Computed tomographic (CT) angiography provides high sensitivity for the detection of coronary stenosis, while its specificity is relatively low. The aim of this study was to determine the incremental value of coronary flow velocity reserve (CFVR) by transthoracic echocardiography when used with CT angiography for detecting stenosis of the major coronary arteries compared with invasive quantitative coronary angiography.
Sixty patients who underwent CFVR measurement before coronary angiography were retrospectively selected, and the cutoff value of CFVR to predict diameter stenosis > 70% was determined using receiver operating characteristic curve analysis. Second, CFVR measurement and CT angiography were prospectively performed in 50 patients who were scheduled to undergo coronary angiography. CT angiography using a 64–detector row scanner and CFVR measurement in the proximal to middle portions of the three major coronary arteries by transthoracic echocardiography were performed on the same day, <48 hours before invasive angiography.
The cutoff values of CFVR were determined to be 2.0 for the left anterior descending coronary artery and 2.1 for the circumflex and right coronary arteries. Using these determined cutoff values, the sensitivity, specificity, and positive and negative predictive value of CFVR to identify diameter stenosis ≥ 70% stenosis on invasive quantitative coronary angiography were determined to be 84%, 87%, 66%, and 95%, respectively, and those of CT angiography were 91%, 80%, 58%, and 97%, respectively, in the prospective study with 50 patients. The combination of ≥70% stenosis on CT angiography and impaired CFVR was 94% specific for ≥70% stenosis, while the presence of <70% stenosis on CT angiography and preserved CFVR was 100% specific for the exclusion of ≥70% stenosis on invasive quantitative coronary angiography.
When the results of CT angiography and CFVR are concordant, the combination is highly accurate in the detection and exclusion of coronary stenosis. CFVR measurement in addition to CT angiography could be helpful in identifying false-positive CT angiographic results.
Computed tomographic (CT) angiography makes it possible to visualize coronary stenosis clearly, and it has been increasingly used as a diagnostic tool to study the coronary anatomy in patients with suspected coronary artery disease. According to previous studies, negative findings on CT angiography could practically exclude coronary atherosclerosis ; nevertheless, when its results are equivocal or unexpected, additional functional assessment could be helpful. In an era of concern about the overuse of testing, there may be undesirable consequences when invasive angiographic results are frequently “negative” for flow-limited coronary artery disease. Therefore, a noninvasive multimodality imaging approach integrating morphologic and functional information is needed.
Transthoracic color Doppler echocardiography allows the measurement of the coronary flow velocity reserve (CFVR), a useful parameter to detect significant narrowing of coronary arteries. In addition, recent improvements in ultrasound capabilities of this technique have made it possible to visualize not only the left anterior descending coronary artery (LAD) but also the left circumflex coronary artery (LCx) and the right coronary arteries (RCA). Therefore, CFVR measurement might provide incremental functional information to CT angiography and consequently might reduce the need for invasive coronary angiography. We aimed to determine the diagnostic accuracy of CFVR and combined CFVR and CT angiography to predict significant coronary stenosis of the major coronary arteries, using invasive coronary artery angiography as the reference method.
We retrospectively selected 60 patients (mean age, 66 ± 10 years; 47 men; mean body mass index, 23 ± 3 kg/m 2 ) who underwent CFVR measurement <1 week before coronary angiography. All were suspected of having coronary artery disease. Patients with previous coronary artery bypass graft surgery, significant valvular disease, and left ventricular hypertrophy (left ventricular mass index >150 g/m 2 in men and >120 g/m 2 in women) were not included. The prevalence rates of diabetes mellitus, hypertension, dyslipidemia, and smoking were 43% ( n = 26), 72% ( n = 43), 53% ( n = 32), and 30% ( n = 18), respectively. We determined the cutoff value of CFVR to predict diameter stenosis (DS) >70% using receiver operating characteristic curve analysis.
Next, we prospectively performed CFVR measurement and CT angiography in 50 patients who were scheduled to undergo coronary angiography. Patients were excluded if they had acute coronary syndromes, congestive heart failure, previous coronary artery bypass graft surgery, significant valvular disease, impaired renal function, left ventricular hypertrophy, contraindications to adenosine (second- or third-degree atrioventricular block, sick sinus syndrome, symptomatic bradycardia, severe asthma, or obstructive pulmonary disease), atrial fibrillation, or frequent extrasystolic beats. CT angiography and CFVR measurement in the proximal to middle portions of the three major coronary arteries by transthoracic echocardiography were performed on the same day, <48 hours before invasive angiography. Patient characteristics are summarized in Table 1 .
|Age (yrs)||67 ± 9|
|BMI (kg/m 2 )||24 ± 3|
|Cardiovascular risk factors|
|Heart rate (beats/min)||70 ± 12/73 ± 13|
|Systolic BP(mm Hg)||122 ± 27/118 ± 19|
|Diastolic BP(mm Hg)||63 ± 14/60 ± 14|
|LVDd (mm)||46 ± 5|
|LVDs (mm)||31 ± 5|
|IVSWT (mm)||9.8 ± 1.4|
|PWT (mm)||9.5 ± 1.3|
|LVEDV (mL)||90 ± 18|
|LVESV (mL)||38 ± 14|
|LVEF (%)||58 ± 8|
|LVMI (g/m 2 )||104 ± 23|
|LAVI (mL/m 2 )||35 ± 12|
All patients’ medications were continued on the day of CFVR and CT angiography and invasive coronary angiography. The study protocol was approved by the local institutional review board, and all patients gave written informed consent.
CT angiography was performed with a 64-slice CT scanner (Brilliance 64; Philips Medical Systems, Eindhoven, The Netherlands). All patients with heart rates >65 beats/min received a β-blocker (intravenous propranolol 2–4 mg) unless their systolic blood pressure was 100 mm Hg or other contraindications were present. All image acquisitions were performed during a single breath-hold in inspiration. A contrast agent (Iopamiron 370; Bayer Schering Pharma AG, Berlin, Germany) was injected intravenously at a rate of 3 to 5 mL/sec. Images were acquired with 64 0.625-mm slice collimation, a gantry rotation time of 420 msec, tube voltage of 120 kV, and an effective tube current of 330 to 500 mA using electrocardiographically correlated tube current modulation.
Transaxial images were reconstructed with a slice thickness of 0.8 mm and increments of 0.4 mm, with a retrospectively electrocardiographically gated half-scan algorithm with a temporal resolution of 210 msec. If necessary, additional reconstructions were performed to minimize motion artifacts. An experienced observer who was blinded to clinical characteristics and echocardiographic findings quantitatively evaluated all CT data sets with the workstation (Virtual Place Lexus; AZE Inc, Tokyo, Japan), which allowed the automatic creation of curved multiplanar reformations along the coronary arteries and maximum-intensity projections. The presence of significant coronary stenosis was defined as a luminal obstruction of ≥70% of the diameter of the reference coronary segment ( Figure 1 ).
Doppler Echocardiographic Studies
All enrolled patients fasted overnight and abstained from any beverage containing significant amounts of flavonoids for 48 hours to avoid of any effect of flavonoids in improving coronary endothelial function. Doppler echocardiography was performed with a Vivid E9 scanner (GE Vingmed Ultrasound AS, Horten, Norway) using a broadband transducer at 3.0 MHz for the LAD and at 2.4 MHz for the RCA and the LCx. For color Doppler flow mapping, the velocity ranged from 12 to 25 cm/sec. The color gain was adjusted to provide optimal images.
To measure LAD flow velocity, we located an acoustic window in the midclavicular line in the fourth or fifth intercostal space in the left lateral decubitus position. After the lower portion of the interventricular sulcus had been located in a long-axis cross-section, the ultrasound beam was inclined laterally. Next, coronary blood flow in the LAD (middle to distal) was measured using color Doppler flow mapping. After a sample volume (length, 1.5 mm) was positioned on the color signal in the LAD, coronary flow velocity was recorded by fast Fourier transformation analysis. We tried to make the ultrasound beam as parallel to the LAD flow as possible. To measure RCA flow, we selected the distal part of the RCA for color Doppler identification and flow velocity measurement. After an optimal two-dimensional image had been obtained in the apical two-chamber view, the transducer was rotated in a counterclockwise manner until the posterior interventricular sulcus was clearly visualized. Next, the linear color signal, which persisted throughout diastole, was searched carefully in the posterior interventricular sulcus under the guidance of Doppler color flow mapping. A sample volume was positioned on the color signal, and the coronary flow velocity was recorded by pulsed-wave Doppler echocardiography. To measure LCx flow, we searched Doppler flow signals in the LCx as the linear color signal persisted during diastole at the basal to the middle portion of the left ventricular lateral region in the apical four-chamber view, avoiding the far apical portion, where a signal recorded could belong to a diagonal artery coming from the LAD. The reference structure was the lateral wall itself, where obtuse marginal arteries run. Next, Doppler spectral tracings of LCx flow velocities were recorded with the sample volume positioned on the visualized color signal. To measure CFVR, we first recorded the baseline spectral Doppler signals in more than five cardiac cycles at end-expiration by transthoracic color Doppler echocardiography. Then, we intravenously administered adenosine triphosphate (0.14 mg/kg body weight/min) for 2 min to record spectral Doppler signals. This allowed us to obtain the peak flow response induced by coronary microvessel dilation. Heart rate was monitored, and electrocardiography was monitored continuously in all patients. Blood pressure was recorded at baseline and every minute during adenosine triphosphate infusion. An experienced investigator who was blinded to all other data measured coronary flow velocities offline by tracing the contour of the spectral Doppler signal (EchoPAC version 6.1; GE Vingmed Ultrasound AS, Horten, Norway). Mean diastolic velocities were measured at baseline and at the peak flow response. Measurements were averaged over five cardiac cycles. CFVR was defined as the ratio of mean diastolic velocity at the peak flow response to that at the baseline ( Figure 1 ).
Invasive Coronary Angiography
Coronary angiography was performed by standard transfemoral or transradial arterial catheterization. All patients received an intravenous bolus injection of 3,000 IU of heparin and intracoronary isosorbide dinitrate (2 mg) before angiography. Quantitative coronary angiography (QCA) was performed with GE automated edge detection software, which calibrates using the coronary guide catheter as its reference diameter (Centricity Cardiology CA1000; GE Healthcare, Dornstadt, Germany). Two observers, who were blinded to patient details, CT angiographic and CFVR results, and patient outcomes, marked the region of interest within a coronary artery they thought was the most severe region of coronary artery disease. The software then calibrated the minimal luminal diameter, reference vessel diameter (before a stenosis), and ideal vessel diameter at stenosis. From these results, the percentage DS was calculated. The two observers performed this calculation for three major coronary arteries per patient, depending on circulatory dominance and number of occluded arteries. Significant coronary stenosis was defined as ≥70% luminal diameter narrowing on the angiogram.
Continuous variables are expressed as mean ± SD and categorical variables as percentages. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated to predict the ability of each imaging modality to identify coronary stenosis. Receiver operating characteristic curve analysis was performed to evaluate the predictive performance of CFVR for DS ≥ 70% on invasive coronary angiography. The Wilcoxon signed-rank test was used to assess the deviation of median values on QCA from 70%. Statistical significance was assessed as P < .05 in a two-tailed probability analysis. Data were analyzed using SPSS for Windows version 13.0 (SPSS, Inc, Chicago, IL).
In the retrospectively selected 60 patients, mean CFVR and angiographic percentage DS were 2.44 ± 0.66 and 51.5 ± 24.3%, respectively. Determined cutoff values of CFVR to predict DS >70% were 2.0 (area under the curve, 0.89; 95% confidence interval, 0.77–1.00) for the LAD, 2.1 (area under the curve, 0.80; 95% confidence interval, 0.64–0.93) for the RCA, and 2.1 (area under the curve, 0.84; 95% confidence interval, 0.72–0.96) for the LCx.
In the prospectively studied 50 patients, coronary CT angiography showed that 30 patients (60%) had significant stenosis, 13 (26%) had one-vessel disease, 12 (24%) had double-vessel disease, and five (10%) had triple-vessel disease. Fifty-two vessels (35%) had significant stenosis. Eight vessels (5.3%) had at least one nonevaluable segment secondary to heavy calcification, and these were considered positive. Regarding CFVR measurement, adequate Doppler signals were obtained in 139 vessels, 49 of 50 (98%) in the LAD, 45 (90%) in the RCA, and 45 (90%) in the LCx. On quantitative coronary angiographic analysis of invasive coronary angiography, 21 patients (42%) had ≥70% stenosis in at least one coronary vessel. Eleven patients had single-vessel disease, nine had double-vessel disease, and one had triple-vessel disease. On a per vessel basis, 32 vessels had ≥70% stenosis, 12 (38%) in the LAD, 11 (34%) in the RCA, and nine (28%) in the LCx. All analyses were performed in 139 coronary vessels.
CT Angiography versus Invasive Angiography
CT angiography agreed with invasive coronary angiography in 29 (90%) of 32 vessels with DS ≥ 70% and 86 (80%) of 107 vessels with DS <70% by QCA. There were three false-negative vessels, one in each coronary vessel, and 21 false-positive vessels, six in the LAD, six in the RCA, and nine in the LCx. The sensitivity, specificity, PPV, NPV, and accuracy to detect vessels with ≥70% DS in all three major vessels on QCA were 91%, 80%, 58%, 97%, and 83%, respectively. On a per vessel basis, CT angiography had sensitivity and specificity of 92% (70%–99%) and 84% (77%–88%) for ≥70% LAD stenosis, 91% (67%–98%) and 82% (75%–85%) for the RCA, and 89% (61%–98%) and 75% (68%–77%) for the LCx.
CFVR versus Invasive Coronary Angiography
A total of 139 vessels from 50 patients were analyzed for the comparison between CFVR and invasive coronary angiography using determined cutoff values of CFVR. The average CFVR in vessels with DS ≥70% ( n = 41) was lower than in those with DS <70% ( n = 98) (1.8 ± 0.5 vs 2.6 ± 0.6, P < .0001; Figure 2 ). CFVR of 2.0 provided sensitivity of 83% and specificity of 92% to detect DS ≥70% in the LAD by invasive coronary angiography. Similarly, CFVR of 2.1 for both the RCA and the LCx provided sensitivity of 82% and 89% and specificity of 85% and 83%, respectively, to detect DS ≥70%. In total, CFVR correctly identified 27 of 32 (84%) DS ≥70% stenoses and 93 of 107 (87%) DS <70% stenoses by QCA of invasive coronary angiography. There were five false-negative vessels, two in the LAD, two in the RCA, and one in the LCx. There were 14 false-positive vessels, three in the LAD, five in the RCA, and six in the LCx. The diagnostic accuracy of CFVR is summarized in Table 2 .
|Assessable||49/50 (98%)||45/50 (90%)||45/50 (90%)|
|AUC||0.89 (0.77–1.00)||0.78 (0.64–0.93)||0.84 (0.72–0.96)|
|Sensitivity (%)||83 (62–94)||82 (58–94)||89 (61–98)|
|Specificity (%)||92 (85–96)||85 (78–89)||83 (76–86)|
|PPV (%)||77 (57–87)||64 (45–74)||57 (39–63)|
|NPV (%)||94 (87–98)||94 (85–98)||97 (89–99)|
|κ||0.7 (0.6–0.9)||0.6 (0.3–0.8)||0.6 (0.3–0.7)|
|Accuracy (%)||90 (79–95)||84 (73–91)||84 (73–88)|
Combined CFVR and CT Angiography versus Invasive Coronary Angiography
The relationship between combined CFVR and CT angiography and invasive coronary angiography is summarized in Figure 3 . Positive findings on CFVR and CT angiography agreed in 31 vessels, in which 24 vessels (77%) were revealed as having ≥70% stenoses on QCA of invasive coronary angiography. On the other hand, 79 instances of agreement on negative findings between CFVR and CT angiography were found, and all of these vessels showed <70% stenosis on QCA. There were 19 vessels with significant stenosis on CT angiography but preserved CFVR, of which five (26%) had ≥70% stenosis on QCA. On the other hand, 10 vessels showed nonsignificant stenosis on CT angiography but impaired CFVR, in which 3 (30%) had ≥70% stenosis on QCA. Overall, when CFVR and CT angiography were considered positive when subtended by impaired CFVR and significant stenosis on CT angiography, sensitivity was 75%, specificity 94%, PPV 77%, NPV 93%, and accuracy 89%. In contrast, when CFVR and CT angiography were considered positive when subtended by impaired CFVR or significant stenosis on CT angiography, sensitivity was 100%, specificity 74%, PPV 53%, NPV 100%, and accuracy 80% ( Table 3 ). In 21 false-positive results on CT angiography, 14 (67%) had preserved CFVR, and in three false-negative results, all showed decreased CFVR. Regarding CFVR findings, there were seven (50%) vessels with <70% stenosis on CT angiography in 14 false-positive results, and all five false-negative vessels were positive on CT angiography.