Contents lists available at ScienceDirect

Journal of Cardiovascular Computed Tomography

journal homepage: www.elsevier.com/locate/jcct

Society of cardiovascular computed tomography expert consensus documenton myocardial computed tomography perfusion imagingAmit R. Patela,∗, Fabian Bambergc, Kelley Branchd, Patricia Carrascosae, Marcus Chenf,Ricardo C. Curyg, Brian Ghoshhajrah, Brian Koi, Koen Niemanj, Francesca Pugliesek,Joseph Schoepfl, Ron Blanksteinb,∗∗

aUniversity of Chicago, USAb Brigham and Women's Hospital, USAcUniversity of Tuebingen, GermanydUniversity of Washington, USAe Diagnostico Maipu, ArgentinafNational Institutes of Health/NHLBI, USAg Radiology Associates of South Florida, USAhMassachusetts General Hospital, USAiMonash Heart, Australiaj Stanford University, USAkQueen Mary University of London, UKlMedical University of South Carolina, USA

1. Introduction

Advances in X-ray computed tomography (CT) over the past 20years have enabled non-invasive imaging of the coronary arteries, aswell as cardiac structure and function. Coronary computed tomographyangiography (CTA) has emerged as an accurate non-invasive imagingmodality for identifying the anatomic severity of coronary stenoses.1–3

Similar to other anatomic coronary imaging, coronary CTA alone hasbeen limited in its ability to define the hemodynamic significance ofcoronary stenoses.4 Since the identification of the presence and severityof ischemia are important parameters for selecting between invasiveand medical management of coronary artery disease,5–7 the assessmentof both coronary anatomy and myocardial perfusion is needed in someclinical scenarios.

The development of X-ray CT-based myocardial perfusion imagingto determine myocardial ischemia began in the late 1980s. Early in-vestigators used electron beam computed tomography and a fast mul-tislice CT scanning platform to perform quantitative myocardial per-fusion imaging with good results.8,9 However, the introduction ofmultidetector CT and the proven feasibility of coronary CTA over 10years ago10–13 also allowed for development and refinement of myo-cardial CT perfusion (CTP) imaging.4,14–16 In single center studies,myocardial CTP imaging has demonstrated high accuracy when com-pared with single photon emission computed tomography (SPECT),cardiovascular magnetic resonance (CMR), invasive coronary

angiography (ICA), positron emission tomography (PET), and invasivefractional flow reserve (FFR).4,14,17–25 Multicenter studies have alsoestablished the accuracy of myocardial CTP with coronary CTA.26,27

These studies suggest that CTP is particularly accurate when interpretedin the context of coronary CTA findings.

Currently, there is large body of evidence that establishes myo-cardial CTP diagnostic accuracy and incremental value over coronaryCTA.28 However, an important limitation to myocardial CTP im-plementation is the heterogeneity within the literature of pharmaco-logic stress agents, imaging sequences, scanner types, acquisition pro-tocols, post-processing, and interpretation of CTP results. Clinicaladoption of myocardial CTP is further hindered by the absence of expertconsensus regarding when and how CTP should be performed. The goalof this document is to bring together international experts in myo-cardial CTP imaging to provide a consensus document that can be usedas a resource for clinical implementation, resource allocation, and fu-ture investigations.

2. General overview of myocardial CTP, patient selection, andlogistics

Myocardial CTP imaging is the contrast enhanced imaging ofmyocardial perfusion using qualitative, semi-quantitative, or quantita-tive methods. Myocardial CTP imaging may be performed at rest andduring pharmacologically induced stress. An additional acquisition can

https://doi.org/10.1016/j.jcct.2019.10.003Received 4 October 2019; Accepted 15 October 2019

∗ Corresponding author.∗∗ Corresponding author.E-mail addresses: amitpatel@uchicago.edu (A.R. Patel), RBLANKSTEIN@bwh.harvard.edu (R. Blankstein).

Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

Available online 11 December 20191934-5925/ © 2020 Published by Elsevier Inc. on behalf of Society of Cardiovascular Computed Tomography

T

be performed 5–10 minutes after vascular contrast washout to detectmyocardial scar. The two primary methods of acquisition are static CTPand dynamic CTP. Static CTP, which has been more extensively studied,involves acquisition of a single image during an infusion of iodinatedcontrast to allow qualitative and/or semi-quantitative CTP assessments.Dual energy CTP is another approach to performing static CTP, which isless widely used, but utilizes different energy photons to better visualizeiodine contrast within the myocardium. Dynamic CTP imaging involvesacquisition of multiple data sets during the first pass of iodinatedcontrast from the coronary arteries into the myocardium. Dynamic CTPallows for a qualitative, semi-quantitative, as well as fully quantitativeassessment of myocardial perfusion.

Regardless of CTP acquisition mode, patient selection is of utmostimportance. Coronary CTA alone has a very high negative predictivevalue to exclude myocardial ischemia in the presence of no CAD or anon-obstructive stenosis (≤50% severity).4,29–33 Therefore, selection ofa myocardial CTP protocol should generally be reserved for situationswhere the determination of the presence or absence of ischemia can behelpful such as in the presence of coronary artery stenoses of unknownhemodynamic significance, severe coronary calcification, or coronarystents.34,35 In the setting of an uninterpretable coronary segment fromany cause, myocardial CTP can improve diagnostic accuracy for bothstenosis and ischemia.36,37 While it is conceivable that patients whohave a higher risk of obstructive CAD are more likely to benefit fromCTP, models to calculate the probability of obstructive CAD are notwidely used and such models often over-estimate the probability ofobstructive disease.38,116

If CTP is used to evaluate for myocardial ischemia, pharmacologicstress is required. Resting myocardial blood flow is only decreased bythe most severe stenosis, and can be difficult to differentiate frommyocardial infarction, making the use of rest CTP alone non-specific forischemia.39,40

Although stress CTP is safe, there are some circumstances where itshould avoided altogether or only used with caution. Absolute and re-lative CTP contraindications are listed in Table 1.

Summary: We recommend that myocardial CTP may be added tocoronary CTA when patients are at high atherosclerotic risk for ob-structive coronary artery disease, including those with prior coronaryintervention or significant calcification, or when there is a stenosis ofindeterminate functional significance. In such cases, myocardial CTP isrecommended if knowledge regarding the presence and severity ofischemia would impact patient management and CTP is feasible.

2.1. Stress CTP procedure logistics

Myocardial CTP imaging with pharmacologic stress testing requiressupervision by a trained medical provider who is experienced in theadministration of pharmacologic stress agents and stress testing. Theseclinicians should also be trained in dealing with emergencies related tocontrast injection and vasodilator stress testing. Proper resuscitativeequipment such as a code cart, defibrillator with transcutaneous pacingcapabilities, aminophylline as a reversal agent, bronchodilator therapy,beta-blockers, sublingual nitroglycerin, and epinephrine and di-phenhydramine should be available in the immediate vicinity.

Prior to starting the study, a focused patient history and physicalexam, review or performance of a 12-lead ECG, and informed consentare required to determine suitability of stress testing, similar to otherstress testing guidelines (39,41,42). Appropriate intravenous accessshould then be obtained, typically a 20 gauge or larger IV that canhandle the high flows required for contrast injection. During the ad-ministration of the pharmacologic stress agent, continuous heart rate/rhythm monitoring and intermittent blood pressure monitoring arerequired. Continuous 12-lead ECG monitoring may be performed usingradiolucent electrodes and leads. Alternatively, 12-lead ECG mon-itoring can be interrupted during scan acquisition while continuousrhythm is still being recorded and monitored by the scanner console.

Upon the completion of imaging, a 12 lead ECG should be performed toensure there are no significant changes from baseline suggestive of post-stress myocardial ischemia or injury.

2.2. The use of pharmacologic stress agents, aminophylline andnitroglycerin

Myocardial CTP stress testing uses only vasodilator stress agents atpresent. Exercise and dobutamine are not practical options for CTP dueto the elevated heart rates that would preclude diagnostic imaging withmost currently available CT platforms. While there are no agents ap-proved for myocardial CTP, adenosine, dipyridamole, and regadenosonare FDA approved for radionuclide myocardial perfusion imaging.Therefore, their use in CTP imaging is off-label. These agents, theiractions and adverse reactions have been extensively reviewed else-where.41 Specific considerations for vasodilator stress CTP include theneed for two IV's for both continuous infusion of adenosine and contrastinjection, while regadenoson and dipyridamole require only one IV foralternating injections. A potential disadvantage of regadenoson is thehigher heart rate compared with other vasodilators, which may lead tomore motion artifacts during CTP acquisition. However, regadenosontends to be better tolerated and has been tested in single center CTPstudies43–45 and a randomized diagnostic accuracy study that comparedCTP and SPECT myocardial perfusion imaging.26 Overall, regadenosonwas similarly tolerated in regards to adverse events between CTP andSPECT imaging.

Nitroglycerin is commonly used prior to coronary CTA to dilate thecoronary arteries and improve coronary artery visualization. Since it isalso a vasodilator, simultaneous administration of nitroglycerin and avasodilator stress agent can lead to hypotension. A 10–20 minute delaybetween the administration of the two agents is recommended to avoidhypotension. Adequate preload will also decrease the risk of hypoten-sion. The mean elimination half-life of nitroglycerin is 2–3 minutes.

2.3. CT contrast agents

In general, during myocardial CTP, images are acquired during thetransit of contrast from the coronary arteries to the myocardium.Because iodinated contrast attenuates x-rays proportionally to theconcentration of iodine, hypoattenuated myocardial areas typicallyrepresent regions with low blood flow, or hypoperfusion. Typically,50–70ml of iodinated contrast is injected at a rate of 5–6 ml/sec. Asaline bolus (40–50ml) following the contrast bolus with the same in-jection rate is recommended to assist with contrast delivery to themyocardium. Patients should be warned and educated about the ‘hotflash’ sensation caused by contrast injection. This sensation can beuncomfortable and cause breathing or patient movement that in turnleads to motion artifacts and deterioration of image quality.

3. CTP imaging strategies overview

Myocardial CT perfusion protocols typically include a rest and astress CT scan. The rest portion of the scan allows for assessment of thecoronary anatomy in addition to rest myocardial perfusion, thus restimages should be acquired according to coronary CTA performance andacquisition guidelines46 and include the use of beta-blockers and ni-trates. However, the optimal sequence of scans (rest first versus stressfirst) has not been determined and either sequence has advantages anddisadvantages. When stress CT imaging is performed first, the myo-cardium does not have residual contrast in the myocardium from anearlier contrast injection potentially making it easier to detect ischemia.In addition, beta-blocker administration may potentially lead to someunderestimation of the ischemic burden, and a stress-first acquisitionwill allow for delay of beta-blocker administration until after the stressacquisition has been performed. Alternatively, if rest coronary CT an-giography acquisition is evaluated first (e.g. prior to stress), patients

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

88

without an obstructive or questionable coronary stenosis would avoidstress CT perfusion. However, areas of myocardial infarction may havedelayed contrast enhancement which can potentially be misinterpretedas “normally perfused” myocardium on the stress CTP images. Neithersequence of testing has been shown to be superior. Regardless, delaying10–20 minutes between the rest and stress CT acquisitions is re-commended to minimize myocardial contrast agent contamination andto eliminate the effects of pharmacologic stress or nitroglycerin ad-ministration.

Summary: Our committee suggests assessment of factors that wouldimpair exclusion of obstructive CAD on coronary CTA (e.g. significantcoronary calcium or known CAD) prior to decision to perform CTP.Those patients with known significant coronary disease or severeamount of CAC may undergo stress myocardial perfusion imaging first;whereas, patients who lack the above features may undergo coronaryCTA first (Fig. 1). The routine use of coronary calcium scoring prior tostress or resting imaging has been tested and compared against standardof care in a prospective randomized trial and may be a practical, cost-effective approach to design a patient-specific imaging protocol.47 Theoptimal order of a comprehensive CT perfusion study has not beendetermined and is a topic of debate.

4. CTP image acquisition

CTP image acquisition is most commonly performed using eitherstatic, dynamic, or dual energy approach. Discussion of each of theacquisition modalities is described in detail below and summarized inTable 2.

4.1. Static CTP image acquisition

Static CTP refers to imaging all or portions of the left ventricularmyocardium during a single period in time as the contrast transits themyocardium. This method uses retrospective ECG-gating (most often

with 64 detector scanners) or prospective ECG-triggering to ensureimaging during specific cardiac phases. Static CTP provides a qualita-tive or semi-quantitative assessment of myocardial perfusion.14,18,48,117

Conveniently, every coronary CTA is a resting static CTP study. Severalstudies have demonstrated that a resting CTP has a high sensitivity andspecificity for the detection of prior myocardial infarction ranging from77 to 91% and 79–97%, respectively compared with the referencestandard MRI.49–53

Static stress myocardial CTP imaging has been well validated for thedetection of coronary stenosis and myocardial ischemia compared tovarious reference standards including pre-clinical validation, invasivecoronary angiography (ICA), single photon emission computed tomo-graphy (SPECT), magnetic resonance perfusion, and fractional flowreserve.4,14,15,19,20,117 Advantages of static CTP are its relative simpli-city of image acquisition, reduced radiation dose, and ease of imageinterpretation. Disadvantages include the potential of imaging myo-cardial contrast off peak enhancement, imaging different myocardialregions at different time points during CT table movement, presence ofimaging artifacts, and limitations of only qualitative or semi-quantita-tive ischemia assessment.

4.1.1. ECG-gating strategyTo date, there is no definitive data available to provide guidance on

whether prospective or retrospective ECG gating is preferred for allpatients. The consensus of this committee is that retrospective ECGgating is favored when smaller z-axis detectors (<8 cm) are utilized toallow for whole heart coverage in order to reduce artifacts at thetransition zones. Alternatively, prospective ECG triggering is favoredfor large Z-axis coverage scanners (≥8 cm) and if the phase acquisitionwindow can be widened to allow for the acquisition of additional car-diac phases (i.e. at least 20% of the RR interval). There is currently nodefinitive data to recommend the optimal portion of the cardiac cycle toacquire images. CTA is generally imaged at 65–75% of the R to R in-terval relative cardiac motion quiescence, or “diastasis”. As the heart

Table 1Absolute and relative contraindications to pharmacologic stress myocardial CT perfusion.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

89

rate increases, such as during the administration of a pharmacologicalstressor, diastasis becomes shorter and occurs later in the cardiac cycle.In heart rates over 80 beats per minute, the phase of relative cardiacquiescence may occur during end-systole (approximately 30–50% of theR to R interval). The systolic phase may be preferred at higher heartrates due to a shorter and more stable duration of about 200 msec and itmay be less affected by extrasystolic events. In addition, the in-traventricular amount of contrast is lower compared with the diastolicphase, which may reduce beam-hardening artifacts and fluctuations inmyocardial CTP measurements.54,55

There is no consensus about whether imaging should be performedat end-systole or during diastole although most larger trials to date haveselected diastole. When performing stress CTP, a multiphase acquisitionwhich includes phases from both systole and diastole can be consideredso that the phase with the least amount of artifact can be selected foranalysis retrospectively.

4.1.2. Contrast timingOptimal CTP contrast timing is determined manually by contrast

timing bolus or using an automated bolus-tracking algorithm.56 If anautomated bolus-tracking algorithm is utilized, a region of interestshould be defined in either the proximal-ascending or mid-descendingthoracic aorta. The automated trigger should be set to a Hounsfield Unit(HU) threshold that is ∼80-100HU above the baseline HU of the re-ference area immediately prior to the contrast injection. After thescanner is triggered, a short pause (∼5 seconds) is required to allow forbreath-hold instructions to be given. Contrast dose and flow rate is ty-pically ∼60–70ml injected at 5–6ml/sec. Higher doses of contrast may

be required if a single source 64 slice CT scanner is being used. Imaging isperformed during suspended breathing after peak hyperemia is achieved.

4.1.3. Static CTP minimum scanner requirementsStatic CTP imaging requires a CT system capable of acquiring high

spatial resolution images of the entire heart within a short time intervalwhile patients are holding their breath. This includes virtually allcontemporary 64 slice or greater CT scanners. The temporal resolutionof a CT scan is determined by the time required to acquire the data forreconstruction of a single transverse section or “slice”. The speed ofgantry rotation is one of the primary determinants of the temporal re-solution of the CT scan. In CTP, high temporal resolution minimizesmotion artifacts related to a fast heart rate or cardiac contractility thattypically occurs during the administration of a pharmacologicalstressor. The minimum gantry rotation time on current generationscanners (the time required to complete a 360° rotation) is between 250and 400ms, depending on the manufacturer and model. Software andhardware improvements have further increased temporal resolution.The routine use of half-scan reconstruction results in an effective tem-poral resolution of approximately one half the time required for thetypical 360° CT rotation, or a reduction to approximately 125–200msper rotation. Dual-source CT scanners equipped with 2 sets of x-raysources and detectors offset by ∼90° from each other achieve temporalresolution of approximately 65ms. Disadvantages of partial- or half-scan reconstructions are decreases in the stability of Hounsfield unitswithin the images due to geometrical and misregistration artifacts.Hybrid reconstruction protocols utilizing a combination of half-scandata for image detail and full-scan (360°) data for contrast resolution

Fig. 1. Comparison of CTP acquisition protocols.

Table 2Comparison of CT perfusion imaging modalities.

Imaging Modality Description Advantages Disadvantages

Static CT perfusion Single image taken at peak myocardialperfusion to visualize hypoattenutation(hypoperfusion)

• Technically simple

• Can use CCTA images for rest CTP

• Low radiation exposure

• Qualitative to semiquantitative analyses only

• Timing of CTP acquisition critical

• Need normal segment to identify abnormal perfusion

• CTP artifacts, such as beam hardening, commonDynamic CT perfusion Repeated images over time to create

time attenuation curves (TAC's)• Semiquantitative myocardial blood flow (MBF)

and volume (MBV) assessment possible

• Modeled, fully quantitative myocardial bloodflow (ml/min/g) assessment possible

• Potential to detect balanced and microvascularischemia

• Increased scanner hardware requirements

• Lower detector systems may require shuttle mode

• CTP artifacts common

• Higher radiation exposure when compared to static CTP

• Additional software requirements

Dual Energy CTperfusion

Separate energy X-ray beams used todistinguish myocardium, contrastattenuation, and other structures

• Higher contrast differentiation than other CTPmodalities

• Beam hardening correction possible

• Additional quantification approaches possible

• Specific scanner hardware requirements

• Additional software requirements

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

90

may provide an advantage.54 The high, nearly isotropic, spatial re-solution with most modern scanner are less important for visualizingperfusion defects in the myocardium as discussed later in the CTP re-construction methods.

4.2. Dynamic CTP image acquisition

Dynamic CT perfusion imaging refers to the imaging of myocardialcontrast enhancement by taking multiple sequential CTP images overtime as the contrast transits through the heart. The myocardial contrastenhancement images can be analyzed visually or their enhancement overtime can be quantified in order to create time-attenuation curves (TAC).These TAC's are proportional to myocardial blood flow and can beevaluated in various ways as explained below. Advantages of dynamicCTP include the ability to quantify myocardial blood flow (MBF), myo-cardial blood volume, and potentially other hemodynamic parameters byapplying mathematic models to vascular contrast enhancement and thetabulated myocardial TAC's (Table 3). The parameters used and howaccurately they can be determined depends strongly on the temporalsampling rate. The disadvantages of dynamic CTP protocols are the ra-diation exposure is higher than that of static CTP, cardiac and patientmotion, and lack of standardized models to quantify blood flow. To re-duce radiation dose, automatic tube current modulation can be usefulwith spiral dynamic perfusion CT57 and iterative reconstruction techni-ques to maintain image quality with lower radiation exposure (58) canbe used. Myocardial misregistration from patient and cardiac motionduring the relatively long dynamic CTP image acquisition can be reducedusing motion correction algorithms.55

4.2.1. Dynamic CTP minimal scanner requirementsDynamic perfusion imaging requires a system capable of acquiring

repeated moderate spatial resolution images that cover the entiremyocardium (which in adults is about 7–10cm) within the 20–30second breath hold required for contrast transit. High temporal re-solution is required to minimize motion artifacts related to a fast heartrate during pharmacologic stress or patient breath hold. Sufficienttemporal sampling is required for TAC generation and quantification orsemi-quantification of myocardial perfusion. Second and third genera-tion dual-source computed tomography (CT) systems and later

generation single-source CT systems fulfill these requirements. Thenumber of longitudinal detector rows/data channels that can simulta-neously measure x-ray attenuation determines the volumetric z-axiscoverage of the CT scanner. Both 64-detector CT and first-generationdual-source CT have limited z-axis coverage per gantry rotation (∼4cm), if the table remains in the same position.59 For these scanners, theCT table and patient need to move back and forth through time (shuttlemode) to cover the entire myocardium. In second- and third-generationdual-source CT and z-axis coverage of 38–40 mm, ECG-triggered se-quential-scanning with table movement back and forth between the twoscanning positions provides a final coverage of 7.3–8cm. This coverageis limited but large enough to collect information about myocardialperfusion in the end-systolic phase, when the size of the heart is smallerand the thickness of the myocardial wall is higher (compared withdiastolic phase).54

Using higher slice systems with larger z-axis coverage (detectorwidth of 12–16 cm) provides whole heart coverage per gantry rotation.However, the lower temporal resolution of these systems increases therisk of motion artifacts with higher heart rates. This may be mitigated,to some extent, by the use of multisector reconstruction.

4.2.2. Timing of the scan and frequency of temporal samplingDynamic perfusion relies on accurately measuring arrival, peak and

washout of the first pass of myocardial contrast and aorta or left ven-tricle over time. CTP images should be acquired during the early por-tion of first-pass circulation, when the iodinated contrast is mainly in-travascular60 Images are then acquired repeatedly at different timepoints yielding the various temporal frames of the dynamic dataset,generally 1–3 seconds apart. Each frame of the dynamic dataset gen-erates a point of the TAC in each voxel of the myocardium, as well as apoint of the arterial input function in the aorta. Other crucial timepoints include at least one baseline (pre-contrast) frame including thewhole myocardium, and adequate frames to define the upslope, peakand early portion of the downslope of the time-attenuation curves.

4.2.3. Dual-source CTDual source CT with detector coverage that cannot cover the entire

myocardium requires “shuttling” the patient back and forth to imagethe myocardium over time. With the shuttle mode technique, CT data

Table 3Approaches for assessment and quantification of myocardial perfusion.

CTP Measure Description Advantages Disadvantages

Static CTP

Qualitative Visual analysis of perfusion defects (fixed vsreversible, transmural vs endocardial)

• Simple

• Rapid analysis• May be more susceptible to artifact and

interpretation errorsTransmyocardial perfusion ratio (TPR) Ratio of endocardial to epicardial attenuation. • Simple adjunct

• Semiquantitative• Dependent on good image quality

• Susceptible to artifact

• Unclear if added value over qualitative assessment

Dynamic CTP

Qualitative Visual analysis of transit of contrast over time • Simple

• Rapid analysis• May be susceptible to artifact and interpretation

errorsMyocardial Perfusion Reserve Index Ratio of stress-over-rest myocardial peak

attenuation or upslope of myocardial attenuationcurves normalized to the corresponding arterialTAC parameter.

• Simple calculation

• Shorter imagingtime

• Semiquantitative

• Requires dynamic CT imaging

• Questionable accuracy

• Imaging of peak attenuation criticalQuantitative myocardial blood flow

(MBF) or myocardial blood volume(MBV)

Modeled calculation of absolute MBF or MBV fromarterial and myocardial time attenuation curves

• Fully quantifiableflow (ml/min/gram)

• Higherquantificationaccuracy

• Longer scan and more susceptibility to artifacts

• Requires dynamic CT imaging

• Errors within model are possible

• Higher radiation than other CTP

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

91

are acquired at two alternate table positions by moving the table backand forth covering the entire myocardium in the end-systolic phase.54

Each stack contains attenuation data of the blood pool in the des-cending aorta or left ventricular cavity as well as attenuation data of themyocardium. Second and third generation dual-source technology al-lows CTP acquisition at every heart beat for up to 63 beats per minute,and every second heart beat above 63 beats per minute, in keeping witha temporal sampling of 2–3 s under pharmacological stress condi-tions.54,61,62 These are generally sufficient to allow the estimation andfit of the TAC while limiting radiation exposure to patients.

This allows construction of the arterial input function with doubletemporal sampling compared to the myocardium, and provides suffi-cient time points for the construction and fitting of the TAC's for both inleft ventricle or aorta (arterial input function) and in each voxel of themyocardium (tissue). The scan typically acquires 10–15 whole myo-cardial volumes corresponding to 10–15 time points (or frames) over20–30s.

4.2.4. Single source CTSingle source CT with smaller (<8 cm) coverage also requires

“shuttling” the patient back and forth over time, similar to dual sourceCT, but may not have the temporal resolution to obtain all needed datapoints for TAC generations. Larger coverage scanners (>8 cm) imagethe whole heart every rotation without table motion.63 This is achievedwith the 320-detector CT scanner that can provide up to 16 cm ofcoverage along the z-axis. This allows image acquisition of the heartwithin a fraction of a single heartbeat, with more homogeneous con-trast attenuation of the myocardium and no reconstruction slab arti-facts. Single source scanner generally require a minimum number ofscans (∼10–15) to generate robust TAC although more recent data arechallenging this requirement.64 Contemporary acquisition protocolscan also vary the sampling rate during the passage of contrast, to obtainmore data points during the important upslope portion of the TAC.

4.2.4.1. Image artifacts for static and dynamic CTP. Beam hardening is awell-known artifact which occurs when x-ray beams pass through high-density objects, in which lower energy beams are absorbed, leading to ahypoenhanced region that may appear as a falsely-positive perfusiondefect. This commonly occurs in the basal inferolateral wall, which isadjacent to the contrast-enhanced descending aorta. There are anumber of reconstruction algorithms that have been described in aneffort to correct for beam hardening, but it remains a problematicartifact in CT-MPI imaging.65–67

The visual interpretation of perfusion defects in CT-MPI can belimited by suboptimal image quality and relatively poor signal-to-noiseratio due to patient body habitus and the desire to minimize radiationexposure to the patient. Iterative reconstruction algorithms have beendeveloped in an effort to reduce image noise in CT. These algorithmsprovide improved signal-to-noise and contrast-to-noise ratios withoutrequiring an increase in radiation. This new reconstruction strategy isespecially promising for CTP imaging, as it offers an effective way toacquire additional images with only minimal radiation exposure,without compromising image quality and diagnostic accuracy58,68

4.3. Dual energy CTP

Overview. Dual energy CT (DECT) imaging produces a spectrum ofphoton strengths (kVp) during CT acquisition that, when measured withmodern CT scanners, are processed into different imaging spectra.DECT has the ability to differentiate iodinated contrast from back-ground tissue, water and fat by utilizing the reactions of each of thesesubstances to photons at the so called “k-edge”. DECT has the ad-vantage of overcoming some challenges of interpreting CTP images

such as beam hardening artifacts and the relatively low contrast tonoise ratio needed to keep radiation exposure judicious over severalheartbeats. DECT scanning techniques vary according to the type ofscanner whether it has one or two x-ray tube sources.

Dual Source DECT. In the dual source scanners, two tubes are placedperpendicular in the same gantry with an angle offset of ∼90°. One ofthe tubes operates at 140 kV and the other at 80–100 kV. A limitation ofthis approach is that when both sources are active at the same time,cross-scatter can result in artifacts and degraded contrast-to-noise ratio.The best achievable temporal resolution in this configuration is 165milliseconds; however, reconstruction algorithms have been describedthat may further improve this.69 DECT perfusion studies with dualsource are acquired with retrospective ECG-gating and ECG-dependenttube current modulation.

Single Source DECT. There are currently two approaches for dualenergy CT using only one x-ray tube source: the rapid kV switchingduring rotation and the dual layer detector approach. With the firstapproach, rapid KV switching occurs within a single tube between lowand high tube potentials every 0.5 milliseconds and are generally 80and 140 keV. Each pair of 80 and 140 kV projections are basicallyobtained from an identical view angle. DECT using this approach onlyoperates in a prospective ECG-triggered mode and a wider padding isrequired for imaging both systolic and diastolic phases. With the secondapproach, a dual layer detector system utilizes two different scintil-lating materials fused together with a single source CT gantry. Theupper layer detects the lower energy photons and permits higher energyX-ray photons to pass through it to be detected by the lower layer. Thetwo signals for the two detectors would resemble two x-rays in twodifferent energy ranges. Advantages of this design is that at each viewthe high- and low-energy projections are exactly registered with respectto each other. A potential disadvantage is less spectral separation be-tween low and high energy projections that can lead to an inadequatematerial decomposition and beam hardening correction.

4.3.1. Image reconstruction and analysis using DECTDECT allows for two types of analyses, either monochromatic ex-

amination or material decomposition, to reduce beam hardening andother artifacts. The monochromatic examination renders each image ata monochromatic energy level over a wide spectrum of energies thatranges between 40 keV and 140 keV. This means that for each image,there are several possible energy levels for discrimination of chemicalcomposition by mapping the differing interactions with x-ray photons.Those with high atomic number has inverse conspicuity, or HU density,to photon energy with DECT. High atomic numbers result in higherdensity (HU) at lower energy levels (for example, 40 keV) and lowerdensity (HU) at higher energy levels (140 keV). Iodinated contrast has ahigh atomic number and thus has a higher density at low energy levelsand a relatively lower conspicuity at high energy level reconstructions.

While low energy levels depict higher contrast enhancement in themyocardium, they are associated with increased noise, and are subjectto increased beam hardening artifacts that can mimic perfusion ab-normalties and increase false positive readings unless dedicated algo-rithms are used.70 Conversely, higher energy levels yield lower myo-cardial and intravascular contrast but lower image noise, and decreasedbeam hardening artifacts. Thus, DECT may reduce or even eliminatebeam hardening by generating monochromatic images. Once the op-timal energy level is determined, grayscale, color mapped, or color-overlaid images can be utilized for analysis. Semi-quantitative andquantitative analysis can be complemented in a similar manner tosingle energy CT as described below.

The second type of DECT evaluation is material decomposition(MD).71 MD is based upon different attenuation coefficients of varioustissues, which in turn depend on the energy levels of the X-ray beam.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

92

MD yields information about tissue atomic number and can provide“mass density maps” for basic materials such as water or iodine. MDallows the visual isolation of a material as well measurement of itsconcentration in a given voxel (mg/Ugr/mm3).56,58,72 In myocardialCTP, the amount of iodine can be measured in order to establish normalor an abnormal perfusion. MD can be portrayed in grayscale or in colorscale.

5. CTP image analysis Overview

We recommend interpretation of coronary CTA and myocardial CTPimages be performed concomitantly. Accordingly, workstations used forCTP interpretation should possess all requirements for standard coronaryCTA interpretation. This constitutes the ability to facilitate two andthree-dimensional display of the coronary arteries and the myocardiumsubtended in all conventional reconstruction formats. These include axialand multiplanar reformations (MPR), maximum intensity projections(MIP), minimal intensity projections (MinIP), and average projections.73

If dynamic CTP is performed, then software that allows visualization ofmyocardial enhancement over time is also required.

CTP images should be viewed using the available cardiac phases toreduce motion-related artifacts. Diagnostic accuracy may be improvedby the use of additional image reconstructions performed at a phasewith the least cardiac motion. For this reason, images may be re-constructed from multiple phases of the cardiac cycle, typically at

5–10% intervals. Raw CTA and CTP data files should be retained untilimage interpretation is complete. Images may be viewed in systolicand/or diastolic phases, depending on acquisition parameters.

CTP images may be viewed in static and/or dynamic CTP modes,depending on the acquisition. If available, qualitative and quantitativeventricular function should be reported including regional wall motionabnormalities to correlate with the CTP findings.

5.1. Static CTP image analysis

In static visual CTP interpretation, myocardial images are typicallyarranged using three standard orthogonal views - the short axis, verticallong axis, and horizontal long axis (Fig. 2).74 CTP images usually use anarrow window width of 200–300 and level setting of 100–150 with anaverage slice thickness of 5–8 mm. Images may be viewed in minimalintensity projection (MinIP) or average intensity projections, althoughthe former may be more sensitive. CTP images should be automaticallyor visually divided into standard American Heart Association 17 myo-cardial segment model.75 Dedicated perfusion software platformsshould also allow side by side display and comparison of rest and stressimages to examine the presence and extent of reversibility of perfusiondefects. The ability to review the perfusion images acquired duringdifferent cardiac phases in order to determine if a potential perfusiondefect persists during multiple phases of the cardiac cycle can helpimprove diagnostic confidence. In the event a true perfusion defect has

Fig. 2. Potential display for static CTP images. A surface rendered image (bottom left) suggests the presence of a stenosis in the left anterior descending artery. StressCTP images are shown in the 2-chambr (top left), 3-chamber (top middle), and short axis (bottom middle) imaging planes and demonstrate the presence of an apicalanterior perfusion defect. The finding is verified on the quantitative bullseye plots (top and bottom right).

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

93

been identified, visual coronary CTA analysis or automated softwareshould enable the interpreter to track the stenosis in the major epi-cardial coronary artery or branch that subtends the myocardial segmentwith the corresponding perfusion defect.

5.1.1. Semi-quantitative analysis for static CTPQualitatively, perfusion defects are visually evaluated using the

American Heart Association 17-segment model. We recommend thatperfusion defects be “scored” using the following:

1) Severity. Hypoperfusion severity should be scored by whether hy-poperfusion is present (binary). In addition, it may be useful todescribe the severity of the defect using a continuous scale of mild,moderate, or severe.

2) Size. The size of the perfusion defect (small, medium, or large)should be described, as well as whether it is transmural (>50%) ornon-transmural (<50%).

3) Reversibility. Perfusion defects should also be described as re-versible (perfusion defect present only on the stress CTP images),fixed (perfusion defect present and similar on both rest and stressCTP images), or partially reversible (perfusion defect present onboth the stress and rest CTP images but less pronounced of the restimages).

CTP software is also available which permits evaluation of othersemi-quantitative metrics such as the transmural perfusion ratio (TPR),absolute contrast map and various other types of perfusion indexes. TPRis the ratio of the subendocardial attenuation of a specific segment ofmyocardium and the entire epicardial attenuation. This index has beenvalidated using quantitative coronary angiography14 and fractionalflow reserve,76 and can be used to guide visual reads. This can be dis-played as an average for the entire myocardial segment or within eachpixel of the segment using a 17-myocardial segment polar plot whicheffectively presents a 34 segment model.14 Another proposed CTP indexnormalizes myocardial subendocardial perfusion to the LV cavity HUdensity resulting in a segmental myocardial perfusion index.77 Thisallows for different “normal” CTP attenuation in certain segments thatare prone to more attenuation due to adjacent structures, such as thebasal inferior wall. A related semi-quantitative CTP method measuresthe 3-dimensional myocardial attenuation in volumetric myocardialsegments, which is normalized to the LV cavity but also compared to ahistogram of a reference normal myocardial segment from a library ofnormal individuals. The hypoenhanced voxels in a myocardial segmentreflects the relative volume of perfusion defect, and the defect severityis estimated by the difference in the positions of peaks of the attenua-tion histogram of the particular segment of interest and the referencesegment. Together, these indices are used to estimate an overall defectvolume and severity. This normalized voxel assessment has been shownto have good agreement with SPECT MPI for rest CT-MPI, with im-proved detection for CAD noted with the addition of stress CT-MPI.45,78

Despite the number of available methods for semi-quantitative assess-ment of myocardial perfusion for static CTP, there is no clear benefit ofusing one method over another and the authors can make no re-commendations for preferential use of any technique at present.

5.2. Semi-quantitative analysis for dynamic CTP

The result of the dynamic perfusion acquisition is a sequence ofvolumetric CT scans, each consisting of overlapping axial slices usuallywith a thickness of 1.5–3mm. The individual CT images are char-acterized by a low contrast-to-noise ratio. Although more difficult tointerpret than static perfusion scans, enhancement heterogeneity mayreadily be observed on the source images. From the time-resolved dataand temporal changes in CT attenuation it is possible to investigate a

range of myocardial perfusion parameters.79 The measured CT at-tenuation values show a linear relation with the iodine concentration.Sampling the myocardium or myocardial region of interest and thencharting the myocardial attenuation values over time create time-at-tenuation curves (TAC's). The time-attenuation curves provide a semi-quantitative measure of myocardial perfusion by normalizing againstslope or peak of the attenuation curve measured in the left ventricle oraorta59 which is related to the arterial blood flow supplying the myo-cardium and can be compared between different myocardial regions.Ischemic myocardium will demonstrate a shallower upslope and lowerpeak compared to normal myocardium.

5.2.1. Post-processing for quantitative analysis for dynamic CTPDepending on the CT scanner and myocardial perfusion package

used, a number of pre-processing steps may be required. Correction ofmyocardial displacement between phases is essential to measurechanges in attenuation. Phases with non-correctable displacement,perhaps due to arrhythmia, may be entirely deleted from the analysis.To measure the change in myocardial attenuation, it is important toacquire one or more non-enhanced datasets, that serve as the baseline,non-enhanced image for the TAC. To quantify the myocardial bloodflow, both a measure of the arterial input function and myocardial floware needed to model blood flow calculations in absolute blood flowunits (ml/min/gram of myocardial tissue). The arterial input functioncan be approximated by placing a region of interest in the descendingthoracic aorta or left ventricular cavity and sampling the attenuationover time. For shuttle-mode acquisitions, it is important to place a re-gion of interest both in the upper and lower part of the scan to increasethe sampling rate. The software application may require further pre-paration steps, such as isolation of the myocardial volume, or adjust-ment of the image filtering to avoid excessive image noise. Semi-au-tomated software applications exist to perform these functions, butrequire visual conformation and possibly manual adjustments by ex-perienced users to optimize quantitative measures.

5.2.2. Quantitative analysis for dynamic CTPCalculation of absolute parameters of myocardial perfusion requires

more complex iodinated contrast kinetic modeling. Generally, two-compartment models are applied that consider the blood flow into themyocardium, the myocardial blood flow itself, and the extra-vascularextraction rate of the contrast into the interstitial space. The most in-vestigated method is the deconvolution model.59,80 By this approach,the arterial input function, the tissue TACs, and the assumed temporalexchange of contrast between the vessel and the interstitial space overtime (assuming a short contrast bolus) are deconvoluted with a CTPmodel to derive quantitative flow.55,59,81 Preferably the sampling in-terval should be less than the shortest transit time, although the numberof CTP images is limited by the scanner configuration and radiationdose. The duration of scanning is further limited by the time of thebreath hold.

Other processing techniques have been reported although nonehave not been widely validated. Patlak plot analysis was reported as anapproach to calculate myocardial blood flow.82 To overcome the lim-ited number of images over time and specifically designed for theshuttle-mode perfusion imaging technique, is a hybrid model com-bining a simplified deconvolution algorithm with the maximum-slopemethod.54 Using this method, myocardial blood flow can be calculatedby dividing the maximum slope of the fit model curve by the peak of thearterial input function.61,62,82 In addition to myocardial blood flow,myocardial blood volume may also be calculated. While other in-dependent or derived hemodynamic parameters may be calculated,which have established value in non-myocardial dynamic contrast en-hanced imaging applications, these parameters have not been ex-tensively investigated for myocardial perfusion imaging. The most

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

94

commonly used approaches for assessing myocardial perfusion usingstatic and dynamic CTP are summarized in Table 3.

5.2.3. Image display for dynamic CTPFor qualitative assessment, the dynamic perfusion images may be

displayed in the same orthogonal cardiac imaging planes as static CTPwith similar settings. Dynamic CTP images should be viewed duringcontrast enhancement as a cine loop to qualitatively view of myocardialcontrast uptake over time. This also allows for visual confirmation ofsignificant patient or cardiac motion that may affect myocardial co-registration over time. The results may also be displayed using color-coded maps, where the shading represent particular myocardial per-fusion values, such as myocardial blood flow or volume. The voxel-based functional imaging maps may be sampled for (averaged) regionalperfusion values. While quantification of myocardial perfusion valuesshould be used to differentiate normal, ischemic, and infarcted myo-cardium, there is no accepted positivity criterion with threshold valuesthat may be applied for interpretation. Thus, currently, relative differ-ences in perfusion values between different myocardial segments hasbeen adopted in clinical practice.

6. CTP clinical studies

6.1. Static CTP clinical studies

Current evidence on the feasibility (efficacy) of static myocardial CTperfusion imaging using contemporary CT scanners is summarized inTable 4. Overall, studies range from proof-of-concept studies to multi-center studies demonstrating a promising performance by static perfu-sion CT for the detection of functionally relevant CAD, with incrementaldiagnostic value over CT angiography alone. The reference standardvaried significantly among the studies and included SPECT, MRI, PET,invasive angiography, and invasive measurement of FFR for hemody-namically significant lesions.

6.2. Dynamic CTP clinical studies

Current evidence on the feasibility of dynamic myocardial CT per-fusion imaging using contemporary CT scanners is summarized inTable 5. Overall, studies were proof-of-concept studies demonstrating apromising performance by dynamic perfusion CT for the detection offunctionally relevant CAD, with incremental diagnostic value over CTangiography alone. The reference standard varied significantly amongthe studies and included SPECT, MRI, PET, invasive angiography, andinvasive measurement of FFR with different FFR thresholds (0.75 vs.0.80) for hemodynamically significant lesions. Also, derived myocardialperfusion parameters (i.e. myocardial blood flow and volume) were notemployed in a standardized fashion to serve as a positivity criterion.The reported associated radiation exposure varied but ranged from 5.3to 13.1 mSv. Reported technical limitations included limited coverageof the perfusion volume and cardiac and patient motion during longbreath holds of up to 30 seconds.

6.3. Dual energy clinical studies

Current evidence on the feasibility (efficacy) of dual energy myo-cardial CT perfusion imaging is summarized in Table 6.

7. CTP interpretation and reporting

The components of a CTP structured report should include in-formation regarding the following:

7.1. Clinical history and indications

This section should include the main clinical indication for the studyand specific signs and symptoms to support the clinical indication.

7.2. Technique or procedure description

This section should include pertinent information of image

Table 4Summary of available studies validating static CTP.

Scanner type Imaging protocol Reference standard Na Radiation Dose (mSv)b Sensitivity (%) Specificity (%)

Kurata et al.16 16 detector rows Retrospective SPECT-MPI 12 Not reported 90 79George et al.14 64 &256 detector rows Retrospective QCA/SPECT-MPI 27 16.8 (64 CT)

21.6 (256 CT)c81 85

Blankstein et al.4 64 DSCT Retrospective QCA/SPECT-MPI 33 9.1 93 74Rocha-Filho et al.83 64 DSCT Retrospective QCA 34 9.8 91 91Okada et al.84 64 DSCT Retrospective SPECT-MPI 47 10.0 Pearson r= 0.56 (rest) and 0.66 (stress)Cury et al.85 64 detector rows Retrospective low resolution QCA 26 3.4 88 79Ko S. et al.86 64 DSCT Retrospective MRI-MPI 41 8.6 91 72Tamarapoo et al.17 64 DSCT Retrospective SPECT-MPI 30 15.7 92 86Weininger et al.87 128 DSCT Retrospective (DECT) MRI-MPI 20 15.2 (DECT) DECT 93 DECT 99Feuchtner et al.19 128 DSCT Prospective High pitch Spiral

modeMRI-MPI 30 0.9 96 95

Ko B. et al.76 320 detector rows Prospective FFR 42 5.3 76 84George et al.18 320 detector rows Prospective CCTCA/SPECT-

MPI50 7.0 100 85

Ko B. et al.88 320 detector rows Prospective FFR 40 4.5 95 95Nasis et al.89 320 detector rows Prospective SPECT MPI 20 9.2 94 98Bettencourt et al.20 64 detector rows Retrospective FFR 101 5.0 (including rest

CCTCA)71 90

Rief et al.37 320 detector rows Prospective QCA 91 7.9 56 78Pontone et al.25 320 detector rows Prospective FFR 130 2.5± 1.1 91 86Rochitte et al.90 320 detector rows Prospective ICA/SPECT 381 9.3 80£ 74d

Cury et al.91 Multivendor Prospective SPECT 110 N/A 90μ 84e

a Number of patients in the study who underwent both CTP and reference standard imaging modality.b Radiation dose only for the stress perfusion acquisition.c Only the radiation dose for both stress and rest perfusion was reported.d Assuming a summed stress score of 4.e Per patient results available only.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

95

acquisition and image reconstruction of coronary CTA and myocardialCT perfusion. Several aspects of image acquisition should be included,such as:

• type of studies (including calcium score, CCTA, rest and stressmyocardial CT perfusion, delayed enhancement CT)• equipment (scanner type – 64-slice, 128-slice, 256-slice, 320-slice ordual-source),• technical acquisition protocol (with reference to both stress andrest/CTA acquisition as well as acquisition modes: prospective vsretrospective, static vs dynamic perfusion),• type and amount of contrast (for both stress and rest/CTA acquisi-tions),• medications use (including stress vasodilator agent and sequence oftesting, such as rest then stress CTP),• clinical parameters during the procedure, including heart rate andany complications or side effects (particularly symptoms or sideeffects during the administration of stress vasodilator agent).• Image Reconstruction should include all the performed reconstruc-tion modes and descriptions for any semi-quantitative or quantita-tive methods used.

7.3. Findings

The study quality needs to be described for each component of theexamination and the presence of any significant artifacts that may affectimage quality and study interpretation. The artifacts specific to cor-onary CTA and myocardial CTP should be included in the report, suchas beam hardening artifacts, motion artifacts, misalignment, etc.

Perfusion defects should be assessed qualitatively by comparing therest and stress acquisitions based on the severity (mild, moderate andsevere), extent (small, medium and large) and reversibility (reversible,partially reversible and fixed) as described above. Quantitative or semi-quantiative information can be included, if such techniques were uti-lized. When available, global and regional ventricular function assess-ment should be included, particularly in association with any myo-cardial perfusion abnormality.

Coronary CTA findings should be described separately in a per-vessel and per segment distribution for stenosis and plaque assessmentfollowing the updated SCCT guidelines for interpretation and re-porting.73 This includes a complete report of the non-coronary cardiacstructures.

Table 6Summary of CT Perfusion studies validating dual energy CTP.

Author Scanner Type Imaging Protocol Reference Standard N Average CT Dose (mSv) Sensitivity (%) Specificity (%)

Ruzsics 2009102 DSCT Shuttle, Rest SPECT 36 14 97 67Bauer 201052 DSCT Shuttle, Rest CMR 36 9.7 77 97Nagao 2010103 DSCT Shuttle, Stress SPECT, ICA 10 NR 86 75Nance 201169 DSCT Shuttle, Rest ICA 12 NR 91 95Ko 201186 DSCT Shuttle, Stress CMR 41 8.6 89 78Wang 2011104 DSCT Shuttle, Rest SPECT 31 10.5 81 92Meyer 2012105 DSCT Shuttle, Rest/Stress/Deleayed CMR 50 13.4 90 71Ko 2012106 DSCT Shuttle, Rest/Stress ICA 45 16.5 89 74Weininger 2012107 DSCT Shuttle, Stress SPECT

CMR20 12.8 94

939899

Delgado 2013108 DSCT Shuttle, Stress CMR 56 8.2 76 99Ko 2014109 DSCT Shuttle, Rest/Stress CMR, ICA 40 4.2

4.64287

8379

Ko 2014110 DSCT Shuttle, Stress CMR 100 4.2 97 36Kim 2014111 DSCT Shuttle, Stress/Rest CMR 50 11.4 94 71Zhao 2014112 DSCT Shuttle, Rest ICA 60 NR 94 91Kido 2014113 DSCT Shuttle, Stress ICA 21 7.7 67 92De Cecco 2014114 DSCT Shuttle, Rest/Stress ICA

SPECT29 5.8

6.695 5091 38

Delgado 2016115 DSCT Shuttle, Stress CMR 36 5.4 100 90

Table 5Summary of available studies on the diagnostic accuracy of dynamic myocardial CT perfusion imaging. CAD: coronary artery disease; cMRI: cardiac magneticresonance imaging; FFR: invasive measurement of fractional flow reserve; PET: positron emission tomography; DSCT: dual-source computed tomography.

Author name Scanner type Imaging protocol Reference standard N Radiation Dose [mSv] Sensitivity (%) Specificity (%)

Bastarrika et al.92 128 DSCT Shuttle cMRI 10 1290.4± 233.3 mGy cm 86.1 98.2Ho et al.93 128 DSCT Shuttle SPECT 35 9.15± 1.32 mSv (stress)

9.09± 1.4 mSv (rest)83 78

Weininger et al.87 128 DSCT Shuttle SPECT 20 12.8± 2.4 mSv 86 98Kono et al.94 128 DSCT Shuttle FFR 42 9.4 mSv 97.8 69.6Bamberg et al.95 128 DSCT Shuttle FFR 33 10±2 93 87Kurata et al.96 256 detector rows Stationary Invasive Angiography 11 10.4 100 78.6Bamberg et al.97 128 DSCT Shuttle cMRI 38 11.08±2.0 100 75Wang et al.98 128 DSCT Shuffle SPECT 30 12.8± 1.6 90 81.4Kim et al.57 128 DSCT Shuttle cMRI 33 5.3± 1.0 82 84Ebersberger et al.99 128 DSCT Shuttle SPECT 37 9.6± 4 86 96Greif et al.100 128 DSCT Shuttle FFR 65 9.7± 2.2 95.1 74.7Kikuchi et al.21 320 Single Temporal Volume O–H2O-PET 32 12.8± 2.9 85.7 92.3Rossi et al.101 128 DSCT Shuttle FFR 80 9.4 84 81Huber et al.63 256 detector rows Stationary FFR 32 9.5 75.9 100Pontone et al.23 256 detector rows Stationary FFR 85 5.3± 0.7 mSv 73% 86%

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

96

7.4. Impression or conclusion

It is recommended to start the impression with an overall statementsummarizing the findings: For example, “severe coronary athero-sclerosis with significant 3-vessel anatomic disease and a medium sizearea of reversible ischemia in the LAD territory only”. It is important torelate any perfusion deficits to the culprit vessel since this may havetreatment implications.

Subsequently, all other available information should be reported,such as: 1) total calcium score, 2) presence or absence of significantcoronary stenosis, 3) presence or absence of myocardial perfusion de-fects, 4) information related to global and regional LV function, 5) Anyother relevant cardiac or extra-cardiac findings.

7.5. Images

Attaching representative images of important pathology for myo-cardial CT Perfusion and CCTA imported from the workstation is sug-gested. For referring physicians, short-axis images of myocardial CTperfusion are often preferable.

8. Summary and final conclusions

Myocardial CTP provides incremental value over coronary CTA.This document provides some consensus regarding the heterogeneitythat currently exists about the use of pharmacologic stress agents,imaging sequences, scanner types, acquisition protocols, post-proces-sing, and interpretation of CTP results. This document represents theconsensus opinion of international experts in myocardial CTP imagingand is meant to serve as a resource for clinical implementation, resourceallocation, and future investigations related to CTP. Key statementsselected by the committee are shown in Table 7.

Conflict of Interest

Grants and Research for the authors:

Amit Patel: General Electric; Philips; Astellas, Speaker's Bureau:Astellas - stress testing; Alnylam - amyloidosis.

Ron Blankstein: Astellas Inc; Amgen, Inc., Consultant: Amgen Inc.Fabian Bamberg: Bayer; Siemens Healthineers, Speakers Bureau: GE;Siemens Healthineers; Bracco.Kelley Branch: Bayer; Kestra; Astellas; Eli Lilly, Consultant: Bayer;Janssen; Astra Zeneca.Ricardo Cury: GE Healthcare, Consultant: Covera Health, Stocks:Cleerly, Royalties: IBM Watson - RASF Innovations.Brian Ghoshhajra: Siemens Healthcare, Stocks: Apple, Inc.Brian Ko: Canon Medical, Speakers Bureau: Medtronic; St. Jude Medical.Koen Nieman: Siemens Healthineers; GE; Bayer; HeartFlow.Francesca Pugliese: Siemens Healthineers.Joseph Schoepf: Astellas; Bayer; GE; Siemens, Consultant: Guerbet,Speaker's Bureau: HeartFlow.Richard George: AstraZeneca, Patents: AstraZeneca, Employee:AstraZeneca.Gianluca Pontone: GE Healthcare; Bracco; HeartFlow, SpeakersBureau: GE Healthcare; Bracco; Bayer; Medtronic; HeartFlow.Gianluca Pontone: GE Healthcare; Bracco; HeartFlow, SpeakersBureau: GE Healthcare; Bracco; Bayer; Medtronic; HeartFlow.

Acknowledgment

Richard T. George.

References

1. Miller JM, Rochitte CE, Dewey M, et al. Diagnostic performance of coronary an-giography by 64-row CT. N Engl J Med. 2008;359:2324–2336.

2. Budoff MJ, Dowe D, Jollis JG, et al. Diagnostic performance of 64-multidetector rowcoronary computed tomographic angiography for evaluation of coronary arterystenosis in individuals without known coronary artery disease: results from theprospective multicenter ACCURACY (Assessment by Coronary ComputedTomographic Angiography of Individuals Undergoing Invasive CoronaryAngiography) trial. J Am Coll Cardiol. 2008;52:1724–1732.

3. Meijboom WB, Meijs MFL, Schuijf JD, et al. Diagnostic accuracy of 64-slice com-puted tomography coronary angiography: a prospective, multicenter, multivendorstudy. J Am Coll Cardiol. 2008;52:2135–2144.

4. Blankstein R, Shturman LD, Rogers IS, et al. Adenosine-induced stress myocardialperfusion imaging using dual-source cardiac computed tomography. J Am CollCardiol. 2009;54:1072–1084.

5. Tonino PA, De Bruyne B, Pijls NH, et al. Fractional flow reserve versus angiographyfor guiding percutaneous coronary intervention. N Engl J Med. 2009;360:213–224.

Table 7Key consensus statements for CTP.

1. We recommend that myocardial CTP may be added to coronary CTA when there is a high likelihood of ischemic heart disease, known CAD, prior coronary intervention, orsignificant calcifications, Additionally, CTP may be added when there is a known stenosis of indeterminate functional significance. In such cases, myocardial CTP isrecommended if knowledge regarding the presence and severity of ischemia would impact patient management and CTP is feasible.

2. Myocardial CTP imaging with pharmacologic stress testing requires supervision by a trained medical provider who is experienced in the administration of pharmacologicstress agents and stress testing.

3. Continuous 12-lead ECG monitoring may be performed using radiolucent electrodes and leads. Alternatively, 12-lead ECG monitoring can be interrupted during scanacquisition while continuous rhythm is still being recorded and monitored by the scanner console. Upon the completion of imaging, a 12 lead ECG should be performed toensure there are no significant changes from baseline suggestive of post-stress myocardial ischemia or injury.

4. If an automated bolus-tracking algorithm is utilized, a region of interest should be defined in either the proximal-ascending or mid-descending thoracic aorta. Theautomated trigger should be set to a Hounsfield Unit (HU) threshold that is ∼80-100HU above the baseline HU of the selected reference area immediately prior to thecontrast injection. After the scanner is triggered, a short pause (∼5 seconds) is required to allow for breath-hold instructions to be given. Contrast dose and flow rate istypically ∼60–70ml injected at 5–6ml/sec.

5. The optimal sequence of scans (rest first versus stress first) has not been determined and either sequence has advantages and disadvantages. Our committee suggestsassessment of factors that could impair exclusion of obstructive CAD on coronary CTA (e.g. significant coronary calcium, prior coronary event or coronary revascularization,high pre-test probability of obstructive CAD) prior to decision to perform CTP. Those patients with a high likelihood or known significant coronary disease or severe amountof CAC may undergo stress myocardial perfusion imaging first; whereas, patients who lack the above features may undergo coronary CTA first.

6. A delay of 10–20 minutes between the rest and stress CT acquisitions is recommended to minimize myocardial contrast agent contamination and to eliminate the potentialadverse interaction of pharmacologic stress agents and nitroglycerin administration.

7. There is no consensus about whether imaging should be performed at end-systole or during diastasis although most larger trials to date have selected diastole. A multiphaseacquisition which includes phases from both systole and diastole can be considered so that the phase with the least amount of artifact can be selected for analysisretrospectively.

8. For optimal image interpretation, images may be reconstructed from multiple phases of the cardiac cycle, typically at 5–10% intervals of the R-R image acquisition windowand raw CTA and CTP data files should be retained until image interpretation is complete.

9. In static visual CTP interpretation, myocardial images are typically arranged using three standard orthogonal views - the short axis, vertical long axis, and horizontal longaxis. CTP images usually use a narrow window width of ∼200–300 and level setting of ∼100–150 with an average slice thickness of 5–8 mm. Images may be viewed inminimal intensity projection (MinIP) or average intensity projections

11. Perfusion defects should be described in terms of size, transmurality, and reversibility in the context of the degree of stenosis in the supplying coronary artery.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

97

6. Hachamovitch R, Hayes SW, Friedman JD, Cohen I, Berman DS. Comparison of theshort-term survival benefit associated with revascularization compared with med-ical therapy in patients with No prior coronary artery disease undergoing stressmyocardial perfusion single photon emission computed tomography. Circulation.2003;107:2900–2907.

7. Hachamovitch R, Rozanski A, Shaw LJ, et al. Impact of ischaemia and scar on thetherapeutic benefit derived from myocardial revascularization vs. medical therapyamong patients undergoing stress-rest myocardial perfusion scintigraphy. Eur HeartJ. 2011;32:1012–1024.

8. Wang T, Wu X, Chung N, Ritman E. Myocardial blood flow estimated by synchro-nous, multislice, high-speed computed tomography. IEEE Trans Med Imaging.1989;8:70–77.

9. Wolfkiel CJ, Ferguson JL, Chomka EV, et al. Measurement of myocardial blood flowby ultrafast computed tomography. Circulation. 1987;76:1262–1273.

10. Garcia MJ, Lessick J, Hoffmann MHK, for the CSI. Accuracy of 16-row multidetectorcomputed tomography for the assessment of coronary artery stenosis. J Am MedAssoc. 2006;296:403–411.

11. Hoffmann U, Moselewski F, Cury RC, et al. Predictive value of 16-slice multidetectorspiral computed tomography to detect significant obstructive coronary artery dis-ease in patients at high risk for coronary artery disease: patient- versus segment-based analysis. Circulation. 2004;110:2638–2643.

12. Kuettner A, Beck T, Drosch T, et al. Diagnostic accuracy of noninvasive coronaryimaging using 16-detector slice spiral computed tomography with 188 ms temporalresolution. J Am Coll Cardiol. 2005;45:123–127.

13. Kuettner A, Trabold T, Schroeder S, et al. Noninvasive detection of coronary lesionsusing 16-detector multislice spiral computed tomography technology: initial clinicalresults. J Am Coll Cardiol. 2004;44:1230–1237.

14. George RT, Arbab-Zadeh A, Miller JM, et al. Adenosine stress 64- and 256-rowdetector computed tomography angiography and perfusion imaging: a pilot studyevaluating the transmural extent of perfusion abnormalities to predict athero-sclerosis causing myocardial ischemia. Circ Cardiovasc Imaging. 2009;2:174–182.

15. George RT, Silva C, Cordeiro MAS, et al. Multidetector computed tomographymyocardial perfusion imaging during adenosine stress. J Am Coll Cardiol.2006;48:153–160.

16. Kurata A, Mochizuki T, Koyama Y, et al. Myocardial perfusion imaging using ade-mosine triphosphate stress milti-slice spiral computed tomography: alternative tostress myocaridla perfusion scintography. Circ J. 2005;69:550–557.

17. Tamarappoo BK, Dey D, Nakazato R, et al. Comparison of the extent and severity ofmyocardial perfusion defects measured by CT coronary angiography and SPECTmyocardial perfusion imaging. JACC Cardiovasc Imaging. 2010;3:1010–1019.

18. George RT, Arbab-Zadeh A, Miller JM, et al. Computed tomography myocardialperfusion imaging with 320-row detector computed tomography accurately detectsmyocardial ischemia in patients with obstructive coronary artery disease. CircCardiovasc Imaging. 2012;5:333–340.

19. Feuchtner G, Goetti R, Plass A, et al. Adenosine stress high-pitch 128-slice dual-source myocardial computed tomography perfusion for imaging of reversiblemyocardial ischemia: comparison with magnetic resonance imaging. Circ CardiovascImaging. 2011;4:540–549.

20. Bettencourt N, Chiribiri A, Schuster A, et al. Direct comparison of cardiac magneticresonance and multidetector computed tomography stress-rest perfusion imagingfor detection of coronary artery disease. J Am Coll Cardiol. 2013;61:1099–1107.

21. Kikuchi Y, Oyama-Manabe N, Naya M, et al. Quantification of myocardial bloodflow using dynamic 320-row multi-detector CT as compared with 15O-H2O PET.Eur Radiol. 2014;24:1547–1556.

22. Pontone G, Andreini D, Guaricci AI, et al. Incremental diagnostic value of stresscomputed tomography myocardial perfusion with whole-heart coverage CT scannerin intermediate- to high-risk symptomatic patients suspected of coronary arterydisease. JACC Cardiovasc Imaging. 2019;12:338–349.

23. Pontone G, Baggiano A, Andreini D, et al. Dynamic stress computed tomographyperfusion with a whole-heart coverage scanner in addition to coronary computedtomography angiography and fractional flow reserve computed tomography de-rived. JACC Cardiovasc Imaging. 2019;12(12):2460–2471. https://doi.org/10.1016/j.jcmg.2019.02.015.

24. Pontone G, Baggiano A, Andreini D, et al. Stress computed tomography perfusionversus fractional flow reserve CT derived in suspected coronary artery disease: thePERFECTION study. JACC Cardiovasc Imaging. 2018;12(8 Pt 1):1487–1497. https://doi.org/10.1016/j.jcmg.2018.08.023.

25. Pontone G, Baggiano A, Andreini D, et al. Diagnostic accuracy of simultaneousevaluation of coronary arteries and myocardial perfusion with single stress cardiaccomputed tomography acquisition compared to invasive coronary angiography plusinvasive fractional flow reserve. Int J Cardiol. 2018;273:263–268.

26. Cury RC, Kitt TM, Feaheny K, et al. A randomized, multicenter, multivendor studyof myocardial perfusion imaging with regadenoson CT perfusion vs single photonemission CT. J Cardiovasc Comput Tomogr. 2015;9 103-112 e1-2.

27. George RT, Mehra VC, Chen MY, et al. Myocardial CT perfusion imaging and SPECTfor the diagnosis of coronary artery disease: a head-to-head comparison from theCORE320 multicenter diagnostic performance study. Radiology. 2015;274:626.

28. Celeng C, Leiner T, Maurovich-Horvat P, et al. Anatomical and functional computedtomography for diagnosing hemodynamically significant coronary artery disease: ameta-analysis. JACC Cardiovasc Imaging. 2018;12(7 Pt 2):1316–1325. https://doi.org/10.1016/j.jcmg.2018.07.022.

29. Hacker M, Jakobs T, Matthiesen F, et al. Comparison of spiral multidetector CTangiography and myocardial perfusion imaging in the noninvasive detection offunctionally relevant coronary artery lesions: first clinical experiences. J Nucl Med.2005;46:1294–1300.

30. Rispler S, Keidar Z, Ghersin E, et al. Integrated single-photon emission computed

tomography and computed tomography coronary angiography for the assessment ofhemodynamically significant coronary artery lesions. J Am Coll Cardiol.2007;49:1059–1067.

31. Di Carli MF, Dorbala S, Curillova Z, et al. Relationship between CT coronary an-giography and stress perfusion imaging in patients with suspected ischemic heartdisease assessed by integrated PET-CT imaging. J Nucl Cardiol. 2007;14:799–809.

32. Schuijf JD, Wijns W, Jukema JW, et al. Relationship between noninvasive coronaryangiography with multi-slice computed tomography and myocardial perfusionimaging. J Am Coll Cardiol. 2006;48:2508–2514.

33. Gaemperli O, Schepis T, Koepfli P, et al. Interaction of caffeine with regadenoson-induced hyperemic myocardial blood flow as measured by positron emission to-mography: a randomized, double-blind, placebo-controlled crossover trial. J AmColl Cardiol. 2008;51:328–329.

34. Wykrzykowska JJ, Arbab-Zadeh A, Godoy G, et al. Assessment of in-stent restenosisusing 64-MDCT: analysis of the CORE-64 multicenter international trial. Am JRoentgenol. 2010;194:85–92.

35. Arbab-Zadeh A, Miller JM, Rochitte CE, et al. Diagnostic accuracy of computedtomography coronary angiography according to pre-test probability of coronaryartery disease and severity of coronary arterial calcification: the CORE-64 (coronaryartery evaluation using 64-row multidetector computed tomography angiography)international multicenter study. J Am Coll Cardiol. 2012;59:379–387.

36. Sharma RK, Arbab-Zadeh A, Kishi S, et al. Incremental diagnostic accuracy ofcomputed tomography myocardial perfusion imaging over coronary angiographystratified by pre-test probability of coronary artery disease and severity of coronaryartery calcification: the CORE320 study. Int J Cardiol. 2015;201:570–577.

37. Rief M, Zimmermann E, Stenzel F, et al. Computed tomography angiography andmyocardial computed tomography perfusion in patients with coronary stents: pro-spective intraindividual comparison with conventional coronary angiography. J AmColl Cardiol. 2013;62:1476–1485.

38. Genders TSS, Steyerberg EW, Alkadhi H, et al. A clinical prediction rule for thediagnosis of coronary artery disease: validation, updating, and extension. Eur HeartJ. 2011;32:1316–1330.

39. Di Carli M, Czernin J, Hoh CK, et al. Relation among stenosis severity, myocardialblood flow, and flow reserve in patients with coronary artery disease. Circulation.1995;91:1944–1951.

40. Uren NG, Melin JA, De Bruyne B, Wijns W, Baudhuin T, Camici PG. Relation be-tween myocardial blood flow and the severity of coronary-artery stenosis. N Engl JMed. 1994;330 1782-80.

41. Henzlova MJ, Duvall WL, Einstein AJ, Travin MI, Verberne HJ. ASNC imagingguidelines for SPECT nuclear cardiology procedures: stress, protocols, and tracers.J Nucl Cardiol. 2016;23:606–639.

42. Douglas PS, Garcia MJ, Haines DE, et al. ACCF/ASE/AHA/ASNC/HFSA/HRS/SCAI/SCCM/SCCT/SCMR 2011 Appropriate Use Criteria for Echocardiography. A Report ofthe American College of Cardiology Foundation Appropriate Use Criteria Task Force. vol.57. American Society of Echocardiography, American Heart Association, AmericanSociety of Nuclear Cardiology, Heart Failure Society of America, Heart RhythmSociety, Society for Cardiovascular Angiography and Interventions, Society ofCritical Care Medicine, Society of Cardiovascular Computed Tomography, andSociety for Cardiovascular Magnetic Resonance <em>Endorsed by the AmericanCollege of Chest Physicians; 2011:1126–1166.

43. Patel AR, Lodato JA, Chandra S, et al. Detection of myocardial perfusion abnorm-alities using ultra-low radiation dose regadenoson stress multidetector computedtomography. Journal of Cardiovascular Computed Tomography. 2011;5:247–254.

44. Baxa J, Milan Hrom, et al. Regadenoson-stress dynamic myocardial perfusion im-proves diagnostic performance of CT angiography in assessment of intermediatecoronary artery stenosis in asymptomatic patients. BioMed Res Int. 2015;2015:7.

45. Mor-Avi V, Lodato JA, Kachenoura N, et al. Quantitative three-dimensional eva-luation of myocardial perfusion during regadenoson stress using multidetectorcomputed tomography. J Comput Assist Tomogr. 2012;36:443–449.

46. Abbara S, Blanke P, Maroules CD, et al. SCCT guidelines for the performance andacquisition of coronary computed tomographic angiography: a report of the societyof cardiovascular computed tomography guidelines committee: endorsed by thenorth American society for cardiovascular imaging (NASCI). J Cardiovasc ComputTomogr. 2016;10:435–449.

47. Lubbers M, Coenen A, Kofflard M, et al. Comprehensive cardiac CT with myocardialperfusion imaging versus functional testing in suspected coronary artery disease: themulticenter, randomized CRESCENT-II trial. JACC Cardiovasc Imaging.2018;11:1625–1636.

48. Cury RC, Magalhaes TA, Paladino AT, et al. Dipyridamole stress and rest transmuralmyocardial perfusion ratio evaluation by 64 detector-row computed tomography.J Cardiovasc Comput Tomogr. 2011;5:443–448.

49. Nikolaou K, Knez A, Sagmeister S, et al. Assessment of myocardial infarctions usingmultidetector-row computed tomography. J Comput Assist Tomogr.2004;28:286–292.

50. Sanz J, Weeks D, Nikolaou K, et al. Detection of healed myocardial infarction withmultidetector-row computed tomography and comparison with cardiac magneticresonance delayed hyperenhancement. Am J Cardiol. 2006;98:149–155.

51. Mahnken AH, Lautenschlager S, Fritz D, Koos R, Scheuering M. Perfusion weightedcolor maps for enhanced visualization of myocardial infarction by MSCT: pre-liminary experience. Int J Cardiovasc Imaging. 2008;24:883–890.

52. Bauer RW, Kerl JM, Fischer N, et al. Dual-energy CT for the assessment of chronicmyocardial infarction in patients with chronic coronary artery disease: comparisonwith 3-T MRI. Am J Roentgenol. 2010;195:639–646.

53. Nikolaou K, Sanz J, Poon M, et al. Assessment of myocardial perfusion and viabilityfrom routine contrast-enhanced 16-detector-row computed tomography of theheart: preliminary results. Eur Radiol. 2005;15:864–871.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

98

54. Mahnken AH, Klotz E, Pietsch H, et al. Quantitative whole heart stress perfusion CTimaging as noninvasive assessment of hemodynamics in coronary artery stenosis:preliminary animal experience. Investig Radiol. 2010;45:298–305.

55. Rossi A, Merkus D, Klotz E, Mollet N, de Feyter PJ, Krestin GP. Stress myocardialperfusion: imaging with multidetector CT. Radiology. 2014;270:25–46.

56. Pelgrim GJ, van Hamersvelt RW, Willemink MJ, et al. Accuracy of iodine quanti-fication using dual energy CT in latest generation dual source and dual layer CT. EurRadiol. 2017;27:3904–3912.

57. Kim SM, Choi J-H, Chang S-A, Choe YH. Detection of ischaemic myocardial lesionswith coronary CT angiography and adenosine-stress dynamic perfusion imagingusing a 128-slice dual-source CT: diagnostic performance in comparison with car-diac MRI. Br J Radiol. 2013;86:20130481.

58. Gramer BM, Muenzel D, Leber V, et al. Impact of iterative reconstruction on CNRand SNR in dynamic myocardial perfusion imaging in an animal model. Eur Radiol.2012;22:2654–2661.

59. George RT, Jerosch-Herold M, Silva C, et al. Quantification of myocardial perfusionusing dynamic 64-detector computed tomography. Investig Radiol.2007;42:815–822.

60. Newhouse JH, Murphy RX. Tissue distribution of soluble contrast: effect of dosevariation and changes with time. Am J Roentgenol. 1981;136:463–467.

61. Bamberg F, Hinkel R, Schwarz F, et al. Accuracy of dynamic computed tomographyadenosine stress myocardial perfusion imaging in estimating myocardial blood flowat various degrees of coronary artery stenosis using a porcine animal model. InvestigRadiol. 2012;47:71–77.

62. Rossi A, Uitterdijk A, Dijkshoorn M, et al. Quantification of myocardial blood flowby adenosine-stress CT perfusion imaging in pigs during various degrees of stenosiscorrelates well with coronary artery blood flow and fractional flow reserve.European Heart Journal - Cardiovascular Imaging. 2013;14:331–338.

63. Huber AM, Leber V, Gramer BM, et al. Myocardium: dynamic versus single-shot CTperfusion imaging. Radiology. 2013;269:378–386.

64. Enjilela E, Lee T-Y, Hsieh J, et al. Ultra-low dose quantitative CT myocardial per-fusion imaging with sparse-view dynamic acquisition and image reconstruction: afeasibility study. Int J Cardiol. 2018;254:272–281.

65. Machida H, Tanaka I, Fukui R, et al. Current and novel imaging techniques incoronary CT. RadioGraphics. 2015;35:991–1010.

66. Stenner P, Schmidt B, Allmendinger T, Flohr T, Kachelrie M. Dynamic iterativebeam hardening correction (DIBHC) in myocardial perfusion imaging using con-trast-enhanced computed tomography. Investig Radiol. 2010;45:314–323.

67. Rodriguez-Granillo GA, Carrascosa P, Cipriano S, et al. Beam hardening artifactreduction using dual energy computed tomography: implications for myocardialperfusion studies. Cardiovasc Diagn Ther. 2015;5:79–85.

68. Bhave NM, Mor-Avi V, Kachenoura N, et al. Analysis of myocardial perfusion fromvasodilator stress computed tomography: does improvement in image quality byiterative reconstruction lead to improved diagnostic accuracy? Journal ofCardiovascular Computed Tomography. 2014;8:238–245.

69. Nance JWJ, Bastarrika G, Kang DK, et al. high-temporal resolution dual-energycomputed tomography of the heart using a novel hybrid image reconstruction al-gorithm: initial experience. J Comput Assist Tomogr. 2011;35:119–125.

70. Lenga L, Albrecht MH, Othman AE, et al. Monoenergetic dual-energy computedtomographic imaging: cardiothoracic applications. J Thorac Imaging.2017;32:151–158.

71. Ruzsics B, Lee HPE, Flohr TG, Costello P, Schoepf UJ. Images in cardiovascularmedicine. Myocardial ischemia diagnosed by dual-energy computed tomography:correlation with single-photon emission computed tomography. Circulation.2008;117:1244–1245.

72. Koonce JD, Vliegenthart R, Schoepf UJ, et al. Accuracy of dual-energy computedtomography for the measurement of iodine concentration using cardiac CT proto-cols: validation in a phantom model. Eur Radiol. 2014;24:512–518.

73. Leipsic J, Abbara S, Achenbach S et al. SCCT guidelines for the interpretation andreporting of coronary CT angiography: a report of the Society of CardiovascularComputed Tomography Guidelines Committee. J cardiovasc Comp Tomogr;8:342-358.

74. Mehra VC, Valdiviezo C, Arbab-Zadeh A et al. A stepwise approach to the visualinterpretation of CT-based myocardial perfusion. Journal of cardiovascular com-puted tomography;5:357-369.

75. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmen-tation and nomenclature for tomographic imaging of the heart. A statement forhealthcare professionals from the cardiac imaging committee of the council onclinical cardiology of the American Heart Association. Circulation. 2002:539–542.

76. Ko BS, Cameron JD, Meredith IT, et al. Computed tomography stress myocardialperfusion imaging in patients considered for revascularization: a comparison withfractional flow reserve. Eur Heart J. 2012;33:67–77.

77. Kachenoura N, Lodato J, Gaspar T, et al. Value of multidetector computed tomo-graphy evaluation of myocardial perfusion in the assessment of ischemic heartdisease: comparison with nuclear perfusion imaging. Eur Radiol.2009;19:1897–1905.

78. Kachenoura N, Gaspar T, Lodato JA, et al. Combined assessment of coronaryanatomy and myocardial perfusion using multidetector computed tomography forthe evaluation of coronary artery disease. Am J Cardiol. 2009;103:1487–1494.

79. Ingrisch M, Sourbron S. Tracer-kinetic modeling of dynamic contrast-enhanced MRIand CT: a primer. J Pharmacokinet Pharmacodyn. 2013;40:281–300.

80. Pelgrim GJ, Handayani A, Dijkstra H, et al. Quantitative myocardial perfusion withdynamic contrast-enhanced imaging in MRI and CT: theoretical models and currentimplementation. BioMed Res Int. 2016;2016:12.

81. So A, Lee T-Y. Quantitative myocardial CT perfusion: a pictorial review and thecurrent state of technology development. J Cardiovasc Comp Tomog.

2011;5:467–481.82. Ichihara T, George RT, Silva C, Lima JAC, Lardo AC. Quantitative analysis of first-

pass contrast-enhanced myocardial perfusion multidetector CT using a patlak plotmethod and extraction fraction correction during adenosine stress. IEEE Trans NuclSci. 2011;58:133–138.

83. Rocha-Filho JA, Blankstein R, Shturman LD, et al. Incremental value of adenosine-induced stress myocardial perfusion imaging with dual-source CT at cardiac CTangiography. Radiology. 2010;254:410–419.

84. Okada DR, Ghoshhajra BB, Blankstein R et al. Direct comparison of rest and ade-nosine stress myocardial perfusion CT with rest and stress SPECT. J NuclCardiol;17:27-37.

85. Cury RC, Magalhaes TA, Borges AC, et al. Dipyridamole stress and rest myocardialperfusion by 64-detector row computed tomography in patients with suspectedcoronary artery disease. Am J Cardiol. 2010;106:310–315.

86. Ko SM, Choi JW, Song MG, et al. Myocardial perfusion imaging using adenosine-induced stress dual-energy computed tomography of the heart: comparison withcardiac magnetic resonance imaging and conventional coronary angiography. EurRadiol. 2011;21:26–35.

87. Weininger M, Schoepf UJ, Ramachandra A, et al. Adenosine-stress dynamic real-time myocardial perfusion CT and adenosine-stress first-pass dual-energy myo-cardial perfusion CT for the assessment of acute chest pain: initial results. Eur JRadiol. 2012;81:3703–3710.

88. Ko BS, Cameron JD, Leung M, et al. Combined CT coronary angiography and stressmyocardial perfusion imaging for hemodynamically significant stenoses in patientswith suspected coronary artery disease: a comparison with fractional flow reserve.JACC Cardiovasc Imaging. 2012;5:1097–1111.

89. Nasis A, Ko BS, Leung MC, et al. Diagnostic accuracy of combined coronary an-giography and adenosine stress myocardial perfusion imaging using 320-detectorcomputed tomography: pilot study. Eur Radiol. 2013;23:1812–1821.

90. Rochitte CE, George RT, Chen MY, et al. Computed tomography angiography andperfusion to assess coronary artery stenosis causing perfusion defects by singlephoton emission computed tomography: the CORE320 study. Eur Heart J.2013;35(17):1120–1130. https://doi.org/10.1093/eurheartj/eht488.

91. Cury R, Kitt TM, Feaheny K, Akin J, George RT. A randomized, multi-center, multi-vnedor study comparing myocardial perfusion imaging with regadenoson stresscomputed tomography perfusion and single photon emission computed tomo-graphy. Journal of Cardiovascular Computed Tomography. 2013;7:S42.

92. Bastarrika G, Ramos-Duran L, Rosenblum MA, Kang DK, Rowe GW, Schoepf UJ.Adenosine-stress dynamic myocardial CT perfusion imaging: initial clinical ex-perience. Investig Radiol. 2010;45:306–313.

93. Ho KT, Chua KC, Klotz E, Panknin C. Stress and rest dynamic myocardial perfusionimaging by evaluation of complete time-attenuation curves with dual-source CT.JACC Cardiovasc Imaging. 2010;3:811–820.

94. Kono AK, Coenen A, Lubbers M, et al. Relative myocardial blood flow by dynamiccomputed tomographic perfusion imaging predicts hemodynamic significance ofcoronary stenosis better than absolute blood flow. Investig Radiol. 2014;49:801–807.

95. Bamberg F, Becker A, Schwarz F, et al. Detection of hemodynamically significantcoronary artery stenosis: incremental diagnostic value of dynamic CT-based myo-cardial perfusion imaging. Radiology. 2011;260:689–698.

96. Kurata A, Kawaguchi N, Kido T, et al. Qualitative and quantitative assessment ofadenosine triphosphate stress whole-heart dynamic myocardial perfusion imagingusing 256-slice computed tomography. PLoS One. 2013;8:e83950.

97. Bamberg F, Marcus RP, Becker A, et al. Dynamic myocardial CT perfusion imagingfor evaluation of myocardial ischemia as determined by MR imaging. JACCCardiovasc Imaging. 2014;7:267–277.

98. Wang Y, Qin L, Shi X, et al. Adenosine-stress dynamic myocardial perfusion imagingwith second-generation dual-source CT: comparison with conventional cathetercoronary angiography and SPECT nuclear myocardial perfusion imaging. AJR Am JRoentgenol. 2012;198:521–529.

99. Ebersberger U, Marcus RP, Schoepf UJ, et al. Dynamic CT myocardial perfusionimaging: performance of 3D semi-automated evaluation software. Eur Radiol.2014;24:191–199.

100. Greif M, von Ziegler F, Bamberg F, et al. CT stress perfusion imaging for detection ofhaemodynamically relevant coronary stenosis as defined by FFR. Heart.2013;99:1004–1011.

101. Rossi A, Dharampal A, Wragg A, et al. Diagnostic performance of hyperaemicmyocardial blood flow index obtained by dynamic computed tomography: does itpredict functionally significant coronary lesions? Eur Heart J Cardiovasc Imaging.2014;15:85–94.

102. Ruzsics B, Schwarz F, Schoepf UJ, et al. Comparison of dual-energy computed to-mography of the heart with single photon emission computed tomography for as-sessment of coronary artery stenosis and of the myocardial blood supply. Am JCardiol. 2009;104:318–326.

103. Nagao M, Kido T, Watanabe K, et al. Functional assessment of coronary artery flowusing adenosine stress dual-energy CT: a preliminary study. Int J Cardiovasc Imaging.2011;27:471–481.

104. Wang R, Yu W, Wang Y, et al. Incremental value of dual-energy CT to coronary CTangiography for the detection of significant coronary stenosis: comparison withquantitative coronary angiography and single photon emission computed tomo-graphy. Int J Cardiovasc Imaging. 2011;27:647–656.

105. Meyer M, Nance JW, Schoepf UJ, et al. Cost-effectiveness of substituting dual-en-ergy CT for SPECT in the assessment of myocardial perfusion for the workup ofcoronary artery disease. Eur J Radiol. 2012;81:3719–3725.

106. Ko SM, Choi JW, Hwang HK, Song MG, Shin JK, Chee HK. Diagnostic performanceof combined noninvasive anatomic and functional assessment with dual-source CTand adenosine-induced stress dual-energy CT for detection of significant coronary

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

99

stenosis. Am J Roentgenol. 2012;198:512–520.107. Weininger M, Schoepf UJ, Ramachandra A, et al. Adenosine-stress dynamic real-

time myocardial perfusion CT and adenosine-stress first-pass dual-energy myo-cardial perfusion CT for the assessment of acute chest pain: initial results. Eur JRadiol. 2012;81:3703–3710.

108. Delgado C, Vázquez M, Oca R, Vilar M, Trinidad C, Sanmartin M. Myocardialischemia evaluation with dual-source computed tomography: comparison withmagnetic resonance imaging. Revista Española de Cardiología (English Edition).2013;66:864–870.

109. Ko SM, Park JH, Hwang HK, Song MG. Direct comparison of stress- and rest-dual-energy computed tomography for detection of myocardial perfusion defect. Int JCardiovasc Imaging. 2014;30:41–53.

110. Ko SM, Song MG, Chee HK, Hwang HK, Feuchtner GM, Min JK. Diagnostic per-formance of dual-energy CT stress myocardial perfusion imaging: direct comparisonwith cardiovascular MRI. Am J Roentgenol. 2014;203:W605–W613.

111. Kim SM, Chang S-A, Shin W, Choe YH. Dual-energy CT perfusion during pharma-cologic stress for the assessment of myocardial perfusion defects using a second-generation dual-source CT: a comparison with cardiac magnetic resonance imaging.J Comput Assist Tomogr. 2014;38:44–52.

112. Zhao R-P, Hao Z-R, Song Z-J. Diagnostic value of Flash dual-source CT coronary

artery imaging combined with dual-energy myocardial perfusion imaging for cor-onary heart disease. Exper Ther Med. 2014;7:865–868.

113. Kido T, Watanabe K, Saeki H, et al. Adenosine triphosphate stress dual-sourcecomputed tomography to identify myocardial ischemia: comparison with invasivecoronary angiography. SpringerPlus. 2014;3:75.

114. De Cecco CN, Harris BS, Schoepf UJ, et al. Incremental value of pharmacologicalstress cardiac dual-energy CT over coronary CT angiography alone for the assess-ment of coronary artery disease in a high-risk population. Am J Roentgenol.2014;203:W70–W77.

115. Delgado Sánchez-Gracián C, Oca Pernas R, Trinidad López C, et al. Quantitativemyocardial perfusion with stress dual-energy CT: iodine concentration differencesbetween normal and ischemic or necrotic myocardium. Initial experience. EurRadiol. 2016;26:3199–3207.

116. Bittencourt MS, Hulten E, Polonsky TS, et al. European Society of Cardiology-re-commended coronary artery disease consortium pretest probability scores moreaccurately predict obstructive coronary disease and cardiovascular events than theDiamond and Forrester Score: the partners registry. Circulation. 2016;134:201–211.

117. Yang J, Dou J, He B, et al. Stress myocardial blood flow ratio by dynamic CTperfusion identifies hemodynamically significant CAD. JACC Cardiovasc Imaging.2019.

A.R. Patel, et al. Journal of Cardiovascular Computed Tomography 14 (2020) 87–100

100