;r..ci@: CoF'@41

@CS)1-,t

to‘0-

0) 0CD@@C- Cs)@-C')

+i: 1111110) F'-Cs)0)

@ oiCthC'i, CO'-C')

CSJCD'0COcsjcs@-

4141414141CsJ―-C')CDO)‘0C'@'0'0C'@‘- CS) ‘0 ‘- Cs)‘0-

LI)

a.ICD'*CD*41a:l1—F-.Ln@o@

I'.-

LI)CS)

aooc,F—‘@

0

i ‘0@C')0 0011@@41410 0 0 Cs) C')F'- CDI')0 00

‘@‘)@C'@1C') CsJC')do cc4141@4141g@qo@0 0 0 0I

E

.@

.@

‘5

u15

§

.@

0

.V?1E

.Cl)

?.@F'-CsJCDc')O

8s8@a00004141414141C')CSJCDOi

dddoo

‘0@

G

@@o+lCOC')0@0

@doo@0

0IO@0)'-

ci

CS)C'Ji-Oe00000c@ctoctct41-H-H-H-H

00000ddc:@c@o

‘0-C')

CD@@°0

@c@c@-H;t!000k

@:@

@IOCO@@

E@ ocdodctc,84141414141

@C')COC')C')0)toCDF'@C')@10@-00000

@ddddd

@0

VCS)10CDIOCCDCDCSJF'-.'0

OCthCthOF-.

4141414141COCSJ0)F'-'00iu@Lti,.:Có,-C@)‘-@-

F—

C'@:C'i-H@; CS)@5@Cs) @;

0)

C@JCs)CsJ It)ItiCó,-1t@cvii-0)0'0C')

41-H-H-H-HF'-.@@@CqIt@0@-00LOCSJ0)10CSJ@

COi,-F'@

CO@-0)+l+l‘@C'iCDCOCO

I-'-

CO0)CO0C'4@@sj,@cö

4141414141LI)C')LI)C')O)

:@c;,@:cocvi‘@ Cs)'-,-.

@D .@ V

2 :@@CSJ10CO10@

C)c@

>@,

.@

a.

Cl)C8

@I

.@@E

@‘0@000

4141414141C')F'-ON-CSJ

@@It)CsJc,ut,-00‘-0Cs)

a.I@CD@ CD-Ha:l@0.or-.@o@,,,- F—

CD0)

@2

0

01-000IqC4JCS)qC')0'@-000

-H-H-H-H-HCOCOF—LOF'-

@F'@@C―4'0-00000

@3II)0)CD0°F@..@;ti-H000 CO

11)11)oo

0COv-0COi-C')'-'-COcc@OdO4141-H-H-HC')@-CSJi-CsJt-.0C')COO‘@0@'

V

CSJLI)CDLI)@C

@...@

,,,

.@@ :@@

‘E°-

:@@i

C@C‘@

€!.@@(D

W@v@

E•@0a.C@ @C9@‘)E@

§‘5to•.1

ç@a

@:

TheJournalof NuclearMedicine1806 Walsh,Geltman,Brownetal

1@

I

::- ‘2 20

‘C DAcute

:- 010 0 Follow-up

@ I.5L-@ 0 Acute@ 008 0 Follow-upI- N

U) 0

@006@ .0U

[1rLthrI.@i ____0

@004w

c@@ 0.5@ 002 U-J> 4

0 Uzc@ @_ _______________________ _______________________

@ I@/ Remote Pen-Infarct Infc@ct Central 0 Remote Pen-Infarct Infarct Central

A Infarct B InfarctFIGURE 7Histogram displaying the myocardial turnover rate constant, k (A), and estimated myocardial oxygen consumption (B)from four patients studied acutely and again at least 7 days after the acute event. There was no change in regional kor MVO2with time in any region.

parallels regional myocardial oxygen consumption overa wide range of physiologic and pathophysiologic conditions including ischemia and reperfusion (7). Comparisons of the myocardial turnover after intravenousadministration of [“Cjacetate with direct measurements of myocardial oxygen consumption have demonstrated the relationship between the myocardial turnover rate constant (k) detected externally with PET,and MVO2 over a wide range (Fig. 1) (8,9). This relationship is not altered by changes in myocardial workor by the pattern of substrate utilization (9). Despitethe fact that acetate is avidly extracted by the myocardium with an extraction fraction of 30—50%,its oxidation does not account for a largefraction of myocardialenergy supply because it circulates in such low concentrations in normal human subjects(80—100Mmol)(26).Nonetheless, the relationship between the myocardialturnover rate constant and MVO2 has not been established directly in human beings. Values of MVO2 inhumans estimated by extrapolation of the relationshipobtained directly in dogs yield values that are somewhatlower than those that have been measured in humansubjects (19—21).We recently reported that the relationship between k and the rate-pressureproduct (anindex of myocardial work and therefore of myocardialoxygen consumption) obtained in healthy human volunteers was indistinguishable from the relationship obtamed in experimental animals (18). Nonetheless, thediscrepancy in values of MVO2 obtained nomnvasivelyin the present study with values obtained by otherinvestigatorsusing invasive techniques may reflect atrue interspecies difference in the relationship betweenk and MVO2. Although the rate-pressureproduct canbe used to estimate MVO2in normal subjects, it reflectsglobal, rather than regional, MVO2. In contrast, thePET approach developed permits sequential assessments of regional MVO2.

In the present study, areas of normal and hypoper

fusion were estimated using H2'5O and a techniquedeveloped previously in our laboratory to correct dataforintravascularradioactivity.Regional perfusion usingthis approach correlates very closely with regional bloodflow estimated with radiolabeledmicrospheres(11-13).In addition, we have demonstrated the utility of thistechnique in patients with coronary artery disease (14).Although tomography with H2150 permitted independent delineation of normal from ischemic myocardium,correction ofmyocardial H2'5O for vascular radioactivity (which requires an independent measure of theblood pool with C'50) results in images of relativelylow signal-to-noise ratios. Although we recently vaildated a mathematical approach to quantitate myocardial perfusion in absolute terms (i.e., ml/g/min) usingH2'5Oand PET which does not requirean independentmeasurement of the blood pool, we chose not to usethe quantitative approach in the current study becauseit is less sensitive to estimates of flow in infarctedmyocardium (27).

Clinical ImplicationsThe resultsofthe present study indicate that positron

emission tomography with [‘‘C]acetate,a tracer of myocardial oxidative metabolism, provides myocardial images of high quality and permits estimation of regionalmyocardial oxygen consumption with results consistentwith those obtained previously by invasive measurements. Oxidative metabolism in the center of zones ofinfarction is markedlydepressed and, in the absence ofacute interventions, does not change with time. Acuteinterventions are likely to restore oxidative metabolismin jeopardized myocardium and can therefore be evaluated with the approach developed here. Accordingly,quantification of myocardial oxygen consumption with[‘‘C]acetateand positron tomography offers promisefor objective, noninvasive characterization of the relative benefits ofdiverse interventions designed to salvage

1807Volume 30 SNumber 11@ November1989

H2'5O.Circulation 1984;70:724—733.12. Knabb RM, Fox KAA, Sobel BE, Bergmann SR.

Characterization ofthe functional significance of subcritical coronary stenoses with H215O and positronemission tomography. Circulation 1985; 71:1271—1278.

13. Knabb RM, Rosamond TL, Fox KAA, Sobel BE,Bergmann SR. Enhancement of salvageof reperfusedischemicmyocardium by diltiazem. JAm Coil Cardiol1986;8:861—871.

14. Walsh MN, Bergmann SR, Steele RL, et al. Deineation of impaired regional myocardial perfusion bypositron emission tomography with H2'50. Circulation1988;78:612—620.

15. Welch MJ, Ter-Pogossian MM. Preparation of shorthalf-lived radioactive gases for medical studies. RadRes 1968;36:580—587.

16. Welch MJ, Kilbourn MR. A remote system for theroutine production of oxygen-iS radiopharmaceuticals. J Label Comp Radiopharmaceuticals 1985; 22:1193—1200.

17. Pike VW, Eakins MN, Allan RM, Selwyn AP. Preparation of [1-―C]acetate—anagent for the study ofmyocardial metabolism by positron emission tomography.IntJApplRadiatlsot1982;33:505—512.

18. Henes GC, Bergmann SR, Walsh MN, Sobel BE,Geltman EM. Assessment of myocardial oxidativemetabolic reserve with positron emission tomographyand carbon-i 1-acetate. J Nuci Med 1989; 30:1489—1499.

19. Kitamura K, Jorgensen CR, Gobel FL, Taylor HL,Wang Y. Hemodynamic correlates ofmyocardial oxygen consumption during upright exercise.JApplPhysiol1972;32:516—522.

20. Nelson RR, Gobel FL, Jorgensen CR, Wang K, WangY, Taylor HL. Hemodynamic predictors or myocardial oxygen consumption during static and dynamicexercise.Circulation1974;50:1179—1189.

21. Gobel FL, Nordstrom LA, Nelson RP, Jorgensen CR,Wang Y. The rate-pressure product as an index ofmyocardial oxygen consumption during exercise inpatients with angina pectoris. Circulation 1978;57:549—556.

22. Knabb RM, Bergmann SR. Fox KAA, Sobel BE. Thetemporal pattern of recovery of myocardial perfusionand metabolism delineated by positron emission tomography after coronary thrombolysis. J Nucl Med1987;28:1563—1570.

23. Brown MA, Nohara R, Vered Z, Perez JE, BergmannSR. The dependence of recovery of stunned myocardium on restoration of oxidative metabolism [Abstract].Circulation1988;78:11—467.

24. Taegtmeyer H, Roberts AFC, Raine AEG. Energymetabolism in reperfused heart muscle: metabolic correlates to return of function. JAm Coil Cardiol 1985;6:864—870.

25. SchwaigerM, SchelbertHR, Ellison D, et al. Sustainedregional abnormalities in cardiac metabolism aftertransient ischemia in the chronic dog model. J AmCoilCardiol1985;6:336—347.

26. In: Lentner C, ed. Geigy Scientific tables. Basle: CibaGeigy, 1984;3:107.

27. Bergmann SR. Herrero P, Markham J, WeinheimerCI, Walsh MN. Noninvasive quantitation of myocardial blood flow in human subjects with oxygen-iSlabeled water and positron emission tomography. JAm Coil Cardiol: in press.

jeopardized, ischemic myocardium. In addition, theapproach developed should be useful for evaluatingregional MVO2 and metabolic oxidative reserve in thehearts of patients with cardiac disease of diverse etiologles and their response to therapy.

ACKNOWLEDGMENTS

The authors thank Theron Bairdand ColleenSchaab,RNfor technical assistance, David Marshall, BS, James Bakke,MS, and the staff of the Washington University MedicalCyclotron for preparationof tracers;and Becky Leonard forpreparationof the typescript.

This work was supported in part by National Heart, Lungand Blood Institute, Grant HL17646, SpecializedCenter ofResearchin IschemicHeart Disease.

REFERENCES

1. Bergmann SR, Fox KAA, Geitman EM, Sobel BE.Positron emission tomography ofthe heart. Prog CardiovascDis1985;28:165—194.

2. Myears DW, Sobel BE, Bergmann SR. Substrate usein ischemic and reperfused canine myocardium: quantitative considerations. Am JPhysiol(Heart CircPhysiol)1987;253:H107—H114.

3. Schelbert HR, Henze E, Schon HR, et al. C-i 1 palmitate for the noninvasive evaluation ofregional myocardial fatty acid metabolism with positron computedtomography. III. In vivo demonstration of the effectsof substrate availability on myocardial metabolism.Am HeartJ 1983;105:492—504.

4. Schelbert HR, Henze E, Sochor H, et a!. Effects ofsubstrateavailabilityon myocardialC-i 1 palmitatekinetics by positron emission tomography in normalsubjectsand patients with ventricular dysfunction. AmHeartf 1986;111:1055—1064.

5. Phelps ME, Hoffman El, Scm C, et al. Investigationof [‘8F]2-fluoro-2-deoxyglucose for the measure ofmyocardial glucose metabolism. J Nuci Med 1978;19:1311—1319.

6. Merhige ME, Ekas R, Mossberg K, Taegtmeyer H,Gould KL. Catecholamine stimulation, substratecompetition, and myocardial glucose uptake in conscious dogs assessed with positron emission tomography. Circ Res 1987; 61:II-l24—II-129.

7. Brown MA, Marshall DR, Sobel BE, Bergmann SR.Delineation ofmyocardial oxygen utilization with carbon- 11labeled acetate. Circulation 1987;76:687—696.

8. Brown MA, Myears DW, Bergmann SR. Noninvasiveassessment of canine myocardial oxidative metabolism with carbon-i 1 acetate and positron emissiontomography. JAm CoilCardiol 1988; 12:1054—1063.

9. Brown MA, Myears DW, Bergmann SR. Validity ofestimates of myocardial oxidative metabolism withcarbon-i 1-acetate and positron emission tomographydespite altered patterns ofsubstrate utilization. J NuciMed 1989;30:187—193.

10. Ter-Pogossian MM, Ficke DC, Yamamoto M, HoodJT. Super PElT I: A positron emission tomographutilizing photon time-of-flight information. IEEETransMedlmagi 1982;3:179—187.

11. Bergmann SR, Fox KAA, Rand AL, et al. Quantification of regional myocardial blood flow in vivo with

1808 Walsh,Geltman,Brownet al TheJournalof NudearMedicine