CLINICAL RESEARCH: CARDIAC IMAGING
Cardiovascular Magnetic Resonance Perfusion Imaging at 3-Tesla for the Detection of Coronary Artery DiseaseA Comparison With 1.5-Tesla
Adrian S.H. Cheng, MBBS, MRCP*, ,
Tammy J. Pegg, MBChB, MRCP*,
Theodoros D. Karamitsos, MD*,,
Nick Searle, DCR(R) ,
Michael Jerosch-Herold, PhD||,
Robin P. Choudhury, DM, MRCP ,
Adrian P. Banning, MD, FRCP, FESC,
Stefan Neubauer, MD, FRCP*,,
Matthew D. Robson, PhD*, and
Joseph B. Selvanayagam, DPhil, FRACP, FESC*,,*
* University of Oxford Centre for Clinical Magnetic Resonance Research, Oxford, United Kingdom
Department of Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom
Department of Radiology
Department of Cardiology, John Radcliffe Hospital, Oxford, United Kingdom; and the ||Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon
Manuscript received September 21, 2006;
revised manuscript received March 2, 2007,
accepted March 6, 2007.
* Reprint requests and correspondence: Dr. Joseph B. Selvanayagam, University of Oxford, Cardiovascular Medicine, Headley Way, Headington, Oxford, United Kingdom (Email: joseph.selvanayagam{at}cardiov.ox.ac.uk).
 |
Abstract
|
|---|
Objectives: This study was designed to establish the diagnostic accuracy of cardiovascular magnetic resonance (CMR) perfusion imaging at 3-Tesla (T) in suspected coronary artery disease (CAD).
Background: Myocardial perfusion imaging is considered one of the most compelling applications for CMR at 3-T. The 3-T systems provide increased signal-to-noise ratio and contrast enhancement (compared with 1.5-T), which can potentially improve spatial resolution and image quality.
Methods: Sixty-one patients (age 64 ± 8 years) referred for elective diagnostic coronary angiography (CA) for investigation of exertional chest pain were studied (before angiogram) with first-pass perfusion CMR at both 1.5- and 3-T and at stress (140 µg/kg/min intravenous adenosine, Adenoscan, Sanofi-Synthelabo, Guildford, United Kingdom) and rest. Four short-axis images were acquired during every heartbeat using a saturation recovery fast-gradient echo sequence and 0.04 mmol/kg Gd-DTPA bolus injection. Quantitative CA served as the reference standard. Perfusion deficits were interpreted visually by 2 blinded observers. We defined CAD angiographically as the presence of  1 stenosis of 50% diameter in any of the main epicardial coronary arteries or their branches with a diameter of 2 mm.
Results: The prevalence of CAD was 66%. All perfusion images were found to be visually interpretable for diagnosis. We found that 3-T CMR perfusion imaging provided a higher diagnostic accuracy (90% vs. 82%), sensitivity (98% vs. 90%), specificity (76% vs. 67%), positive predictive value (89% vs. 84%), and negative predictive value (94% vs. 78%) for detection of significant coronary stenoses compared with 1.5-T. The diagnostic performance of 3-T perfusion imaging was significantly greater than that of 1.5-T in identifying both single-vessel disease (area under receiver-operator characteristic [ROC] curve: 0.89 ± 0.05 vs. 0.70 ± 0.08; p < 0.05) and multivessel disease (area under ROC curve: 0.95 ± 0.03 vs. 0.82 ± 0.06; p < 0.05). There was no difference between field strengths for the overall detection of coronary disease (area under ROC curve: 0.87 ± 0.05 vs. 0.78 ± 0.06; p = 0.23).
Conclusions: Our study showed that 3-T CMR perfusion imaging is superior to 1.5-T for prediction of significant single- and multi-vessel coronary disease, and 3-T may become the preferred CMR field strength for myocardial perfusion assessment in clinical practice.
|
Abbreviations and Acronyms
| | CAD = coronary artery disease | | CMR = cardiovascular magnetic resonance | | CNR = contrast-to-noise ratio | | IV = intravenous | | LAD = left anterior descending | | LV = left ventricle | | MI = myocardial infarction | | PET = positron emission tomography | | RCA = right coronary artery | | ROC = receiver operating characteristic | | ROI = region of interest | | SI = signal intensity | | SNR = signal-to-noise ratio | | SPECT = single-photon emission computed tomography | | T = Tesla |
|
Myocardial perfusion deficits appear early in the ischemic cascade (1). In coronary artery disease (CAD), the revascularization strategy is guided by evaluation of the coronary anatomy, combined with functional assessment of the hemodynamic significance of stenoses and the extent and distribution of ischemia.
The most widely used techniques in the assessment of myocardial perfusion have significant limitations. Although single-photon emission computed tomography (SPECT) myocardial perfusion imaging is a useful tool for noninvasive diagnosis of myocardial ischemia, it is susceptible to attenuation artifacts (2). Positron emission tomography (PET) imaging corrects for attenuation but has limited availability and higher expense. Both SPECT and PET involve exposure to ionizing radiation. Cardiovascular magnetic resonance imaging (CMR) is emerging as an alternative noninvasive approach for the assessment of myocardial perfusion, because it offers superior spatial resolution (3) that potentially allows differentiation between subendocardial and subepicardial perfusion, does not involve ionizing radiation, and, like PET, has the potential to measure myocardial perfusion in absolute terms (4,5). Contrast agents can be tracked as they traverse the myocardium after intravenous (IV) injection, allowing assessment of myocardial perfusion at rest and during pharmacologic stress with vasodilators such as adenosine (6,7).
Although studies have shown that visual and semiquantitative assessment of perfusion CMR at 1.5-Tesla (T) has moderate accuracy for detection of CAD (8–11), it remains principally limited by low differences in contrast enhancement between normal and underperfused myocardium. Compared to perfusion CMR at 1.5-T, 3-T systems provide increased signal-to-noise ratio (12) and contrast enhancement (12,13), which can be used to improve spatial resolution and image quality. The aim of this study was to assess the clinical utility of CMR perfusion at 3-T in a population of patients with suspected CAD.
|
Methods
|
|---|
Population.
We prospectively recruited 65 consecutive patients with suspected CAD who were awaiting diagnostic cardiac catheterization as part of routine clinical care. Patients were excluded if they were medically unstable, had had a myocardial infarction (MI) in the preceding 2 weeks, or had any contraindications to magnetic resonance imaging examination (metallic implants such as pacemakers, defibrillators, cerebral aneurysm clips, ocular metallic deposits, severe claustrophobia) or to adenosine (second- or third-degree atrioventricular block, history of asthma). Patients underwent CMR imaging in the 2 weeks before coronary angiography. No clinical cardiac events occurred in the period between the CMR scan and coronary angiography. The institutional research ethics committee approved our study. All participants gave written informed consent.
CMR protocol.
Patients were asked to avoid agents that could potentiate (dipyridamole) or antagonize (e.g., caffeine, methylxanthines) the effects of adenosine for at least 24 h before CMR. Each patient underwent CMR at both 1.5-T (Sonata, Siemens Medical Solutions, Erlangen, Germany) and 3-T (Trio, Siemens Medical Solutions) on the same day. The scanners had gradient systems with the same specifications, with optimal and identical settings for echo and repetition time in each instance, so that for a segmented gradient echo sequence, the differences in intrinsic signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) between the 2 field strengths were accurately reflected. Perfusion imaging on each system was performed at least an hour apart and in randomized order. Images were acquired with the patient supine, using coils optimized for body imaging by Siemens: anterior phased-array surface coils and either posterior phased-array surface coils (at 3-T) or 2 elements of the integrated spine coil (at 1.5-T). An active Siemens electrocardiogram lead set was used, with the 3 leads positioned over the anterior chest wall.
Cine CMR.
Cine CMR was performed at 1.5-T. From standard pilot images, short-axis cine images covering the entire left ventricle (LV) were acquired using an electrocardiogram-gated steady-state free precession sequence (echo time 1.5 ms, repetition time 3 ms, flip angle 60o).
Perfusion CMR.
Throughout each CMR scan, patients were monitored by electrocardiography, noninvasive sphygmomanometry, and pulse oximetry. After 4 min of IV adenosine infusion (140 µg/kg/min; Adenoscan, Sanofi-Synthelabo, Guildford, United Kingdom), or earlier if angina was provoked, a gadolinium-based contrast agent (Gadodiamide, Omniscan, GE Healthcare, Amersham, United Kingdom) was administered at a dose of 0.04 mmol/kg body weight (IV injection rate 6 ml/s), followed by a 15-ml saline flush at the same rate. The contrast dose was chosen to allow repeat measurement after a delay of at least 1 h and to be within the dose range used previously for semiquantitative analysis of perfusion (9). Perfusion images were acquired every cardiac cycle during the first pass of contrast, using a T1-weighted fast-gradient echo (turbo fast low-angle shot) sequence (echo time 1.04 ms, repetition time 2 ms, saturation recovery time 100 ms, voxel size 2.1 x 2.6 x 8 mm; flip angle 18o at 1.5-T and 17o at 3-T) while patients held their breath for as long as possible in end-inspiration. The last 38 patients underwent parallel imaging with 2-fold acceleration at both field strengths. Four short-axis slices, positioned from the base to the apex of the LV, were obtained. The same imaging sequence was repeated 15 min later without adenosine to obtain perfusion images at rest.
Coronary angiography.
Less than 2 weeks after the CMR examination, all patients underwent coronary angiography using standard techniques. Images of the coronary arteries were obtained in multiple projections, and, as much as possible, overlap of side branches and foreshortening of relevant coronary stenoses was avoided.
CMR analysis.
For each patient, using Argus software (version VA60C, Siemens AG), left ventricular end-diastolic and -systolic volumes, ejection fraction, and mass were calculated in the standard way by manually tracing the endocardial and epicardial contours in end-diastolic and -systolic images obtained at 1.5-T.
Perfusion CMR scans were interpreted in random order by 2 observers acting in consensus, blinded to all data, including clinical information, field strength, and angiographic and cine CMR findings. For visual grading of perfusion deficits, stress and rest perfusion scans were magnified and displayed simultaneously. The presence of a perfusion deficit was defined by the following criteria: 1) the regional hypoenhancement persisted for more than 3 phases after maximal enhancement of segments that appeared most normal, 2) the area of hypoenhancement corresponded with a coronary artery territory (left anterior descending artery: anterior or septal segments; left circumflex artery: lateral segments; right coronary artery: inferior segments), and 3) regional signal intensity or contrast enhancement was reduced or delayed compared to other myocardial segments in the same slice. Within this definition, perfusion deficits could be either fixed (present at stress and rest) or inducible (present at stress only). Perfusion analysis of each myocardial segment (except the apex) was performed using the 17-segment model recommended by the American Heart Association (14). Each myocardial segment was graded as 0 (normal), 1 (possible perfusion deficit), or 2 (definite perfusion deficit). A third observer interpreted all segments that were graded as "possible perfusion deficits" and the majority decision of all 3 observers was used to determine the presence or absence of a perfusion deficit. The presence of artifacts was graded according to type (respiratory, dark rim, reconstruction, and so forth) and severity.
In 30 patients, the mid-ventricular slice of the rest scan was selected by an observer blinded to patient information and field strength. For all phases in each selected slice, using Argus software, the observer measured the mean signal intensity (SI) in a region of interest (ROI) at least 12 pixels in size, placed in the anteroseptal myocardial segment, as well as the standard deviation of the noise in a circular region outside the body and anterior to the chest wall. Care was taken to avoid epicardial fat and ventricular cavity blood pool. Image SNR was calculated as (peak SI in the ROI)/(SD of noise), and CNR of the SI traces was calculated as (peak SI in the ROI – mean before-contrast SI in the ROI)/(SD of before-contrast SI in the ROI).
Angiographic analysis.
The diagnostic X-ray angiogram served as the reference standard in defining the degree of coronary stenosis. Each myocardial segment was ascribed a coronary artery territory according to standard criteria (14). Any lesion with a diameter stenosis of >40% by handheld caliper was subsequently analyzed by computerized quantitative methods using the Siemens Quantcor Coronary Analysis software. The contrast-filled catheter was used for image magnification calibration. Significant CAD was defined angiographically as the presence of at least 1 stenosis of 50% diameter in any of the main epicardial coronary arteries or their branches with a diameter of 2 mm.
Statistical analysis.
Data analysis was performed using SPSS 14.0 for Windows (SPSS Inc., Chicago, Illinois). Continuous data were compared using t tests, which were paired where appropriate (e.g., adenosine infusion time, pulse rate, systolic and diastolic blood pressure, rate pressure product, SNR, CNR). Chi-square tests were used to compare discrete data. The Fisher exact test was used when the assumptions of the chi-square test were not met. Using computer software (Medcalc 9.1.0.1
[EC]
, Mariakerke, Belgium), receiver-operating characteristic (ROC) curve analyses were performed to compare the diagnostic performance of perfusion CMR at each field strength (15,16). Statistical tests were 2-tailed, and p < 0.05 was considered to be significant.
|
Results
|
|---|
Study population.
A total of 65 consecutive patients were enrolled in the study. In 3 patients, both 1.5- and 3-T CMR scans could not be completed: 2 patients were claustrophobic, and 1 could not tolerate adenosine. Hence, 62 of 65 (95%) of the study cohort successfully completed both CMR scan protocols. One patient successfully completed both CMR scans, but X-ray coronary angiography was cancelled for clinical reasons unrelated to the CMR findings. Thus, 61 patients (94% of the cohort) were included in the final analysis, of whom 34 (56%) had their first perfusion CMR scan performed at 1.5-T.
Table 1
outlines the baseline characteristics of the study cohort. Seventy-five percent of the cohort were men with a mean age of 64 years, reflecting the referral pattern for diagnostic angiography in our hospital. Although women represented only one-quarter of the study group, they accounted for more than one-half of the normal studies in our cohort. X-ray coronary angiography demonstrated significant coronary artery stenoses in 66% of patients (40 of 61). There was an even distribution of patients with no significant CAD (n = 21) and single-vessel (n = 19) or multivessel (n = 21) disease. In terms of the anatomic location of the coronary stenoses, 30 patients (49%) had significant left anterior descending (LAD) coronary artery stenosis, 20 (33%) had significant left circumflex coronary artery stenosis, and 25 (41%) had significant right coronary artery (RCA) stenosis. Although the prevalence of CAD was high, and more than one-half of patients with significant coronary disease had more than 1 affected vessel, mean LV ejection fraction was well preserved at 68 ± 9%. There was a trend (p = 0.08) to increased mass index in those patients with CAD, possibly related to the predominance of men in this group.
Adenosine was infused for a similar period of time at both field strengths, and there was no difference between the field strengths in the subsequent hemodynamic response (Table 2). Only 1 patient had significant symptoms (dyspnea) as a result of adenosine that required premature termination of the examination. Nine scans (7%) were complicated by a very short period of heart block (lasting <5 s), all of which resolved on stopping the infusion and did not prevent completion of both CMR scans.
Detection of CAD.
When all patients were considered, 3-T CMR perfusion imaging provided a higher (although not statistically significant) diagnostic accuracy (90% vs. 82%, p = 0.33), sensitivity (98% vs. 90%, p = 0.38), specificity (76% vs. 67%, p = 0.73), positive predictive value (89% vs. 84%, p = 0.55), and negative predictive value (94% vs. 78%, p = 0.34) for the detection of significant coronary stenoses compared to 1.5-T CMR perfusion imaging (Table 3). Perfusion imaging (3-T) failed to detect coronary disease in only 1 patient, who had a 60% stenosis of the distal RCA. At a threshold for significant coronary stenosis of 70% (rather than 50%), 3-T continued to provide higher diagnostic accuracy (77% vs. 69%), sensitivity (100% vs. 90%), and specificity (55% vs. 48%) than 1.5-T. The diagnostic performance of 3-T was higher than at 1.5-T, irrespective of the order in which the scans were performed. When imaging was performed at 1.5-T first, the area under ROC curve was 0.85 ± 0.07 vs. 0.78 ± 0.08, respectively. Similarly, when imaging was performed at 3-T first, the area under ROC curve was 0.91 ± 0.06 vs. 0.79 ± 0.09, respectively. Excluding patients with a history of prior MI reduced sensitivity (97% at 3-T and 88% at 1.5-T) and increased specificity at 3-T to 83%. Figures 1 and 2
 demonstrate patient examples at both field strengths.
View this table:
[in this window]
[in a new window]
|
Table 3 Diagnostic Performance of 1.5- and 3-T Perfusion Imaging for the Detection of Significant Coronary Artery Stenosis
|
|
View larger version (87K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1 Examples of Agreement Between the 2 Field Strengths for the Identification of Significant Coronary Artery Disease
The angiogram of Patient #1 (top row) in the right lateral (right anterior oblique 90o) projection demonstrates a subtotally occluded left anterior descending (LAD) coronary artery after the bifurcation of the first diagonal (between the 2 full black arrows). The LAD continues partially filled (indicated by the 3 intermittent black arrows) and the LAD territory is partially collateralized from the diagonal branch. There are corresponding stress perfusion deficits in the anteroseptal and inferoseptal segments at both field strengths (white arrows). Patient #2 (bottom row) had a 70% stenosis in the right coronary artery, with corresponding perfusion deficits at 1.5- and 3-Tesla (T) imaging (white arrows).
|
|
View larger version (85K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2 Examples of the Superiority of 3-T Imaging When Compared to 1.5-T
Patient #1 (top row) has a 70% stenosis of the left anterior descending (LAD) coronary artery. Whereas 3-Tesla (T) imaging demonstrates a corresponding stress perfusion deficit in the anteroseptal segment, visual assessment of the 1.5-T scan failed to delineate a deficit. Patient #2 (bottom row) had significant multivessel coronary artery disease (left system not shown). Although 1.5-T imaging demonstrates a perfusion deficit of the anterior segment (white arrow), corresponding to a LAD coronary stenosis, it fails to identify a perfusion deficit relating to the 90% stenosis in the right coronary artery. The 3-T imaging detects reversible perfusion deficits in the anterior, anteroseptal, inferoseptal, and inferior segments (white arrows), corresponding to significant stenoses in the left anterior descending and right coronary arteries.
|
|
The diagnostic performance of 3-T perfusion imaging was significantly greater than that of 1.5-T in identifying both single-vessel disease (area under ROC curve: 0.89 ± 0.05 vs. 0.70 ± 0.08; p < 0.05) and multivessel disease (area under ROC curve: 0.95 ± 0.03 vs. 0.82 ± 0.06; p < 0.05). However, there was no statistical difference between field strengths for the overall detection of coronary disease (area under ROC curve: 0.87 ± 0.05 vs. 0.78 ± 0.06; p = 0.23) (Table 4). The ROC curves for this data are presented in Figure 3.
View this table:
[in this window]
[in a new window]
|
Table 4 Diagnostic Performance of 1.5- and 3-T Perfusion Imaging for the Detection of the Extent and Distribution of Significant Coronary Artery Disease
|
|
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3 Receiver-Operating Characteristic Curves
(A) Receiver-operating characteristic curves for visual assessment of 1.5- and 3-Tesla (T) perfusion imaging for the correct identification of significant coronary artery disease. There was no significant difference in diagnostic performance of perfusion imaging between the field strengths (p = 0.23). (B) Receiver-operating characteristic curves for visual assessment of 1.5- and 3-T perfusion imaging for the correct identification of single vessel disease. The diagnostic performance of 3-T perfusion imaging was significantly greater (p < 0.05). (C) Receiver-operating characteristic curves for visual assessment of 1.5- and 3-T perfusion imaging for the correct identification of multivessel disease. The diagnostic performance of 3-T perfusion imaging was significantly greater (p < 0.05). AUC = area under the curve.
|
|
Detection of anatomic location of CAD.
We found that 3-T CMR perfusion imaging identified the anatomic location of coronary disease with high accuracy. For the LAD coronary artery, 3- and 1.5-T perfusion imaging had sensitivities of 97% versus 87%, specificities of 87% versus 71%, and diagnostic accuracies of 92% versus 79% for each field strength, respectively (Figs. 1 and 2). There was a strong trend to improved diagnostic performance at 3-T for detection of significant LAD stenoses compared to 1.5-T (area under ROC curve: 0.92 ± 0.04 vs. 0.79 ± 0.06, respectively; p = 0.05). Disease of the circumflex coronary artery was more difficult to detect at both field strengths. The large majority of stenosed circumflex coronary arteries (19 of 20) occurred in the context of multi-vessel disease, which might have accounted for the poorer performance of both field strengths when assessed visually. The sensitivity, specificity, and diagnostic accuracy was 30%, 98%, and 75%, respectively, for 1.5-T imaging. Diagnostic performance of 3-T imaging for detection of significant circumflex stenoses was significantly greater than that at 1.5-T (area under ROC curve: 0.84 ± 0.06 vs. 0.64 ± 0.08, respectively; p < 0.05), because 3-T imaging detected a further 8 lesions in this cohort. This improved sensitivity to 70%, specificity to 98%, and diagnostic accuracy to 89%. There was no significant difference in diagnostic performance of 3- and 1.5-T for detection of significant RCA stenoses (area under ROC curve: 0.90 ± 0.04 vs. 0.91 ± 0.04, respectively; p = 0.98), with diagnostic accuracy of 90% vs. 92%, sensitivity of 92% vs. 84%, and specificity of 89% vs. 97% for 3- and 1.5-T, respectively.
Image analysis and artifacts.
All images acquired were of sufficient quality for analysis, and no images were excluded. There was no significant difference in the number of scans classified as having a "possible perfusion deficit" that required interpretation by a third reviewer: 7 of 61 (11.5%) scans at 3-T and 4 of 61 (6.6%) scans at 1.5-T (p = 0.30). Dark-rim artifacts (17) (hypoenhanced zone in the subendocardial layer before contrast material arrival) were observed more commonly at 1.5-T compared with 3-T (23% patients at 1.5-T vs. 8% of patients at 3-T, p < 0.01). Reconstruction artifacts attributable to parallel imaging in the region of the heart occurred in 4 patients (6%) at 3-T and 2 patients (3%) at 1.5-T (p = 0.4).
SNR and CNR.
As expected, 3-T perfusion imaging provided a significant increase in both SNR (17 ± 6 at 3-T vs. 11 ± 2 at 1.5-T; p < 0.01) and CNR (17 ± 10 at 3-T vs. 11 ± 4 at 1.5-T; p < 0.01) compared with 1.5-T (Fig. 4). Although the order of the scans was randomized in a balanced manner, a further analysis was performed to elucidate whether the order of the scans influenced SNR or CNR. We found that SNR was significantly higher at 3-T compared with 1.5-T in those who had their first scan performed at 1.5-T (15.9 ± 6.25 vs. 10.6 ± 4.68, respectively, p < 0.01) and also in those who had their first scan performed at 3-T (17.8 ± 6.34 vs. 12.1 ± 2.51, respectively, p < 0.01). Similarly, CNR was higher at 3-T compared with 1.5-T in those who had their first scan performed at 1.5-T (16.3 ± 10.3 vs. 10.6 ± 4.68, respectively, p = 0.08) and also in those who had their first scan performed at 3-T (17.7 ± 9.97 vs. 11.8 ± 3.72, respectively, p < 0.05).
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4 Boxplot Graphs Demonstrating a Significant Increase in Both SNR and CNR at 3-T Compared With 1.5-T
(A) Signal-to-noise ratio (SNR). (B) Contrast-to-noise ratio (CNR). p < 0.01 for both. T = Tesla.
|
|
|
Discussion
|
|---|
The principal finding in this prospective study is that 3-T perfusion CMR is superior to 1.5-T perfusion CMR in the diagnosis of significant single-vessel and multivessel coronary disease. In a routine clinical setting, with more than 95% of the enrolled patients successfully completing both CMR studies, 3-T perfusion CMR had high sensitivity (98%) and negative predictive value (94%), indicating significant clinical utility as a screening tool. Our findings have important implications for the noninvasive assessment of myocardial ischemia and for defining the role of high-field CMR systems in current clinical practice.
To the best of our knowledge, this is the first study to assess systematically the utility of 3-T CMR perfusion imaging in diagnosing CAD in contemporary clinical practice. Prior studies at 1.5-T have shown that semiquantitative assessment of perfusion has good accuracy for detection of CAD, with a sensitivity of 91% and a specificity of 94% for detecting CAD as defined by PET (8) and a sensitivity of 87% to 88% and specificity of 85% to 90% compared to quantitative coronary angiography (8,9). However, visual assessment of perfusion CMR at 1.5-T has usually performed less well compared to semiquantitative assessment, with a sensitivity of 82% to 91% and specificity of 62% to 63% (11,18). Direct comparison between these studies is confounded by variation in pulse sequence, contrast dose, and angiographic definition of significant CAD. Paetsch et al. (18), in a study of 79 patients, of whom 67% had significant coronary disease using the same >50% coronary stenosis definition as we did, found that 1.5-T CMR perfusion imaging had a sensitivity of 91%, a specificity of 62%, and diagnostic accuracy of 81%, findings similar to the 1.5-T arm of our study.
Our significant increases in SNR and CNR for perfusion CMR at 3-T, compared with 1.5-T, are consistent with earlier work (12,13). The significantly higher diagnostic accuracy of 3-T perfusion found in our study is attributed to a combination of increased SNR, allowing detection of reductions in endocardial perfusion (the most sensitive parameter for a reduction of overall coronary blood flow [19]) and reduced occurrence of dark-rim artifact at 3-T. The latter can often be confused as a real perfusion deficit and probably explains the lower specificity seen at 1.5-T.
We used visual analysis in our cohort, as this is the most commonly used method in clinical practice and we wanted to compare the clinical utility of each field strength in the same population. Absolute and semiquantitative measurements of perfusion CMR have an important role in assessing myocardial blood flow in a research setting (5,20), but require time and labor-intensive afterprocessing and are not performed routinely in clinical practice.
Recently, Klem et al. (11) proposed that a combined CMR approach of stress-rest perfusion imaging followed by delayed enhancement imaging (DE-CMR) at 1.5-T is superior to perfusion imaging alone for depicting clinically significant coronary artery stenosis. Their study population differs from ours in that the prevalence of CAD is lower (40% vs. 66%), and they excluded patients with previous revascularization and MI. Notwithstanding these important differences, the performance of perfusion analysis alone in their study is lower than the performance of 1.5-T in our study, while using the interpretative algorithm incorporating DE-CMR in addition to perfusion markedly increases specificity in their cohort. Although DE-CMR was not performed in the current study, we have recently shown, in a separate population of patients with confirmed MI, that there was strong agreement between 1.5- and 3-T in the mass and transmural extent of hyperenhancing myocardium when using the same pulse sequence (21). Hence, the addition of DE-CMR to the imaging protocol at 3-T is clinically robust and might increase the specificity of 3-T perfusion imaging alone in this clinical setting.
As expected, we found greatly reduced specificity for detecting stenoses 70% at both field strengths, whereas sensitivity was increased at 3-T. This mirrors the findings of Klem et al. (11), who found similar changes in diagnostic performance dependent on the severity of coronary stenosis (detection of coronary stenosis 70%, as opposed to 50%, resulted in a reduction in specificity from 63% to 58% and an increase in sensitivity from 82% to 84%). It is salient to note that all 8 patients in our study who had lesions that were quantified between 50% and 69% had corresponding perfusion deficits; of these, 6 proceeded to coronary revascularization, and 2 were treated medically. Hence, despite being unaware of the CMR findings, the treating interventional cardiologist or cardiac surgeon considered these lesions to be clinically significant.
We attempted to avoid before-test referral bias by recruiting only those patients referred for diagnostic coronary angiography and not "healthy" volunteers, and after-test referral bias by performing X-ray coronary angiography after and independent of CMR findings. Our data at 3-T compare favorably with the published work on SPECT imaging under similar clinical circumstances. After adjusting for selection bias, Cecil et al. (22) found a corrected sensitivity and specificity of 82% and 59%, respectively, and Miller et al. (23) found a corrected sensitivity and specificity of 65% and 67%, respectively, for the detection of CAD by SPECT imaging. Compared with these results, the accuracy of visual assessment of perfusion CMR at 3-T indicates clinical utility, particularly taking into account its additional advantages of safety, tolerability, lack of ionizing radiation, and short examination time.
We found that perfusion imaging at either field strength had lower sensitivity for detection of significant coronary disease in the circumflex coronary artery compared to the LAD and RCA. Almost all (19 of 20) stenosed circumflex coronary arteries in our cohort were in the context of multi-vessel disease, and it is more difficult to detect hypoperfused myocardium in the case of "balanced" disease, especially by nonquantitative means. This difficulty might have been most acute in the circumflex territory because it is farthest from the radiofrequency coil and because visual, unlike semiquantitative, assessment of perfusion imaging cannot correct signal intensity for distance from the coil. We expect that semiquantitative assessment of images obtained by perfusion CMR at 3-T will provide even greater diagnostic accuracy for detection of significant coronary stenosis, most improved in the circumflex territory (9).
Residual contrast at the time of acquisition of the second set of perfusion images in each particular patient may have influenced the diagnostic accuracy or the SNR or CNR measured. However, in healthy human volunteers, Weinmann et al. (24) demonstrated that plasma concentration of gadolinium declines according to a bi-exponential function with a distribution phase with a mean half-life of 12 ± 7.8 min and concluded that its efficiency to enhance signal intensity will last only up to about 1 h after intravenous injection of 0.1 mmol/kg body mass. Accordingly, in our CMR protocol, the time interval between the sets of images acquired on each system was at least 1 h for each patient, and randomization of the order of the scans was balanced to avoid bias. Furthermore, our data indicate that the order of the scans did not influence diagnostic accuracy, SNR, or CNR at either field strength.
Study limitations.
Although we found a significant superiority of 3-T over 1.5-T for the diagnosis of single-vessel and multivessel disease, we did not find a statistically significant difference between the 2 field strengths in the overall results. This is likely to be due to the small number of patients in this study, with consequent lower statistical power. Furthermore, ROC analysis indicated higher area under the curve values for 3-T in the overall analysis, as well as individual artery territories. We did not compare perfusion CMR with other myocardial perfusion techniques, such as SPECT and PET. The X-ray coronary angiogram is an imperfect reference standard but remains the established technique for making decisions regarding revascularization in clinical practice. Unlike perfusion CMR, X-ray coronary angiogram does not provide evaluation of hemodynamic severity of a stenosis and, consequently, the sensitivity and specificity of perfusion CMR might be reduced in certain predictable situations. The presence of a coronary stenosis of at least 50% supplying an area of myocardium that is well collateralized might fail to produce a perfusion deficit, even at stress, leading to a false-negative result. Similarly, perfusion CMR may yield false-positive results in microvascular disease and where a coronary artery supplying an area of MI has opened spontaneously or after therapeutic intervention. This was highlighted by the increase in specificity when patients with a history of prior MI were excluded from the analysis.
The true apex was not visualized in the perfusion images using our approach. However, a study by Elkington et al. (25) showed that long-axis views did not provide helpful information over short-axis slices alone and did not visualize inducible perfusion deficits in the true apex that had been demonstrated with SPECT studies. We therefore decided to maximize the number of short-axis slices, within the limits of the scanning sequence, to achieve maximum coverage of the left ventricle.
|
Conclusions
|
|---|
We found that 3-T CMR perfusion imaging is clinically feasible and is superior to 1.5-T CMR perfusion imaging for identification of significant single-vessel and multivessel coronary disease in patients suspected of CAD, most likely because of an increase in SNR. Quantitative/semiquantitative perfusion analysis and addition of other sequence techniques, such as function and delayed enhancement imaging, as part of the comprehensive CMR protocol at 3-T could further improve diagnostic accuracy.
|
Appendix
|
|---|
For accompanying videos, please see the online version of this article.
|
Footnotes
|
|---|
This work is supported by a project grant from the British Heart Foundation (BHF). Dr. Cheng is funded by a grant from the Oxfordshire Health Services Research Committee. Drs. Pegg, Karamitsos, and Selvanayagam are funded by the British Heart Foundation. Dr. Karamitsos is also funded by a scholarship from Hellenic Cardiological Society. Drs. Cheng and Pegg have contributed equally to this work.
|
References
|
|---|
1. Nesto RW, Kowalchuk GJ. The ischemic cascade: temporal sequence of hemodynamic, electrocardiographic and symptomatic expressions of ischemia Am J Cardiol 1987;59:23C-30C.[CrossRef][Medline]2. Hendel RC, Berman DS, Cullom SJ, et al. Multicenter clinical trial to evaluate the efficacy of correction for photon attenuation and scatter in SPECT myocardial perfusion imaging Circulation 1999;99:2742-2749.[Abstract/Free Full Text] 3. Zenovich A, Muehling OM, Panse PM, Jerosch-Herold M, Wilke N. Magnetic resonance first-pass perfusion imaging: overview and perspectives Rays 2001;26:53-60.[Medline] 4. Jerosch-Herold M, Wilke N, Stillman AE. Magnetic resonance quantification of the myocardial perfusion reserve with a Fermi function model for constrained deconvolution Med Phys 1998;25:73-84.[CrossRef][Web of Science][Medline] 5. Selvanayagam JB, Jerosch-Herold M, Porto I, et al. Resting myocardial blood flow is impaired in hibernating myocardium: a magnetic resonance study of quantitative perfusion assessment Circulation 2005;112:3289-3296.[Abstract/Free Full Text] 6. 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] 7. 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] 8. 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] 9. 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] 10. Arai AE, Gaither 3rd CC, Epstein FH, Balaban RS, Wolff SD. Myocardial velocity gradient imaging by phase contrast MRI with application to regional function in myocardial ischemia Magn Reson Med 1999;42:98-109.[CrossRef][Web of Science][Medline] 11. Klem I, Heitner JF, Shah DJ, et al. Improved detection of coronary artery disease by stress perfusion cardiovascular magnetic resonance with the use of delayed enhancement infarction imaging J Am Coll Cardiol 2006;47:1630-1638.[Abstract/Free Full Text] 12. Gutberlet M, Noeske R, Schwinge K, Freyhardt P, Felix R, Niendorf T. Comprehensive cardiac magnetic resonance imaging at 3.0 Tesla: feasibility and implications for clinical applications Invest Radiol 2006;41:154-167.[CrossRef][Web of Science][Medline] 13. Araoz PA, Glockner JF, McGee KP, et al. 3 Tesla MR imaging provides improved contrast in first-pass myocardial perfusion imaging over a range of gadolinium doses J Cardiovasc Magn Reson 2005;7:559-564.[Web of Science][Medline] 14. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association Circulation 2002;105:539-542.[Free Full Text] 15. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve Radiology 1982;143:29-36.[Abstract/Free Full Text] 16. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases Radiology 1983;148:839-843.[Abstract/Free Full Text] 17. Di Bella EV, Parker DL, Sinusas AJ. On the dark rim artifact in dynamic contrast-enhanced MRI myocardial perfusion studies Magn Reson Med 2005;54:1295-1299.[CrossRef][Web of Science][Medline] 18. 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] 19. Bache R, Schwartz J. Effect of perfusion pressure distal to a coronary stenosis on transmural myocardial blood flow Circulation 1982;65:928-935.[Abstract/Free Full Text] 20. Wang L, Jerosch-Herold M, Jacobs Jr. DR, Shahar E, Folsom AR. Coronary risk factors and myocardial perfusion in asymptomatic adults: the Multi-Ethnic Study of Atherosclerosis (MESA) J Am Coll Cardiol 2006;47:565-572.[Abstract/Free Full Text] 21. Cheng AS, Robson MD, Neubauer S, Selvanayagam JB. Irreversible myocardial injury: assessment with cardiovascular delayed-enhancement MR imaging and comparison of 1.5 and 3.0 T—initial experience. Radiology 2007;242:735–42. 22. Cecil MP, Kosinski AS, Jones MT, et al. The importance of work-up (verification) bias correction in assessing the accuracy of SPECT thallium-201 testing for the diagnosis of coronary artery disease J Clin Epidemiol 1996;49:735-742.[CrossRef][Web of Science][Medline] 23. Miller TD, Hodge DO, Christian TF, Milavetz JJ, Bailey KR, Gibbons RJ. Effects of adjustment for referral bias on the sensitivity and specificity of single photon emission computed tomography for the diagnosis of coronary artery disease Am J Med 2002;112:290-297.[CrossRef][Web of Science][Medline] 24. Weinmann HJ, Laniado M, Mutzel W. Pharmacokinetics of GdDTPA/dimeglumine after intravenous injection into healthy volunteers Physiol Chem Phys Med NMR 1984;16:167-172.[Web of Science][Medline] 25. Elkington AG, Gatehouse PD, Prasad SK, Moon JC, Firmin DN, Pennell DJ. Combined long- and short-axis myocardial perfusion cardiovascular magnetic resonance J Cardiovasc Magn Reson 2004;6:811-816.[CrossRef][Web of Science][Medline]
This article has been cited by other articles:
|
|
|
|
 
T. D. Karamitsos, J. M. Francis, S. Myerson, J. B. Selvanayagam, and S. Neubauer
The Role of Cardiovascular Magnetic Resonance Imaging in Heart Failure
J. Am. Coll. Cardiol.,
October 6, 2009;
54(15):
1407 - 1424.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
O. Bruder, S. Schneider, D. Nothnagel, T. Dill, V. Hombach, J. Schulz-Menger, E. Nagel, M. Lombardi, A. C. van Rossum, A. Wagner, et al.
EuroCMR (European Cardiovascular Magnetic Resonance) Registry: Results of the German Pilot Phase
J. Am. Coll. Cardiol.,
October 6, 2009;
54(15):
1457 - 1466.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. F. Christian, S. P. Bell, L. Whitesell, and M. Jerosch-Herold
Accuracy of Cardiac Magnetic Resonance of Absolute Myocardial Blood Flow With a High-Field System: Comparison With Conventional Field Strength
J. Am. Coll. Cardiol. Img.,
September 1, 2009;
2(9):
1103 - 1110.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. G. Camici and O. E. Rimoldi
The Clinical Value of Myocardial Blood Flow Measurement
J. Nucl. Med.,
July 1, 2009;
50(7):
1076 - 1087.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Lockie, E. Nagel, S. Redwood, and S. Plein
Use of Cardiovascular Magnetic Resonance Imaging in Acute Coronary Syndromes
Circulation,
March 31, 2009;
119(12):
1671 - 1681.
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Plein, J. Schwitter, D. Suerder, J. P. Greenwood, P. Boesiger, and S. Kozerke
k-Space and Time Sensitivity Encoding-accelerated Myocardial Perfusion MR Imaging at 3.0 T: Comparison with 1.5 T
Radiology,
November 1, 2008;
249(2):
493 - 500.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Plein, S. Kozerke, D. Suerder, T. F. Luescher, J. P. Greenwood, P. Boesiger, and J. Schwitter
High spatial resolution myocardial perfusion cardiac magnetic resonance for the detection of coronary artery disease
Eur. Heart J.,
September 1, 2008;
29(17):
2148 - 2155.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
I. Klem, S. Greulich, J. F. Heitner, H. Kim, H. Vogelsberg, E.-M. Kispert, S. R. Ambati, C. Bruch, M. Parker, R. M. Judd, et al.
Value of cardiovascular magnetic resonance stress perfusion testing for the detection of coronary artery disease in women.
J. Am. Coll. Cardiol. Img.,
July 1, 2008;
1(4):
436 - 445.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. Selvanayagam
Women with chest pain expanding the diagnostic armamentarium.
J. Am. Coll. Cardiol. Img.,
July 1, 2008;
1(4):
446 - 449.
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. K. Kuhl, F. Traber, J. Gieseke, W. Drahanowsky, N. Morakkabati-Spitz, W. Willinek, M. von Falkenhausen, C. Manka, and H. H. Schild
Whole-Body High-Field-Strength (3.0-T) MR Imaging in Clinical Practice * Part II. Technical Considerations and Clinical Applications
Radiology,
April 1, 2008;
247(1):
16 - 35.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. S.H. Cheng, J. B. Selvanayagam, M. Jerosch-Herold, W. J. van Gaal, T. D. Karamitsos, S. Neubauer, and A. P. Banning
Percutaneous treatment of chronic total coronary occlusions improves regional hyperemic myocardial blood flow and contractility insights from quantitative cardiovascular magnetic resonance imaging.
J. Am. Coll. Cardiol. Intv.,
February 1, 2008;
1(1):
44 - 53.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
Top
Abstract
Methods