DOI: 10.1148/radiol.2432060331
(Radiology 2007;243:377-385.)
© RSNA, 2007
Contrast-enhanced Cardiac MR Imaging in the Detection of Reduced Coronary Flow Velocity Reserve1
Achim A. Barmeyer, MD,
Alexander Stork, MD,
Kai Muellerleile, MD,
Claudia Tiburtius,
Anne K. Schofer, MD,
Thomas A. Heitzer, MD,
Thomas Hofmann, MD,
Gerhard Adam, MD,
Thomas Meinertz, MD, and
Gunnar K. Lund, MD
1 From the Clinic of Cardiology/Angiology, Heart Center (A.A.B., C.T., A.K.S., K.M., T.A.H., T.H., T.M., G.K.L.), and Clinic of Diagnostic and Interventional Radiology (A.S., G.A.), University Hospital Hamburg-Eppendorf, Martini-strasse 52, 20246 Hamburg, Germany. Received February 20, 2006; revision requested April 24; revision received May 31; accepted June 21; final version accepted September 1. Supported in part by Bracco-Altana Pharma, Konstanz, Germany.
Address correspondence to A.A.B. (e-mail: barmeyer{at}uke.uni-hamburg.de).
 |
ABSTRACT
|
|---|
Purpose: To prospectively evaluate the accuracy of contrast material–enhanced cardiac magnetic resonance (MR) imaging for determining impaired coronary flow velocity reserve (CFR) by using Doppler flow measurement as the reference standard.
Materials and Methods: The study was approved by the institutional ethics committee, and all patients gave written informed consent. Eligible patients underwent contrast-enhanced cardiac MR imaging and invasive measurement of CFR. For contrast-enhanced MR imaging, a three-section single-shot saturation recovery gradient-recalled echo sequence with steady-state free precession was used. Sections were divided into six segments. For each segment, a transmural and subendocardial myocardial perfusion reserve index (MPRI) was calculated by using the upslope of the signal intensity–time curve during the first pass of contrast material at rest and during adenosine infusion (140 µg per kilogram body weight per minute). MPRIs of vascular regions were compared with the corresponding CFR. Receiver operating characteristic (ROC) analysis was performed to find the number of segments needed for best diagnostic accuracy of MPRI and to find a cutoff value for MPRI in the detection of a reduced CFR.
Results: Thirty-five patients were evaluated (male-to-female ratio, 27:8; mean age ± standard deviation, 63.5 years ± 8.2; mean body mass index, 28.8 kg/m2 ± 3.8), and 43 vascular regions were analyzed. A linear correlation was found between the MPRI and CFR (r = 0.44, P < .05). The MPRI was significantly lower in vascular regions with a CFR of less than 2.00 than in regions with a CFR of 2.00 or greater (P < .05). Detection of a CFR of less than 2.00 was more accurate with subendocardial MPRI measurements than with transmural measurements. The mean subendocardial MPRI of the segments with the three lowest MPRIs of a vascular region showed the best diagnostic performance in the detection of a CFR of less than 2.00 (area under the ROC curve, 0.85; sensitivity, 84%; specificity, 75%) by using a cutoff value of 1.21.
Conclusion: The diagnostic accuracy of subendocardial perfusion analysis in contrast-enhanced cardiac MR imaging is higher than that of transmural analysis.
© RSNA, 2007
|
INTRODUCTION
|
|---|
Invasive Doppler ultrasonographic (US) measurement of coronary flow velocity reserve (CFR) provides information about the functional status of myocardial perfusion independently of the underlying cause (eg, epicardial coronary stenosis or microvascular disease). The sensitivity and specificity of Doppler US–determined CFR in the detection of stress-induced perfusion impairment are high compared with those of nuclear imaging. Several studies have shown that a cutoff value for CFR of 2.00 helps in the reliable detection of stress-induced perfusion impairment (1–3). In addition, a strong linear correlation was shown between Doppler US–determined CFR and myocardial perfusion reserve as determined with [15O]H2O positron emission tomography (PET) (4). All methods for functional evaluation of coronary flow and myocardial perfusion, however, are disadvantageous in several respects. Intracoronary Doppler measurement is invasive and is hampered by the risk of complications. Scintigraphic examinations are limited by the problems of generation, storage, and application of radioactive tracers and are not well suited for follow-up examinations. For clinical decision-making and follow-up studies, it is, therefore, desirable to have methods that enable noninvasive imaging of myocardial perfusion without the previously mentioned disadvantages.
Contrast material–enhanced cardiac magnetic resonance (MR) imaging is a noninvasive method for assessment of myocardial perfusion. Researchers in previous studies who used the first pass of a contrast agent through the myocardium for this assessment have shown that semiquantitative measurement of myocardial perfusion can be performed in animal models of coronary occlusion (5,6), healthy subjects (7), and patients with coronary artery disease (CAD) (8–11). Investigators in other studies have evaluated contrast-enhanced cardiac MR imaging with respect to single photon emission computed tomography (12–16) and PET (17,18). In addition, comparative studies of contrast-enhanced cardiac MR imaging and quantitative coronary angiography (8,10,19,20) have been conducted. A myocardial perfusion reserve index (MPRI) calculated with the upslope of the signal intensity–time curves in the left ventricle and myocardium during the first pass of the contrast material at rest and stress has been proposed as an accurate parameter of myocardial perfusion (8,10,21). To our knowledge, however, in no study to date has diagnostic performance of contrast-enhanced cardiac MR imaging been compared with that of invasive measurement of CFR in patients who are suspected of having CAD.
Thus, the purpose of our study was to prospectively evaluate the accuracy of contrast-enhanced cardiac MR imaging for the determination of impaired CFR, with Doppler flow measurement as the reference standard.
|
MATERIALS AND METHODS
|
|---|
The study was approved by the institutional ethics committee. All patients provided written informed consent. Partial financial support was given by Bracco-Altana Pharma, Konstanz, Germany; however, the authors had complete control of the data and the information submitted for publication.
Patients
Consecutive patients who met our study criteria, were clinically suspected of having CAD, and were referred to our center for coronary angiography were included in the study, without other specific knowledge of their coronary status. Patients were included from October 2002 until March 2004. Inclusion criteria were chest pain or dyspnea at exercise or a positive finding at stress testing. Patients with clinical signs of severe congestive heart failure (New York Heart Association class 
III) or angina at low-level exercise (Canadian Cardiovascular Society class III), previous acute myocardial infarction within less than 3 months, severe valvular dysfunction, atrioventricular (Mobitz [Wenckeback]) block of grade II or higher, atrial flutter and/or fibrillation, severe chronic obstructive pulmonary disease, contraindications to MR imaging, or pregnancy were excluded from the study. All patients were advised not to smoke or drink tea or coffee within 24 hours before the examinations.
MR Imaging
Patients were examined in the supine position with a 1.5-T unit (Symphony Vision; Siemens, Erlangen, Germany) equipped with quantum gradients (30 mT/m). A four-element phased-array radiofrequency coil was used for signal reception. First, the double-oblique angle of the left ventricular axis was determined with a series of scout views. Then, a series of parallel short-axis cine views (8-mm-thick sections, 2-mm gap) were obtained that included the entire left ventricle from the apex to the base of the heart. For perfusion measurements at rest, three short-axis sections (basal, center, and apical) were determined by choosing one section in the center of the left ventricle and two sections at a distance of 20 mm from this reference point toward the base and apex, respectively. Three-section perfusion imaging was performed by using a single-shot saturation-recovery gradient-recalled-echo sequence with steady-state free precession (repetition time msec/echo time msec, 2.2/1.1; flip angle, 50°; field of view, 35.0 x 26.3 cm; matrix, 128 x 88; resulting pixel size, 2.7 x 3.0 mm, interpolated to 1.3 x 1.5 mm; section thickness, 8 mm). The saturation prepulse was not section selective and was applied for the acquisition of each section. The time between the saturation prepulse and the first excitation pulse was 10 msec, and this factor yielded an effective inversion time of 140 msec. During and after injection of a dose of 0.05 mmol per kilogram body weight gadobenate dimeglumine (MultiHance; Bracco-Altana) in an antecubal vein, flushed with saline at a flow rate of 5 mL/sec by using a power injector (Spectris; Medrad, Indianola, Pa), three images per heartbeat were acquired during a breath hold in expiration during 60 consecutive heartbeats. Patients were instructed to hold their breath as long as possible and then were allowed to breathe with shallow breaths.
Twenty minutes later, stress measurements were performed during infusion of adenosine (Adenoscan; Sanofi-Synthelabo, Paris, France) at a dose of 140 µg per kilogram body weight per minute. After 3 minutes of adenosine infusion, perfusion imaging was repeated as described earlier. Heart rate and blood pressure were monitored continuously during the entire study.
Coronary Angiography with Doppler Measurement as Reference Standard
All patients underwent coronary angiography within a mean ± standard deviation of 20.8 hours ± 7.5 (range, 12–42 hours) after contrast-enhanced cardiac MR imaging. After angiograms were obtained of all coronary arteries in multiple projections, intracoronary Doppler measurements (reference standard) were obtained by placing the tip of a 0.014-inch Doppler-tipped guidewire (FloWire; Jomed, Haan, Germany) distal to the stenosis in the target vessel. If stenotic lesions were found in two coronary arteries, intracoronary Doppler measurements were obtained in both. The average peak velocity was measured at rest and after 3 minutes of adenosine infusion of the dose mentioned before. The CFR was calculated by dividing the average peak velocity during adenosine infusion by the av-erage peak velocity at rest. A CFR of less than 2.00 was considered indicative of impaired myocardial perfusion. Doppler measurements were performed by two individuals (T.A.H. and T.H., with 7 and 10 years of experience with this procedure, respectively), without knowledge of the findings at contrast-enhanced MR im-aging. Heart rate and blood pressure were monitored continuously during the entire study.
Image Analysis
Short-axis sections were divided into six equiangular segments starting in a clockwise direction from the anterior septal insertion of the right ventricle (Fig 1). Segments were assigned to vascular regions according to the segmental model of the American Society of Echocardiography, with modifications to correct for variable coronary dominance as previously described (19). Segments 6, 1, and 2 were assigned to the left anterior descending artery; segments 2, 3, and 4 were assigned to the circumflex artery; and segments 4 and 5 were assigned to the right coronary artery. Within each segment, signal intensity was measured by defining regions of interest that comprised the entire myocardium for transmural analysis and the inner two rows of pixels of the myocardium for subendocardial analysis. Transmural and subendocardial signal intensity–time curves were generated for all segments by obtaining signal intensity on consecutive images before and during the arrival of contrast material (Fig 2a). The signal intensity–time curve for the left ventricle was generated on the basal section as a measure of the input function. The maximal initial upslope of every signal intensity–time curve was determined by using a sliding window with a four-point linear fit for the myocardium (Fig 2b) and a three-point linear fit for the left ventricle, as described earlier (19). Myocardial upslopes were then divided by left ventricular upslopes to correct them for the arrival speed of the bolus of contrast material. To calculate the MPRI for transmural analysis and the MPRI for subendocardial analysis, the resulting value at stress was divided by that at rest. We calculated an MPRI for every segment of a vascular region (comprising all three sections). Among these segments, we identified the four segments where the MPRI was lowest. The MPRIs of these segments were used to identify the lowest MPRI and to calculate the mean MPRI of the two, three, and four lowest segmental MPRIs of a vascular region. Also, the MPRI of all segments (or segmental MPRIs) of a vascular area was calculated. With this approach, all segments of a vascular region were considered, regardless of whether they were adjacent or not. Measurements were obtained independently and without knowledge of the Doppler results by two observers (C.T. and A.K.S.) with 2 years of experience in cardiac MR imaging. The mean values of both measurements were used for analysis.
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1: Diagrams show segments assigned to vascular regions. On every section, segments 6, 1, and 2 were assigned to the left anterior descending artery (LAD); segments 2, 3, and 4, to the circumflex artery (CFX); and segments 4 and 5, to the right coronary artery (RCA).
|
|
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a: Example of perfusion measurement in a patient with a CFR of 1.7 in the right coronary artery. (a) Sequential cardiac MR images (2.2/1.1, 50° flip angle°) obtained with a short-axis view in the center section show the first pass of a bolus of contrast material during rest (top row) and stress (bottom row). A region of delayed contrast material inflow is identified by a dark area in the inferior and inferoseptal segments (arrowheads). (b) Signal intensity–time curves obtained in segment 5 at rest and stress show that the delayed contrast material inflow results in a reduced upslope during stress compared with rest. AU = arbitrary units, SI = signal intensity.
|
|
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b: Example of perfusion measurement in a patient with a CFR of 1.7 in the right coronary artery. (a) Sequential cardiac MR images (2.2/1.1, 50° flip angle°) obtained with a short-axis view in the center section show the first pass of a bolus of contrast material during rest (top row) and stress (bottom row). A region of delayed contrast material inflow is identified by a dark area in the inferior and inferoseptal segments (arrowheads). (b) Signal intensity–time curves obtained in segment 5 at rest and stress show that the delayed contrast material inflow results in a reduced upslope during stress compared with rest. AU = arbitrary units, SI = signal intensity.
|
|
Statistical Analysis
Data are expressed as means ± standard deviations. Bland-Altman analysis was used to assess interobserver variability. Linear regression analysis was used to compare the MPRI and CFR, and correlation was performed with the Kendall 
rank correlation. The t test for unpaired samples was used to test for statistically significant differences of MPRI in vascular regions with and without a reduction in CFR. The paired-samples t test was used to analyze differences in heart rate, systolic blood pressure, and rate-pressure product (calculated as the product of heart rate and systolic blood pressure) between Doppler measurement and contrast-enhanced MR imaging and to analyze differences in these parameters between studies obtained at rest and stress. Receiver operating characteristic (ROC) analysis was used to analyze the effect of the number of segments used for calculation of the MPRI on diagnostic accuracy. Areas under the ROC curve (AUCs) were used to compare the accuracies of the different approaches. Sensitivity and specificity were calculated for all coordinates of each curve, and coordinates with best values for sensitivity and specificity were determined as cutoff values. A difference with a P value of less than .05 was considered statistically significant. Calculations were performed by using a standard statistical software package (SPSS, version 13.0; SPSS, Chicago, Ill).
|
RESULTS
|
|---|
Patients
A total of 40 patients were eligible for enrollment in the study. Five of these 40 patients were not included in the evaluation because (Fig 3) a high-grade stenosis of the left main stem was found in one patient, the Doppler wire could not be placed in the target vessel in two patients, contrast-enhanced MR images were of poor quality in one patient, and there was loss of data in one patient. The remaining 35 patients included 27 men and eight women, with a mean age of 63.5 years ± 8.2 and a mean body mass index of 28.2 kg/m2 ± 3.8. Ten (28%) of 35 patients had a myocardial infarction more than 3 months earlier. In 35 patients, arterial hypertension was present in 27 (77%), hyperlipoproteinemia was present in 27 (77%), diabetes mellitus was present in eight (23%), and ongoing smoking was present in six (17%). In eight patients, measurements were performed in two vascular regions. As a result, a total of 43 measurements were performed: Twenty-six measurements were in the left anterior descending artery, eight were in the circumflex artery, and nine were in the right coronary artery. According to American Heart Association classification, 12 stenoses were localized in a proximal segment, 26 were in a middle segment, and five were in a distal segment.
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3: Flow diagram of the patients eligible for participation in the study. Five patients were not included in the final evaluation for various reasons. ce-CMR = contrast-enhanced cardiac MR imaging.
|
|
Interobserver Variability
Good agreement between both observers was found in the measurement of MPRI. The mean difference in the MPRI between both observers was 0.01 ± 0.23 for transmural measurements and 0.01 ± 0.27 for subendocardial measurements (Fig 4).
Hemodynamic Data
No statistically significant differences were found when we compared contrast-enhanced cardiac MR imaging with Doppler measurements obtained at rest and stress (Table 1). The heart rate and rate-pressure product obtained at stress were significantly higher than those obtained at rest. The systolic blood pressure at stress was significantly lower than that at rest.
Comparison of Contrast-enhanced Cardiac MR Imaging and Doppler Flow Measurement
Linear correlations were found between all calculated MPRIs and CFRs (mean r = 0.43 ± 0.04; P < .05 for each correlation). When we eliminated the influence of possible outliers with the Kendall rank correlation, we also found a weak but statistically significant positive correlation (r = 0.33, P = .01) between subendocardial MPRI 3 and the CFR. The MPRI was significantly lower in vascular regions with a CFR of less than 2.00 than in vascular regions with a CFR of 2.00 or greater (Figs 5, 6; Table 2). Sensitivity, specificity, and AUC were higher in subendocardial than in transmural measurements, a finding that indicates that subendocardial measurements are more accurate than transmural measurements in the detection of a reduced CFR (Table 3). For transmural and subendocardial measurements, the AUC increased with the increasing number of segments used for calculating MPRI until it reached best results with use of the mean MPRI of the lowest MPRI of three segments of a vascular region for analysis. ROC curves are shown for the lowest MPRI of three segments for both the transmural and subendocardial analyses (Fig 7). The use of more than three segments for calculating the MPRI resulted in a decrease in the AUC at ROC analysis (Table 3). The cutoff value was 1.33 for transmural MPRI 3 and 1.21 for subendocardial MPRI 3.
View larger version (5K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5: Graph shows the linear correlation between the CFR and MPRI in the three lowest segments of a vascular region with subendocardial analysis (MPRIsub3), which is calculated with 0.943 + (0.128 · CFR) (r = 0.44, P < .05).
|
|
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6: Box plots (indicating minimum, 25% quartile, median, 75% quartile, and maximum values) of transmural (left plot) and subendocardial (right plot) perfusion measurements in vascular regions with a normal CFR of at least 2.00 and a reduced CFR of less than 2.00. The mean MPRI for the segments with the three lowest transmural MPRI values (MPRItrans3) and those with the three lowest subendocardial MPRI values (MPRIsub3), are shown. The difference between regions with normal and reduced CFR was statistically significant (P < .001).
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7: ROC curves of mean MPRI for segments with the three lowest transmural MPRI values (MPRItrans3) and mean MPRI for segments with the three lowest subendocardial MPRI values (MPRIsub3) for prediction of CFR of less than 2.00. The diagnostic accuracy of mean MPRI for segments with the three lowest subendocardial MPRI values (AUC, 0.85; sensitivity, 84%; specificity, 75%; cutoff value, 1.21) was better than that of mean MPRI for segments with the three lowest transmural MPRI values (AUC, 0.79; sensitivity, 79%; specificity, 58%; cutoff value, 1.33).
|
|
|
DISCUSSION
|
|---|
The results of our study show that the MPRI derived from noninvasive contrast-enhanced cardiac MR imaging is linearly correlated with CFR, as determined with invasive intracoronary Doppler flow measurement. A reduction in the CFR can be detected with MPRI, with subendocardial measurement showing a better diagnostic accuracy than transmural measurement. Best diagnostic accuracy was achieved by averaging MPRIs of the three lowest segments of the vascular region, with use of a cutoff value of 1.33 for transmural measurement and that of 1.21 for subendocardial measurement.
To our knowledge, our study is the first in which contrast-enhanced cardiac MR imaging as a method for noninvasive measurement of myocardial perfusion is compared with invasive intracoronary Doppler flow measurement in patients with CAD. We found a moderate linear correlation between all calculated MPRIs and CFRs. The observation of a linear correlation between CFR and a method for the functional evaluation of myocardial perfusion has been made in an earlier study (4). In that study, good correlation (r = 0.76; P < .001) was found between the CFR obtained with intracoronary Doppler measurement and a myocardial perfusion reserve derived from [15O]H2O PET. Patients in that previous study, however, were highly selected, with isolated proximal stenosis of the left coronary artery and exclusion of diabetes and hypertensive heart disease. Recent studies have shown that microvascular perfusion is severely altered in animals and humans with diabetes mellitus (22,23), hypertension (24), and obesity (25). Therefore, the moderate correlation between epicardial flow obtained with invasive CFR and myocardial blood flow MPRI obtained with contrast-enhanced MR imaging found in our study is most likely related to the high prevalence of hypertension and diabetes mellitus in the examined patients.
Contrast-enhanced cardiac MR imaging has previously been evaluated with respect to CFR, as measured with an external Doppler flowmeter in dogs with artificial partial coronary occlusion (26). In that study, a regional reduction in vasodilated flow in viable myocardium was detected with high accuracy over the range of one to five times resting flow. It was concluded that contrast-enhanced cardiac MR imaging can depict and aid in the quantification of reductions in CFR that result from coronary stenosis. The results of our study support these findings and extend the value of contrast-enhanced cardiac MR imaging to the detection of impaired CFR in humans with CAD. The diagnostic performance of the method suggests that contrast-enhanced cardiac MR imaging may be useful in the screening and follow-up of patients who are suspected of having CAD.
In our study, analysis of subendocardial perfusion with contrast-enhanced cardiac MR imaging provided better diagnostic accuracy in the detection of a reduced CFR than did analysis of transmural perfusion. This finding is consistent with results from previous studies. In an earlier study with microsphere-assessed myocardial blood flow, it was demonstrated that contrast-enhanced cardiac MR imaging reflects intramyocardial redistribution of blood flow and that perfusion abnormalities can be identified most distinctly in the subendocardium (27). In one study, the performance of MPRI in subendocardial layers was better for the detection of hemodynamically significant CAD than was the transmural MPRI (18). Researchers in another study demonstrated that patients with syndrome X can be differentiated from healthy subjects with analysis of subendocardial MPRI but not with analysis of transmural MPRI (28). Some investigators even excluded part of the outer ventricular wall from analysis to achieve a more subendocardial weighting of the perfusion analysis (10). Subendocardium is the most remote territory of the vascular bed with the smallest vascular branches. In the case of microvascular abnormalities, the smallest branches are the first to be affected. In addition, wall stress in the myocardium is higher in inner than in outer regions, and this factor leads to a demand for a higher perfusion pressure for the subendocardium than for the subepicardium (29). With ischemia, the subendocardial layers are, therefore, the first to be adversely affected, and, in case of acute nontransmural myocardial infarction, scarring of the myocardium is predominantly located in the subendocardium (30). Therefore, analysis of subendocardial perfusion enables the detection of myocardial perfusion impairment with a higher sensitivity than does analysis of transmural perfusion.
With contrast-enhanced cardiac MR imaging, however, discrimination of regional perfusion inhomogeneities cannot be provided with CFR measurement. To correct for this methodological difference, we analyzed the effect of the number of segments used to calculate MPRI on diagnostic performance. In our subjects, we achieved best diagnostic performance by calculating mean MPRI of the three segments with the lowest MPRI of a vascular region. Calculation of the mean MPRI in the lowest segments of a vascular region is an alternative to previously proposed methods. Earlier studies defined clinically significant CAD as a reduction of MPRI in an arbitrarily defined region of interest within a vascular region (28) or in at least one segment of a vascular region (19). Investigators in another study (18) analyzed contrast material wash-in in one to two segments per section in the center of a vascular region. In yet another study (10), the highest sensitivity, specificity, and accuracy in the detection of relevant CAD were achieved with analysis of only the second smallest value of the MPRI in a vascular region to reduce the negative influence of outliers. By calculating the mean of several lowest segments of a vascular region, we propose a method that, at the same time, corrects for the methodologic difference between contrast-enhanced cardiac MR imaging and Doppler flow measurement and prevents the risk of incorrect test results caused by outliers.
Our study revealed that a cutoff value of 1.21 for subendocardial MPRI 3 was the best for detection of impaired CFR. In another study, a cutoff value of 1.1 was found to enable the detection of clinically significant CAD (10). In that earlier study, only the second smallest MPRI of three sections was used for analysis, and this factor could have resulted in the lower cutoff value compared with that in our study. Researchers in another study (19) described a higher cutoff value of 1.5 for MPRI. In contrast to values found in our study, that higher value was derived from non-ischemic reference segments without stenosis in the supplying coronary artery. Cutoff values in our study were higher for transmural measurements than they were for subendocardial measurements. In segments in which perfusion impairment was predominantly located in the subendocardium, the amount of myocardium with normal perfusion was higher in transmural measurements than it was in subendocardial measurements. This difference resulted in a steeper upslope of the signal intensity–time curve at stress, a higher MPRI, and higher cutoff values for transmural measurements.
The use of a retrospectively determined cutoff value in ROC analysis may lead to a beneficial evaluation of the diagnostic performance. The value of a prospective application of this cutoff value to future patients is yet to be determined. Cutoff values determined in this study are directly related to the method. It is possible that other methods, such as a dual-bolus perfusion method (31), could lead to different values in the same patients. For the statistical analysis, we assumed that measurements in two different vascular regions in the same patient were not dependent. A dependency, however, cannot be definitely excluded. A drawback for contrast-enhanced cardiac MR imaging—and this includes all modalities used to image myocardial perfusion—is that it is impossible to differentiate whether perfusion impairment results from stenosis of an epicardial vessel for which coronary intervention is needed or from a microvascular abnormality for which coronary intervention is not needed.
The results of our study demonstrate that an impaired CFR can be detected with an optimized MPRI derived from contrast-enhanced cardiac MR imaging. In the detection of a reduction in CFR, the diagnostic accuracy obtained with analysis of subendocardial perfusion was better than that obtained with analysis of transmural perfusion. This method is potentially useful as a tool to detect impaired myocardial perfusion.
|
ADVANCES IN KNOWLEDGE
|
|---|
- Impaired coronary flow velocity reserve can be determined with contrast-enhanced cardiac MR imaging and the use of an optimized myocardial perfusion reserve index.
- The accuracy of subendocardial perfusion analysis with contrast-enhanced cardiac MR imaging is higher than that of transmural perfusion analysis.
|
FOOTNOTES
|
|---|
Abbreviations: AUC = area under the ROC curve • CAD = coronary artery disease • CFR = coronary flow velocity reserve • MPRI = myocardial perfusion reserve index • ROC = receiver operating characteristic
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantors of integrity of entire study, A.A.B., A.S., G.K.L.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, A.A.B., G.K.L.; clinical studies, A.A.B., A.S., K.M., C.T., A.K.S., T.A.H., T.H., G.K.L.; statistical analysis, A.A.B., G.K.L.; and manuscript editing, A.A.B., A.S., G.A., T.M., G.K.L.
|
References
|
|---|
- Joye JD, Schulman DS, Lasorda D, Farah T, Donohue BC, Reichek N. Intracoronary Doppler guide wire versus stress single-photon emission computed tomographic thal-lium-201 imaging in assessment of intermediate coronary stenoses. J Am Coll Cardiol 1994;24:940–947.[Abstract]
- Porenta G, Binder T, Moertl D, et al. Functional assessment of coronary arteries by poststenotic intravascular Doppler ultrasound. J Vasc Res 2000;37:594–602.[CrossRef][Medline]
- Chamuleau SA, Meuwissen M, van Eck-Smit BL, et al. Fractional flow reserve, absolute and relative coronary blood flow velocity reserve in relation to the results of technetium-99m sestamibi single-photon emission computed tomography in patients with two-vessel coronary artery disease. J Am Coll Cardiol 2001;37:1316–1322.[Abstract/Free Full Text]
- Miller DD, Donohue TJ, Wolford TL, Kern MJ. Assessment of blood flow distal to coronary artery stenoses: correlations between myocardial positron emission tomography and poststenotic intracoronary Doppler flow reserve. Circulation 1996;94:2447–2454.[Abstract/Free Full Text]
- Miller DD, Holmvang G, Gill JB, et al. MRI detection of myocardial perfusion changes by gadolinium-DTPA infusion during dipyridamole hyperemia. Magn Reson Med 1989;10:246–255.[Medline]
- Schaefer S, Lange RA, Gutekunst DP, Parkey RW, Willerson JT, Peshock RM. Contrast-enhanced magnetic resonance imaging of hypoperfused myocardium. Invest Radiol 1991;26:551–556.[CrossRef][Medline]
- Atkinson DJ, Burstein D, Edelman RR. First-pass cardiac perfusion: evaluation with ultrafast MR imaging. Radiology 1990;174:757–762.[Abstract/Free Full Text]
- Cullen JH, Horsfield MA, Reek CR, Cherryman GR, Barnett DB, Samani NJ. A myocardial perfusion reserve index in humans using first-pass contrast-enhanced magnetic resonance imaging. J Am Coll Cardiol 1999;33:1386–1394.[Abstract/Free Full Text]
- Penzkofer H, Wintersperger BJ, Knez A, Weber J, Reiser M. Assessment of myocardial perfusion using multisection first-pass MRI and color-coded parameter maps: a comparison to 99mTc Sesta MIBI SPECT and systolic myocardial wall thickening analysis. Magn Reson Imaging 1999;17:161–170.[CrossRef][Medline]
- Nagel E, Klein C, Paetsch I, et al. Magnetic resonance perfusion measurements for the noninvasive detection of coronary artery disease. Circulation 2003;108:432–437.[Abstract/Free Full Text]
- Eichenberger AC, Schuiki E, Kochli VD, Amann FW, McKinnon GC, von Schulthess GK. Ischemic heart disease: assessment with gadolinium-enhanced ultrafast MR imaging and dipyridamole stress. J Magn Reson Imaging 1994;4:425–431.[Medline]
- Matheijssen NA, Louwerenburg HW, van Rugge FP, et al. Comparison of ultrafast dipyridamole magnetic resonance imaging with dipyridamole sestaMIBI SPECT for detection of perfusion abnormalities in patients with one-vessel coronary artery disease: assessment by quantitative model fitting. Magn Reson Med 1996;35:221–228.[Medline]
- Keijer JT, van Rossum AC, van Eenige MJ, et al. Magnetic resonance imaging of regional myocardial perfusion in patients with single-vessel coronary artery disease: quantitative comparison with (201)thallium-SPECT and coronary angiography. J Magn Reson Imaging 2000;11:607–615.[CrossRef][Medline]
- Klein MA, Collier BD, Hellman RS, Bamrah VS. Detection of chronic coronary artery disease: value of pharmacologically stressed, dynamically enhanced turbo-fast low-angle shot MR images. AJR Am J Roentgenol 1993;161:257–263.[Abstract/Free Full Text]
- Lauerma K, Virtanen KS, Sipila LM, Hekali P, Aronen HJ. Multislice MRI in assessment of myocardial perfusion in patients with single-vessel proximal left anterior descending coronary artery disease before and after revascularization. Circulation 1997;96:2859–2867.[Abstract/Free Full Text]
- Ishida N, Sakuma H, Motoyasu M, et al. Noninfarcted myocardium: correlation between dynamic first-pass contrast-enhanced myocardial MR imaging and quantitative coronary angiography. Radiology 2003;229:209–216.[Abstract/Free Full Text]
- Bremerich J, Buser P, Bongartz G, et al. Noninvasive stress testing of myocardial ischemia: comparison of GRE-MRI perfusion and wall motion analysis to 99 mTc-MIBI-SPECT, relation to coronary angiography. Eur Radiol 1997;7:990–995.[CrossRef][Medline]
- Schwitter J, Nanz D, Kneifel S, et al. Assessment of myocardial perfusion in coronary artery disease by magnetic resonance: a comparison with positron emission tomography and coronary angiography. Circulation 2001;103:2230–2235.[Abstract/Free Full Text]
- Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myocardial ischemia from perfusion reserve based on cardiovascular magnetic resonance. Circulation 2000;101:1379–1383.[Abstract/Free Full Text]
- Paetsch I, Jahnke C, Wahl A, et al. Comparison of dobutamine stress magnetic resonance, adenosine stress magnetic resonance, and adenosine stress magnetic resonance perfusion. Circulation 2004;110:835–842.[Abstract/Free Full Text]
- Al-Saadi N, Gross M, Bornstedt A, et al. Comparison of different semiquantitative parameters for the assessment of a myocardial perfusion index by magnetic resonance imaging. Z Kardiol 2001;90:824–834.[CrossRef][Medline]
- Iltis I, Kober F, Dalmasso C, Cozzone PJ, Bernard M. Noninvasive characterization of myocardial blood flow in diabetic, hypertensive, and diabetic-hypertensive rats using spin-labeling MRI. Microcirculation 2005;12:607–614.[CrossRef][Medline]
- Sander GE, Wilklow FE, Giles TD. Heart failure in diabetes mellitus: causal and treatment considerations. Minerva Cardioangiol 2004;52:491–503.[Medline]
- Rodriguez-Porcel M, Zhu XY, Chade AR, et al. Functional and structural remodeling of the myocardial microvasculature in early experimental hypertension. Am J Physiol Heart Circ Physiol 2006;290:H978–H984. [Published correction appears in Am J Physiol Heart Circ Physiol 2006;290:H2163.]
- de Jongh RT, Serne EH, IJzerman RG, de Vries G, Stehouwer CD. Impaired microvascular function in obesity: implications for obesity-associated microangiopathy, hypertension, and insulin resistance. Circulation 2004;109:2529–2535.[Abstract/Free Full Text]
- Klocke FJ, Simonetti OP, Judd RM, et al. Limits of detection of regional differences in vasodilated flow in viable myocardium by first-pass magnetic resonance perfusion imaging. Circulation 2001;104:2412–2416.[Abstract/Free Full Text]
- Keijer JT, van Rossum AC, Wilke N, et al. Magnetic resonance imaging of myocardial perfusion in single-vessel coronary artery disease: implications for transmural assessment of myocardial perfusion. J Cardiovasc Magn Reson 2000;2:189–200.[Medline]
- Panting JR, Gatehouse PD, Yang GZ, et al. Abnormal subendocardial perfusion in cardiac syndrome X detected by cardiovascular magnetic resonance imaging. N Engl J Med 2002;346:1948–1953.[Abstract/Free Full Text]
- Hoffman JI. Heterogeneity of myocardial blood flow. Basic Res Cardiol 1995;90:103–111.[CrossRef][Medline]
- Jennings RB, Steenbergen C Jr, Reimer KA. Myocardial ischemia and reperfusion. Monogr Pathol 1995;37:47–80.[Medline]
- Christian TF, Rettmann DW, Aletras AH, et al. Absolute myocardial perfusion in canines measured by using dual-bolus first-pass MR imaging. Radiology 2004;232:677–684.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
