Author + information
- Received February 2, 2017
- Revision received June 12, 2017
- Accepted June 13, 2017
- Published online August 7, 2017.
- Shingo Kato, MD, PhDa,∗ (, )
- Naka Saito, MTa,
- Tatsuya Nakachi, MDa,
- Kazuki Fukui, MDa,
- Tae Iwasawa, MDb,
- Masataka Taguri, PhDc,
- Masami Kosuge, MDd and
- Kazuo Kimura, MDd
- aDepartment of Cardiology, Kanagawa Cardiovascular and Respiratory Center, Yokohama, Kanagawa, Japan
- bDepartment of Radiology, Kanagawa Cardiovascular and Respiratory Center, Yokohama, Kanagawa, Japan
- cDepartment of Biostatistics, Yokohama City University, Yokohama, Kanagawa, Japan
- dDepartment of Cardiology, Yokohama City University Medical Center, Yokohama, Kanagawa, Japan
- ↵∗Address for correspondence:
Dr. Shingo Kato, Department of Cardiology, Kanagawa Cardiovascular and Respiratory Center, 6-16-1, Tomiokahigashi, Tokohama, Kanagawa 2360051, Japan.
Background Phase-contrast (PC) cine magnetic resonance imaging (MRI) of the coronary sinus is a noninvasive method to quantify coronary flow reserve (CFR).
Objectives This study sought to compare the prognostic value of CFR by cardiac magnetic resonance (CMR) and stress perfusion CMR to predict major adverse cardiac events (MACE).
Methods Participants included 276 patients with known coronary artery disease (CAD) and 400 with suspected CAD. CFR was calculated as myocardial blood flow during adenosine triphosphate infusion divided by myocardial blood flow at rest using PC cine MRI of the coronary sinus.
Results During a median follow-up of 2.3 years, 47 patients (7%) experienced MACE. Impaired CFR (<2.0) and >10% ischemia on stress perfusion CMR were significantly associated with MACE in patients with known CAD (hazard ratio [HR]: 5.17 and HR: 5.10, respectively) and suspected CAD (HR: 14.16 and HR: 6.50, respectively). The area under the curve for predicting MACE was 0.773 for CFR and 0.731 for stress perfusion CMR (p = 0.58) for patients with known CAD, and 0.885 for CFR and 0.776 for stress perfusion CMR (p = 0.059) in the group with suspected CAD. In patients with known CAD, sensitivity, specificity, and positive and negative predictive values to predict MACE were 64%, 91%, 38%, and 97%, respectively, for CFR, and 82%, 59%, 15%, and 97%, respectively, for stress perfusion CMR. In the suspected CAD group, these values were 65%, 99%, 80%, and 97%, respectively, for CFR, and 72%, 83%, 22%, and 98%, respectively, for stress perfusion CMR.
Conclusions The predictive values of CFR and stress perfusion CMR for MACE were comparable in patients with known CAD. In patients with suspected CAD, CFR showed higher HRs and areas under the curve than stress perfusion CMR, suggesting that CFR assessment by PC cine MRI might provide better risk stratification for patients with suspected CAD.
- coronary artery disease
- coronary flow reserve
- magnetic resonance imaging
- prognostic value
- stress perfusion CMR
Cardiovascular disease remains the leading global cause of death, accounting for more than 17.3 million deaths per year (1). Invasive x-ray coronary angiography is the gold standard to assess the atherosclerotic burden in patients with coronary artery disease (CAD), but this modality offers only limited ability to predict favorable responses with medical therapy or revascularization. Measurement of fractional flow reserve is an established method to identify functional ischemic vessel (2,3), but it cannot assess coronary microvascular function. Quantitative coronary flow reserve (CFR), as estimated from myocardial perfusion positron-emission tomography (PET), can provide an accurate, noninvasive assessment of the hyperemic reactivity of myocardium to vasodilatory agents. Recent studies have shown that PET-derived CFR represents a strong prognostic marker and can effectively reclassify the risk of future cardiovascular events (4–6). Such studies have shown the importance of assessing CFR for accurate risk stratification of known or suspected CAD patients.
Noncontrast phase-contrast (PC) cine magnetic resonance imaging (MRI) of the coronary sinus has emerged as a noninvasive method for quantifying global left ventricular (LV) myocardial blood flow (MBF) (7–9). Without the use of radioactive tracers, CFR can be calculated as the MBF during vasodilator infusion divided by the MBF at rest by PC cine MRI of the coronary sinus. A previous study showed that CFR measured by PC cine MRI correlates well with CFR measured by PET-derived CFR (7). However, no data are available regarding the prognostic value of CFR quantified by PC cine MRI.
The aims of the present study were to analyze the prognostic value of CFR by PC cine MRI for patients with known or suspected CAD, and to compare the predictive abilities of CFR and stress perfusion cardiac magnetic resonance (CMR) for the occurrence of major adverse cardiac events (MACE).
A total of 709 patients were enrolled between 2009 and 2016 from the Kanagawa Cardiovascular and Respiratory Center. Inclusion criteria included age >18 years and clinical suspicion of myocardial ischemia at the discretion of the referring clinician. Exclusion criteria consisted of contraindications to CMR (e.g., metallic hazards, pregnancy, severe renal dysfunction, and claustrophobia) and contraindications to vasodilator stress testing. A history of CAD was defined as evidence of myocardial infarction (MI), previous percutaneous coronary intervention or coronary artery bypass graft, or angiographically significant coronary stenosis (>70% stenosis in any epicardial coronary artery or >50% of the left main coronary artery). Previous MI was confirmed by definitive clinical evidence in the medical record or presence of pathological Q waves according to published criteria (10). This study was approved by the institutional review board at Kanagawa Cardiovascular and Respiratory Center, and all patients provided informed consent to participate before enrollment in this study.
CMR image acquisition
Acquisition of MRI was performed using a 1.5-T MRI scanner equipped with 32-channel cardiac coils. The CMR protocol is shown in Figure 1. Cine steady-state free-precession (repetition time: 4.1 ms; echo time: 1.7 ms; flip angle: 55°; field of view: 350 × 350 mm; acquisition matrix: 128 × 128; slice thickness: 10 mm; number of phases per cardiac cycle: 20) was acquired for imaging ventricular size and function.
For the accurate configuration of imaging slices for coronary sinus blood flow, cine images in the axial plane were obtained through the atrioventricular groove to detect the location of the coronary sinus. The imaging plane for blood flow measurement by PC cine images was positioned perpendicular to the coronary sinus at 2 cm from the ostium of the coronary sinus on axial cine CMR images (Figure 2A).
PC cine MRI of the coronary sinus was acquired during breath holding using a vector-electrocardiogram-triggered gradient echo sequence (repetition time: 7.3 ms; echo time: 4.4 ms; flip angle: 10°; field of view: 380 × 228 mm; acquisition matrix: 160 × 160; reconstruction matrix: 256 × 256; reconstruction resolution: 1.48 × 1.48 mm; number of phases per cardiac cycle: 20; velocity encoding: 50 cm/s; slice thickness: 6 mm) (Figures 2B and 2C).
Pharmacological stress was achieved by continuous injection of adenosine triphosphate (ATP) (140 μg/kg/min) into the left antecubital vein. Fast-pass perfusion CMR images were acquired with a turbo field echo sequence (4 short-axis slices/2 RR intervals; repetition time: shortest; echo time: shortest; flip angle: 40°; field of view: 360 × 324 mm; acquisition matrix: 192 × 172; reconstruction matrix: 256 × 230; slice thickness: 8 mm).
Immediately after the sequence for perfusion CMR started, gadolinium contrast was injected into the right antecubital vein at a dose of 0.05 mmol/kg and a flow rate of 4 ml/s, followed by a 20 ml saline flush. The interval between stress and resting perfusion CMR image acquisition was at least 10 min. All patients were asked to refrain from caffeinated beverages for at least 24 h before the MRI procedure. After the acquisition of rest perfusion MRI, gadolinium contrast was injected into the patient at a total dose of 0.15 mmol/kg. Fifteen minutes after this injection, late gadolinium-enhanced (LGE) images were acquired in the same planes as cine images using inversion recovery-prepared gradient-echo sequences (repetition time: 4.3 ms; echo time: 1.3 ms; flip angle: 15°; field of view: 380 × 380 mm; acquisition matrix: 256 × 180; slice thickness: 10 mm).
CMR image analysis
Two observers used an Extend MR WorkSpace workstation (Philips Healthcare, Best, the Netherlands) to analyze cine, perfusion, PC cine, and LGE images. To measure LV volumes and mass, epicardial and endocardial borders on short-axis cine images were manually traced with exclusion of the papillary muscles at each anatomical level that encompassed the left ventricle. LV mass was calculated by the consensus decision of the 2 observers as the sum of myocardial volume areas multiplied by the specific gravity (assumed to be 1.05 g/ml) of the myocardial tissue (11).
The contours of the coronary sinus were manually traced on each frame of all PC cine images to quantify blood flow in the coronary sinus, and velocity in the adjacent myocardium was measured for phase-offset correction (Figures 2D and 2E). Phase-offset correction was performed by subtracting the background velocity of adjacent myocardium from the velocity of the coronary sinus blood flow for each acquired phase. Peak velocity of the coronary sinus was 24.5 ± 6.2 cm/s at rest, increasing to 34.3 ± 11.2 cm/s during stress (p = 0.035). Peak velocity of the myocardium was 6.5 ± 1.9 cm/s at rest and 6.8 ± 1.3 cm/s during stress (p = 0.74). Blood flow in the coronary sinus was calculated by integrating the product of the cross-sectional area and mean velocity in the coronary sinus, and then corrected using mean velocity in the adjacent tissue for all cardiac phases in the cardiac cycle (Figure 2F). MBF was calculated as follows:
Resting MBF correlates linearly with rate pressure product at rest, but hyperemic MBF does not correlate linearly with rate pressure product (12). Therefore, we corrected resting MBF by resting rate pressure product at rest from each subject using the following formulas (MBF during ATP infusion was not corrected by rate pressure product during stress).
The average rate pressure product at rest was 7,500 from healthy controls with mean age of 50.1 ± 9.7 years reported in a previous study (13).
ΔMBF and CFR were calculated as:
The presence and segmental extent of perfusion defects on perfusion images were confirmed by the consensus decision of 2 independent reviewers. Perfusion defects were defined as hypo-enhanced regions that persisted for at least 3 phases after peak contrast enhancement and followed a coronary distribution (14). Inducible ischemia was defined by the presence of a stress perfusion defect in any segment without LGE. We defined the presence of >10% ischemia as inducible ischemia in ≥3 of 32 subsegments (endocardial and epicardial sectors for each of the 16 segments). This >10% ischemia criterion has been observed as reflecting high risk in the nuclear sub-study of the COURAGE (Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation) trial (15). Interobserver variability of CFR measurements and evaluation of ischemia on stress perfusion CMR were evaluated in 50 randomly selected patients. LGE mass was measured on short-axis slices using manual planimetry. Total mass of the LGE area was calculated by summing the LGE masses of all sections, with the ratio of LGE areas expressed as:
Follow-up of cardiac events
Prognostic information was obtained using electronic medical records or telephone interviews. Cardiac mortality was defined by any death preceded by acute MI, decompensated heart failure, or ventricular arrhythmia. Nonfatal MI was defined by the presentation of acute coronary syndrome and elevation of cardiac biomarkers consistent with acute injury. MACE was defined as cardiovascular death, acute MI, unstable angina, hospitalization for heart failure or ventricular tachyarrhythmia necessitating defibrillation. In our study, revascularization (percutaneous coronary intervention or coronary bypass graft surgery) was not included in MACE. The first event after the CMR examination was recorded. Noncardiac death was not included in MACE.
Data were statistically analyzed using SPSS version 17.0 software (SPSS Inc., Chicago, Illinois) and MedCalc for Windows version 14.8.1 software (MedCalc Software, Ostend, Belgium). Continuous values are presented as mean ± SD. Normality was determined using the Shapiro-Wilk test. Normally distributed values were compared using an unpaired Student t test, and non-normally distributed values were compared using the Mann-Whitney U test. The significance of differences in categorical variables was calculated using the chi square test. Univariate associations with MACE were determined by Cox proportional hazards regression, and event-free survival stratified by CFR was estimated with Kaplan-Meier survival methods. Clinical variables showing values of p < 0.05 on univariate models were incorporated into multivariate Cox regression analysis. We calculated clinical risk using multivariable logistic regression models (age + sex + hypertension + dyslipidemia + diabetes mellitus + family history of CAD + current smoking + body mass index [BMI]). For known CAD patients, clinical risk model was calculated by the following formula: −3.509 + 0.073 × age + 0.364 × sex + 0.032 × hypertension − 1.623 × dyslipidemia + 1.095 × diabetes mellitus − 0.569 × family history of CAD + 0.250 × current smoking − 0.167 × BMI (cutoff value: −2.54). For suspected CAD patients, the clinical risk model was calculated by −10.74 + 0.101 × age + 0.127 × sex − 0.141 × hypertension + 0.218 × dyslipidemia + 0.652 × diabetes mellitus + 0.307 × family history of CAD − 18.11 × current smoking + 0.016 × BMI (cutoff value: −2.55). A cutoff value for clinical risk model was calculated by receiver-operating characteristic curve that minimized the difference between sensitivity and specificity. A 2-sided p value of <0.05 was considered significant.
Among 709 patients, 10 patients were excluded due to exclusion criteria (6 patients with claustrophobia, 4 patients with intolerable symptoms during ATP infusion). Nine patients were excluded due to nonanalyzable images: 4 PC cine images with severe cardiac motion; 2 PC cine images from patients having persistent left-sided vena cava; and 3 perfusion CMR images with poor image quality. Follow-up information was obtained from 676 of 690 patients (98%). Characteristics of these 676 patients are shown in Table 1. In known CAD patients, the symptoms of patients referred for CMR were typical chest pain (n = 101; 36.6%), atypical chest pain (n = 110; 39.9%), nonanginal symptoms (dyspnea: n = 28; 10.1%; and palpitation: n = 2; 0.7%), and electrocardiographic abnormalities (n = 35; 12.6%). In patients with suspected CAD, the indication for CMR was typical chest pain (n = 31; 7.8%), atypical chest pain (n = 141; 35.2%), nonanginal symptoms (dyspnea: n = 98; 24.5%; palpitation: n = 45; 11.3%; and syncope: n = 5; 1.3%), and electrocardiographic abnormalities (n = 80; 20%).
Table 2 shows the comparison of MBF, CFR, and stress perfusion CMR. At rest, MBF was 0.94 ± 0.26 ml/min/g in patients with known CAD and 0.88 ± 0.26 ml/min/g in those with suspected CAD. Significant augmentation of MBF was achieved on ATP infusion. However, CFR did not differ significantly between patients with known and suspected CAD (2.69 ± 0.60 vs. 2.87 ± 0.61; p = 0.22). MBF at rest was significantly higher in patients with compared to without MACE (0.91 ± 0.24 ml/min/g vs. 1.25 ± 0.35 ml/min/g; p < 0.001 for known CAD; 0.86 ± 0.23 ml/min/g vs. 1.25 ± 0.32 ml/min/g; p < 0.001 for suspected CAD). Significant differences were also found in ΔMBF and CFR between patients with and without MACE (both in patients with known CAD and patients with suspected CAD). Prevalence of >10% ischemia was significantly higher in patients with MACE in comparison to those without (regardless of CAD status) (Table 2). High reproducibility was observed for CFR measurement (mean difference: 1.97%; limit of agreement: −6.13% to 10.06% by Bland and Altman plot). Kappa value was 0.898 (95% confidence interval: 0.700 to 1.000) to identify CFR <2.0 and 0.878 (95% confidence interval: 0.744 to 1.000) to detect >10% ischemia on stress perfusion CMR.
MACE and event-free survival and mortality
During a median follow-up of 2.3 years, 47 of 676 patients (6.9%) experienced MACE (cardiovascular death: n = 18; acute MI: n = 1; unstable angina: n = 16; heart failure hospitalization: n = 11; ventricular tachyarrhythmia necessitating defibrillation: n = 1) (Online Figure 1). In patients with known CAD, univariate Cox regression analysis identified age, current smoking, BMI, LV ejection fraction, extensive LGE (>15% of LV mass), ischemia on perfusion CMR, and CFR as significant predictors for MACE (Online Table 1). In patients with suspected CAD, age, diabetes mellitus, LV ejection fraction, extensive LGE, ischemia on perfusion CMR, and CFR were all significant predictors for MACE (Online Table 2). In Kaplan-Meier analysis, patients with impaired CFR and patients with >10% ischemia showed a significantly higher rate of MACE both in known and suspected CAD (Figure 3).
Annual event rates stratified by CFR and ischemia on perfusion CMR are shown in Figure 4. Regardless of CAD status, annual rates of MACE were low in patients with preserved CFR and in patients with ischemia ≤10% and higher rates in patients with impaired CFR and those with >10% ischemia on perfusion CMR.
Diagnostic and prognostic value of CFR and perfusion MRI
The Central Illustration shows the comparison of the area under the curve (AUC) for CFR and perfusion MRI for predicting MACE. No significant differences in AUC were found between CFR and perfusion MRI in patients with known CAD (0.773 vs. 0.731; p = 0.58) and in those with suspected CAD (0.885 vs. 0.776; p = 0.059). Although there was no statistically significant difference, higher AUC of CFR in comparison to that of stress perfusion CMR was observed in the suspected CAD group.
On multivariable analysis, the adjusted HRs for MACE of >10% ischemia on stress perfusion CMR and CFR <2.0 were 5.10 and 5.17 in known CAD patients (Online Table 1) and 6.50 and 14.16 in suspected CAD patients (Online Table 2). Table 3 showed predictive accuracy of clinical risk model, stress perfusion CMR, and CFR for predicting MACE.
This study was the first to show the prognostic value of MRI-derived CFR for patients with known or suspected CAD. The principal findings of our study were: 1) impairment of CFR was a significant independent predictor for MACE in both patients with known and suspected CAD; 2) the predictive values of CFR and stress perfusion CMR for the occurrence of MACE were comparable in patients with known CAD; and 3) in patients with suspected CAD, CFR showed a higher hazard ratio (HR) and AUC compared to stress perfusion CMR. The results suggested that the large portion of patients with suspected CAD should be preferably evaluated by CFR for risk stratification.
Stress perfusion CMR is superior to myocardial single-photon emission computed tomography for detection of ischemia (16). Several studies have shown the strong prognostic value of stress perfusion CMR for patients with known or suspected CAD (14,17–19). These studies were different in terms of number of study subjects (244 to 792 participants), follow-up duration (1.4 to 5 years), and proportion of patients with previous MI (8% to 24%). Compared to these studies, our study had the second-highest number of patients (n = 674). Follow-up duration (2.3 years) and proportion of previous MI (21%) were similar to that reported by Jahnke et al. (17). Despite these differences, all studies showed strong prognostic value of stress perfusion CMR (HR of stress perfusion CMR: 2.34 to 10.57) (14,17–19). These results indicated that stress perfusion CMR is a robust method to predict future cardiac events for known or suspected CAD.
Among these studies, ours was the only study that investigated MRI-derived CFR. Therefore, the main purpose of this study was to directly compare the predictive values of CFR and stress perfusion CMR. Our results have shown that the predictive value of CFR was comparable to that of stress perfusion CMR in patients with known CAD. In those with suspected CAD, higher HR and AUC were shown. In addition, patients with normal CFR showed a low annual event rate, regardless of history of CAD (Figure 4). This finding is important because the goal of stress imaging is not only to identify high-risk patients, but also to separate those patients with low cardiac event rates. Observed strong prognostic value of CFR is conceivably due to the characterization of different pathophysiological alterations by CAD; that is, microvascular dysfunction could be assessed only by CFR. Many studies have already shown that, even in patients without obstructive CAD, CFR is impaired under conditions such as hypertension (20), diabetes (21), smoking (22), dyslipidemia (23), and chronic kidney disease (24) compared to healthy subjects. These studies have suggested that coronary microvascular dysfunction precedes the flow-limiting atherosclerotic plaque burden of epicardial coronary arteries. This could explain why CFR by PC cine MRI provided independent prognostic value over stress perfusion CMR. Also, rest MBF was significantly higher in patients with MACE than in patients without MACE, but CFR was significantly lower in patients with MACE in comparison to those without MACE (Table 2). One possible explanation for this phenomenon is that MBF is already elevated at rest to account for ischemia by epicardial coronary stenosis or microvascular dysfunction or both, whereas the reserve capacity for pharmacological stress is decreased in patients with MACE.
Our observations suggested that CFR by PC cine MRI can provide new insights for noninvasive risk stratification in known or suspected CAD. Importantly, we found that CFR was prognostic even for patients with suspected CAD. Although patients with known CAD were found to have already commenced secondary prevention therapy with antiplatelet therapy, statins, and other pharmacotherapies, patients with suspected CAD were not routinely taking such medications. Indeed, CAD prevention might be more feasible in suspected CAD patients. In other words, risk stratification based on CFR assessment could be applied to such patients to identify high-risk individuals who would then be started on pharmacotherapy. Consequently, we believe the most important finding in this study to be the fact that CFR assessment aided in risk stratification of patients with suspected CAD.
Another benefit of CFR assessment by PC cine MRI compared to stress perfusion CMR was that acquisition of CFR did not necessitate any injection of contrast media, allowing patients with renal dysfunction to undergo CFR assessment. Because of the high event rate of this population, we believe that CFR assessment by PC cine MRI would be beneficial for patients with renal dysfunction. Another strength of CFR was the small number of patients who were excluded due to issues related to imaging and the equally small number of patients excluded due to reduced image quality. This point is important if this approach is to be incorporated into clinical practice.
In this study, we compared the predictive value of a clinical risk model and CFR for predicting future MACE. Clinical risk modelling was based on information from history taking and clinical examination (age, sex, hypertension, dyslipidemia, diabetes mellitus, family history of CAD, current smoking, and BMI). Hence, clinical risk models can be calculated without any additional cost, and can guide physicians in deciding upon further diagnostic tests. However, it should be noted that predictive accuracy of clinical risk modelling is limited, and our data showed a substantially lower AUC of a clinical risk model in comparison to that of CFR, both in patients with known CAD (0.661 vs. 0.773; p = 0.12) and suspected CAD (0.756 vs. 0.885; p = 0.013). In addition, calculation of clinical risk is complicated because it is based on many clinical parameters. Therefore, it would be desirable to have 1 reproducible test that accurately identifies patients at risk. In this point of view, CMR-derived CFR would be preferable due to its high accuracy, noninvasive nature, and high reproducibility. Further study is necessary to clarify which is better: to start with CFR assessment and omit the clinical risk model or start with clinical risk modelling before CFR assessment, for effective risk stratification of patients with known or suspected CAD.
This was a single-center, observational study of a limited number of patients; thus, the main limitation of the study was that it was underpowered to show significantly different AUCs between CFR and stress perfusion CMR in patients with suspected CAD, although there was a clear trend with a higher AUC of CFR. Additionally, the exact mechanisms relating noninvasive measures of CFR to increased cardiac mortality cannot be determined from this study. Furthermore, CFR can evaluate global CFR, but cannot locate the region of coronary territory with reduced CFR. In addition, although the correlation between CFR and the result of x-ray coronary angiography is important, we do not have sufficient data yet regarding x-ray coronary angiography. Therefore, the data comparing CFR and severity of stenosis on x-ray coronary angiography was not shown in our study. Finally, quantitative evaluation of absolute MBF by stress perfusion CMR is becoming more clinically feasible. A comparison of absolute MBF by stress perfusion CMR and CFR would therefore be interesting.
Noninvasive quantitative assessment of CFR with PC cine MRI of the coronary sinus offers a powerful, independent predictor of MACE for patients with known or suspected CAD. The predictive values of CFR and stress perfusion CMR for MACE were comparable in patients with known CAD. In those with suspected CAD, CFR showed higher HRs than stress perfusion CMR, suggesting that CFR assessment by PC cine MRI might be useful for better risk stratification compared to stress perfusion CMR for patients with suspected CAD.
COMPETENCY IN MEDICAL KNOWLEDGE: CFR assessed by MRI is a noninvasive and significant prognostic factor for known or suspected CAD patients.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: The predictive values of CFR and stress perfusion CMR for MACE were comparable in known CAD patients. In suspected CAD patients, CFR showed much higher HR than stress perfusion CMR, suggesting that CFR assessment by PC cine MRI might be useful for better risk stratification over stress perfusion CMR for suspected CAD patients.
TRANSLATIONAL OUTLOOK 1: Because of its noninvasiveness (no radiation exposure, no contrast injection), young patients or patients with renal dysfunction can undergo CFR assessment using PC cine MRI.
TRANSLATIONAL OUTLOOK 2: This study is a single-center study with limited number of patients. Therefore, a large-scale, multicenter study is necessary to confirm our results.
The authors thank Yukihiro Ishii, RT, for his efforts in acquiring MRI scans.
For supplemental tables and a figure, please see the online version of this paper.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Kato and Saito contributed equally to this work.
- Abbreviations and Acronyms
- area under the curve
- coronary artery disease
- coronary flow reserve
- cardiac magnetic resonance
- hazard ratio
- late gadolinium-enhanced
- major adverse cardiac event(s)
- myocardial blood flow
- magnetic resonance imaging
- Received February 2, 2017.
- Revision received June 12, 2017.
- Accepted June 13, 2017.
- 2017 American College of Cardiology Foundation
- Benjamin E.J.,
- Blaha M.J.,
- Chiuve S.E.,
- et al.
- Murthy V.L.,
- Naya M.,
- Foster C.R.,
- et al.
- Ziadi M.C.,
- Dekemp R.A.,
- Williams K.A.,
- et al.
- Herzog B.A.,
- Husmann L.,
- Valenta I.,
- et al.
- Schwitter J.,
- DeMarco T.,
- Kneifel S.,
- et al.
- Kato S.,
- Saito N.,
- Kirigaya H.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- White H.D.
- Czernin J.,
- Muller P.,
- Chan S.,
- et al.
- Naya M.,
- Tsukamoto T.,
- Morita K.,
- et al.
- Shah R.,
- Heydari B.,
- Coelho-Filho O.,
- et al.
- Shaw L.J.,
- Berman D.S.,
- Maron D.J.,
- et al.
- Jahnke C.,
- Nagel E.,
- Gebker R.,
- et al.
- Steel K.,
- Broderick R.,
- Gandla V.,
- et al.
- Greenwood J.P.,
- Herzog B.A.,
- Brown J.M.,
- et al.
- Laine H.,
- Raitakari O.T.,
- Niinikoski H.,
- et al.
- Prior J.O.,
- Quinones M.J.,
- Hernandez-Pampaloni M.,
- et al.
- Campisi R.,
- Czernin J.,
- Schoder H.,
- Sayre J.W.,
- Schelbert H.R.
- Kaufmann P.A.,
- Gnecchi-Ruscone T.,
- Schafers K.P.,
- Luscher T.F.,
- Camici P.G.
- Charytan D.M.,
- Shelbert H.R.,
- Di Carli M.F.