Author + information
- Received October 16, 2017
- Revision received November 20, 2017
- Accepted November 24, 2017
- Published online January 29, 2018.
- Roel S. Driessen, MDa,
- Wijnand J. Stuijfzand, MDa,
- Pieter G. Raijmakers, MD, PhDb,
- Ibrahim Danad, MDa,
- James K. Min, MD, PhDc,
- Jonathon A. Leipsic, MD, PhDd,
- Amir Ahmadi, MD, PhDe,
- Jagat Narula, MD, PhDe,
- Peter M. van de Ven, PhDf,
- Marc C. Huisman, PhDb,
- Adriaan A. Lammertsma, PhDb,
- Albert C. van Rossum, MD, PhDa,
- Niels van Royen, MD, PhDa and
- Paul Knaapen, MD, PhDa,∗ ()
- aDepartment of Cardiology, VU University Medical Center, Amsterdam, the Netherlands
- bDepartment of Radiology, Nuclear Medicine & PET Research, VU University Medical Center, Amsterdam, the Netherlands
- cInstitute for Cardiovascular Imaging, Weill-Cornell Medical College, New York-Presbyterian Hospital, New York, New York
- dDepartment of Medicine and Radiology, University of British Columbia, Vancouver, Canada
- eDivision of Cardiology, Icahn School of Medicine at Mount Sinai, New York, New York
- fDepartment of Epidemiology and Biostatistics, VU University Medical Center, Amsterdam, the Netherlands
- ↵∗Address for correspondence:
Dr. Paul Knaapen, Department of Cardiology, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands.
Background Atherosclerotic plaque characteristics may affect downstream myocardial perfusion, as well as coronary lesion severity.
Objectives This study sought to evaluate the association between quantitative plaque burden and plaque morphology obtained using coronary computed tomography angiography (CTA) and quantitative myocardial perfusion obtained using [15O]H2O positron emission tomography (PET), as well as fractional flow reserve (FFR) derived invasively.
Methods Two hundred eight patients (63% men; age 58 ± 8.7 years) with suspected coronary artery disease were prospectively included. All patients underwent 256-slice coronary CTA, [15O]H2O PET, and invasive FFR measurements. Coronary CTA-derived plaque burden and morphology were assessed using commercially available software and compared with PET perfusion and FFR.
Results Atherosclerotic plaques were present in 179 patients (86%) and 415 of 610 (68%) evaluable coronary arteries. On a per-vessel basis, traditional coronary plaque burden indexes, such as plaque length and volume, minimal lumen area, and stenosis percentage, were significantly associated with impaired hyperemic myocardial blood flow (MBF) and FFR. In addition, morphological features, such as partially calcified plaques, positive remodeling (PR), and low attenuation plaque, displayed a negative impact on hyperemic MBF and FFR. Multivariable analysis revealed that the morphological feature of PR was independently related to impaired hyperemic MBF as well as an unfavorable FFR (p = 0.004 and p = 0.007, respectively), next to stenosis percentage (p = 0.001 and p < 0.001, respectively) and noncalcified plaque volume (p < 0.001 and p = 0.010, respectively).
Conclusions PR and noncalcified plaque volume are associated with detrimental downstream hyperemic myocardial perfusion and FFR, independent of lesion severity.
- coronary artery disease
- coronary computed tomography angiography
- fractional flow reserve
- myocardial perfusion
- positron emission tomography
Coronary computed tomography angiography (CTA) is an established tool for the noninvasive assessment of coronary artery stenosis (1–4). However, the relationship between anatomic coronary stenosis severity and myocardial ischemia is complex. Whereas approximately one-fifth of ≥70% stenotic lesions do not cause ischemia, attenuation of flow can occur even in apparently nonobstructive coronary arteries due to diffuse atherosclerosis (5–9). Based on an increasing awareness of this discrepancy, other atherosclerotic factors are under investigation to improve the prediction of ischemia.
In line with the search beyond traditional stenosis severity, coronary CTA also allows for noninvasive evaluation of quantitative atherosclerotic plaque burden and plaque morphology, which closely matches results obtained using invasive intravascular ultrasound and optical coherence tomography (4,10–14). Although adverse plaque characteristics (APCs) such as positive remodeling (PR) and low attenuation are linked to plaque rupture and acute coronary syndromes (15), plaque morphology has also been associated with hemodynamic consequences of those lesions (16–19). Consequently, the additive value of coronary CTA-derived atherosclerotic plaque burden and morphology might be beneficial for both diagnostic and prognostic purposes. Therefore, the aim of this study was to investigate the associations of coronary CTA-derived plaque burden and morphology, myocardial perfusion as assessed with [15O]H2O positron emission tomography (PET), as well as the clinical reference standard for coronary revascularization: fractional flow reserve (FFR).
This was a post hoc substudy comprising all patients from the PACIFIC (Prospective Comparison of Cardiac PET/CT, SPECT/CT Perfusion Imaging and CT Coronary Angiography With Invasive Coronary Angiography; NCT01521468) trial involving 208 patients suspected of having coronary artery disease (CAD), who underwent coronary CTA and PET with interrogation by FFR of all coronary arteries (20). Patients suspected of having stable CAD with an intermediate pre-test likelihood of CAD and normal left ventricular function were included. Exclusion criteria incorporated a documented history of CAD, signs of previous myocardial infarction, atrial fibrillation, renal failure, second- or third-degree atrioventricular block, symptomatic asthma, or pregnancy. The study complied with the principles of the Declaration of Helsinki, and the study protocol was approved by the VUMC Medical Ethics Review Committee. All patients provided written informed consent.
Coronary computed tomography angiography
All patients underwent coronary CTA using a 256-slice computed tomography (CT) scanner (Philips Brilliance iCT, Philips Healthcare, Best, the Netherlands). Sublingual nitroglycerine was administered to all patients before the scanning protocol. Metoprolol was administered only if necessary, aiming for a heart rate <65 beats/min. CT scans were performed using a standard scanning protocol, with section collimation of 2 × 128 × 0.625 mm, gantry rotation time of 270 ms, and tube current between 200 and 360 mA at 120 kV, primarily adjusting current based on body habitus. The scan was triggered, using an automatic bolus-tracking technique with a region of interest placed in the descending thoracic aorta, at a threshold of 150 Hounsfield units (HU). Prospective electrocardiogram gating (Step & Shoot Cardiac, Philips Healthcare) was used at 75% of the R-R interval. In case of persistent heart rate >65 beats/min, 4 scans were performed using a retrospective helical protocol. An intravenous 100-ml bolus of iodinated contrast agent (XenetiX 350, Guerbet, Brussels, Belgium) was injected at a flow rate of 5.7 ml/s, directly followed by a 50-ml saline flush.
Assessment of coronary plaques
Using dedicated semiautomated software (Comprehensive Cardiac Analysis, Philips Healthcare), all coronary segments with a diameter ≥2 mm were assessed by an experienced reader blinded to both PET and FFR results. The coronary tree was evaluated according to a 17-segment coronary artery model using axial, multiplanar reformation, maximum intensity projection, and cross-sectional images (slice thickness 0.9 mm, increment 0.50 mm). Manual corrections to centerline and vessel contours by the reader were allowed. Stenosis severity was graded as 0%, 1% to 50%, 50% to 70%, or >70%. Quantitative coronary atherosclerotic burden was assessed automatically within manually designated regions. Total plaque lengths and volumes were calculated per vessel by summing the lengths and volumes of separate plaques along each coronary artery. A scanner-specific threshold of 150 HU was used for distinguishing noncalcified from calcified plaque components, which subsequently was used to determine noncalcified and calcified plaque volumes. At the site of maximal stenosis, lumen area, wall area, and maximal plaque burden (plaque area/vessel area × 100%) were assessed in the short axis of the vessel. Coronary lesions were also analyzed for qualitative atherosclerotic plaque characteristics. The remodeling index was computed as the ratio of vessel area at the site of the maximal lesion to that of a proximal reference point, with an index >1.1 representing PR (17). Low-attenuation plaque (LAP) was defined as a plaque containing any voxel <30 HU (17). Spotty calcification (SC) was characterized by a calcified plaque composing <90° of the vessel circumference and <3 mm in length (17). Napkin ring sign was defined by a plaque core with low CT attenuation surrounded by a rim-like area of higher CT attenuation (21).
Positron emission tomography
Patients were scanned using a hybrid PET-CT device (Philips Gemini TF 64, Philips Healthcare). The scanning protocol has been described in detail previously (22). In summary, a dynamic PET perfusion scan was performed during resting conditions using 370 MBq of [15O]H2O. A 6-min emission scan was started simultaneously with the administration of [15O]H2O. This dynamic scan sequence was followed immediately by a low-dose CT scan for attenuation correction. After a 10-min interval to allow for decay of radioactivity, an identical PET sequence was performed during hyperemic conditions induced by an intravenous adenosine infusion (140 μg/kg/min), initiated 2 min before the stress scan for maximal vasodilation. Images were reconstructed using the 3-dimensional row action maximum likelihood algorithm and applying all appropriate corrections. Parametric myocardial blood flow (MBF) images were generated and quantitatively analyzed using software developed in-house (Cardiac VUer) (23). MBF was expressed in ml/min/g of perfusable tissue. In addition to calculating both resting and hyperemic MBF for the left ventricle as a whole, MBF was calculated for each of the 3 vascular territories derived from standard segmentation: left anterior descending, left circumflex, and right coronary artery. Coronary flow reserve (CFR) was defined as the ratio between hyperemic and resting MBF. For comparison with CT characteristics, hyperemic MBF was predominantly used because of its greater ability to identify ischemia compared with resting MBF or CFR using [15O]H2O as a tracer (22,24).
Invasive coronary angiography and FFR measurements
Selective invasive coronary angiography (ICA) was performed by standard catheterization in accordance with the American College of Cardiology recommendations for coronary angiography (25). Intracoronary nitroglycerin was administered before contrast injection, inducing epicardial coronary vasodilation. All major coronary arteries and side branches >2.0 mm were interrogated with FFR. Operators refrained from FFR measurement in tight lesions >90% to avoid the risk of a coronary dissection by the pressure wire. FFR was measured using a 0.014-inch sensor-tipped guidewire, introduced through a 5- or 6-F guiding catheter. To induce maximal coronary hyperemia, adenosine was infused by either intravenous or intracoronary administration at a dose of 140 μg/kg/min or 150 μg, respectively. FFR was calculated as the ratio of the mean distal intracoronary pressure, measured by the pressure wire and the mean arterial pressure measured by the coronary catheter (26). FFR ≤0.80 was considered hemodynamically significant (5).
Continuous variables are given as mean ± SD and categorical variables as number and percentage. Mixed effects models and generalized estimating equations (GEEs) were used in SPSS version 20.0 (IBM Corp., Armonk, New York) to identify factors associated with reduced hyperemic MBF, CFR, and FFR ≤0.80, respectively. Mixed effects models and GEEs were used rather than standard linear regression and logistic regression to account for correlation of outcomes between multiple vessels within the same patient. Mixed effects models included a random effect for patients, and the GEEs assumed an exchangeable correlation matrix to account for within-patient correlation. Predicted values for the outcomes based on the fixed parts of the obtained models were exported to MedCalc version 11.0 (MedCalc Software, Ostend, Belgium) to perform the receiver-operating characteristic (ROC) analyses. MBF and CFR data were logarithmically transformed to correct for rightward skew. All candidate predictors were first tested for association in univariable analyses. Additional analyses were performed with correction for luminal stenosis severity based on 4 pre-defined categories: no stenosis, 0% to 50% stenosis, 50% to 70% stenosis, and >70% stenosis. Subsequently, a forward selection procedure was to find sets of independent predictors for MBF and FFR values. Only predictors with p < 0.10 in univariable analyses were considered in the multivariable analysis. Forward selection was used rather than backward elimination because the latter required estimation of a model containing all predictors, which was hampered by multicollinearity. Regression coefficient and odds ratio with 95% confidence interval are reported as effect sizes for mixed effects models and GEEs, respectively. One-way analysis of variance with Bonferroni correction for multiple pairwise comparisons was used to compare differences in MBF between particular numbers of APCs as well as plaque features between subgroups of combined MBF, CFR, and FFR. In addition, area under the curve (AUC) was used to evaluate the additive value of plaque characteristics next to stenosis severity in predicting ischemia (FFR ≤0.80). Correlations of both reference standards and plaque features were compared using the William test. Interobserver reproducibility of plaque characteristics was quantified using the intraclass correlation coefficient (ICC). A 2-sided p < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS and MedCalc software.
The final study population comprised 204 patients and 610 evaluable coronary arteries, after 14 vessels were excluded because of nonevaluable coronary CTA scans. The median number of evaluable arteries per patient was 3 (range 2 to 3). A total of 596 of these coronary arteries were interrogated by FFR or showed (sub)total occlusion on ICA. Vascular territories could be related to PET-derived MBF in 203 patients. Baseline characteristics are given in Table 1. Overall, 303 vessels (50%) were found to be obstructive with coronary stenosis ≥50% on coronary CTA, whereas invasive FFR was ≤0.80 in only 158 coronary arteries (26%). Table 2 provides an overview of CT findings in the total population.
Quantitative plaque burden and morphology related to hyperemic MBF
The association between plaque characteristics and hyperemic MBF is given in Table 3. Using univariable regression analysis, all quantitative plaque parameters showed a significant association with reduced hyperemic MBF. With respect to plaque morphology, only the presence of noncalcified plaques and calcified plaques was not significantly related to MBF. After adjusting for luminal stenosis, all quantitative plaque parameters, except lumen area, still predicted a significantly lower MBF. Likewise, the morphological features LAP, PR, and SC remained significantly related. Upon subsequent use of multivariable regression analysis with forward selection, only noncalcified plaque volume (NCPV) and PR persisted to be significantly associated with impaired hyperemic MBF, next to stenosis severity. Multivariable analysis to predict PET-derived CFR also revealed NCPV and PR as independent predictors next to stenosis severity (Online Table 1). Figure 1 shows the gradual reduction in hyperemic MBF in relation to the number of adverse plaque characteristics.
Quantitative plaque burden and morphology related to FFR
Table 4 lists the results of regression analysis between plaque characteristics and FFR. Univariable analysis showed that all quantitative and qualitative plaque characteristics had a significant relationship with FFR. Applying an adjustment for luminal stenosis excluded calcified plaque volume, wall area, and noncalcified and calcified plaques along with napkin ring sign as significant predictors. Multivariable analysis yielded inverse relationships with FFR for noncalcified plaque volume, LAP, PR, and SC next to luminal stenosis severity.
Additive value of plaque characteristics next to stenosis severity in identifying ischemia
Next to stenosis severity, the only plaque characteristics that independently predicted both impaired hyperemic MBF and FFR were noncalcified plaque volume and PR. Whereas LAP and SC remained significantly related to FFR, this was not the case for the comparison with hyperemic MBF (Central Illustration). Noncalcified plaque volume affected perfusion impairment and a significant reduction in FFR comparably (p = 0.29 for difference in correlation). However, PR and luminal stenosis had a greater impact on a significant hyperemic pressure drop than on impaired perfusion (p < 0.001 and p = 0.005, respectively).
Figure 2 shows the quantitative relationship between MBF and FFR; Online Figure 1 shows the relationship for PET CFR and FFR. Most plaque features were significantly detrimental for concordant abnormal MBF and FFR results compared with concordant normal MBF and FFR results (Online Table 2). Among discrepancy subgroups, impaired FFR in combination with normal MBF was characterized by higher-grade stenosis severity, smaller lumen areas, and more severe maximal plaque burden compared with the group with preserved FFR in the presence of impaired MBF. Conversely, the latter group displayed smaller plaque lengths and similar plaque volumes. Plaque characteristics, however, did not significantly differ between the subgroups with discrepant hyperemic MBF and CFR results (Online Figure 2, Online Table 3).
Figure 3 shows ROC analyses for identifying a positive FFR (≤0.80). Compared with luminal stenosis alone, the addition of noncalcified plaque volume and, subsequently, PR improved the predictive value for FFR ≤0.80. Accordingly, AUC increased from 0.86 to 0.89 and 0.90, respectively. Figure 4 shows an example of an ischemia-causing lesion that was not obstructive, yet showed multiple plaque characteristics (noncalcified plaque with LAP and PR).
A total number of 21 patients and 63 vessels (10% of study population) were reanalyzed to test interobserver variability. For stenosis severity, interobserver reproducibility expressed as ICC was 0.78. Agreement was also good for plaque features NCPV and PR, with ICC of 0.81 and 0.66, respectively.
The present substudy of the PACIFIC trial was conducted to explore whether, and if so, to what extent, quantitative plaque burden and morphology are related to hyperemic MBF in patients undergoing coronary CTA. Various traditional lesion severity indexes such as plaque length and volume showed an inverse relationship with hyperemic MBF, even after adjustment for luminal stenosis. Morphological features such as noncalcified plaque volume and PR also attenuated flow independent of stenosis severity.
Although coronary CTA displays good agreement with ICA, the relatively poor predictive value for ischemia using stenosis severity alone derived from either invasive or noninvasive angiography is well established (5–8). Apparently, hemodynamics are affected not only by luminal stenosis but also by other geometric variables such as plaque length and volume (27,28). Recent studies have suggested that plaque morphology also may affect blood flow (16–18,29). Given the recent technological developments that have advanced the ability of coronary CTA to quantify plaque burden and evaluate plaque characteristics, a comprehensive assessment could potentially improve the discrimination between ischemic and nonischemic CAD.
In line with the majority of previous studies, the present study showed an association between multiple atherosclerotic plaque characteristics and ischemia. Some of the previous studies used myocardial perfusion imaging as the reference method. For example, Naya et al. (16) reported only a modest effect of plaque geometry and morphology on CFR using 82Rb PET. Interestingly, PR was not correlated with myocardial perfusion when a multivariable model was used, which is in contrast with the present study. Partially calcified plaques, however, were associated with impaired myocardial flow reserve. A study by Shmilovich et al. (29) described that the presence of LAP and PR in severely stenotic plaques on coronary CTA was predictive of myocardial hypoperfusion. Their finding of PR as an independent predictor of hypoperfusion is in agreement with the current study using [15O]H2O PET. A reason why this does not completely account for LAP could be that NCPV is a more robust feature and an even stronger predictor containing both quantitative and morphological features. Furthermore, the fact that total plaque volume was not incorporated in the present multivariable models, which could at least partially be explained by collinearity with NCPV, does not detract from its intuitively important role in causing ischemia and cardiac events.
In addition to the assessment of myocardial perfusion using PET, the present study included FFR measurements. The latter parameter is considered a more clinical useful tool to guide revascularization decision making, which has been shown to improve outcome (30). NCPV, LAP, PR, and SC seemed to be independently associated with a significant reduction in FFR. A few other studies have investigated the relationship between plaque morphology and FFR. Nakazato et al. (27) found that aggregated plaque volume demonstrated incremental discriminatory power over luminal stenosis for lesions with FFR ≤0.80. This is in agreement with the present findings, especially for the noncalcified part of the plaque volume. Other plaque characteristics, however, were not evaluated. Park et al. (17) reported that PR, LAP, percentage aggregated plaque volume, and the number of APCs were independent predictors of ischemia using FFR as a reference. In contrast to those 2 studies, in the current study SC also seemed to be significantly associated with adverse FFR. Although the analysis by Nakazato et al. (27) and Park et al. (17) did not include specific plaque composition volumes (i.e., calcified and noncalcified), a recently published substudy of the NXT (Analysis of Coronary Blood Flow Using CT Angiography: Next Steps) trial did differentiate calcified and noncalcified plaque volumes (18). Compared to the present study, similar predictors of ischemia were found, including NCPV, plaque length, and particularly low-density NCPV (18).
Although several studies, including the present one, observed interactions between multiple plaque characteristics and ischemia, the underlying mechanisms have not yet been clarified. A number of potential pathophysiological causes might, at least in part, explain the connection between certain plaque characteristics and reduced MBF. First, the presence of (compensatory) PR might represent a stage of atheroma development during which the atherosclerotic burden exceeds a certain threshold that is associated with ischemia. This is supported by findings of De Bruyne et al. (9), who reported that early-stage coronary atherosclerosis is often associated with abnormal resistance of the coronary arteries, even before angiography reveals a focal high-grade stenosis. PR together with LAP and SC may be anatomic markers of focal or systemic pathophysiological processes. These processes include vascular inflammation, endothelial dysfunction, and altered shear stress patterns, which are associated with the inability of coronary arteries to vasodilate and consequently cause ischemia (31,32). Previous studies have reported an association between coronary endothelial dysfunction and necrotic core plaques (33), which can be recognized as LAP and PR using coronary CTA. PR had a greater impact on a reduction in FFR than in hyperemic MBF. It could imply that this adverse plaque morphology has a closer agreement specifically with epicardial pathology than overall downstream consequences, as FFR tends to evaluate the epicardial lesion specific hemodynamics, whereas MBF using quantitative PET reflects both the epicardial and microvascular bed. However, it could be postulated that microvascular disease, as defined by abnormal PET despite a normal FFR, cannot be fully determined by these coronary CTA abnormalities (Online Table 2).
Combining both endpoints of hyperemic MBF and FFR revealed that noncalcified plaque volume and PR affected both parameters in an independent manner. ROC analysis indicated that adding these morphological features to diameter stenosis alone improved diagnostic accuracy for obstructive CAD taking FFR as a reference. Although statistically significant, the incremental value appeared to be only modest starting from a high baseline AUC value (AUC increased from 0.86 to 0.90; p < 0.001). As can be derived from ROC analysis, with a fixed sensitivity of 88.2%, specificity increased from 64.9% for stenosis severity alone to 72.1% when adding NCPV and PR. Using a second fixed sensitivity of 65.2%, specificity merely increased from 92.2% to 93.5% with the addition of the aforementioned plaque features. The increase in diagnostic accuracy in absolute values may appear modest but could be considered of clinical significance. This is further highlighted in Figure 1, which shows great overlap between blood flow in arteries with and without adverse plaque characteristics. In other words, although addition of adverse plaque characteristics may hold prognostic value and enhance diagnostic accuracy for identifying myocardial ischemia, at present the clinical relevance remains limited.
In the present study, stenosis severity was graded categorically, as commonly used in clinical practice. However, the use of stenosis severity as a continuous variable could have provided additional information and possibly altered the current findings. Similarly, dichotomized FFR was used in present models, according to an established threshold with proven prognostic value (30). For statistical and methodological reasons, continuous FFR values could not be used in the current analysis, even though it might have provided additional insights into the relationship between plaque and FFR. As a consequence, the difference in association of plaque features with dichotomized FFR and continuous MBF should be interpreted with caution given the slightly different statistical approach. Furthermore, some predictors, such as lumen area and maximal plaque burden, were found to be associated with perfusion in univariable analyses but were also strongly associated with stenosis severity. Therefore, they were not included in the final multivariable model because their predictive value when added to stenosis severity was only small. Another limitation involves standard segmentation for vascular territories, which was used to derive MBF data from the PET scans. Therefore, individual variations of coronary anatomy and analysis of atherosclerotic major side branches by FFR and coronary CTA may not exactly coincide with the assigned PET segments and may have diluted the potential influence of coronary lesions on flow parameters. Finally, it should be noted that reproducibility of the semiautomated plaque analysis remains suboptimal with modest ICCs of 0.66 to 0.81.
Coronary CTA-derived morphological adverse plaque characteristics are associated with detrimental hyperemic myocardial perfusion and invasive FFR values, independent of lesion severity. Including these atherosclerotic plaque characteristics in a comprehensive coronary CTA evaluation modestly improves the prediction of ischemic CAD compared with luminal stenosis assessment alone.
COMPETENCY IN MEDICAL KNOWLEDGE: Positive remodeling and noncalcified coronary atherosclerotic plaque volume reduce both MBF and FFR independent of luminal stenosis, although these plaque characteristics have limited additional value as predictors of ischemia.
TRANSLATIONAL OUTLOOK: Additional studies are required to clarify the mechanisms underlying the associations between plaque morphology and myocardial perfusion.
Dr. Min serves as a consultant to HeartFlow and Abbott Vascular; serves on the scientific advisory board of Arineta; and has an equity interest in MDDX. Dr. Leipsic has core laboratory contracts with Edwards Lifesciences for which he receives no direct compensation; and has served as a consultant for and received stock options from Circle CVI and HeartFlow. Dr. Lammertsma has received research grants from AVID, Philips Healthcare, F. Hoffmann La Roche Ltd., and the European Commission. Dr. van Royen has received educational grants from Abbott, Baxter, Biotronik, and Philips Healthcare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Ron Blankstein, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- adverse plaque characteristic
- area under the curve
- coronary artery disease
- computed tomography angiography
- coronary flow reserve
- computed tomography
- fractional flow reserve
- generalized estimating equation
- Hounsfield unit
- invasive coronary angiography
- intraclass correlation coefficient
- low attenuation plaque
- myocardial blood flow
- noncalcified plaque volume
- positron emission tomography
- positive remodeling
- receiver-operating characteristic
- spotty calcification
- Received October 16, 2017.
- Revision received November 20, 2017.
- Accepted November 24, 2017.
- 2018 American College of Cardiology Foundation
- Meijboom W.B.,
- Meijs M.F.,
- Schuijf J.D.,
- et al.
- Shaw L.J.,
- Hausleiter J.,
- Achenbach S.,
- et al.
- Leber A.W.,
- Knez A.,
- von Z.F.,
- et al.
- Tonino P.A.,
- Fearon W.F.,
- De Bruyne B.,
- et al.
- Meijboom W.B.,
- van Mieghem C.A.,
- van P.N.,
- et al.
- Shaw L.J.,
- Berman D.S.,
- Maron D.J.,
- et al.
- Gould K.L.,
- Johnson N.P.,
- Bateman T.M.,
- et al.
- De Bruyne B.,
- Hersbach F.,
- Pijls N.H.,
- et al.
- Hoffmann U.,
- Moselewski F.,
- Nieman K.,
- et al.
- Pundziute G.,
- Schuijf J.D.,
- Jukema J.W.,
- et al.
- Motoyama S.,
- Ito H.,
- Sarai M.,
- et al.
- Naya M.,
- Murthy V.L.,
- Blankstein R.,
- et al.
- Park H.B.,
- Heo R.,
- Hartaigh B.,
- et al.
- Ahmadi A.,
- Stone G.W.,
- Leipsic J.
- Danad I.,
- Raijmakers P.G.,
- Driessen R.S.,
- et al.
- Maurovich-Horvat P.,
- Hoffmann U.,
- Vorpahl M.,
- Nakano M.,
- Virmani R.,
- Alkadhi H.
- Danad I.,
- Uusitalo V.,
- Kero T.,
- et al.
- Bashore T.M.,
- Balter S.,
- Barac A.,
- et al.
- Nakazato R.,
- Shalev A.,
- Doh J.H.,
- et al.
- Stone P.H.,
- Coskun A.U.,
- Kinlay S.,
- et al.
- Lavi S.,
- Bae J.H.,
- Rihal C.S.,
- et al.