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 . 4 , N O . 8 , 2 0 1 1

© 2 0 1 1 B Y T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N I S S N 1 9 3 6 - 8 7 8 X / $ 3 6 . 0 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 1 . 0 4 . 0 1 7

S T A T E - O F - T H E - A R T P A P E R

Stress Myocardial CT PerfusionAn Update and Future Perspective

Tust Techasith, BS,*† Ricardo C. Cury, MD‡

Boston, Massachusetts; and Miami, Florida

Coronary computed tomography angiography (CTA) has been shown by several multicenter trials to have

excellent diagnostic accuracy in the detection and exclusion of significant coronary stenosis. However, a

major limitation of coronary CTA is that the physiological significance of stenotic lesions identified is often

unknown. Stress myocardial computed tomography perfusion (CTP) is a novel examination that provides

both anatomic and physiological information (i.e., myocardial perfusion). Multiple single-center studies

have established the feasibility of stress myocardial CTP. Furthermore, it has been illustrated that a

combined CTA/CTP protocol improves the diagnostic accuracy to detect hemodynamic significant steno-

sis as compared with CTA alone; this combined protocol can also be accomplished at a radiation dose

comparable to nuclear myocardial perfusion imaging exams. Although initial results hold some promise,

stress myocardial CTP is a modality in its infancy. Further research is required to define, validate, and

optimize this new technique. However, it is a modality with significant potential, particularly in the

evaluation of chest pain patients, given the advantages of short exam time and comprehensive data

acquisition. This review highlights how to perform and interpret stress myocardial CTP, summarizes the

current literature, and discusses some future directions. (J Am Coll Cardiol Img 2011;4:905–16) © 2011 by

the American College of Cardiology Foundation

diacculavant.

Recent advances in computed tomography(CT) technology and the rapid evolution ofmultidetector row CT (1), have allowed cardiacimaging to flourish on this platform. CardiacCT has been proven to have numerous clini-cally relevant applications, including coronaryartery calcium scoring, coronary computed to-mography angiography (CTA), global and re-gional left ventricular function assessment, andmost recently, the assessment of myocardialCT perfusion (CTP) (2–6).

The majority of cardiac CT exams performedtoday are coronary CTAs, aiming to elucidate the

From the *Harvard Medical School, Boston, Massachusetts; †CarHospital, Boston, Massachusetts; and the ‡Baptist Cardiac and VasFlorida. Mr. Techasith has reported that he has no relationships relereceives research grants from Astellas Pharma, and GE Healthcare

Manuscript received January 12, 2011; revised manuscript received Ma

patient’s coronary anatomy noninvasively. Coro-nary CTA has been shown by several multicentertrials to have excellent diagnostic accuracy in thedetection and exclusion of significant coronarystenosis (CAD) as compared with invasive coro-nary angiography (6–8).

Despite its successes, coronary CTA hasseveral inherent limitations restricting its use toa specific population. First and foremost, cor-onary CTA only provides anatomic informa-tion. Many studies have established that func-tional information is essential in both guidingclinical management and long-term prognosis

MR PET CT Program, Massachusetts Generalr Institute and Baptist Hospital of Miami, Miami,to the contents of this paper to disclose. Dr. Cury

rch 31, 2011, accepted April 7, 2011.

sradtmatoAudcNnoagatmc

isim2simlcfqsmoes

v

M

M

c

M

P

Q

a

S

computed tomography

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

906

(9–11). The presence of anatomic lesions does notnecessarily correlate with functional abnormality—that is, decreased myocardial perfusion (12). An-other shortcoming of coronary CTA is its tendencyto overestimate the extent of CAD in high-riskpatients due to the presence of calcified plaques andstents (13–14).

Given these shortcomings, many attempts havebeen made to combine coronary CTA with func-tional studies such as positron emission tomographyor single-photon emission computed tomography(SPECT), illustrating complementarity betweenthe techniques (15,16). However, in such scenarios,the patient is required to undergo at least 2 diagnosticstudies in order to arrive at the final result, which not

only is inconvenient but also subjects thepatient to increased radiation exposure. Assuch, it is evident that a CT-based exam thatcombines the information provided by anat-omy and perfusion, namely stress myocardialcomputed tomography perfusion (CTP) im-aging, can have many potential benefits.

Recently, several single-center studieshave illustrated the feasibility of stress myo-cardial CTP and have shown that stressmyocardial CTP adds incremental value totraditional coronary CTA in the detection ofsignificant coronary stenosis, potentiallyovercoming the limitations of coronaryCTA alone (17–20). Although initial resultshold some promise, stress myocardial CTPis a modality in its infancy, with only a smallnumber of subjects evaluated overall, withno clear optimized protocol as of yet. Fur-ther research is required to define, validate,and optimize this technique. However, it is amodality with significant potential, particu-larly in the evaluation of chest pain patients,

given the advantages of short exam time and compre-hensive data acquisition. This review will address pro-tocol setup and interpretation of stress myocardial CTP,current literature, and future directions.

Protocol Setup and Interpretation of StressMyocardial CTP

Overview. Myocardial CTP protocol is composedof a stress phase acquisition and a rest phaseacquisition, similar to a nuclear myocardial perfu-sion imaging (MPI) exam. These acquisitions areevaluated for myocardial perfusion information butalso coronary anatomy as well, generally using the rest

se

hy

hy

ted

w

alue

lue

y

ssion

acquisition. Iodinated contrast is administered both in d

the stress and rest acquisition (60 to 75 ml for eachacquisition), for a total contrast dose of approximately130 to 150 ml. Stress phase imaging is performedunder pharmacological administration of stress agents,such as adenosine, dipyridamole, or regadenoson,similar to nuclear medicine MPI. Additionally, a thirdoptional delayed-phase acquisition can be performedin cases where late contrast enhancement evaluationfor myocardial scar is desired.Patient preparation. In addition to the standardetup of a coronary CTA, myocardial CTP protocolequires a few additional components: namely, 1dditional intravenous catheter if adenosine/ipyridamole infusion is being used, a 12-lead elec-rocardiogram (ECG) machine, and a blood pressureonitor. Two 18- to 20-gauge intravenous catheters

re inserted into the patient’s antecubital veins: 1 forhe delivery of iodinated contrast material and thether for infusion of the pharmacological stress agent.ctual gauge and site of insertion may vary dependingpon the patient’s anatomy. Contrast is prepared in aual-syringe contrast injection system. The pharma-ological stress agent is prepared in an infusion pump.ote is made that newer stress agents, such as regade-

oson, may require only 1 intravenous site for deliveryf the stress agent in a single bolus of 10 s followed bysaline flush. The contrast administration can be

iven after 1 to 2 min of regadenoson infusion. ECGnd blood pressure measurements are performed prioro scan acquisition to establish a baseline for patient

onitoring, during and after the procedure isompleted.Choice of stress agents. Although it has been shownn many studies that pharmacological and exercisetress testing have comparable diagnostic character-stics, exercise is the preferred method of stress in

yocardial perfusion imaging when possible (21–5). This is due to its physiological mechanism andome reports of greater extent, severity, and revers-bility of defects with exercise compared with phar-

acological stress (26,27). The main significantimitation of exercise is that many patients, espe-ially in a specific population that would benefitrom a stress test, are not able to exercise ade-uately. Additionally, all published literature ontress myocardial CTP to date has employed phar-acological stress. As such, this review will focus

n pharmacological stress agents; however, usingxercise stress in myocardial CTP is an area thathould be explored in future studies.

Pharmacological stress agents that have beenalidated via feasibility trials include adenosine and

A B B R E V I A T I O N S

A N D A C R O N YM S

CAD � coronary artery disea

CTA � computed tomograp

angiography

CTP � computed tomograp

perfusion

DECT � dual energy compu

tomography

ECG � electrocardiogram

BF � myocardial blood flo

DCT � multidetector

omputed tomography

PI � myocardial perfusion

imaging

NPV � negative predictive v

PV � positive predictive va

CA � quantitative coronar

ngiography

PECT � single-photon emi

ipyridamole (17–19). Both agents are coronary

sia

gacDeictta

lfmmcctmebbbsia

rsbitcipam

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

907

vasodilators and work via coronary steal phenome-non where the normal coronary arteries dilate morethan their diseased counterpart, leading to a differ-ence in flow in the downstream myocardium andhighlighting the perfusion defect in the territorysupplied by the diseased artery. Exogenously ad-ministered adenosine acts directly on adenosinereceptors, causing coronary vasodilation. Dipyri-damole increases the level of endogenous adenosineby decreasing the level of reuptake mechanism byendothelial cells (28). Both adenosine and dipyri-damole have been shown to have good sensitivityand specificity for the detection of myocardialperfusion defect with stress myocardial CTP. Theyboth have similar side effects profiles that includereflex tachycardia and are similarly contraindicated inpatients with asthma, chronic obstructive pulmonarydisease, advanced atrioventricular block in the absenceof pacemaker, and caffeine use. Adenosine has a shorthalf-life in the seconds and is rapidly removed fromthe body upon administration. Thus, it has a rapidonset and short duration of action, and continuousinfusion is required for the purpose of myocardialCTP. Dipyridamole is also administered using aninfusion pump; however, its effects last longer and,unlike adenosine, may require reversal with an aden-osine receptor antagonist such as aminophylline. Bothagents require weight-based dosing.

Other pharmacological stress agents have beenused for stress echocardiograms and nuclear stressMPI but have not been fully validated in studiesevaluating myocardial CTP. For example, regade-noson is a selective adenosine receptor agonist,preferentially acting on the A2A receptor, leading toelective coronary vasodilation and, therefore, lead-ng to fewer side effects in patients with reactiveirway diseases. Regadenoson is easy to administer,

Table 1. Various Stress Agents and Their Advantages/Disadvan

Agent Advantage

Exercise PhysiologicalLeast expensive

Pharmacological

Adenosine* Good sensitivity and specificity

Dipyridamole* Good sensitivity and specificityInexpensive

Regadenoson Easier dosing (10-s bolus)Fewer side effects in patients with CO

Dobutamine Physiological mechanismIncrease oxygen consumption in the

*Stress myocardial CTP examinations with adenosine and dipyridamole have b

COPD � chronic obstructive pulmonary disease; CTP � computed tomography pe

iven as a standard dose with no need for weightdjustment, but has a prolonged effect. It may alsoause tachycardia, even more so than adenosine.obutamine has been utilized mainly in stress

chocardiography and works via a more physiolog-cal mechanism, by increasing the physiologicalonsumption of oxygen in the myocardium leadingo induction of functional abnormalities. Advan-ages and disadvantages of pharmacological stressgents are summarized in Table 1.Image acquisition. Scout images are acquired toocalize the heart position (Fig. 1). Contrast timingor image acquisition is determined by using a contrastaterial test bolus of 10 to 15 ml (flow rate: 4 to 5l/s) followed by 20 ml of saline flush. The timing is

alculated by adding 2 to 4 s to the time of peakontrast enhancement in the ascending aorta, in ordero achieve a sufficient level of perfusion of contrastaterial throughout the myocardium, from the sub-

picardial to the subendocardial layer and from thease to the apex of the heart. Alternatively, a triggerolus method may be used where an attenuation-ased threshold (Hounsfield unit [HU]) is set at apecific location, for example, 180 HU at the descend-ng aorta, and image acquisition commences once thettenuation is met at that particular location.

There are 2 ways in which to set up a stress andest myocardial CTP protocol based on the order ofcan acquisition, namely, stress phase first followedy rest phase, or vice versa. The main considerations that the first scan will be a “clean” acquisition, andhat the contrast used in the first acquisition canross-contaminate the second acquisition if thenterval between the scans is short (less than ap-roximately 20 to 30 min) (Table 2). When doingstress phase acquisition first, the detection ofyocardial ischemia is optimized by not having

s

Disadvantage

Motion ¡ Not validated using MDCTEffort-dependent

Mild tachycardiaContraindicated in patients with COPD, asth

TachycardiaMay requires aminophylline antagonismSimilar contraindications to adenosine

and asthmaProlonged effectsTachycardia

cardiumLower sensitivity and specificity for perfusioTachycardiaCan provoke ischemia

used in preliminary trials.

tage

ma, and caffeine use

PD

myon defects

een

rfusion; MDCT � multidetector computed tomography.

arnsCriCth

Cosp

stCtfwsic(

CTA � computed tomogr

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

908

contamination of contrast; however, the secondscan can underestimate the presence of infarct inthe myocardium if a short scan interval is used.Iodinated contrast has been shown to have similarpharmacokinetics to gadolinium-based compoundsin magnetic resonance imaging (29). The contrastfrom the stress scan would accumulate in an area ofmyocardial infarct due to the slow wash-out phe-nomenon, leading to possible underestimation ofmyocardial infarction (perfusion defect during therest scan) if the second scan is done too early afterthe first one (i.e., �10 min apart). A coronary CTAcquisition can be acquired simultaneously with theest acquisition, and beta-blockers and sublingualitroglycerin can be given to optimize the secondcan. For the patient population that myocardialTP targets, it can be argued that obtaining accu-

ate stress phase images over rest phase is of utmostmportance, because the primary aim of myocardialTP is to detect myocardial ischemia. This impor-

ance of stress phase myocardial perfusion imagingas been illustrated in previous studies evaluating

Figure 1. Full CTP Protocol

The protocol includes patient preparation and post-exam check (inare as listed in sequence. Heart rhythm and symptoms are monitorsion; ECG � electrocardiogram; IV � intravenous.

and Disadvantages of Different CTP Protocol Sequences

Advantages

Better sensitivity of stress scan (ability to detect ischemia)Coronary CTA can be optimized with second acquisition by giving

medications without interfering with perfusion assessment

Ability to stop protocol after rest phase (if no or minimal diseaseis evident)

Better sensitivity of rest scan (ability to detect infarct)

aphy angiography; CTP � computed tomography perfusion.

SPECT (30); going forward, it will be important toverify these findings for myocardial CTP as well.Moreover, there are subsets of patients that wouldbenefit from a stress perfusion study first, such aspatients with intermediate to high pre-test proba-bility of CAD; patients with high calcium scores(i.e., calcium score �400); patients with known

AD; patients with prior myocardial infarct; or anyther situation where simultaneous evaluation of atenotic coronary vessel leading to ischemia is im-ortant to define revascularization options.On the other hand, performing rest phase acqui-

ition/CTA first could make sense in clinical prac-ice if the decision to proceed to stress myocardialTP would be guided by the CTA results, meaning

hat stress myocardial CTP would only be per-ormed if a moderate or severe coronary stenosisould be found or if there is any nondiagnostic

egment (calcified lesions or stents). The potentialssues of this approach are 2-fold: first, the cross-ontamination of contrast in the second acquisitionstress phase), potentially masking areas of ischemic

ing an optional delayed phase acquisition (shaded box). All stepshroughout the entire exam. CTP � computed tomography perfu-

Disadvantages

Contrast contamination, leading to appearance of late contrastenhancement during rest acquisition (decreased sensitivityfor infarct)

Contrast contamination, leading to appearance of late contrastenhancement during stress acquisition (decreased sensitivityfor ischemia)

Beta-blocker given during first acquisition can underestimatemyocardial ischemia

cluded t

Table 2. Advantages

Sequence

Stress ¡ Rest

Rest ¡ Stress

bCui3aTrafmrwowoCb

p1am

impTslwb

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

909

myocardium (reversible stress perfusion defect) asillustrated in Figure 2; and second, the use ofeta-blockers to optimize image quality of coronaryTA in the first acquisition could also lead tonderestimation of myocardial ischemia. The firstssue can be avoided by having an interval of at least0 min between the rest and stress acquisition tollow washout of contrast from the myocardium.he second issue can be avoided if the temporal

esolution of CT scanners keeps improving, allevi-ting the need to use beta-blockers. In summary, foruture implementation of stress myocardial CTP, itay be advantageous to stratify patients into low-

isk and high-risk groups. The low-risk groupould undergo a rest phase scan first with the aimf ruling out CAD, whereas the high-risk groupould start with the stress phase scan for detectionf myocardial ischemia and then an optimizedTA/rest myocardial CTP exam with a beta-locker and nitroglycerin can be acquired.For the stress phase acquisition, infusion of the

harmacological stress agent is initiated (with a rate of40 �g/kg/min for adenosine infused over 3 to 4 minnd 0.56 mg/kg of dipyridamole infused over 4 to 6in). When the pharmacological stress agent reaches

Figure 2. Contrast Contamination of Stress Acquisition Leading

Patient underwent a computed tomography perfusion (CTP) protocShows the rest acquisition revealing a perfusion defect inferiorly. (Btrast administration, leading to the appearance of a smaller area ofdefect on single-photon emission computed tomography (SPECT). Aticular example, it is easy to imagine that an area of stress-induced

formed second. Of note, there is gastric uptake in (C); however, this did

ts peak, 60 to 70 ml of contrast is delivered at 4 to 5l/s. The stress phase scan is then acquired using the

reviously derived timing or the trigger bolus method.he stress phase scan can be acquired using a retro-

pectively gated scan with ECG tube current modu-ation. Tube voltage of 100 kV is used for patientsith body mass index of �30 kg/m2 and 120 kV forody mass index �30 kg/m2. Tube current can also be

adopted depending on the patient’s body habitus.Throughout the entire process, the patient’s symp-toms, heart rate, blood pressure, and ECG are closelymonitored. The pharmacological stress agent isstopped immediately after the stress phase acquisition(or reversed in case of dipyridamole).

If the rest phase acquisition follows the stressphase acquisition, there should be a period of delay(at least 20 min) in order for the effects of thepharmacologic stress to resolve. Resolution ofsymptoms and return of heart rate to baseline can beused as a proxy to judge when the effects haveresolved. For the rest acquisition, another 60 to 70ml of contrast is again delivered at 4 to 5 ml/s;the acquisition employs similar timing as the stressacquisition. Unlike the stress acquisition, the restacquisition is performed via prospective ECG

a False Negative

ith rest acquisition followed by stress acquisition 10 min later. (A)veals a stress acquisition that is contaminated by previous con-ect during stress. (C,D) Reveal a fixed, inferior wall perfusionugh SPECT did not reveal any ischemic myocardium in this par-emia can be missed on CTP if the stress phase acquisition is per-

to

ol w) Redeflthoisch

not interfere with the interpretation in this case.

frd0a4Tmtjss(arF(a

adrvclevsp(srtt

tiuipepdhum

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

910

triggering at mid-diastole to minimize radiationdose, but the tube voltage and current settings weresimilar to those used for the stress scan. This phasecan be used to assess both the coronary arteries andrest myocardial CTP. Beta-blockade with intrave-nous metoprolol can be administered, as in the caseof traditional CTA protocol, to optimize the visu-alization of coronary anatomy by decreasing the heartrate. Studies evaluating the feasibility of myocardialCTP have been conducted with and without the useof beta-blockers, all showing comparable results (17–19). The administration of nitroglycerin can also beused to achieve coronary vasodilation. However, itshould be noted that nitroglycerin is known to affectmyocardial perfusion, and thus may interfere with theaccuracy of the study (31).

A third scan can be added to assess for delayedenhancement (i.e., myocardial infarct) This acqui-sition is performed approximately 5 to 10 min afterthe completion of the second (either stress or rest)acquisition. No contrast injection is needed for thisphase, as it employs previously injected contrast inthe stress and rest phases. The acquisition is pro-spectively ECG-triggered at mid-diastole of thecardiac cycle, similar to the rest phase scan. Tubevoltage is set to 100 kV; tube current settings aresimilar to those used for the stress and rest scan.Finally, post-exam ECG and blood pressure mea-surements are performed to ensure patient’s safety.

The necessity and the value of the delayed phasescan as part of myocardial CTP has not been wellestablished at this point. Gerber et al. (29) illus-trated the similar pharmacokinetics between iodin-ated contrast and gadolinium in infarcted myocar-dium over a decade ago. Since then, it has beenshown in patients that CT delayed enhancementcan be detected in infarcted myocardium, but withrelatively poorer signal-to-noise and contrast-to-noise ratios (32). Blankstein et al. (18). found thatdelayed enhancement on CT correlated well withrest perfusion defect on SPECT as a part ofcomprehensive myocardial CTP protocol; however,this comparison was done only in subjects withfavorable image quality. There is also additionalradiation exposure to the patient, though this isrelatively low (approximately 1 mSv or less) whenprospective triggering and low tube voltage is used.Overall, there is not enough literature to definitivelysupport the incorporation of a delayed phase scan asa part of myocardial CTP protocol; however, fur-ther research on the use of delayed enhancement

CT for viability evaluation is needed.

Image processing and interpretation. Raw scan datarom the 2 or 3 phases (stress, rest, and delayed) areeconstructed into axial image datasets at mid-iastole, using a smooth filter with slice thickness of.7 mm and overlap of 0.4 mm. The axial imagesre then used to create short-axis, 2-chamber, and-chamber views by multiplanar reconstruction.hicker reconstruction (�8 to 10 mm) of averageultiplanar reformation is preferred for the detec-

ion of perfusion defect. Minimum-intensity pro-ections may be more sensitive in bringing outubtle perfusion defects; however, it is also moreusceptible to increased noise and false positives33). Maximum intensity projections are not useful,s they mask perfusion defects. Images are co-egistered and read in a side-by-side fashion (seeig. 3 for an example). Narrow window width�100 to 200 HU) should be used centered atttenuation of around 100 HU (window level).

There is a direct relationship between themount of iodinated contrast within the myocar-ium and myocardial attenuation (HU). As such, aelative difference in myocardial perfusion can beisualized as differing attenuations; areas of myo-ardium with decreased blood flow would have aower attenuation—a perfusion defect. When pres-nt, a perfusion defect can usually be detectedisually; confirmation with region of interest mea-urements can be helpful. The interpretation oferfusion defect while under stress and/or at resti.e., ischemia and/or infarct) in myocardial CTP isimilar to the interpretation of nuclear MPI, witheversibility serving as the key distinguishing fea-ure. A simple guide to myocardial CTP interpre-ation is shown in Figure 4.

For the interpretation of delayed phase acquisi-ion, the fundamental guiding principle is the sim-lar pharmacokinetics between iodinated contrastsed in CT and the gadolinium-based contrast usedn magnetic resonance. Time-dependent contrastatterns characterizing myocardial infarct have beenstablished with both contrast agents, with first-ass hypoenhancement due to decreased contrastelivery to the infarcted region and delayed en-ancement secondary to increased distribution vol-me and altered contrast pharmacokinetics in deadyocardial tissue (29,34).

Review of the Literature

The concept of CT examination for myocardialperfusion defect was first investigated in 1978

soon after the advent of the modality (35).

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

911

However, the reliable use of CT to detect myo-cardial perfusion defect has only been achievedwith the recent advancements in CT technology.The first attempt to use multidetector CT tocharacterize myocardial perfusion defects in hu-man under stress was first explored by Kurata etal. in 2005 (36), using adenosine stress on a16-detector multidetector computed tomography(MDCT) scanner. Although the study showedgood agreement between myocardial CTP and

Figure 3. Comparison of CTP, SPECT, and Invasive Coronary An

Shown is a 55-year-old man with history of hypertension and hypeexertion. SPECT myocardial perfusion imaging illustrated partially reels: top � stress, bottom � rest). CTP illustrates similar defects (mition, occlusion in the left anterior descending (bottom right) and sare seen. Abbreviations as in Figure 2.

Figure 4. Basic Guide to Interpretation of CTP

A perfusion defect under pharmacological stress that reverses at reis characteristic of myocardial infarction. Note that myocardial infar

late gadolinium enhancement in cardiac magnetic resonance. CTP � co

thallium-201 myocardial perfusion scintigraphy,only 48% of coronary segments were evaluableunder stress (compared with 89% at rest, p �0.05), illustrating that myocardial CTP was notfeasible at that time due to the limitations in thetemporal resolution. Since then, numerous stud-ies have been performed to evaluate the feasibilityof stress myocardial CTP in both animal modelsand humans with newer generation of scannersshowing improved results.

raphy

emia who presented with symptoms of substernal chest pain onible anterior, anteroseptal, and lateral perfusion defect (left pan-panels: top � stress, bottom � rest). Upon cardiac catheteriza-

ficant stenosis in the left circumflex (top right) coronary arteries

, by definition, stress-induced ischemia. A fixed (irreversible defect)also exhibits late contrast enhancement in a fashion similar to

giog

rlipidversddleigni

st isction

mputed tomography perfusion.

DSCT � dual source comp ardi

abbreviations as in Tables

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

912

The first animal studies illustrating feasibility ofmyocardial CTP using MDCT were performed byGeorge et al. (37,38) in canine models (Table 3).These studies established the semiquantitative rela-tionship between CT attenuation and blood flowbased on microsphere myocardial blood flow(MBF). Further work in a porcine model hasconfirmed that coronary flow reserve measurementsby microsphere MBF and myocardial CTP correlatewell with each other (39). In human studies, resultsfrom multiple cohorts employing different scannersand pharmacological stress agents (Table 4) seem toconfirm the initial findings of animal studies. Itshould be noted that these initial studies are con-ducted at various institutions with differences inprotocols and reference standards. The unifyingtheme, however, is that perfusion defects on myo-cardial CTP seem to correlate well with perfusion

Animal Studies Evaluating Myocardial Stress CTP

al Model CTP Protocol Reference Standard

e (n � 8)tenosis

Adenosine stress64-detector MDCT

Microsphere MBF Lineasigattatt

e (n � 6)stenosis

Adenosine stress64-detector MDCT

Microsphere MBF MDCTwip �

ne (n � 8) Adenosine stress64-detector MDCT

Microsphere MBF Signicobe(r

ne (n � 10)stenosisal control

Adenosine stress128-detector DSCTDynamic scan mode

— No siMBcosigad

uted tomography; LAD � left anterior descending coronary artery; MBF � myoc

Human Studies Evaluating Myocardial Stress CTP

ef. #) Scanner Type Comparison* Steno

) 64/256-MDCT CTA/CTP vs. QCA/SPECT �

(18) DSCT CTP and SPECT vs. QCA �

0 (20) DSCT CTA/CTP vs. QCA �

64-MDCT CTP and SPECT vs. QCA �

2nd-generation DSCT Dynamic CTP vs. SPECT/QCA

DSCT DECT vs. CMR/QCA

5

ns and reference standards vary widely. †Stenosis (%) refers to the stenosis threresonance; NPV � negative predictive value; PPV � positive predictive value; Q

1 and 3.

defects on SPECT, stenoses on quantitative coro-nary angiography (QCA), or both. For furtherdetails on reference standards used, refer to Table 4.George et al. (17) used adenosine as the pharma-cological stress agent, beta-blockers for heart ratecontrol, and 64-detector MDCT or 256-detectorMDCT for image acquisition. The combined anal-ysis of all patients (including both scanner types) inthis study showed a per-vessel territory sensitivity,specificity, positive predictive value (PPV), andnegative predictive value (NPV) of 75%, 87%, 60%,and 93%, respectively, when compared with QCAand SPECT. The potential limitations of this studyare nonuniformity of protocols and use of beta-blockers, which may have blunted the degree ofdifference in perfusion between the ischemic andthe nonischemic heart. Blankstein et al. (18) ac-quired myocardial CTP images under adenosine

Results Conclusion

ationship between myocardialdensity ratio (myocardialation/left ventricularation)

Adenosine stress MDCT providessemi-quantitative measurementsof myocardial perfusion incanine model of LAD stenosis

rived MBF strongly correlatedicrospheres (R � 0.91,0001)

MDCT MBF measurements usingupslope and model-baseddeconvolution methodscorrelate well withmicrosphere MBF

t correlation betweenry flow reserve measurementsen microsphere MBF and CT94, p � 0.0001)

CT first-pass myocardialperfusion imaging is feasibleusing a simple semi-quantitative analysis whichprovides reasonableestimates of MBF.

cant differences in CT-basedetween the stenotic andl group at rest; however,ant difference undersine stress (p � 0.0024).

DSCT permits quantitativewhole heart perfusionimaging, with ability to showhemodynamic effect of highgrade coronary arterystenosis.

al blood flow; other abbreviations as in Tables 1 and 2.

(%)† n mSv Sn (%) Sp (%) PPV (%) NPV (%)

27 16.8 (64)21.6 (256)

86 92 92 85

33 12.7 92 67 89 75

34 11.8 96 100 100 91

36 14.7 94 75 89 86

35 18.2 95 65 78 79

50 — 89 76 — —

215 11.8–21.6 86–96 67–100 78–100 79–91

used in studies which employed QCA as a part of the reference standard.� quantitative coronary angiography; Sn � sensitivity; Sp � specificity; other

Table 3. Summary of

First Author,Year (Ref. #) Anim

George et al.,2006 (37)

CaninLAD s

r relnalenuenu

George et al.,2007 (38)

CaninLAD

-deth m

0.

Christian etal.,2010 (39)

Porci ficanronatwe� 0.

Mahnken etal.,2010 (43)

PorciLADNorm

gnifiF bntronificeno

Table 4. Summary of

First Author, Year (R sis

George et al., 2009 (17 50

Blankstein et al., 2009 50

Rocha-Filho et al., 201 50

Cury et al., 2010 (19) 70

Ho et al., 2010 (44) —

Ko et al., 2011 (46) —

Total 0–70

*Note that the study desig sholdCMR � cardiac magnetic CA

FpsicSstFitiPbtcacf5beirls87

ndsmwtwdgetpD8

fnasDtoos

Cbawnawsmas8QNapHfsif

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

913

stress using a dual source CT scanner, which hashigher temporal resolution, obviating the need forbeta blockade. The comprehensive myocardial CTPprotocol also uniformly included a rest and a de-layed phase acquisition in all study subjects. Theresults showed that myocardial CTP is equivalentto SPECT in detecting coronary artery stenosis byQCA, with per-vessel territory sensitivity and spec-ificity of 79% and 80% for myocardial CTP and67% and 83% for SPECT, respectively (p � NS).

urthermore, the CTA obtained during the stresshase of the scan was excellent for diagnosingtenosis of 70% or more on QCA. The study alsollustrated that the full myocardial CTP protocolan be done at similar radiation exposure toPECT MPI (12.7 mSv for both). Subsequenttudies based on the same cohort further confirmedhe feasibility and value of myocardial CTP. Rocha-ilho et al. (20) demonstrated that adding perfusion

nformation obtained from stress myocardial CTPo coronary CTA improves all diagnostic character-stics of CTA alone, particularly specificity andPV. Okada et al. (40) illustrated the concordanceetween myocardial CTP and SPECT in the de-ection and evaluation of size and severity of myo-ardial perfusion defects at stress and at rest. Innother cohort of patients, Tamarappoo et al. (41)ompared the extent and severity of perfusion de-ects between myocardial CTP and SPECT using a-point scale and a total perfusion deficit score. Foroth methods, good agreement was achieved. Curyt al. (19) employed a different stress agent, dipyr-damole, on a 64-detector MDCT scanner. Theiresults also demonstrated that myocardial CTP is ateast equivalent to SPECT in the detection oftenosis found on QCA (sensitivity and specificity:8% and 79% for myocardial CTP and 69% and1% for SPECT, p � NS).Recently introduced dual energy mode and dy-

amic scan mode have also been investigated for theetection of perfusion defects (42–44). Dual energycan mode takes advantage of the fact that differentaterials have different spectral characteristicshen penetrated by different x-ray energy levels;

his allows for the mapping of iodine concentrationithin the myocardium—potentially reflecting theegree of myocardium perfusion (45). Initial investi-ation by Ruzsics et al. (42) demonstrated that dualnergy computed tomography (DECT) had a sensi-ivity and specificity of 96% and 95% in detecting fixederfusion defect seen on SPECT. Furthermore,ECT had a sensitivity and specificity of 88% and

9%, respectively, for the detection of reversible per-

usion defect (42), although the physiological mecha-ism behind such a concordance remains controversials DECT was not performed under stress in thetudy. Most recently, Ko et al. (46), in a study ofECT in the setting of adenosine stress, illustrated

hat DECT had a sensitivity, specificity, and accuracyf 89%, 78%, and 82%, respectively, for the detectionf myocardial segments with perfusion defect, usingtress perfusion MRI as the gold standard.

Preliminary experiences with dynamic myocardialT perfusion imaging also showed good promise inoth animal models and human patients. Mahnken etl. (43) illustrated the ability of dynamic quantitativehole-heart perfusion imaging to detect the hemody-amic effect of high-grade coronary artery stenosis incanine model. Furthermore, in an early experienceith a patient cohort, Ho et al. (44) demonstrated that

tress and rest dynamic perfusion imaging can detectyocardial perfusion defect with good diagnostic

ccuracy when compared with SPECT MPI (per-egment sensitivity, specificity, PPV, and NPV of3%, 78%, 79%, and 82%, respectively) and withCA (per-segment sensitivity, specificity, PPV, andPV of 95%, 65%, 78%, and 79%, respectively) and

llows for defining time-attenuation curves with theotential for quantification of myocardial blood flow.owever, it should be noted that the radiation dose

or this protocol (around 20 mSv) is much higher thantatic myocardial CTP. Furthermore, the use of qual-tative SPECT as a reference standard is suboptimalor such analysis (47).Study limitations. Myocardial CTP as a modalitystill has a number of limitations, some of whichwill eventually be overcome as CT technologyimproves. There are a few CT-related artifactsthat the physician should recognize and try tominimize when performing and interpreting myo-cardial CTP, namely beam-hardening and motionartifacts.

Beam hardening is a phenomenon that occurswhen x-ray beams pass through objects of highdensity, leading to a selective attenuation of lower-energy beams and increased mean energy of theremaining beams. The resulting appearance is ahypoenhanced region that may mimic areas of trueperfusion defect. There are a few characteristics thatphysicians can use to help identify beam-hardeningartifacts. The hypoenhanced region is usually trian-gular in shape, appears to originate from the regionof high attenuation next to it, and does not conformto vascular territories (37). A particularly commonlocation includes the basal inferolateral wall, due to

proximity to the descending aorta with iodinated

CK

r

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

914

contrast and dense vertebral bodies. Attempts todevelop an algorithm to minimize beam-hardeningartifacts are ongoing, with the use of iterativereconstruction.

Myocardial CTP is also prone to motion artifactssimilar to coronary CTA, particularly during thestress phase acquisition, due to the increased heartrate. Cardiac motion during the acquisition leads tohypoenhanced areas that can mimic true perfusiondefects. Motion artifacts are becoming less of anissue as the temporal resolution of CT continues toimprove. Moreover, by assessing multiple phases ofthe cardiac cycle, one might be able to differentiatea true perfusion defect from an artifact as theperfusion defect will persist and the artifact willdisappear or change in location. Regional wallmotion abnormalities can also be detected associ-ated with true perfusion defects. Table 5 listshelpful tips in distinguishing artifacts from trueperfusion defects.

Conclusions and Future Directions

The feasibility of stress myocardial CTP as a clinicalentity has thus far been evaluated in animal modelsand single-center prospective patient cohort studies

nguish Artifacts From True Perfusion Defects

Tips

s not correspond to a vascular territorys not correspond with wall motion abnormality in the sameegments not persist across different phases of the cardiac cycle

e as motion artifact aboveusion defect in early arterial phase due to the presence of highose of dense contrast in the ventricleusion defect in the basal inferolateral wall due to the proximity ofhe dense spine and contrast in the descending aortausion defect that is triangular in shape, originating from theegion of high attenuation in the proximity

prospective triggering with dual- assessment and in ev

64-MDCT and newer scanners have been promis-ing, showing that myocardial CTP has comparablediagnostic characteristics to SPECT MPI in thedetection of myocardial perfusion defects. Further-more, myocardial CTP protocol allows for thesimultaneous acquisition of coronary anatomy andmyocardial perfusion, and a combined CTA/CTPprotocol has been shown to have better diagnosticcharacteristics than CTA alone. The radiation ex-posure in such protocols is similar to that of atraditional SPECT MPI exam; the radiationexposure is also likely to decrease as CT technol-ogy advances (18). These data suggest that myo-cardial CTP has the potential to become a robustclinical tool for the evaluation of chest painpatients. However, it should be noted that themodality is in a very early stage of development.Published studies are composed only of single-center preliminary experiences. All evaluated co-horts thus far are based on a referral population ofpatients with intermediate to high pre-test prob-ability of CAD and patients referred to othertests such as SPECT MPI and cardiac catheter-ization. Furthermore, no standardized protocolcurrently exists for myocardial CTP studies as thefeasibility studies all utilized varying protocols.More research is needed in order to furtherdefine, optimize, and validate the modality. Par-ticularly, a prospective multicenter, multivendortrial evaluating the diagnostic characteristics ofmyocardial CTP in a wider patient population isneeded to confirm the findings of these promisingpreliminary trials in order to establish myocardialCTP as a viable clinical option.

Reprint requests and correspondence: Dr. Ricardo C.ury, Baptist Cardiac and Vascular Institute, 8900 Northendall Drive, Miami, Florida 33176. E-mail:

as listed in Tables 3 and 4. The initial results using rcury@baptisthealth.net.

R E F E R E N C E S

1. Ohnesorge B, Flohr T, Becker C, etal. Cardiac imaging by means of elec-trocardiographically gated multisec-tion spiral CT: initial experience. Ra-diology 2000;217:564–71.

2. Achenbach S, Ulzheimer S, Baum U,et al. Noninvasive coronary angiogra-phy by retrospectively ECG-gatedmultislice spiral CT. Circulation2000;102:2823–8.

3. Blankstein R, Shah A, Pale R, et al.Radiation dose and image quality of

source cardiac computed tomography.Am J Cardiol 2009;103:1168–73.

4. Cury RC, Nieman K, Shapiro MD, etal. Comprehensive assessment of myo-cardial perfusion defects, regional wallmotion, and left ventricular function byusing 64-section multidetector CT. Ra-diology 2008;248:466–75.

5. Greenland P, Bonow RO, BrundageBH, et al. ACCF/AHA 2007 clinicalexpert consensus document on coronaryartery calcium scoring by computed to-mography in global cardiovascular risk

aluation of patients

with chest pain: a report of the Ameri-can College of Cardiology FoundationClinical Expert Consensus Task Force(ACCF/AHA Writing Committee toUpdate the 2000 Expert ConsensusDocument on Electron Beam Com-puted Tomography). J Am Coll Cardiol2007;49:378–402.

6. Meijboom WB, Meijs MF, SchuijfJD, et al. Diagnostic accuracy of 64-slice computed tomography coronaryangiography: a prospective, multi-center, multivendor study. J Am Coll

Table 5. Tips to Disti

Artifacts

Motion DoeDoe

sDoe

Beam-hardening SamPerf

dPerf

tPerf

Cardiol 2008;52:2135–44.

1

1

1

2

2

2

2

2

2

2

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

915

7. Miller JM, Rochitte CE, Dewey M, etal. Diagnostic performance of coro-nary angiography by 64-row CT.N Engl J Med 2008;359:2324–36.

8. Budoff MJ, Dowe D, Jollis JG, et al.Diagnostic performance of 64-multidetector row coronary computedtomographic angiography for evalua-tion of coronary artery stenosis inindividuals without known coronaryartery disease: results from the pro-spective multicenter ACCURACY(Assessment by Coronary ComputedTomographic Angiography of Indi-viduals Undergoing Invasive CoronaryAngiography) trial. J Am Coll Cardiol2008;52:1724–32.

9. Shaw LJ, Berman DS, Maron DJ, et al.Optimal medical therapy with or with-out percutaneous coronary interventionto reduce ischemic burden: results fromthe Clinical Outcomes Utilizing Revas-cularization and Aggressive Drug Eval-uation (COURAGE) trial nuclear sub-study. Circulation 2008;117:1283–91.

10. Boden WE, O’Rourke RA, Teo KK,et al. Optimal medical therapy with orwithout PCI for stable coronary dis-ease. N Engl J Med 2007;356:1503–16.

11. Hachamovitch R, Hayes SW, Fried-man JD, Cohen I, Berman DS. Com-parison of the short-term survival ben-efit associated with revascularizationcompared with medical therapy in pa-tients with no prior coronary arterydisease undergoing stress myocardialperfusion single photon emissioncomputed tomography. Circulation2003;107:2900–7.

12. Meijboom WB, Van Mieghem CA,van Pelt N, et al. Comprehensive as-sessment of coronary artery stenoses:computed tomography coronary an-giography versus conventional coro-nary angiography and correlation withfractional flow reserve in patients withstable angina. J Am Coll Cardiol2008;52:636–43.

13. Pflederer T, Marwan M, Renz A, etal. Noninvasive assessment of coro-nary in-stent restenosis by dual-sourcecomputed tomography. Am J Cardiol2009;103:812–7.

14. Pugliese F, Weustink AC, VanMieghem C, et al. Dual source coro-nary computed tomography angiogra-phy for detecting in-stent restenosis.Heart 2008;94:848–54.

15. Gaemperli O, Husmann L, SchepisT, et al. Coronary CT angiographyand myocardial perfusion imaging todetect flow-limiting stenoses: a poten-tial gatekeeper for coronary revascu-larization? Eur Heart J 2009;30:2921–9.

16. Di Carli MF, Dorbala S, Curillova Z,et al. Relationship between CT coro-

nary angiography and stress perfusion

imaging in patients with suspectedischemic heart disease assessed by in-tegrated PET-CT imaging. J NuclCardiol 2007;14:799–809.

7. George RT, Arbab-Zadeh A, MillerJM, et al. Adenosine stress 64- and256-row detector computed tomogra-phy angiography and perfusion imag-ing: a pilot study evaluating the trans-mural extent of perfusion abnormalitiesto predict atherosclerosis causing myo-cardial ischemia. Circ Cardiovasc Imag-ing 2009;2:174–82.

8. Blankstein R, Shturman LD, RogersIS, et al. Adenosine-induced stressmyocardial perfusion imaging usingdual-source cardiac computed tomog-raphy. J Am Coll Cardiol 2009;54:1072–84.

9. Cury RC, Magalhaes TA, Borges AC,et al. Dipyridamole stress and restmyocardial perfusion by 64-detectorrow computed tomography in patientswith suspected coronary artery disease.Am J Cardiol 2010;106:310–5.

0. Rocha-Filho JA, Blankstein R, Shtur-man LD, et al. Incremental value ofadenosine-induced stress myocardialperfusion imaging with dual-sourceCT at cardiac CT angiography. Radi-ology 2010;254:410–9.

1. Varma SK, Watson DD, Beller GA.Quantitative comparison of thallium-201 scintigraphy after exercise anddipyridamole in coronary artery dis-ease. Am J Cardiol 1989;64:871–7.

2. Kong BA, Shaw L, Miller DD, Chai-tman BR. Comparison of accuracy fordetecting coronary artery disease andside-effect profile of dipyridamolethallium-201 myocardial perfusionimaging in women versus men. Am JCardiol 1992;70:168–73.

3. Nishimura S, Mahmarian JJ, BoyceTM, Verani MS. Equivalence be-tween adenosine and exercisethallium-201 myocardial tomography:a multicenter, prospective, crossovertrial. J Am Coll Cardiol 1992;20:265–75.

4. Santos-Ocampo CD, Herman SD,Travin MI, et al. Comparison of ex-ercise, dipyridamole, and adenosine byuse of technetium 99m sestamibi to-mographic imaging. J Nucl Cardiol1994;1:57–64.

5. Levine MG, Ahlberg AW, Mann A,et al. Comparison of exercise, dipyri-damole, adenosine, and dobutaminestress with the use of Tc-99m tetro-fosmin tomographic imaging. J NuclCardiol 1999;6:389–96.

6. David N, Marie PY, Angioi M, et al.Dipyridamole and exercise SPET pro-vide different estimates of myocardialischaemic areas: role of the severity ofcoronary stenoses and of the increase

in heart rate during exercise. EurJ Nucl Med 2000;27:788–99.

27. Chow BJ, Beanlands RS, Lee A, et al.Treadmill exercise produces largerperfusion defects than dipyridamolestress N-13 ammonia positron emis-sion tomography. J Am Coll Cardiol2006;47:411–6.

28. Knabb RM, Gidday JM, Ely SW,Rubio R, Berne RM. Effects of dipyr-idamole on myocardial adenosine andactive hyperemia. Am J Physiol 1984;247:H804–10.

29. Gerber BL, Belge B, Legros GJ, et al.Characterization of acute and chronicmyocardial infarcts by multidetectorcomputed tomography: comparisonwith contrast-enhanced magnetic res-onance. Circulation 2006;113:823–33.

30. Gibson PB, Demus D, Noto R, Hud-son W, Johnson LL. Low event ratefor stress-only perfusion imaging inpatients evaluated for chest pain. J AmColl Cardiol 2002;39:999–1004.

31. Zoghbi GJ, Dorfman TA, IskandrianAE. The effects of medications onmyocardial perfusion. J Am Coll Car-diol 2008;52:401–16.

32. Nieman K, Shapiro MD, Ferencik M,et al. Reperfused myocardial infarc-tion: contrast-enhanced 64-SectionCT in comparison to MR imaging.Radiology 2008;247:49–56.

33. Rogers IS, Cury RC, Blankstein R, etal. Comparison of postprocessingtechniques for the detection of perfu-sion defects by cardiac computed to-mography in patients presenting withacute ST-segment elevation myocar-dial infarction. J Cardiovasc ComputTomogr 2010;4:258–66.

34. Flacke SJ, Fischer SE, Lorenz CH.Measurement of the gadopentetatedimeglumine partition coefficient inhuman myocardium in vivo: normaldistribution and elevation in acute andchronic infarction. Radiology 2001;218:703–10.

35. Siemers PT, Higgins CB, SchmidtW, Ashburn W, Hagan P. Detec-tion, quantitation and contrast en-hancement of myocardial infarctionutilizing computerized axial tomog-raphy: comparison with histochemicalstaining and 99mTc-pyrophosphateimaging. Invest Radiol 1978;13:103–9.

36. Kurata A, Mochizuki T, Koyama Y, etal. Myocardial perfusion imaging us-ing adenosine triphosphate stressmulti-slice spiral computed tomogra-phy: alternative to stress myocardialperfusion scintigraphy. Circ J 2005;69:550–7.

37. George RT, Silva C, Cordeiro MA, etal. Multidetector computed tomogra-phy myocardial perfusion imagingduring adenosine stress. J Am Coll

Cardiol 2006;48:153–60.

4

4

4

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 . 4 , N O . 8 , 2 0 1 1

A U G U S T 2 0 1 1 : 9 0 5 – 1 6

Techasith and Cury

Stress Myocardial CTP

916

38. George RT, Jerosch-Herold M, SilvaC, et al. Quantification of myocardialperfusion using dynamic 64-detectorcomputed tomography. Invest Radiol2007;42:815–22.

39. Christian TF, Frankish ML, Sise-moore JH, et al. Myocardial perfusionimaging with first-pass computed to-mographic imaging: Measurement ofcoronary flow reserve in an animalmodel of regional hyperemia. J NuclCardiol 2010;17:625–30.

40. Okada DR, Ghoshhajra BB, Blank-stein R, et al. Direct comparison ofrest and adenosine stress myocardialperfusion CT with rest and stressSPECT. J Nucl Cardiol 2010;17:27–37.

41. Tamarappoo BK, Dey D, NakazatoR, et al. Comparison of the extent andseverity of myocardial perfusion de-

fects measured by CT coronary an-giography and SPECT myocardial

perfusion imaging. J Am Coll CardiolImg 2010;3:1010–9.

2. Ruzsics B, Schwarz F, Schoepf UJ, etal. Comparison of dual-energy com-puted tomography of the heart withsingle photon emission computed to-mography for assessment of coronaryartery stenosis and of the myocardialblood supply. Am J Cardiol 2009;104:318–26.

3. Mahnken AH, Klotz E, Pietsch H, etal. Quantitative whole heart stressperfusion CT imaging as noninvasiveassessment of hemodynamics in coro-nary artery stenosis: preliminary ani-mal experience. Invest Radiol 2010;45:298–305.

4. Ho KT, Chua KC, Klotz E, PankninC. Stress and rest dynamic myocardialperfusion imaging by evaluation ofcomplete time-attenuation curves

with dual-source CT. J Am Coll Car-diol Img 2010;3:811–20.

45. Johnson TR, Krauss B, Sedlmair M,et al. Material differentiation by dualenergy CT: initial experience. Eur Ra-diol 2007;17:1510–7.

46. Ko SM, Choi JW, Song MG, et al.Myocardial perfusion imaging usingadenosine-induced stress dual-energycomputed tomography of the heart:comparison with cardiac magnetic res-onance imaging and conventional cor-onary angiography. Eur Radiol 2011;21:26–35.

47. Blankstein R, Jerosch-Herold M.Stress myocardial perfusion imagingby computed tomography a dynamicroad is ahead. J Am Coll Cardiol Img2010;3:821–3.

Key Words: cardiac CT ymyocardial perfusion y

pharmacologic stress.