| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the Departments of Cardiology (B.L.G., A.P., D.V., J.L.J.V.) and Radiology (E.C., E.K., C.G., B.E.V.B.), Cliniques Universitaire St Luc UCL, Av Hippocrate 10/2806, B-1200 Woluwe St Lambert, Brussels, Belgium. Received August 19, 2003; revision requested October 31; final revision received March 5, 2004; accepted April 15. B.L.G. supported by a grant from the Fondation de la Recherche Scientifique of the Belgian government (FRSM 3.4557.02). Address correspondence to B.L.G. (e-mail: bernhard.gerber@clin.ucl.ac.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: All patients gave written informed consent for the study, which was approved by the local ethics committee. Twenty-seven patients underwent multi–detector row CT and 3D navigator free-breathing MR imaging a mean of 5 days before undergoing invasive coronary angiography. The acquired multi–detector row CT and MR images were graded for the presence of greater than 50% stenosis in vessels larger than 1.5 mm in diameter. The diagnostic accuracies of the two examinations were compared with that of quantitative coronary angiography (QCA) by using the McNemar test.
RESULTS: Owing to claustrophobia, MR images were not acquired in one patient; thus, 26 patients were included for analysis. According to QCA findings, 21 of the 26 patients had significant coronary artery disease and 58 (20%) of a total of 294 coronary artery segments larger than 1.5 mm in diameter had significant (>50%) stenosis. Multi–detector row CT had significantly higher sensitivity (46 [79%] of 58 segments) than MR imaging (36 [62%] segments, P < .05) for detection of segments with significant stenosis. Conversely, MR imaging had significantly higher specificity (198 [84%] of 236 segments) than did CT (168 [71%] segments, P < .001) for exclusion of segmental coronary artery stenosis. Both examinations had high negative predictive value for exclusion of segmental stenosis: 93% (168 of 180 segments) for CT and 90% (198 of 220 segments) for MR imaging. The overall diagnostic accuracy of MR imaging (80% [234 of 294 segments]) was significantly higher than that of CT (73% [214 segments], P < .05).
CONCLUSION: MR imaging had significantly higher diagnostic accuracy than multi–detector row CT in the evaluation of coronary artery stenosis. Both techniques have high negative predictive value, making them particularly useful for ruling out coronary artery disease in symptomatic patients.
© RSNA, 2004
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Considerable progress in the field of noninvasive coronary imaging has been achieved by using magnetic resonance (MR) imaging (4,5) or multi–detector row spiral coronary computed tomography (CT) (6,7). The results of several studies (8–10) suggest that these two modalities have similar diagnostic accuracies. Yet, to our knowledge, until now, there has been no direct comparison of these two imaging approaches in the same patients. Thus, the purpose of our study was to prospectively compare the diagnostic accuracy of multi–detector row CT and of three-dimensional (3D) navigator-gated MR imaging in patients referred for conventional invasive coronary angiography for the detection of coronary artery stenosis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
The availability of both CT and MR imaging units allowed the recruitment of one patient into the study per week. Of 128 patients who underwent primary screening, 36 did not meet the inclusion criteria because they had previously undergone revascularization procedures. Six patients were excluded because of atrial fibrillation; five patients, because of heart failure; 21 patients, because of contraindications to multi–detector row CT (14 patients had possible allergies to iodinated contrast media, and seven had decreased renal function); and three patients, because of contraindications to MR imaging (pacemaker implantation). Ten patients were unavailable at the time of screening, and seven patients were not approached because another patient had already agreed, on the given day, to participate in the study. Thirteen patients refused to participate in the study. Overall, 27 patients agreed to participate in the study and underwent CT and MR imaging. All patients gave written informed consent to the study protocol, which had been approved by our local ethics committee.
Study Protocol
Patients underwent 3D navigator-gated MR imaging and multi–detector row CT in random order on the same day. Both examinations were performed a mean of 5 days ± 5 (standard deviation) before conventional coronary angiography was performed. Whenever the heart rate was greater than 65 beats per minute, the patient was given 50 mg of atenolol (Tenormin; Astra-Zeneca, Brussels, Belgium) orally at least 1 hour before the multi–detector row CT examination.
MR Imaging
MR imaging was performed by using a 1.5-T magnet (Intera CV; Philips Medical Systems, Best, the Netherlands) that has high performance gradients (60 mT/m, 150 mT/m/sec rise time) and by using dedicated cardiac software (release 9.1, part of Intera CV unit). Gated imaging was performed by using a five-segment phased-array coil (part of Intera CV unit) and a vectocardiographic monitoring system.
Two scout localizer MR imaging examinations were performed for coronary artery localization and navigator placement, as previously described (11). The first localizer examination consisted of a vectocardiographically gated free-breathing two-dimensional multisection gradient-echo three-plane scout sequence that enabled the acquisition of seven transverse, sagittal, and coronal MR images of the thorax. These scout images were used to localize the volume for subsequent 3D transverse localizer scout image acquisition and to position the navigator. The navigator was placed on the dome of the right hemidiaphragm, and all subsequent imaging was performed with a 5-mm gating window for acceptance end-diastolic navigator gating and a constant superior-inferior correction factor of 0.6. The second localizer examination consisted of a vectocardiographically triggered free-breathing navigator-gated multisection 3D segmented transverse balanced turbo fast field-echo sequence performed with T2 preparation and the sensitivity-encoding parallel imaging technique.
The MR imaging parameters were 3.7/1.8 (repetition time msec/echo time msec), a 90° flip angle, a 270-mm field of view, a 160 x 160 x 70 matrix reconstructed to 256 x 256 x 70 pixels, and a sensitivity-encoding acceleration factor of two. A total of 70 1.7-mm-thick overlapping sections were acquired.
Next, coronary MR images were acquired by using a vectocardiographically triggered free-breathing navigator-gated multisection 3D segmented transverse balanced turbo fast field-echo sequence with T2 preparation. The imaging parameters were 5.8/2.9, a 110° flip angle, a 270-mm field of view, a 272 x 272 x 10 matrix (with an 80% rectangular field of view) reconstructed to 512 x 512 x 20 pixels, and a turbo fast field-echo factor (ie, number of shots per heartbeat) of 16, which resulted in an acquired spatial resolution of 1 x 1 x 3 mm reconstructed to 0.5 x 0.5 x 1.5 mm. The MR images were acquired during mid-diastole for 90 milliseconds. Four double-oblique section locations were prescribed by using a three-point-plan imaging tool (Fig 1).
|
Multi–Detector Row Coronary CT
Multi–detector row CT was performed by using a four-section system (MX 8000; Philips Medical Systems, Cleveland, Ohio) with a 0.5-second rotation time and dedicated cardiac reconstruction software. Oxygen was administered at a flow rate of 5 L/min by using a mask to facilitate breath holds during data acquisition. A 140-mL bolus of nonionic contrast medium (ioversol, Optiray 350; Guerbet, Roissy, France) was injected at a rate of 3 mL/sec by using an infusion pump. Scanning was initiated after automatic detection of the contrast agent bolus in a region of interest placed over the ascending aorta. The patient was then instructed to maintain an inspiratory breath hold for ±40 seconds (mean, 44 seconds ± 3 [standard deviation]; range, 38–50 seconds), during which a three-lead electrocardiographic trace and a 15-cm-long transverse stack of multi–detector row CT data were acquired. Acquisition parameters were as follows: 4 x 1-mm detector collimation (1.2-mm effective section thickness), 3.75 mm/sec table feed, 120-kV tube voltage, and a 225-mA tube current.
Synchronized to the recorded electrocardiographic signal, transverse sections were reconstructed from the acquired multi–detector row CT data by using an algorithm that involves the use of only the data from a half gantry rotation per section, which results in a temporal resolution of 250 msec. The CT images were reconstructed to a 512 x 512 matrix by using a field of view of 28 cm, a section thickness of 1.2 mm, and a reconstruction interval of 0.6 mm. To determine the optimal position of the image reconstruction window relative to the cardiac cycle, eight image data sets were reconstructed at 0%, 12.5%, 25.0%, 37.5%, 50.0%, 62.5%, 75.0%, and 82.5% of the cardiac cycle. For each patient, the data set that contained the fewest motion artifacts was selected for further evaluation. In 20 (74%) patients, the optimal reconstruction window was at 75% of the cardiac cycle, whereas in the remaining seven (26%) patients it was at 50% of the cardiac cycle.
Coronary Angiogram Acquisition and Interpretation
Selective biplanar coronary angiography was performed by using a femoral approach with 6-F catheters according to a modified Judkins technique. Coronary angiograms were acquired after intracoronary injection of nitrates. Multiple orthogonal projections were used. A cardiologist (A.P.) with 7 years experience in coronary angiography evaluated the data in a blinded fashion by using a quantitative coronary angiography (QCA) system (Cardiovascular Angiographic Analysis System II; Pie Medical Equipment, Maastricht, the Netherlands) (12,13), catheter-based image calibration, and automated vessel contour detection. The images were calibrated by measuring the size of the 6-F catheter. The reference vessel diameter and the diameter and length of the stenosis were computed automatically. As shown in Figure 2, the American Heart Association standard classification system was used (14). Only segments with a reference diameter of greater than 1.5 mm were evaluated. Segments with a smaller diameter and segments distal to a proximal occlusion were considered to be absent. Significant vessel stenosis was defined as a greater than 50% reduction in luminal diameter at QCA.
|
A custom "soap bubble" postprocessing software (15) was used to reconstruct 3D Delaunay triangulation projections for the MR and CT images. This tool was also used by the observer (B.L.G.) to measure the length of the reconstructed coronary segments on anonymous multi–detector row CT and MR images. Sets of anonymous MR and CT images were visually evaluated by two reviewers—a cardiologist experienced in cardiac imaging (B.L.G.) and an experienced chest radiologist (E.C.)—both of whom had approximately 2 years experience with both coronary imaging modalities. For image interpretation, the reviewers had access to all raw MR and multi–detector row CT images, to the oblique reformatted images, and to the soap bubble projection images. Both reviewers graded the images for the presence and severity of stenosis in each of the 14 predefined coronary artery segments on the anonymous MR and CT images.
The two reviewers interpreted the images independently, and analyses of the MR and CT image findings were performed at least 30 days apart to avoid interpretation bias. Interpretation differences were later resolved by consensus. Segmental image quality (ie, quality of segment depiction) was graded on a five-point scale, on which grade 1 meant very poor; grade 2, fair; grade 3, satisfactory; grade 4, good; and grade 5, excellent image quality. For each segment, stenosis severity was judged to be nonsignificant (<50% luminal diameter reduction) or significant (>50% luminal diameter reduction).
Statistical Analyses
Statistical analyses were performed by using SPSS 10 software (SPSS, Chicago, Ill). Values are reported as means ± 1 standard deviation. Measurements of vessel length, CNR, and subjectively graded image quality were compared between MR and multi–detector row CT by using paired t tests. Interobserver concordance regarding segmental stenosis severity was evaluated by using 
statistics, with a value of less than 0.40 indicating poor agreement; a value of 0.40 or greater but less than 0.60, fair agreement; and a value of 0.60 or greater, good agreement. In the assessment of the diagnostic accuracy of MR imaging and of multi–detector row CT, discordances between the two readers were resolved by means of consensus image reading. The diagnostic accuracies of the two examinations were compared by using QCA as the reference standard. Differences in accuracy between MR imaging and multi–detector row CT were compared by using the McNemar test and the Z test. At Z testing, estimates of variance and covariance were weighted for data clustering in patients. All statistical tests were two sided, and P < .05 was considered to indicate statistical significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Coronary Artery Stenosis Severity at QCA
According to QCA findings, a total of 294 segments (10.9 per patient) had a reference diameter of greater than 1.5 mm. Fifty-eight (20%) of these segments were shown to have greater than 50% luminal diameter stenosis (mean stenosis degree, 71% ± 18). The mean stenosis length was 8.5 mm ± 5.7. Six patients had single-vessel disease, six had two-vessel disease, 10 had three-vessel disease, and five were considered to be free of significant CAD. The location and severity of the coronary stenoses are shown in Table 1.
|
|
|
|
|
|
|
= 0.65) with 3D navigator MR imaging and 81% ( = 0.58) with multi–detector row CT. Segment-by-segment analysis.—The diagnostic accuracy in the prediction of stenosis severity computed in all segments larger than 1.5 mm in diameter is shown in Figure 8. The sensitivity for detecting greater than 50% stenosis was better with multi–detector row CT (79% [46 of 58 segments]) than with MR imaging (62% [36 of 58 segments]; P < .05 at McNemar test, P = .07 at Z test with results corrected for data clustering in patients). On the other hand, MR imaging had significantly better specificity (84% [198 of 236 segments]) than multi–detector row CT (71% [168 of 236 segments]; P < .001 at McNemar test, P < .005 at Z test with results corrected for data clustering in patients).
|
The reasons for the false-positive and false-negative MR and multi–detector row CT image readings are summarized in Table 3. The main reason for the false-positive MR image readings was poor opacification or small vessel size, as well as low CNR in vessels distal to a proximal stenosis. The main reason for the false-positive CT image readings was the presence of substantial amounts of coronary calcium. Other reasons were poor opacification or small vessels and overestimations of the degree of narrowing in borderline stenoses. Vessels with borderline stenosis had a mean percentage of stenosis of 43% at QCA. False-negative MR image readings were infrequent and were related to small vessels and poor CNR. In nine cases, the stenosis was not visualized. False-negative CT image readings were even rarer and were most often related to small vessels or poor opacification.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Motion correction for coronary artery MR imaging is performed by means of prospective triggering of the image acquisition to the detection of the R wave of a vectocardiogram by acquiring images with a temporal resolution of 90 msec in mid-diastole. For multi–detector row CT, cardiac motion correction is performed by means of retrospective rearrangement of multisection partial scan data relative to an electrocardiographic signal that is recorded during scan acquisition (7). In the present study, the effective temporal resolution was 250 msec. Correction forrespiratory motion at MR imaging is performed by using a navigator pulse, which tracks the position of the diaphragm in real time (11) and gates the image acquisition to end expiration while the patient is breathing freely. Respiratory motion at multi–detector row CT is prevented by performing imaging during a breath hold.
The coronary MR imaging pulse sequence did not require contrast agent injection. Blood in the coronary arteries appears bright because of its high T2-T1 ratio, while the signal from adjacent tissue is suppressed by selective fat saturation and T2 preparation pulses. In contrast, differentiating between the coronary arteries and the adjacent tissue at multi–detector row CT requires intravenous injection of an iodinated contrast agent. Although both MR imaging and multi–detector row CT yield submillimeter in-plane resolution, the CT examination yields a smaller and more isotropic voxel size (0.8 x 0.8 x 1.2 mm3) compared with MR imaging (1 x 1 x 3 mm3) and thus enables better 3D reformatting. Such postprocessing of images is mandatory for multi–detector row CT depiction of the oblique course of the coronary arteries in the body, since these scans are acquired transversely. In contrast, MR imaging allows one to directly acquire images in the double-oblique orientation of the coronary arteries and therefore requires less postprocessing.
Differences in CNR, Spatial Resolution, and Image Quality between Multi–Detector Row CT and MR Imaging
The present study results show that the CNRs measured with multi–detector row CT were significantly higher than those measured with MR imaging. These results, together with the slightly better spatial resolution at CT, likely explain why this technique permitted better visualization of small coronary artery segments, such as side branches and distal segments, than did MR imaging. Despite having these technical advantages, multi– detector row CT did not allow more complete visualization of the length of the major coronary arteries, nor did its use result in better image quality compared with the use of MR imaging.
Image quality was judged to be significantly better with MR imaging than with multi–detector row CT, mainly because fewer motion artifacts occurred with the MR technique. Indeed, the coronary arteries have high motion during the cardiac cycle, mainly during systole and early diastole. Imaging is performed during the period of diastasis in mid-diastole, when there is the least motion. To prolong this period, patients were premedicated with ß-blockers. However, this premedication did not appear to be sufficient to prevent motion artifacts at CT, probably because the temporal resolution (250 msec) was too low.
Variable cardiac cycle length caused by, for instance, arrhythmic beats was another potential cause of motion artifacts. Because cardiac gating at multi–detector row CT is performed retrospectively, without the possibility for arrhythmic heartbeat rejection, multi–detector row CT is more susceptible to arrhythmia-related artifacts than is MR imaging, which involves the use of a prospective gating algorithm with which arrhythmic cycles can be rejected. Although in the present study patients with known arrhythmias were excluded, premature heartbeats still occurred occasionally in some patients and were responsible for the occurrence of motion artifacts. Not surprisingly, these artifacts mostly affected the RCA and the LCX artery, which have the highest velocity of motion during the cardiac cycle (16). In addition, these vessels are oriented perpendicularly to the multi–detector row CT transverse acquisition plane. Therefore, if the motion in these vessels is too fast or if the cardiac cycle length is too variable, the likelihood of these arteries being misaligned between adjacent planes increases, with the result being steplike offset artifacts after reformatting in oblique directions.
Diagnostic Accuracy of Multi– Detector Row CT and of MR Imaging in Detection of Coronary Artery Stenosis
The salient finding in the present study was probably the observation that despite having similar overall diagnostic accuracy in the detection of significant CAD, MR imaging and multi–detector row CT significantly differed in their capability to depict diseased and nondiseased segments. On a segmental or per-vessel basis, CT had higher sensitivity for the detection of significant coronary stenoses, whereas MR imaging had better capability to depict nondiseased vessels and thus had higher specificity.
The higher sensitivity of multi–detector row CT was undoubtedly related to the better spatial resolution and higher CNRs. It also probably reflected the greater capability of CT to depict calcified plaques, which usually generate a strong signal on multi–detector row CT images and are therefore quite easily identified. In contrast, these lesions often appear as a signal void on MR images, which makes their identification more difficult. On the other hand, MR imaging had greater capability to depict normal or nonsignificantly diseased vessels and thus had better specificity than multi–detector row CT. This probably reflected the greater propensity of the CT findings to lead to overestimated stenosis severity, particularly in the presence of coronary calcium, which causes blurring of the vessel lumen.
The diagnostic accuracy of coronary artery MR imaging and of multi–detector row CT has been reported separately by several authors. In a large multicenter trial that included 109 patients (17), coronary MR imaging had somewhat higher sensitivity (93%) but lower specificity (42%) and diagnostic accuracy (72%) compared with the values in our study. It is probable, although speculative, that these differences in accuracy are due to our inclusion of distal vessels and side branches, which were excluded from the multicenter trial; the acquisition of two more planes for the LAD and LCX arteries in our study; and our use of a more recently developed pulse sequence. Results of the initial experience of Bogaert et al (18), who used a sequence similar to the one that we used and reported similar sensitivity (44%–55%), specificity (83%–95%), and accuracy (79%–85%), support this contention.
The diagnostic accuracy of multi–detector row CT in our study was quite similar to that reported in prior works. However, the specificity was somewhat lower than those previously reported. In a series of 35 patients, Nieman et al (8) reported a specificity of 97%. Yet, in that study, side branches and distal vessels were not evaluated. In a study that included all side branches larger than 2 mm in diameter, Achenbach et al (19) reported a specificity of 84%. The differences in results between their study and ours can probably be explained by their exclusion of a large number of segments (32%) that were not considered to be evaluatable. Similar findings were also reported by Knez et al (20) and Vogl et al (21). More recently, with use of 16-section multi–detector row CT, Ropers et al (22) and Nieman et al (23) reported a sensitivity of 92% and a specificity of 93%. Again, when the so-called nonevaluatable segments were included, sensitivity was reduced to 73%.
To our knowledge, direct head-to-head comparisons between MR imaging and multi–detector row CT of native coronary vessels had not been reported before the present study. The only study in which MR imaging and CT were directly compared was one in which MR imaging enhanced with gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Montville, NJ) was compared with electron-beam CT for the detection of restenosis after angioplasty (24). Results indicated higher accuracy with electron-beam CT (71%) than with MR imaging (53%) (24).
Clinical Implications
The present study results show that neither MR imaging nor multi–detector row CT is ready to be a replacement for conventional coronary angiography at this time. The findings nonetheless suggest that these modalities may be able to compete with other noninvasive examinations that are used to detect CAD, such as exercise stress testing. This would be particularly true in situations in which exercise electrocardiography has either low sensitivity, such as in women and in patients with limited exercise capability, or low specificity, such as in patients with preexisting electrocardiographic abnormalities (25). Because of their high negative predictive value, MR imaging and multi–detector row CT might be particularly useful in helping to rule out the presence of significant CAD and thus avoid the acquisition of unnecessary coronary angiograms in patients in whom exercise electrocardiography might have high false-positive rates (26).
The choice between multi–detector row CT and MR imaging will probably depend on local equipment availability and operator expertise. The choice will also probably be influenced by individual patient characteristics. For example, MR imaging will probably be preferred for patients with known allergies to contrast agents and those at risk for renal insufficiency after contrast agent administration, as well as for elderly patients, in whom the high prevalence of extensive coronary calcifications may hamper the interpretation of stenosis severity at multi–detector row CT. On the other hand, multi–detector row CT may be the method of choice in younger patients, in whom renal toxicity from contrast agents and the presence of severe coronary calcifications are less common. Multi–detector row CT will also probably be preferred for anxious or claustrophobic patients given the associated shorter acquisition time, as compared with that for MR imaging.
Study Limitations
We examined a relatively small group of patients who had a high prevalence of severe CAD. Images were interpreted by two reviewers who were fairly experienced with both imaging techniques. The study findings may not necessarily be applicable to patients with less severe CAD, and they may not necessarily reflect the readings performed by other, less experienced reviewers.
Finally, in the present study, we used four-section multi–detector row CT, which represented the state of the art in multi–detector row CT technology at the time the study was initiated. More recently, multi–detector row CT scanners with 16 detector rows have been introduced. This new generation of multi–detector row CT scanners yields higher temporal and spatial resolution and thus might have better diagnostic accuracy for the detection of CAD than the four-section machine used in the present study. Since the field of noninvasive coronary imaging is rapidly evolving, we anticipate that upcoming improvements will result in additional increases in the diagnostic accuracy of both imaging techniques.
Present study results show that despite having similar overall diagnostic accuracy for the detection of significant CAD, 3D navigator MR imaging and multi–detector row CT differ significantly in their capability to depict diseased and nondiseased vessels. On a segmental or vessel basis, multi–detector row CT has higher sensitivity for the detection of significant coronary artery stenosis, whereas MR imaging has greater capability to depict nondiseased vessels and thus has higher specificity.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, B.L.G., E.C., J.L.J.V.; study concepts, B.L.G., E.C., B.E.V.B., J.L.J.V.; study design, B.L.G., E.C., J.L.J.V.; literature research, B.L.G., E.C.; clinical studies, B.L.G., E.C., A.P., E.K., D.V., C.G., B.E.V.B.; data acquisition, B.L.G., E.C.; data analysis/interpretation, B.L.G., E.C., A.P., E.K., D.V., C.G., B.E.V.B.; statistical analysis, B.L.G., E.C.; manuscript preparation, B.L.G., E.C., A.P., E.K., D.V., B.E.V.B., J.L.J.V.; manuscript definition of intellectual content, B.L.G., E.C., J.L.J.V., B.E.V.B.; manuscript editing, B.L.G., J.L.J.V.; manuscript revision/review and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. J. de Feyter and R.-J. van Geuns Noninvasive Coronary Imaging: The Contest Between Magnetic Resonance and Computed Tomographic Coronary Angiography Circ Cardiovasc Imaging, September 1, 2008; 1(2): 89 - 91. [Full Text] [PDF]
|
|
|
|
|
|
A.-C. Pouleur, J.-B. le Polain de Waroux, J. Kefer, A. Pasquet, J.-L. Vanoverschelde, and B. L. Gerber Direct Comparison of Whole-Heart Navigator-Gated Magnetic Resonance Coronary Angiography and 40- and 64-Slice Multidetector Row Computed Tomography to Detect the Coronary Artery Stenosis in Patients Scheduled for Conventional Coronary Angiography Circ Cardiovasc Imaging, September 1, 2008; 1(2): 114 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Pugliese, N. R. Mollet, M. G. M. Hunink, F. Cademartiri, K. Nieman, R. T. van Domburg, W. B. Meijboom, C. Van Mieghem, A. C. Weustink, M. L. Dijkshoorn, et al. Diagnostic Performance of Coronary CT Angiography by Using Different Generations of Multisection Scanners: Single-Center Experience Radiology, February 1, 2008; 246(2): 384 - 393. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Liu, X. Zhao, J. Huang, C. J. Francois, D. Tuite, X. Bi, D. Li, and J. C. Carr Comparison of 3D Free-Breathing Coronary MR Angiography and 64-MDCT Angiography for Detection of Coronary Stenosis in Patients with High Calcium Scores Am. J. Roentgenol., December 1, 2007; 189(6): 1326 - 1332. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. K. Vanhoenacker, M. H. Heijenbrok-Kal, R. Van Heste, I. Decramer, L. R. Van Hoe, W. Wijns, and M. G. M. Hunink Diagnostic Performance of Multidetector CT Angiography for Assessment of Coronary Artery Disease: Meta-analysis Radiology, August 1, 2007; 244(2): 419 - 428. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
