| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the BHF Cardiac Magnetic Resonance Unit (S.P., D.B., J.P.G., M.U.S.) and Department of Medical Physics (A.R., J.P.R.), the General Infirmary at Leeds, Room 170, D-floor, Jubilee Bldg, Great George St, Leeds LS1 3EX, England; and Institute of Cardiovascular Research, University of Leeds, Leeds, England (S.G.B.). Received March 9, 2004; revision requested May 20; revision received June 24; accepted July 27. S.P. supported by British Heart Foundation. Address correspondence to S.P. (e-mail: sven.plein@leedsth.nhs.uk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: All 102 subjects provided written informed consent, and the local ethics committee approved the study. A saturation-recovery segmented k-space gradient-echo pulse sequence was combined with SENSE to allow dynamic acquisition of myocardial perfusion data on four parallel short-axis MR image sections at every heartbeat. This technique was evaluated in 10 healthy volunteers and in 92 patients scheduled to undergo conventional coronary angiography. Gadopentetate dimeglumine was peripherally injected at rest and during adenosine-induced stress. The maximal upslope of the signal intensity–time profiles was plotted for 16 myocardial segments defined on three MR image sections, and a myocardial perfusion reserve index (MPRI) between stress and rest, normalized to the input function from the blood pool of the most basal section, was calculated. Areas under receiver operating characteristic curves (AUCs) were used to assess the diagnostic performance of cardiac MR imaging for depiction of greater than 70% CAD seen at coronary angiography, the reference standard.
RESULTS: In volunteers, the mean myocardial enhancement was 2.1 ± 1.2 (standard deviation), with homogeneous signal intensity distribution across the segments. The diagnostic accuracy of MPRI measurements was high (AUC, 0.908; sensitivity, 88% [52 of 59 patients]; specificity, 82% [27 of 33 patients]). Diagnostic performance was similar among separate analyses of the three coronary territories and among separate analyses of data in the patients with diabetes mellitus, left ventricular hypertrophy, or myocardial infarction.
CONCLUSION: Multisection myocardial perfusion MR imaging with SENSE is feasible and has high diagnostic accuracy in the detection of CAD.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
To overcome these limitations and to allow data acquisition in multiple sections at each or alternate heartbeats, very fast hybrid gradient-echo echo-planar imaging pulse sequences have been used (7,13–16). Although initial results have been encouraging, the long echo trains that are required for these techniques may lead to impaired image quality due to ghosting and magnetic susceptibility artifacts and can impose restrictions on the available trade-offs between signal intensity and contrast-to-noise ratios versus the number of sections acquired.
The relatively recently developed parallel acquisition methods, including sensitivity encoding (SENSE) (18), represent alternative approaches to shortening data acquisition times at cardiac MR imaging. With SENSE, data acquisition can be accelerated by a factor of two or more, as compared with the data acquisition times with conventional pulse sequences. In addition, data are corrected for the nonuniform signal intensity caused by variations in the sensitivity of the radiofrequency receiver coils. These properties of SENSE should be particularly useful in myocardial perfusion MR imaging for two reasons: First, faster data acquisition allows multisection perfusion imaging to be performed without use of echo-planar imaging, and second, the inherent signal intensity uniformity correction applied with SENSE reconstruction should facilitate semiquantitative and quantitative analyses.
We hypothesized that the use of SENSE would allow multisection myocardial perfusion cardiac MR imaging at a high sampling rate without the use of echo-planar imaging. Thus, the purpose of our study was to evaluate the technical performance of SENSE-accelerated myocardial perfusion cardiac MR imaging and prospectively assess the diagnostic accuracy of this examination for the depiction of significant CAD.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
To meet these objectives, we designed a dynamic segmented k-space gradient-echo pulse sequence combined with SENSE. This sequence consisted of saturation-recovery T1-weighted turbo field-echo MR imaging with the following parameters: 3.3/1.6 (repetition time msec/echo time msec); a 15° flip angle; a single nonselective saturation pulse; a maximum field of view, chosen as required to avoid image aliasing, of up to 400 mm; an acquisition matrix of 160 x 112 reconstructed to 256 x 256; one signal acquired; an 8-mm section thickness; and a SENSE factor of two. This pulse sequence enables four image sections to be acquired at every heartbeat up to a heart rate of 100 beats per minute. For this study, we used a single saturation-recovery preparation pulse and acquired four sections in the left ventricular short-axis orientation sequentially, from the base to the apex of the left ventricle.
The saturation-recovery times between the preparation pulse and the center of k-space for the four section acquisitions were 42, 164, 287, and 410 msec. Use of this approach optimizes section 1 for the determination of left ventricular input function. The position of this section nearest the outflow tract ensures that no papillary muscles or trabeculation will be included in the blood pool. Furthermore, the short preparation pulse delay of this imaging section facilitates the most linear signal intensity dependence at the high gadolinium concentrations encountered in the blood pool signal. Sections 2–4 were intended for analysis of myocardial perfusion; 16 myocardial segments were defined on these sections in accordance with American Heart Association recommendations (19). The apical section in the classification was not considered in this analysis.
Study Population
First-pass myocardial perfusion MR imaging was performed in 102 subjects. Ten subjects (seven men, three women; mean age, 31 years; age range, 24–35 years) were healthy volunteers who had no risk factors for CAD, no history or symptoms of cardiac disease, normal physical examination results, and normal at-rest electrocardiograms. Ninety-two subjects (68 men, 24 women; mean age, 58 years; age range, 42–78 years) were patients who were suspected of having or known to have CAD and were scheduled to undergo conventional coronary angiography owing to clinical referral. Nineteen of these patients had a history of myocardial infarction, as indicated by clinical information, Q waves at electrocardiography, and wall motion abnormalities at echocardiography. In eight of these 19 patients, a myocardial infarction had occurred less than 1 week before the study (acute myocardial infarction), and in the remaining 11 patients, a myocardial infarction had occurred at least 6 months before study recruitment (chronic myocardial infarction). Additional patient characteristics are given in Table 1.
|
MR Imaging
Cardiac MR imaging was performed in all 102 subjects while they were in the supine position by using a whole-body 1.5-T cardiac MR imaging system (Gyroscan Intera CV; Philips Medical Systems, Best, the Netherlands) equipped with 30 mT/m peak gradients and a 150 mT/sec slew rate. Cardiac MR imaging signals were received by a five-element cardiac phased-array coil, and electrocardiographic gating and triggering were performed by using a vectorcardiographic method (21). Fast survey images were acquired to determine the true short axis of the left ventricle. Then, a low-spatial-resolution reference image was acquired to serve as the three-dimensional fast field-echo coil sensitivity map (8/0.51, 7° flip angle, 530 x 530-mm field of view, 32 x 26 matrix, two stacks of 50 coronal sections) that is required with use of SENSE.
First, at-rest perfusion was assessed by using the described segmented k-space gradient-echo saturation-recovery T1-weighted turbo field-echo pulse sequence. The four MR image sections were distributed to cover the heart between the outflow tract and the apex by adjusting the gap between the sections. A dynamic series of images was acquired continuously for 40 seconds, with 10 baseline images obtained during an initial expiratory breath hold, which was followed by two respiratory cycles and an additional breath hold that typically lasted 20 seconds and was timed to correspond to the interval of the first pass of the contrast material bolus. During the inspiratory phase of the second breath, a bolus of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) was injected, by using a power injector (Spectris; Medrad, Indianola, Pa), into an antecubital vein at a dose of 0.05 mmol per kilogram of body weight and an injection rate of 6 mL/sec and followed by a 10 mL-saline flush.
After 30 minutes, to allow adequate clearance of the first bolus of the contrast agent, adenosine was infused at a dose of 140(µg · kg–1)/min for up to 6 minutes. A second perfusion image acquisition was commenced in the same plane after 4 minutes of the adenosine infusion and by using the same acquisition sequence that was used to obtain the at-rest images. During adenosine infusion, electrocardiographic activity was monitored continuously and blood pressure and heart rate measurements were obtained every minute.
Between the two perfusion image acquisitions, a multisection multiple-phase data set of 10–12 short-axis cine image sections to cover the entire left ventricle was acquired by using a balanced steady-state free precession technique. Left ventricular mass was calculated by using a standard method (21).
Cardiac MR Image Analysis
For initial analysis, cardiac MR imaging data were downloaded to a SPARC 10 workstation (Sun Microsystems, Mountain View, Calif) equipped with commercially available analysis software (MASS 5.0; Medis, Leiden, the Netherlands) and analyzed by an observer who had 5 years experience in cardiac MR imaging (S.P.) and was blinded to all clinical data. All images were initially assessed for the presence and extent of respiratory, susceptibility, and other artifacts; and the data that contained substantial artifacts were excluded from further analysis. On a representative image from the dynamic series acquired at rest, the observer then traced the epicardial and endocardial contours in sections 2–4 and placed a 200-mm2 circular region of interest in the left ventricular blood pool depicted on the most basal section (section 1). In each section, a reference point was placed at the anterior septal insertion of the right ventricle. All contours were then automatically copied to the other dynamic images of each section, and their positions were reviewed and manually adjusted to correct for respiratory motion during data acquisition.
On sections 2–4, 16 segments were defined to correspond to segments 1–16 of the 17-segment model recommended by the American Heart Association (Fig 1) (19). Each segment was further divided into endocardial and epicardial halves. The mean signal intensity of each transmural segment and its corresponding epicardial and endocardial halves was calculated for each image in the dynamic series. The same analysis was then performed for the stress images, with care taken to ensure identical positioning of the reference points between the two image groups (at-rest and stress images). The total processing time for analysis was approximately 1 hour per patient.
|
Pulse Sequence Evaluation
To assess the technical performance of the pulse sequence, the signal intensity properties of the stress examination in the volunteer group were evaluated (by A.R.). To assess the effects of SENSE reconstruction on signal intensity uniformity, the baseline signal intensity values for each segment were calculated as the mean signal intensity for 10 dynamic images acquired before contrast material injection. For comparison of the signal response between segments and comparison of our data with those obtained by using previously described pulse sequences, the myocardial enhancement seen between the periods of baseline and peak signal intensity was calculated for each segment. Finally, the MPRIs of the 16 segments were compared.
Conventional Angiography
After undergoing cardiac MR imaging, all 92 patients were examined with left-sided cardiac catheterization with conventional coronary angiography by using a standard Judkins technique. A mean of 4.3 days ± 12 (standard deviation) separated the two examinations. The left and right coronary arteries were selectively catheterized, and multiple images in the left and right anterior oblique projections and lateral projections were acquired. Two cardiologists (one of the treating physicians with 8–22 years and M.U.S. with 15 years of experience in coronary angiography), who were blinded to the results of the cardiac MR image analyses, interpreted the angiographic examination findings. In cases of disagreement between the two observers, a consensus decision was reached. Coronary artery lesions of more than 70% luminal stenosis in the main coronary arteries or their first-order branches were considered to be significant. The coronary arteries of the volunteers were presumed to be normal and thus were not studied at angiography. For comparison of the angiographic data with the cardiac MR imaging data, the coronary supply to each segment was determined according to the American Heart Association criteria (Fig 1).
Prospective Evaluation
For prospective evaluation, the volunteer data were added to the patient data to increase the number of normal cases and consequently improve the reliability of the specificity data. Receiver operating characteristic (ROC) analysis was performed to assess the diagnostic performance of MPRI analysis in the detection of significant CAD at conventional angiography. This analysis was performed for patients as a whole and for individual coronary artery territories, and cardiac perfusion MR imaging findings were regarded as abnormal if the MPRI in one or more segments from any of the three sections was abnormal. This part of the analysis was performed only in the subendocardial half of the segments, which has previously been shown to yield the best diagnostic accuracy (7).
At subgroup analyses, additional ROC analyses were performed to compare the diagnostic performances of the three sections (sections 2–4) when they were analyzed separately. Further subgroup analyses were performed with patients who had a history of myocardial infarction, diabetes mellitus, or left ventricular hypertrophy. Left ventricular hypertrophy at cardiac MR imaging was defined according to findings on left ventricular functional images acquired at cardiac MR imaging by using the method of Alfakih et al (20).
Statistical Analyses
Unless otherwise stated, continuous values are cited as means ± standard deviations. Comparisons of diagnostic performance between analyses were performed by using area under the ROC curve statistics (Analyze-it Software, Leeds, England). For all analyses, P < .05 was considered to indicate significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
During adenosine-induced stress, the subjects’ mean heart rate increased from 61 beats per minute ± 10 to 81 beats per minute ± 9. Mean systolic blood pressure was 125 mm Hg ± 22 at rest and 122 mm Hg ± 17 during adenosine-induced stress. No arrhythmias or other marked side effects were observed during stress.
Pulse Sequence Evaluation
None of the images had to be excluded because of susceptibility or ghosting artifacts. In three of the 10 volunteers, thin subendocardial susceptibility artifacts were seen on the apical section during arrival of the contrast material bolus in the left ventricle. These artifacts disappeared during myocardial enhancement. The baseline signal intensity was similar among segments depicted on the same section, with no significant differences (Fig 2, top). Between sections, the baseline signal intensity increased progressively from the base to the apex of the left ventricle, reflecting the increasing delay in data acquisition following the single preparation pulse. The amplitude of myocardial enhancement was similar among segments depicted on the same section but showed lower values overall on sections 2 and 3 (Fig 2, middle). The mean myocardial enhancement for all sections was 2.1 ± 1.2. MPRIs were similar in all 16 segments, although standard deviations were relatively wide (Fig 2, bottom). Differences in MPRI between the three sections were small. As an illustration of the input functions obtained in the basal section, Figure 3 shows examples from three randomly selected volunteers.
|
|
The diagnostic performance of the main analysis, which involved the use of data from all three sections, was high, with an area under the ROC curve of 0.908 (95% confidence interval [CI]: 0.841, 0.975) (Fig 4, left). ROC analysis showed that an MPRI threshold of 1.02 yielded the highest sensitivity and specificity. At this threshold, sensitivity was 88% (52 of 59 patients); specificity, 82% (27 of 33 patients); the positive predictive value, 90% (52 of 58 patients); and the negative predictive value, 79% (27 of 34 patients). Separate analyses of the sections yielded poorer diagnostic performance than analysis of all sections combined (Table 2). Of the three sections, section 2 yielded the best diagnostic performance, with an area under the ROC curve of 0.864 (95% CI: 0.786, 0.942), as compared with 0.807 (95% CI: 0.715, 0.898) for section 1 and 0.798 (95% CI: 0.705, 0.892) for section 3; differences were not significant (Fig 4, right, Table 2).
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
In this study, we used the SENSE parallel imaging technique to shorten data acquisition times and perform multisection myocardial perfusion MR imaging. With SENSE, the spatial information included in the coil sensitivities of the multielement array is used for signal localization. This information is determined from a low-spatial-resolution reference image, and imaging times are reduced by acquiring several parallel lines of k-space data (17). Data acquisition can thus be accelerated without the potential adverse effects associated with using a long echo-planar train, a method that previously has been used to enable multisection myocardial perfusion MR imaging (7,13–16). Furthermore, by relating data to the information from the reference image, SENSE corrects for the signal intensity variations related to the radiofrequency coil sensitivities to result in a uniform signal throughout the imaging plane.
Our study results show that it is feasible to use SENSE to accelerate data acquisition at cardiac perfusion MR imaging and that this faster data acquisition can be used to acquire four parallel sections at every heartbeat. The spatial resolution (approximately 3 x 3 mm in-plane) and myocardial enhancement (2.0–2.3) achieved by using our acquisition strategy were comparable to those reported in previous studies in which echo-planar imaging–based pulse sequences were used (15).
Apart from the small artifacts seen in the subendocardial layer before contrast material arrival, we observed no susceptibility or ghosting artifacts in our study. Minor SENSE reconstruction–related artifacts occurred in only two patients. Furthermore, the coil sensitivity correction that is inherent to SENSE facilitated semiquantitative analysis because it resulted in homogeneous baseline signal intensity, signal intensity amplitude, and MPRIs in the imaging sections. Therefore, further normalization of the data to the baseline signal intensity was not required. Our results thus suggest that SENSE is an attractive myocardial perfusion MR imaging technique that allows simultaneous acquisition of multiple sections and yields high-quality data that are well suited for semiquantitative analysis.
Although we did not directly compare our data with those obtained by using an echo-planar imaging pulse sequence, we speculate that SENSE represents an attractive alternative to the previously described techniques for accelerated data acquisition in myocardial perfusion MR imaging. A potential limitation of SENSE is that the signal-to-noise ratio is reduced by the square root of the SENSE factor; in our study, this led to a reduction in signal-to-noise ratio of approximately 40%. New pulse sequences, such as balanced steady-state free precession, that provide inherently higher signal-to-noise ratios and can be combined with SENSE may help overcome this limitation.
In this study, we used three of the acquired sections for myocardial perfusion assessment and used the fourth section, which was positioned near the left ventricular outflow tract and acquired with a very short prepulse delay, to define the input function. The purpose of this approach was to optimize the calculation of the input function while maintaining myocardial coverage, as proposed in current American Heart Association guidelines (19). In previous studies (6,7), the left ventricular input function usually was derived from the same imaging plane in which the perfusion assessment was performed. This approach can be problematic because the signal intensity changes in the left ventricular cavity are several-fold higher compared with those in the myocardium owing to the higher concentration of the contrast agent.
Because the acquisition pulse sequence and contrast agent doses generally are optimized for the myocardial rather than ventricular blood pool signal, the signal intensity profiles of the left ventricle can be truncated owing to nonlinear dependence of the MR imaging signal at higher contrast agent concentrations. This is especially true when residual contrast agent from a prior injection is present. Because the relationship between contrast agent concentration and signal intensity is dependent on the delay following the preparation pulse, the section acquired with the shortest prepulse delay is the most suitable for assessment of the left ventricular input function but the least suitable for myocardial analysis.
At prospective analysis, the diagnostic performance of our examinations was similar to that in other relatively recent studies of myocardial perfusion MR imaging in which different acquisition and analysis strategies were used. However, the patient populations in those studies generally were smaller and more strictly selected. Al-Saadi and colleagues (6) calculated a threshold for the MPRI between volunteers and patients and prospectively applied this threshold to a second group of patients. Data were acquired at a single midventricular short-axis location by using a fast segmented k-space gradient-echo pulse sequence. The reported sensitivity for the detection of myocardial segments of stenotic coronary arteries was 87%, and the reported specificity was 90%. The study population was restricted to patients with proximal vessel disease and no history of myocardial infarction, and the contrast agent was injected through a central venous catheter, the use of which facilitates a more compact contrast material bolus than does peripheral injection.
Schwitter et al (7) calculated thresholds for the upslopes of signal intensity profiles acquired during dipyridamole-induced stress in healthy volunteers and applied these thresholds to a group of patients and volunteers. In their study, data were acquired by using a hybrid multisection gradient-echo echo-planar technique, and the sensitivity and specificity for the detection of significant CAD at conventional angiography were 90% and 88%, respectively. Patients with myocardial infarction were excluded from that study, and if patients were taking ß-blockers, these were omitted before the examinations. Nagel et al (15) recently used a multisection gradient-echo echo-planar technique and obtained results similar to those reported by Al-Saadi et al (6) and Schwitter et al (7). Compared with the study populations described in the initial reports, a larger and less strictly selected patient population was examined in our study. We also injected contrast material into a peripheral vein, and this approach is more likely to be used in routine clinical practice than is central venous injection. We acquired data at every heartbeat; this method is more robust than acquisitions at alternate heartbeats (17). Our results should therefore be more representative of those observed in actual clinical practice compared with the results obtained in most previous studies, and they should be a more accurate representation of the diagnostic yield that can be expected from myocardial perfusion MR imaging.
Our results also suggest that cardiac MR imaging can be performed in patients with diabetes mellitus, left ventricular hypertrophy, or a previous myocardial infarction. In our study, the accuracy of cardiac perfusion MR imaging in the patients with these risk factors was similar to that in the entire study population, although the small number of patients in these subgroups prevented a formal statistical evaluation. Like investigators in previous studies (15), we observed similar diagnostic accuracy for all three coronary artery territories. This uniform accuracy is one of the main advantages that cardiac MR imaging has over nuclear perfusion techniques, which can be affected by attenuation artifacts that can lead to false readings in the territory supplied by the circumflex or right coronary arteries.
Most previous studies of myocardial perfusion MR imaging have been limited to the assessment of a single imaging section. However, in recent recommendations for cardiac imaging (19), coverage of the left ventricle on three image sections is proposed as the minimum requirement for myocardial perfusion assessment. In our study, analysis of the three sections combined led to improved diagnostic accuracy compared with analyses of individual sections. When analyzed separately, the three sections showed minor nonsignificant differences in diagnostic performance. These differences could have been the result of differences in the prepulse delays and signal intensity amplitudes for each section. Other factors may have included the potentially greater sensitivity to through-plane motion and partial volume effects in the apical section.
Like investigators in most previous studies, we used conventional angiography to determine the presence of significant CAD in our patient population. Although conventional angiography yields only an indirect estimate of the flow limitation caused by a coronary stenosis, it is the most clinically important reference examination whose results are the basis of treatment management. Ideally, however, our data also would have been compared with those obtained by using a reference examination for myocardial perfusion assessment, such as positron emission tomography.
We used a single saturation pulse to obtain all sections; this approach results in different saturation times for each section. This technique may have favored those sections that were acquired with a longer prepulse delay, as discussed earlier. Interestingly, the different signal intensity properties of the three sections did not affect the semiquantitative analysis because such differences are largely removed when the MPRI is calculated. An advantage of using a single prepulse was that it allowed accurate measurement of the left ventricular input function from the basal section, as outlined earlier.
We have not performed a direct comparison of the described SENSE technique with other myocardial perfusion MR imaging acquisition strategies. Any head-to-head comparison of different approaches with this very fast myocardial perfusion MR imaging technique will be hampered by the different factors required to optimize each pulse sequence and the many variables that can be changed to achieve this optimization. To address this limitation, we obtained a range of data from our pulse sequence evaluation that should aid in comparisons with other techniques.
In conclusion, multisection myocardial perfusion cardiac MR imaging with SENSE is feasible, and semiquantitative data analysis performed by using an MPRI had high diagnostic accuracy in the detection of significant coronary artery stenosis in our population of patients who were suspected of having or known to have CAD.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, S.P.; study concepts, all authors; study design, S.P., A.R., M.U.S., S.G.B.; literature research, S.P.; clinical studies, S.P., J.P.G., D.B.; data acquisition, S.P., J.P.G., D.B.; data analysis/interpretation, S.P., J.P.G., A.R.; statistical analysis, S.P., A.R.; manuscript preparation, definition of intellectual content, revision/review, and final version approval, all authors; manuscript editing, S.P.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W P Bandettini and A E Arai Advances in clinical applications of cardiovascular magnetic resonance imaging Heart, November 1, 2008; 94(11): 1485 - 1495. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Gebker, C. Jahnke, R. Manka, A. Hamdan, B. Schnackenburg, E. Fleck, and I. Paetsch Additional Value of Myocardial Perfusion Imaging During Dobutamine Stress Magnetic Resonance for the Assessment of Coronary Artery Disease Circ Cardiovasc Imaging, September 1, 2008; 1(2): 122 - 130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Gebker, C. Jahnke, I. Paetsch, B. Schnackenburg, S. Kozerke, A. Bornstedt, E. Fleck, and E. Nagel MR Myocardial Perfusion Imaging with k-Space and Time Broad-Use Linear Acquisition Speed-up Technique: Feasibility Study Radiology, December 1, 2007; 245(3): 863 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R. Nandalur, B. A. Dwamena, A. F. Choudhri, M. R. Nandalur, and R. C. Carlos Diagnostic Performance of Stress Cardiac Magnetic Resonance Imaging in the Detection of Coronary Artery Disease: A Meta-Analysis J. Am. Coll. Cardiol., October 2, 2007; 50(14): 1343 - 1353. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S Watkins, K G Oldroyd, and S Frohwein Magnetic resonance myocardial perfusion imaging: a new era in the detection of reversible myocardial ischaemia Heart, January 1, 2007; 93(1): 7 - 10. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Kramer When Two Tests Are Better Than One: Adding Late Gadolinium Enhancement to First-Pass Perfusion Cardiovascular Magnetic Resonance J. Am. Coll. Cardiol., April 18, 2006; 47(8): 1639 - 1640. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
