Author + information
- Received February 12, 2016
- Revision received April 27, 2016
- Accepted May 4, 2016
- Published online April 3, 2017.
- Sara Gaur, MD, PhDa,∗ (, )
- Charles A. Taylor, PhDb,c,
- Jesper M. Jensen, MD, PhDa,
- Hans Erik Bøtker, MD, PhD, DMSca,
- Evald H. Christiansen, MD, PhDa,
- Anne K. Kaltoft, MD, PhDa,
- Niels R. Holm, MDa,
- Jonathon Leipsic, MDd,
- Christopher K. Zarins, MDb,e,
- Stephan Achenbach, MD, PhDf,
- Sophie Khem, MSb,
- Alan Wilk, BSb,
- Hiram G. Bezerra, MD, PhDg,
- Jens F. Lassen, MD, PhDa and
- Bjarne L. Nørgaard, MD, PhDa
- aDepartment of Cardiology, Aarhus University Hospital, Aarhus, Denmark
- bHeartFlow, Inc., Redwood City, California
- cDepartment of Bioengineering, Stanford University, Stanford, California
- dDepartment of Radiology and Division of Cardiology, St. Paul’s Hospital, Vancouver, British Columbia, Canada
- eDepartment of Surgery, Stanford University, Stanford, California
- fDepartment of Cardiology, University of Erlangen, Erlangen, Germany
- gCardiovascular Imaging Core Laboratory, Harrington Heart and Vascular Institute, Case Medical Center, Cleveland, Ohio
- ↵∗Address for correspondence:
Dr. Sara Gaur, Department of Cardiology, Aarhus University Hospital, Palle Juul-Jensens Boulevard 99, Aarhus N, Denmark 8200, Denmark.
Objectives This study sought to determine the diagnostic performance of noninvasive fractional flow reserve (FFR) derived from coronary computed tomography angiography (CTA) (FFRCT) for the diagnosis of lesion-specific ischemia in nonculprit vessels of patients with recent in ST-segment elevation myocardial infarction (STEMI).
Background In patients with stable angina, FFRCT has high diagnostic performance in identification of ischemia-causing lesions. The potential value of FFRCT for assessment of multivessel disease in patients with recent STEMI has not been evaluated.
Methods Coronary CTA with calculation of FFRCT and invasive coronary angiography with FFR were performed 1 month after STEMI in patients with multivessel disease. Coronary CTA and invasive coronary angiography stenosis >50% were considered obstructive. Lesion-specific ischemia was assumed if FFRCT was ≤0.80. FFR ≤0.80 was the reference standard. To evaluate the influence of vessel size, the total coronary vessel lumen volume relative to left ventricular mass (volume-to-mass ratio) was calculated and compared with that of patients with stable angina.
Results The study evaluated 124 nonculprit vessels from 60 patients. Accuracy, sensitivity, and specificity of FFRCT were 72%, 83%, and 66% versus 64% (p = 0.033), 93% (p = 0.15), and 49% (p < 0.001) for CTA and 72% (p = 1.00), 76% (p = 0.46), and 70% (p = 0.54) for invasive coronary angiography. Following STEMI, median volume-to-mass ratio was lower than in patients with stable angina, 53 versus 65 mm3/g (p = 0.009). In patients with volume-to-mass ratio ≥65 mm3/g (upper tertile) accuracy, sensitivity, and specificity of FFRCT were all 83% versus 56% (p = 0.009), 75% (p = 0.61), and 44% (p = 0.003) in patients with <49 mm3/g (lower tertile).
Conclusions The diagnostic performance of FFRCT for staged detection of ischemia in STEMI patients with multivessel disease is moderate. STEMI patients have a smaller vessel volume than do patients with stable angina. The diagnostic performance of FFRCT is influenced by the volume-to-mass ratio. This study does not support routine use of FFRCT in the post-STEMI setting. (Assessment of Coronary Stenoses Using Coronary CT-Angiography and Noninvasive Fractional Flow Reserve; NCT01739075)
- coronary computed tomography angiography
- fractional flow reserve
- nonculprit lesion
- ST-segment elevation myocardial infarction
Primary percutaneous coronary intervention (PCI) is well established in patients with ST-segment elevation myocardial infarction (STEMI) (1,2). Approximately 50% of patients with STEMI have multivessel disease, a condition associated with increased incidence of recurrent ischemia and augmented mortality (3,4). Whether to perform revascularization of nonculprit lesions in STEMI—and when to do it—remains controversial (1,5). Current guidelines from the European Society of Cardiology/European Association for Cardio-Thoracic Surgery recommend staged revascularization of nonculprit lesions based on symptoms or evidence of ischemia (1). In the recently updated guidelines from the American College of Cardiology/American Heart Association, PCI of nonculprit vessels in STEMI may be considered, either at the time of primary PCI or as a staged procedure (2). Although fractional flow reserve (FFR) following STEMI may be affected by vessel remodeling (6,7), it has been shown that a strategy of staged complete FFR-guided revascularization compared with a culprit lesion–only strategy is associated with improved clinical outcomes (8). A method of noninvasive computation of FFR from standard acquired coronary computed tomography angiography (CTA) datasets (FFRCT) has been developed (9). The diagnostic performance of FFRCT for identifying ischemia-causing lesions using FFR as the reference standard is high and superior to anatomical interpretation in patients with stable angina (10–12). We hypothesized that noninvasive FFRCT assessment of nonculprit lesions following STEMI may be useful to guide staged revascularization. Thus, the aim of this study was to determine the diagnostic performance of FFRCT for identification of ischemia-causing nonculprit lesions following STEMI using FFR as the reference standard. Furthermore, we aimed to investigate differences in variables potentially affecting the diagnostic performance of FFRCT between patients with recent STEMI and patients suspected of stable angina including data from the recent NXT (HeartFlow Analysis of Coronary Blood Flow Using CT Angiography: Next Steps) trial (12).
This was a prospective, single center study comprising STEMI patients with multivessel disease. The local ethics committee approved the study protocol. All patients provided written informed consent.
STEMI patients undergoing primary PCI with ≥1 potentially significant lesion in ≥1 nonculprit vessel determined by the interventionalist were eligible for inclusion. STEMI was defined as chest pain of <12-h duration and new ST-segment elevation ≥0.1 mV (≥0.2 mV in V1 to V3) involving at least 2 contiguous leads. Study exclusion criteria were hemodynamic instability, allergy to iodinated contrast, plasma creatinine ≥125 μmol/l, atrial fibrillation, contraindications to beta-receptor blocking agents, nitroglycerin or adenosine, age <18 years, body mass index >35 kg/m2, or pregnancy.
Coronary CTA image acquisition and analysis
Coronary CTA was performed 1 month after STEMI using a Siemens SOMATOM Definition Flash Dual Source scanner (Erlangen, Germany) (12). Image acquisition was performed in accordance with societal guidelines (13). Intravenous beta-blockers were administered if necessary, targeting a heart rate <60 beats/min, and all patients received 0.8 mg sublingual nitroglycerin. An initial 120-kV high-pitch spiral nonenhanced scan for calcium scoring was performed. Further details are provided in the Online Appendix. Scans were assessed using axial images and multiplanar reconstructions by 2 experienced cardiologists blinded to patient characteristics and other test results. In an 18-segment coronary model, vessel segments ≥2 mm were evaluated for lumen narrowing (12). Coronary stenosis >50% was considered obstructive.
Computation of FFRCT
FFRCT analysis was performed in a blinded fashion using software version 1.4 (HeartFlow Inc., Redwood City, California) (12). Based on the coronary CTA images, an individual quantitative 3-dimensional anatomical model of the epicardial coronary arteries, the aortic root, and the myocardium was generated. All stents deployed at the primary PCI procedure were assumed patent. Using computational fluid dynamics principles, coronary blood flow and pressure were computed under simulated hyperemic conditions (9). Lesion-specific ischemia was defined as FFRCT ≤0.80.
Calculation of volume-to-mass ratio
The volumes of all vessels in the image-based anatomical model were assessed. Moreover, to extend the latter down to pre-capillary levels, coronary morphometric data were generated using branching laws (14). The total coronary artery lumen volume was calculated from the combined image-based and morphometric data. The volume of the myocardium extracted from the image data was multiplied by an average value of myocardial tissue density (1.05 g/ml) to calculate the left ventricular myocardial mass (9,15). To adjust the vessel lumen volume for differences in size of the supplied myocardium, the vessel volume relative to myocardial mass (volume-to-mass ratio) was computed by dividing the total coronary artery lumen volume by the left ventricular myocardial mass.
Quantitative coronary angiography
Invasive coronary angiography (ICA), scheduled 1 day after coronary CTA, was performed in accordance with standard practice (12). Two projections were obtained in each major epicardial coronary artery with angles of projection optimized based on cardiac position. Stenosis severity was determined by standard quantitative coronary angiography in a blinded fashion (QAngio XA 7.3, Medis Medical Imaging Systems, Leiden, the Netherlands) (16). Stenosis severity >50% was considered obstructive.
During the staged ICA procedure, FFR interrogation was recommended in all major epicardial vessels. However, the final decision regarding FFR interrogation was at the discretion of the interventionalist. Intracoronary nitroglycerin was administered to ensure epicardial vasodilation. Hyperemia was attained by administration of adenosine (140 to 180 μg/kg/min) in a femoral or antecubital vein. A pressure-monitoring guidewire (PrimeWire Prestige PLUS, Volcano Corporation, San Diego, California) was advanced to the distal part of the vessel. Two independent blinded readers interpreted the FFR tracings. In case of discrepancy between readers, consensus was made after reinspection of the FFR tracings.
Integration of coronary CTA and FFR
The FFRCT core laboratory provided a blank 3-dimensional computer model of the coronary arteries. On this model, the location(s) that corresponded to the location(s) of the distal pressure sensor was marked. The model was returned to the FFRCT core laboratory, which reported the FFRCT values at the respective sites.
Cumulative radiation exposures were reported in millisievert. For coronary CTA, the formula used was: mSv = (dose length product) × 0.014, for ICA, a conversion factor of 0.18 mSv/(Gy ∙ cm2) was applied (17).
Comparisons between patients with STEMI and patients with stable angina
Variables associated with coronary CTA image quality (body mass index, heart rate, and Agatston score) were compared between the present study population and the population of the NXT trial (12). Cohorts from the present study population and the NXT trial matched 1:1 with respect to age, sex, and Agatston score were identified and compared with regard to body mass index and heart rate. Based on a strikingly low vessel size despite administration of nitroglycerin in most STEMI patients observed during the study period, we hypothesized that the volume-to-mass ratio would provide insight into the diagnostic performance of FFRCT. Thus, post hoc analyses of the volume-to-mass ratio between the present study and patients from the NXT trial were performed.
Twelve months follow-up
Data on unplanned (clinically driven) ICA and revascularization procedures and repeat myocardial infarction, respectively, were obtained from patient files and registries (18) within 12 months after the ICA 1-month study procedure.
Continuous variables are presented as mean ± SD and range or medians (interquartile range [IQR]) as appropriate. Categorical variables are presented as numbers and percentages. Data were compared by paired or unpaired Student t test, the Wilcoxon signed rank test, or the chi-square test as appropriate. The relationship between FFRCT and FFR was assessed by Pearson correlation coefficient. The reference standard for lesion-specific ischemia was FFR ≤0.80 (8,10–12). The per-vessel diagnostic performance of coronary stenosis >50% determined by coronary CTA or ICA and FFRCT ≤0.80 was assessed by accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. Identification of ischemia by coronary CTA, FFRCT, and ICA was assessed by the area under the receiver-operating characteristics curve. Normal-based bootstrapping with 10,000 samples was used for adjustment for clustering effects in the 95% confidence intervals (CI), and for comparison of diagnostic performance estimates. The primary diagnostic performance analyses comprised all nonculprit vessels. Secondary analyses included assessment of the diagnostic performance of FFRCT in relation to volume-to-mass ratio and Agatston score. Subanalyses relating per-vessel coronary CTA, ICA, and FFRCT to downstream revascularization were performed. Statistical analysis was performed with Stata software version 12 (StataCorp, College Station, Texas).
Sixty patients recruited between February 2012 and February 2014 comprised the study population (Figure 1). Interrogation by FFR was performed in 168 vessels, of which 124 (74%) were of nonculprit origin. Discrepancies in FFR interpretation between the 2 readers were solved by consensus in 3 vessels (2%) from 3 patients (5%).
Baseline characteristics are shown in Table 1. Details regarding the primary PCI procedure and post-procedural medication are shown in Online Table 1. No patients were readmitted to hospital or underwent unscheduled ICA between primary PCI and coronary CTA. Characteristics of study procedures are provided in Online Table 2. At the time of coronary CTA, left ventricular ejection fraction was ≥50% in 51 patients (85%). The mean interval between primary PCI and coronary CTA was 38 days, and mean interval between coronary CTA and ICA was 1 day. For patients not undergoing PCI after follow-up ICA (n = 30), mean estimated effective radiation dose was 6.2 ± 1.7 mSv for the ICA versus 2.8 ± 1.3 mSv for the coronary CTA procedure (p < 0.001). Per-patient and per-vessel characteristics according to the coronary CTA, FFRCT, ICA, and FFR results are shown in Table 2. All stents were patent at follow-up ICA. Mean FFRCT and FFR across nonculprit lesions were 0.79 ± 0.11 (range 0.49 to 0.96) and 0.85 ± 0.13 (range 0.32 to 1.00) (p < 0.001). Overall, mean heart rate during coronary CTA and Agatston score, median volume-to-mass ratio in the present cohort in comparison with the 254 patients from the NXT trial were 58 ± 8 beats/min versus 63 ± 10 beats/min (p = 0.001), 398 ± 431 versus 302 ± 467 (p = 0.015), and 53 mm3/g (IQR: 39 to 72 mm3/g) versus 65 mm3/g (IQR: 45 to 94 mm3/g) (p = 0.009), respectively. In the present study cohort, median volume-to-mass ratio was 53 (IQR: 39 to 79) among smokers and 54 (IQR: 34 to 61) among nonsmokers (p = 0.51).
Diagnostic performance of coronary CTA, FFRCT, and ICA
Diagnostic performance estimates of FFRCT, coronary CTA, and ICA are shown in Table 3. Compared with coronary CTA alone, FFRCT improved accuracy, specificity, and positive predictive value, whereas there was no difference between FFRCT and ICA estimates. Overall, FFRCT did not improve the discrimination of lesion-specific ischemia when compared with anatomical assessment by coronary CTA or ICA (Online Figure 1). Agreement in detection of lesion-specific ischemia between coronary CTA, FFRCT, and ICA is shown in Online Table 3. Median volume-to-mass ratio in vessels correctly or incorrectly classified by FFRCT were 56 mm3/g (IQR: 49 to 84 mm3/g) and 49 mm3/g (IQR: 33 to 57 mm3/g) (p = 0.02), respectively. Figure 2 displays a patient case.
In an analysis comprising all (including culprit) vessels (n = 168), accuracy, sensitivity, specificity, and area under the receiver-operating characteristics curve of FFRCT were unaltered. The diagnostic performance of FFRCT remained equivalent to that of ICA when analyses were restricted to vessels with lesions adjudicated by the interventionalist during the primary PCI procedure as potentially significant (n = 82). In vessels with or without lesions adjudicated as significant during primary PCI, FFR was ≤0.80 in 39 of 82 (48%) and 3 of 42 (7%), respectively. In the latter 3 vessels, FFRCT was ≤0.80. Findings were consistent independent of the arterial distribution and number of stents. FFRCT was superior to coronary CTA for prediction of downstream revascularization (Online Table 4).
In a matched analysis (n = 53) comparing the present study cohort with the NXT cohort, mean body mass index was 26 ± 4 kg/m2 versus 26 ± 3 kg/m2 (p = 0.50) and mean heart rate was 58 ± 8 beats/min versus 62 ± 9 beats/min (p = 0.08). In the matched analysis, 2 patients (4%) from the NXT cohort versus 100% in the STEMI study cohort had a history of previous myocardial infarction. The median volume-to-mass ratio tended to be lower in the STEMI study cohort than in the matched NXT cohort, 53 mm3/g (IQR: 37 to 79 mm3/g) versus 61 mm3/g (IQR: 47 to 87 mm3/g) (p = 0.08). The diagnostic performance of FFRCT in the present study was highest in the volume-to-mass ratio upper tertile (Figure 3) with accuracy of 83% (95% CI: 71% to 95%), sensitivity 83% (95% CI: 60% to 100%), and specificity 83% (95% CI: 68% to 99%) in comparison with 71% (95% CI: 59% to 84%) (p = 0.12), 100% (95% CI: 74% to 100%) (p = 0.15), and 60% (95% CI: 45% to 75%) (p = 0.016) for coronary CTA and 81% (95% CI: 68% to 94%) (p = 0.75), 92% (95% CI: 75% to 100%) (p = 0.59), and 77% (95% CI: 60% to 93%) (p = 0.42) for ICA. Overall, results were consistent when following a volume-to-mass ratio binary analysis approach. In vessels with low to intermediate volume-to-mass ratio and high Agatston score, the diagnostic performance of FFRCT was poor (Figure 4). If the latter group of vessels was excluded from analysis, accuracy, sensitivity, and specificity of FFRCT in 95 vessels increased to 80% (95% CI: 72% to 88%), 90% (95% CI: 79% to 100%), and 75% (95% CI: 64% to 86%), respectively.
Over 12 months of follow-up, 8 patients (13%) underwent unplanned ICA with revascularization (PCI in all) performed in 6 (10%). PCI was performed in 10 lesions, of which 7 were assessed at the 1-month coronary CTA and ICA studies. In 6 and 4 lesions, stenosis was >50% by coronary CTA and ICA, respectively. In 5 and 2 lesions, FFRCT and FFR were ≤0.80. Median FFRCT and FFR were 0.76 (IQR: 0.67 to 0.81; range 0.50 to 0.88) and 0.87 (IQR: 0.69 to 0.91; range 0.68 to 0.93), respectively. Median volume-to-mass ratio in these patients was 40 mm3/g (IQR: 28 to 56 mm3/g; range 25 to 97 mm3/g). No patients suffered a new myocardial infarction.
Correlation between FFRCT and FFR
There was a moderate direct correlation between FFRCT and FFR with Pearson correlation coefficient 0.57 (95% CI: 0.44 to 0.68; p < 0.001). In the highest tertile of volume-to-mass ratio, the Pearson correlation coefficient was 0.60 (95% CI: 0.36 to 0.76; p < 0.001) compared with 0.43 (95% CI: 0.14 to 0.65; p = 0.005) in the lowest tertile. The agreement between FFRCT and FFR is illustrated in Figure 5.
Three previous multicenter prospective trials including patients with suspected or known stable coronary artery disease have shown high diagnostic performance of FFRCT using FFR as the reference standard (10–12). The most recent NXT trial demonstrated superior per-vessel accuracy (86%) and specificity (86%) of FFRCT for detection of ischemia compared with anatomical interpretation by coronary CTA or ICA (12). In STEMI patients, a strategy of ischemia testing prior to decision making regarding revascularization of nonculprit lesions is in accordance with guidelines (1). The fact that these patients frequently are asymptomatic encourages a noninvasive diagnostic approach (5). In the present study, the diagnostic performance of FFRCT for staged assessment of nonculprit lesions following STEMI was investigated for the first time. The overall diagnostic performance of FFRCT for detection of ischemia-causing lesions was superior to anatomical assessment by coronary CTA and equivalent to invasive assessment by ICA. Surprisingly, the diagnostic performance of FFRCT was lower than previously shown in patients with stable angina (10–12). As a potential explanation, post hoc we demonstrated that the STEMI cohort had a substantially lower volume-to-mass ratio than did patients with stable angina. Moreover, we exhibited that increasing volume-to-mass ratio in the post-STEMI setting was positively associated with the diagnostic performance of FFRCT. In fact, in vessels with the highest tertile of volume-to-mass ratio (≥65 mm3/g, comparable to the ratio in patients with stable angina from the NXT trial), the FFRCT diagnostic performance was comparable to the findings in the NXT trial (accuracy 83%, sensitivity 83%, specificity 83%). In contrast, in vessels with the lowest volume-to-mass ratio (<49 mm3/g), FFRCT demonstrated poor diagnostic performance (accuracy 56%, specificity 44%). These observations as well as the clinical benefit and safety in the post-STEMI setting of a noninvasive anatomical-functional approach by coronary CTA and FFRCT need further delineation in future and larger studies.
To the best of our knowledge, no previous study has assessed changes in vessel lumen volume for an extended period after STEMI. However, several factors may modulate coronary remodeling following STEMI. Low coronary blood flow causes adaptive reduction in vessel size (9). In the acute setting of STEMI, a reduction in coronary blood flow has been demonstrated in both culprit and nonculprit vessels (19). During the convalescence phase following STEMI, patients may be less physically active than stable patients. The resultant decrease in myocardial oxygen demand may reduce the coronary blood flow and/or the vasodilatory capacity. Alpha-adrenergic blockade attenuates vasoconstriction in nonculprit vessels after STEMI (20). Furthermore, a reduced microvasculatory vasodilator response has been observed up to 6 months following myocardial infarction in the perfusion beds of both culprit and nonculprit vessels (6,7). In a recent study, there was a decrease in FFR (mean 0.92 to 0.89) and an increase in coronary flow reserve (mean 2.3 to 3.1) from day 1 to 6 months after STEMI that could not be explained by changes in anatomy (7). Thus, adaptation to low flow, local neurohumoral reflexes resulting in epicardial vasoconstriction, and attenuated vasodilatory capacity are possible mechanisms for the reduced volume-to-mass ratio in the present setting. It may be speculated that the microvascular resistance has not stabilized in the 1-month time period between STEMI and CT/FFR assessment in the present study. Therefore, the overall low specificity of FFRCT in this study may be explained by the assumption of a normal vasodilatory response in the computational model (9), a prerequisite that may not be valid in the post-STEMI setting. Inherently, FFRCT may be more indicative of the functional significance of coronary lesions once the microvasculature has fully recovered from STEMI. Moreover, it may be speculated that inadequate vasodilation resulting in falsely elevated FFR values may contribute to the lower than expected correlation between FFR and FFRCT. Our results are certainly hypothesis generating, thus future studies are needed to delineate the mechanisms responsible for the small volume-to-mass ratio and/or attenuated response to vasodilators following STEMI.
The finding in this study of declining diagnostic performance of both coronary CTA and FFRCT in the event of low volume-to-mass ratio is in accord with previous findings showing that epicardial vasodilation mediated through administration of nitroglycerin results in recruitment of evaluable segments together with an increase in the diagnostic accuracy of coronary CTA and FFRCT (21,22). Thus, despite the use of subvoxel techniques to overcome the limitations in CT spatial resolution and automated segmentation algorithms for calculation of FFRCT, low coronary vessel size and/or attenuated vasodilatory responsiveness following STEMI seem to compromise accurate computation of FFR. It should be acknowledged that because the resistance in vessels can be approximated by Poiseuille's law as inversely related to vessel diameter to the fourth power, inadequate vessel segmentation may introduce errors in stenosis diameter and even larger errors in stenosis resistance, pressure drop, and computed FFR (9).
This was single-center study. Because of the low number of patients, analyses were restricted to the per-vessel level. Inherently our findings may not be widely applicable. However, baseline characteristics were comparable to other studies of contemporary STEMI patients (7,8), supporting the generalizability of these findings. Study inclusion was limited by the necessity of obtaining consent within a few hours following primary PCI, as the patients, due to rules from the ethics committee, were not to be contacted regarding study inclusion after discharge from the PCI center. However, potential selection bias was caused solely by logistical reasons and was unrelated to conditions with potential effect on FFRCT performance. Left ventricular function was normal or near normal in the majority of patients, thus the influence of this metric on FFRCT diagnostic performance could not be assessed. As maximum troponin T levels were not recorded, the effect of biomarker levels on FFRCT diagnostic performance could not be assessed. Matching for all variables with potential significance on FFRCT diagnostic performance was not possible. In this study, the volume-to-mass ratio was assessed during the FFRCT computation process. In the future, volume-to-mass ratio may potentially be assessed locally and used as a screening tool before FFRCT assessment. Survival 12-month follow-up data were not available.
The overall diagnostic performance of FFRCT for staged detection of ischemia in nonculprit vessels of STEMI patients with multivessel disease is modest. Relative to patients with stable angina, STEMI patients have a reduced vessel volume. In patients with recent STEMI, the diagnostic performance of FFRCT is influenced by the volume-to-mass ratio. Currently there is no evidence to support FFRCT assessment in the post-STEMI setting. The clinical utility of FFRCT in patients with recent acute coronary syndromes needs further investigation.
COMPETENCY IN MEDICAL KNOWLEDGE: FFRCT is a novel method for noninvasive computation of FFR derived from standard coronary CTA datasets using computational fluid dynamics principles. In patients with stable angina, FFRCT has shown high accuracy for detection of lesion-specific ischemia compared with invasively measured FFR. In this study we demonstrated that the overall diagnostic performance of FFRCT for staged detection of ischemia in nonculprit vessels of STEMI patients with multivessel disease was modest. The median coronary vessel lumen volume relative to myocardial mass (volume-to-mass) ratio was lower in the STEMI cohort than in a cohort of patients with stable angina. In patients with recent STEMI, the diagnostic performance of FFRCT is influenced by the volume-to-mass ratio.
TRANSLATIONAL OUTLOOK: Currently, there is no evidence to support the use of FFRCT in the post-STEMI setting. The clinical benefit and safety of noninvasive anatomical-functional assessment of nonculprit lesions by coronary CTA and FFRCT need further investigation.
For supplemental materials, please see the online version of this article.
Drs. Taylor and Zarins are founders, employees, and shareholders of HeartFlow. Dr. Jensen has received speaker's honorarium from Bracco Imaging. Dr. Christiansen has received research grants from St. Jude Medical. Dr. Holm has received institutional research grants from St. Jude Medical, Medis Medical Imaging, and Boston Scientific. Dr. Leipsic serves as a consultant for GE Healthcare, Edwards Lifesciences, HeartFlow, Samsung, Philips, and Circle Cardiovascular Imaging. Dr. Achenbach has received institutional research grants from Siemens Healthcare and Abbott Vascular. Ms. Khem and Mr. Wilk are employees of HeartFlow. Dr. Lassen has received research grants from St. Jude Medical, Biosensors, Biotronik, Boston Scientific, Radi, Terumo, and Volcano. Dr. Nørgaard has received institutional research grants from Siemens Healthcare, Edwards Lifesciences, and HeartFlow. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- computed tomography angiography
- fractional flow reserve
- fractional flow reserve derived from computed tomography angiography datasets
- invasive coronary angiography
- interquartile range
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Received February 12, 2016.
- Revision received April 27, 2016.
- Accepted May 4, 2016.
- 2017 American College of Cardiology Foundation
- Windecker S.,
- Kolh P.,
- Alfonso F.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Blankenship J.C.,
- et al.
- Sorajja P.,
- Gersh B.J.,
- Cox D.A.,
- et al.
- Cuisset T.,
- Noc M.
- Cuculi F.,
- De Maria G.L.,
- Meier P.,
- et al.
- Engstrom T.,
- Kelbaek H.,
- Helqvist S.,
- et al.,
- for the DANAMI-3—PRIMULTI Investigators
- Koo B.K.,
- Erglis A.,
- Doh J.H.,
- et al.
- Nørgaard B.L.,
- Leipsic J.,
- Gaur S.,
- et al.,
- for the NXT Trial Study Group
- Kassab G.S.,
- Rider C.A.,
- Tang N.J.,
- Fung Y.C.
- Einstein A.J.,
- Moser K.W.,
- Thompson R.C.,
- Cerqueira M.D.,
- Henzlova M.J.
- Nielsen L.H.,
- Norgaard B.L.,
- Tilsted H.H.,
- et al.
- Gibson C.M.,
- Ryan K.A.,
- Murphy S.A.,
- et al.
- Gregorini L.,
- Marco J.,
- Kozakova M.,
- et al.
- Takx R.A.,
- Sucha D.,
- Park J.,
- Leiner T.,
- Hoffmann U.