| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 British Heart Foundation—Cardiac MRI Unit, Rm. 170, D Fl., Jubilee
Bldg., Leeds General Infirmary, Great George St., Leeds LS1 3EX, United
Kingdom.
2 Department of Medical Physics, Leeds General Infirmary, Leeds, LS1 3EX, United
Kingdom.
Received May 16, 2002;
accepted after revision July 15, 2002.
S. Plein is supported by a junior fellowship grant from the British Heart
Foundation.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We implemented an acquisition technique that uses subject-specific acquisition windows in the cardiac cycle and a motion-adapted gating window for respiratory navigator gating. Cardiac acquisition windows and trigger delays were determined individually from a coronary motion scan. Motion-adapted gating used a 2-mm acceptance window for the central 35% of k-space and a 6-mm window for the outer 65% of k-space. In 10 subjects, three-dimensional coronary MR angiograms of the right and left coronary systems were acquired with this technique (the "adaptive technique") as well as a conventional acquisition method, and the scanning times and image quality were compared. The adaptive technique was then applied prospectively to 40 patients who underwent coronary radiographic angiography.
RESULTS. Scanning times with the adaptive technique were reduced by a factor of 2.3 for the right coronary artery and by a factor of 2.2 for the left coronary artery system compared with the conventional technique, mainly because we were able to use longer subject-specific acquisition windows in patients with low heart rates. Subjective and objective measurements of image quality showed no significant differences between the two techniques. Prospective evaluation of MR angiograms yielded a sensitivity and specificity of 74.3% and 88.2%, respectively, to detect significant coronary artery stenoses.
CONCLUSION. Coronary MR angiography with subject-specific acquisition windows and motion-adapted respiratory gating reduces scanning times while maintaining image quality and provides high diagnostic accuracy for the detection of coronary artery stenosis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In the respiratory cycle, data acquisition is performed either during breath-holding [1,2,3] or using respiratory navigator echoes [4,5,6,7,8,9,10]. Navigator-gated methods have recently shown particularly promising results because these methods allow the acquisition of three-dimensional (3D) data sets independent of the breath-hold capability of the patient. However, these techniques require long scanning times because data are accepted during only a narrow and fixed window of diaphragmatic position at end-expiration [4,5,6,7,8,9,10]. This limitation can lead to low scanning efficiencies especially long acquisitions when the probability of diaphragmatic drift is high [14].
We hypothesized that scanning times for navigator-gated 3D coronary MR angiography could be reduced using a flexible and subject-specific approach to both cardiac and respiratory gating. The duration of mid-diastolic diastasis has previously been shown to vary at least as much between individuals as its onset [15]. Particularly long rest periods are seen in subjects with low heart rates. Thus, a direct assessment of the motion of the coronary arteries may allow the definition of an acquisition window that is specific to each subject. In subjects with long coronary rest periods, extended acquisition windows could potentially be used with consequently shortened scanning times.
For respiratory gating in navigator-gated acquisitions, a narrow acceptance window is particularly important for the center of k-space, which determines image contrast, but is less important for the peripheral parts of k-space. With adaptive respiratory gating methods, such as motion-adapted gating, the outer lines of k-space can be filled with data from a wider acquisition window than the central lines of k-space [16, 17]. When applied to coronary MR angiography, motion-adapted gating can be used either to improve navigator efficiency and shorten scanning times if overall wider acceptance windows are used or to improve image quality by using a narrow acceptance window for the center of k-space.
The aim of this study was to investigate whether the combination of subject-specific cardiac gating and motion-adapted respiratory gating can shorten scanning times in coronary MR angiography while preserving image quality.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Exclusion criteria for both groups were contraindications to MR imaging, claustrophobia, pregnancy, active congestive heart failure, and known severe aortic stenosis or obstructive cardiomyopathy. The study was approved by the local ethics committee, and all subjects gave informed written consent.
MR Imaging
MR imaging examinations were performed with the patient in the supine
position on a commercial 1.5-T system (Gyroscan Intera CV; Philips Medical
Systems, Best, The Netherlands) equipped with "Master" gradients
(peak gradient, 30 mT/m; slew rate, 150 m/T per second). A five-element phased
array cardiac synergy coil and a vectorcardiographic method for ECG gating
[18] were used.
Coronary Motion Assessment
From the initial localizing scans, multiple-phase cine images were planned
in the vertical long-axis, short-axis, and horizontal long-axis orientations.
The images were acquired with a steady-state free precession pulse sequence
(balanced fast field echo TR/TE, 2.8/1.4; flip angle, 55°; field of view,
360 x 288 mm; partial Fourier acquisition; matrix, 187 x 144;
slice thickness, 7 mm; ECG-triggered; 18 phases per R-R interval for the
vertical long-axis and short-axis scans). The horizontal long-axis scan was
aligned perpendicular to the mid portion of the right coronary artery and
circumflex artery and was acquired with 30 phases per R-R interval
(acquisition duration, approximately 8 sec).
The motion of both vessels was assessed by scrolling through the cardiac phases of this data set (Fig. 1A,1B). The period in mid-diastole, when both vessels were without perceptible motion, was identified visually. This assessment was aided, when necessary, by placing a circular region of interest with a 1-mm radius on the center of the vessel and copying it to subsequent phases. Perceptible motion was then defined as the vessel moving outside the region of interest between two subsequent phases. A subject-specific trigger delay and an acquisition window in which the motion of both vessels was minimal were then determined for subsequent coronary acquisitions.
|
|
Coronary Localizing Scan
The course of the coronary arteries was determined from a navigator-gated
transverse 3D segmented echoplanar localizing scan (16/4.9; flip angle,
40°; T2-weighted preparation and fat-suppression prepulses; magnetization
transfer; real-time navigator echo acquired perpendicular to the right
hemidiaphragm) that was obtained while the patient was breathing freely. A
double oblique 3D volume was planned for the right coronary artery with a
three-point planscan tool and by placing points on the proximal, mid, and
distal right coronary artery. The resulting imaging plane was reviewed and
adjusted to ensure that it covered the length of the entire right coronary
artery. For the left coronary system, a double oblique transverse 3D volume
was planned by placing points on the proximal left main, the mid left anterior
descending, and mid circumflex coronary arteries. The resulting imaging plane
was reviewed and adjusted to ensure coverage of at least the proximal and mid
aspects of both vessels.
High-Resolution Coronary MR Angiography
In the comparison group, a total of four high-resolution 3D coronary
acquisitions were performed: two data sets of the right coronary artery and
two data sets of the left coronary artery system were acquired using both a
conventional technique and the adaptive acquisition technique. Identical
scanning orientations were used to allow comparison of the two techniques, and
scans were obtained in random order. In the prospective group, data sets of
the two coronary systems were acquired with only the adaptive approach.
The conventional pulse sequence used acquisition parameters that are similar to a previously described 3D segmented k-space gradient-echo technique [19]; however, we made minor modifications to ensure that the conventional technique was comparable to the adaptive technique. For the conventional technique for our study, eight contiguous 3-mm-thick slices were acquired and interpolated during reconstruction to 16 slices of 1.5-mm thickness, and the following parameters were used: 7.4/2.5; flip angle, 30°; T2-weighted and fat-saturation preparation prepulses; field of view, 400 x 300 mm; matrix, 512 x 384; in-plane spatial resolution, 1.04 x 0.78 mm.
A navigator echo was acquired from a cylindrical region generated by a
two-dimensional excitation pulse perpendicular to the right hemidiaphragm.
Prospective gating of the 3D coronary acquisition based on the navigator
position was applied with a fixed acceptance window of 5 mm. For data accepted
within this window, real-time correction of the 3D volume in the craniocaudal
direction was applied. The temporal data acquisition window was approximately
73 msec in all patients, with slight variations according to the angulation of
the imaging plane. The trigger delay (TD) was set according to the
patient's heart rate using an empirically modified algorithm to determine
mid-diastole, which is based on the following formula suggested by Stuber et
al. [10]:
 |
Acquisition with the adaptive technique was performed with an identical sequence and scanning parameters except for the use of a subject-specific data acquisition window and the motion-adapted technique for respiratory gating. Because the two 3D volumes aligned along the left coronary system and the right coronary artery can overlap partially, thus allowing analysis of segments of both coronary systems, we used the same subject-specific acquisition window for both 3D coronary scans. To account for possible variations in the coronary rest period, we limited the maximum duration of the acquisition window to 200 msec. For the motion-adapted gating technique, we used a 2-mm acceptance window for the central 35% of k-space and a 6-mm window for the outer 65% of k-space. The selection of these parameters was based on pilot work that had shown them to minimize respiratory motion artifacts while achieving similar acquisition efficiency as a conventional fixed acceptance window of 5 mm. We also allowed continuous correction of the gating window position to adjust for drift of the mean diaphragmatic position during the acquisition.
Analysis
Acquisition parameters, which included scanning time, data acquisition
window, trigger delay, and navigator efficiency (i.e., number of accepted
shots / total number of heart beats), were recorded for each coronary
acquisition. The trigger delays determined from the coronary motion assessment
were displayed in comparison with the trigger delays that would have been used
if based on the following heart rate—dependent formulas
[10]:
 |
|
|
Further analysis was performed on a workstation (UltraSparc 10; Sun Microsystems, Santa Clara, CA) equipped with commercial image post-processing software (Easy Vision version 4.0; Philips). The overall image quality of each data set was assessed by consensus of two observers who were unaware of which acquisition sequence had been used. The observers used the following scale to rate image quality: 1 (poor), the coronary vessels could not be identified because of image artifacts; 2 (moderate), the coronary arteries could be identified but image quality was impaired due to artifacts; 3 (good), the coronary vessels were clearly outlined and there were no image artifacts; and 4 (excellent), the definition of the coronary vessels was exceptionally clear.
In the comparison group, objective measurements were then carried out by
one observer who was unaware of which acquisition sequence had been used. The
observer measured the length of each coronary artery with a multiplanar
reformatting tool. The signal intensity (SI) and the standard deviation (SD)
of the signal intensity in a standard-sized region of interest in the aorta,
proximal coronary vessel, and adjacent myocardium (only on left coronary data
sets) were measured. From these measurements, the signal-to-noise ratio (SNR)
of the aorta, coronary arteries, and myocardium were calculated using the
following formula:
 |
The contrast-to-noise ratio (CNR) between coronary vessel and myocardium
was calculated for the left coronary data sets using the following formula:
 |
Finally, the data sets acquired with the adaptive technique in all 50 subjects were analyzed by two observers who were unaware of the clinical data and results from previous analyses. Images were reviewed by scrolling through the individual slices of the 3D data sets, and stenosis in the coronary artery lumen of 70% or more was recorded. Data sets that were rated as poor were excluded. To determine the diagnostic accuracy of the adaptive acquisition, we compared these results with the results of radiographic coronary angiography. Data sets of the five healthy volunteers were included in this analysis, and we assumed that their coronary arteries were normal.
Radiographic Angiography
Radiographic angiography of the coronary arteries was performed in the
standard manner and was reported independently by an experienced
interventional cardiologist who was unaware of the results from the MR
angiogrphy. The presence of any coronary artery lesions of 70% or more luminal
stenosis was reported.
Statistical Analysis
Mean values and SDs were calculated for all acquisition and image quality
parameters. The variability of the subject-specific trigger delays was
determined by calculating the SD of the difference of the individual trigger
delays from the mean trigger delay at each heart rate (rounded to the nearest
5 beats per minute). In the comparison group, results of the two acquisition
strategies were compared using a paired samples t test. The
sensitivity, specificity, diagnostic accuracy, and positive and negative
predictive values for coronary MR angiography to detect luminal stenosis of
70% or more on coronary radiographic angiography in the main coronary vessels,
excluding any side branches, were determined.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Coronary Motion and Subject-Specific Acquisition Windows
The trigger delay for data acquisition determined from the coronary motion
scan ranged from 430 to 920 msec with an inverse relationship to the heart
rate (Fig. 2). Compared with
the original formula described by Stuber et al.
[10], the subject-specific
trigger delay was longer in all but two acquisitions. Compared with the
modified formula, subject-specific trigger delays were shorter in all but one
acquisition. Figure 3 shows the
mean trigger delay used for heart rates rounded to the nearest 5 beats per
minute; error bars indicate the SDs. The mean value of these SDs was 60.5
msec.
|
|
The individually determined length of the acquisition windows ranged from 78 to 200 msec and was thus longer than the conventional acquisition window in all subjects (Fig. 4). An inverse relationship was seen between the duration of the acquisition window and the heart rate. An acquisition window of approximately 200 msec was used in 15 patients, mainly those with a heart rate of 60 beats per minute or lower.
|
Comparative Study
The mean acquisition window with the adaptive acquisition was significantly
longer by a factor of 1.88 for the right coronary artery and 1.9 for the left
coronary system compared with the conventional acquisition. Navigator
efficiencies were slightly higher with the adaptive acquisition, but the
difference compared with the conventional acquisition was not statistically
significant. The mean scanning time using the adaptive technique was
significantly shorter by a factor of 2.3 for the right coronary artery and 2.2
for the left coronary system compared with conventional acquisition
(Table 1).
|
Image quality was comparable between the two techniques (Table 1 and Figs. 5A,5B,5C,6A,6B,6C,7). Objective image quality and vessel parameters showed no statistically significant differences except for a higher CNR obtained with the conventional acquisition (10.6 vs 12.3; p < 0.05) (Table 1).
|
|
|
|
|
|
|
Prospective Study
In the prospective group, the acquisition parameters (acquisition window,
trigger delay, navigator efficiency, and scanning time) were similar to those
used for the adaptive acquisitions in the comparative group
(Table 2).
|
Five data sets (two of the left system and three of the right coronary artery) were graded as being of insufficient quality for analysis. On radiographic angiography, 23 patients were found to have significant coronary artery disease, with a total of 35 lesions (two left main, 15 left anterior descending, five circumflex, and 13 right coronary artery). All the lesions in the left main coronary artery were detected on MR angiography, and sensitivity for left anterior descending, right coronary, and circumflex arteries was 73.3%, 76.9%, and 60%, respectively. The sensitivity of MR angiography to reveal significant coronary artery lesions in any vessel was 74.3% with a specificity of 88.2% (Table 3).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The reductions in scanning time were achieved mainly by using longer, subject-specific acquisition windows in the cardiac cycle. Kim et al. [12] have shown that subject-specific trigger delays are superior to trigger delays estimated from heart rate—dependent formulas for acquisition of coronary MR angiography. One reason for this finding is that apart from the heart rate, coronary motion in the cardiac cycle also depends on factors such as the ventricular contraction pattern and the anatomic location of the coronary vessel imaged. Coronary motion therefore varies substantially between individuals even at the same heart rate [14]. These observations were confirmed in our study by the wide variation of subject-specific trigger delays even in individuals with similar heart rates. As a consequence of this inherent inaccuracy of heart rate—dependent formulas, short acquisition windows must be used in conjunction with such formulas to avoid image blurring caused by residual coronary motion during data acquisition. This method of compensation can lead to long scanning times of up to 15 min per coronary data set [21].
We have expanded the application of an individual assessment of coronary motion over previous reports to determine not only the optimal trigger delay but also the maximal possible duration of the acquisition window specifically for each individual. This approach allowed the use of acquisition windows of on average twice the length recommended for a conventional acquisition method with fixed acquisition windows, leading to a reduction of more than twofold in scanning times. In 13 patients we were able to use acquisition windows of almost 200 msec. Despite these long acquisition windows, image quality parameters were similar to those of conventional acquisition. This finding indicates that the advantages of a direct assessment of coronary motion may compensate for the potential disadvantages of using a longer acquisition window. The results suggest further both that the visual assessment of the coronary rest period was accurate and that the motion in the extended acquisition window was minimal.
We saw the most significant reductions in scanning times in patients with heart rates of 60 beats per minute or lower. Because the TR between shots in coronary MR angiography is defined by the R-R interval, a fixed acquisition window will lead to an inverse relationship between heart rate and total scanning time. Coronary MR angiography with fixed acquisition windows is therefore most time-consuming in individuals with low heart rates. In clinical practice, most patients referred for coronary MR angiography have low heart rates as a result of medication prescribed for known or suspected coronary heart disease (e.g., ß-blockers). With subject-specific acquisition windows scanning times can be reduced—particularly in these patients—because the longer diastolic rest period at low heart rates allows the use of longer acquisition windows [14]. Subject-specific acquisition windows may thus be of substantial clinical use.
In this study, we assessed coronary motion from a cine scan aligned perpendicular to the mid portion of the main coronary vessels. This approach has been suggested previously [11, 12] and is attractive because of its simplicity and because it requires little additional scanning or processing time. To improve the technique described in previous reports, we used a steady-state free precession pulse sequence to acquire the coronary motion scan. These techniques provide better image contrast in shorter scanning times than spoiled gradient-echo techniques [22, 23]. The short scanning time ensures that even in patients with poor breath-holding ability, the contamination of data is minimal because fewer cardiac cycles are required for signal averaging. In this study, the coronary motion scan was acquired in a breath-hold of 6-8 sec and yielded 30 phases per R-R interval.
The combination of high temporal resolution and good image quality provided by steady-state free precession pulse sequences ensures that the motion of the coronary arteries can be clearly visualized throughout the cardiac cycle and that the coronary rest period can be confidently determined. However, the assessment of coronary motion in a single imaging plane is limited because the 3D movement of the vessels—in particular, motion through the imaging plane and the motion at different levels of the vessel—cannot be completely assessed. To account for possible underestimation of coronary motion that could result from this incomplete assessment, we have limited the maximal acquisition window to 200 msec. As an alternative, a second imaging plane in an orthogonal orientation could be used to improve the assessment of 3D coronary motion.
A second feature of the presented technique was motion-adapted respiratory gating. In this study, the gating and acceptance windows were optimized to improve image quality rather than navigator efficiency, and the mean acceptance window was in fact comparable to the conventional technique. Consequently, the use of motion-adapted gating did not improve the navigator efficiency significantly. However, half of the subjects in the comparison group were healthy volunteers, and the navigator efficiencies in these subjects were generally high. More important, in the clinical population of the prospective group, the mean navigator efficiencies were also high and compare favorably with those described by another group of researchers [19]. This finding may in part be attributable to the use of motion-adapted gating, but may also reflect the shorter scanning time of our sequence compared with those used in previous studies. Using a shorter scanning time reduces the likelihood of drift of the mean end-expiratory position. Motion-adapted gating may also have contributed to the good image quality of our data, but we could not differentiate between the relative effects of motion-adapted gating and subject-specific acquisition windows on image quality. Furthermore, we used empirically defined window widths in motion-adapted gating and have not systematically assessed different acquisition windows. This assessment will have to be the subject of future work.
The diagnostic accuracy of the acquired data was similar to published data [21, 24]. A recent multicenter study in which a comparable pulse sequence was used but with fixed acquisition windows and heart rate—dependent trigger delays yielded an overall diagnostic accuracy of 72% [21]. Analysis was limited to the proximal segments of the coronary vessels. Similar to our study, that study found that sensitivity was highest for the left anterior descending and right coronary arteries and lowest for the circumflex artery. In another recent report, Bunce et al. [24] used a free-breathing technique with hybrid ordered phase-encoding and reported a sensitivity of 89%, 76%, and 50% for the detection of stenoses in the proximal left anterior descending, right coronary, and circumflex arteries, respectively. The accuracy yielded in our study was comparable to these results, but unlike the previous investigators, we analyzed the full length of the coronary arteries. Because scanning for our study was achieved in a substantially shorter time, our technique may have a role in future clinical practice.
The relatively low sensitivity of MR angiography to detect disease in the circumflex artery has been attributed to the posterior location and small caliber of this vessel, resulting in lower SNR [21]. Another reason may be that the imaging plane used in previous studies as well as in our study is not optimal for visualizing the circumflex artery. In our study, the imaging plane for the left coronary system was planned to include at least the proximal portion and mid portion of the vessel. In addition, the circumflex artery was often also seen on the second MR image aligned along the right coronary artery. However, the distal portion of the vessel may not have been imaged in some patients, potentially contributing to a lower sensitivity for detecting stenoses in these areas.
In conclusion, coronary MR angiography using subject-specific cardiac acquisition windows and motion-adapted respiratory gating is feasible in volunteers and patients. We found that this technique provides shorter scanning times while maintaining image quality. This technique yields high diagnostic accuracies for the detection of coronary artery disease that are comparable to those of previously reported techniques that are more time-consuming to perform.
Acknowledgments
We thank Matthias Stuber and Rene Botnar from Philips Medical Systems for
their support. This work was carried out at a British Heart Foundation Heart
Research Centre.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D H Kwon, N G Smedira, Z B Popovic, B W Lytle, R M Setser, M Thamilarasan, P Schoenhagen, S D Flamm, H M Lever, and M Y Desai Steep left ventricle to aortic root angle and hypertrophic obstructive cardiomyopathy: study of a novel association using three-dimensional multimodality imaging Heart, November 1, 2009; 95(21): 1784 - 1791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. D. Scott, J. Keegan, and D. N. Firmin Motion in Cardiovascular MR Imaging Radiology, February 1, 2009; 250(2): 331 - 351. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ustun, M. Desai, K. Z. Abd-Elmoniem, M. Schar, and M. Stuber Automated Identification of Minimal Myocardial Motion for Improved Image Quality on MR Angiography at 3 T Am. J. Roentgenol., March 1, 2007; 188(3): W283 - W290. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Sakuma, Y. Ichikawa, S. Chino, T. Hirano, K. Makino, and K. Takeda Detection of Coronary Artery Stenosis With Whole-Heart Coronary Magnetic Resonance Angiography J. Am. Coll. Cardiol., November 21, 2006; 48(10): 1946 - 1950. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Dewey, F. Teige, D. Schnapauff, M. Laule, A. C. Borges, K.-D. Wernecke, T. Schink, G. Baumann, W. Rutsch, P. Rogalla, et al. Noninvasive detection of coronary artery stenoses with multislice computed tomography or magnetic resonance imaging. Ann Intern Med, September 19, 2006; 145(6): 407 - 415. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Muthupillai, J. Smink, S. Hong, R. Ravindran, V. V. Lee, and S. D. Flamm SENSE or k-MAG to accelerate free breathing navigator-guided coronary MR angiography. Am. J. Roentgenol., June 1, 2006; 186(6): 1669 - 1675. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Sakuma, Y. Ichikawa, N. Suzawa, T. Hirano, K. Makino, N. Koyama, M. Van Cauteren, and K. Takeda Assessment of Coronary Arteries with Total Study Time of Less than 30 Minutes by Using Whole-Heart Coronary MR Angiography Radiology, October 1, 2005; 237(1): 316 - 321. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Fuster and R. J. Kim Frontiers in Cardiovascular Magnetic Resonance Circulation, July 5, 2005; 112(1): 135 - 144. [Full Text] [PDF]
|
|
|
|
|
|
P. R. Moreno and V. Fuster The year in atherothrombosis J. Am. Coll. Cardiol., December 7, 2004; 44(11): 2099 - 2110. [Full Text] [PDF]
|
|
|
|
|
|
S. Plein, J. P. Greenwood, J. P. Ridgway, G. Cranny, S. G. Ball, and M. U. Sivananthan Assessment of non-ST-segment elevation acute coronary syndromes with cardiac magnetic resonance imaging J. Am. Coll. Cardiol., December 7, 2004; 44(11): 2173 - 2181. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. G. Danias, A. Roussakis, and J. P.A. Ioannidis Diagnostic performance of coronary magnetic resonance angiography as compared against conventional x-ray angiography: A meta-analysis J. Am. Coll. Cardiol., November 2, 2004; 44(9): 1867 - 1876. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
