Duerinckx, André J. M.D., Ph.D
Coronary arteries are small tortuous vessels subjected to significant physiologic motions (respiration and cardiac contractions). With the development of a new group of ultrafast imaging sequences, magnetic resonance angiography (MRA) of the coronary arteries has recently become possible. Coronary MRA is noninvasive, does not require iodinated contrast agents or ionizing radiation, and therefore has the potential to become a very important cardiac screening tool.
Early coronary MRA pulse sequences relied on a combination of segmental acquisition of the data (in k-space) to minimize cardiac motion artifacts and the use of patient breath-holding to minimize respiratory motion artifacts (Table 1). More recently developed techniques have eliminated the need for breath-holding, thus opening the way for higher resolution coronary MRA with increased patient comfort. Three generations of techniques have already been described; all can be used with or without contrast agents.
Most of the clinical applications of coronary MRA have been validated by using first-generation techniques with breath-holding. It is important to know when MRA can add value to the evaluation of certain patients and to assess our level of confidence for obtaining a definitive diagnosis with MRA. Clinical applications of two-dimensional coronary MRA include coronary lesion detection, delineation of congenital coronary artery anomalies, characterization of previously known coronary lesions, assessment of coronary bypass graft patency, evaluation of vessel patency after coronary stent placement, coronary anatomy after heart transplantation, and quantification of coronary flow reserve.
However, the use of coronary MRA for blind prospective detection of coronary lesions is being evaluated, and improved coronary MRA techniques may become an integral part of the clinical evaluation and screening of patients with ischemic heart disease. It is still premature to judge the ultimate capability of coronary MRA in screening for coronary artery disease.
Imaging Coronary Artery Lesions by MRA
The detection of coronary artery lesions by MRA is probably the most widely discussed and investigated future application of coronary MRA. Interestingly enough, it is still the least mature and most experimental application of MRA for evaluation of the coronary arteries. Coronary lesion detection represents the “Holy Grail” of cardiac MR imaging. From the earliest description of coronary MRA it was hoped that coronary MRA would become a noninvasive screening tool for the detection of all (or most) coronary artery lesions. This is still our hope, at least for proximal lesions in the main coronary arteries and their larger side branches. The fact that after 6 years of preclinical trials of coronary MRA techniques their sensitivity and specificity for coronary lesion detection still does not compare with traditional contrast coronary angiography should not discourage us. There are many reasons why we have not yet seen “perfect” results, such as 99% sensitivity for detection of all coronary lesions. First, the MR technology and coronary MRA pulse sequences changed and improved so quickly over the last 5 years that few large preclinical trials were ever performed with any single technique. Second, second-generation coronary MRA techniques, as compared with first-generation techniques, required fewer technical skills but much longer acquisition times, few investigators were willing to take on the challenge. Third, many people wanted coronary MRA to perform like the established invasive gold standard, conventional contrast coronary angiography, even though this was never the case for other noninvasive techniques such as nuclear cardiology or exercise stress testing. There is a need to reassess unrealistic expectations about what coronary MRA should be able to do. The ability to detect 70% to 90% of all significant proximal coronary artery lesions is already a very worthwhile goal. It is not necessary for a noninvasive imaging technique for patients with ischemic heart disease to detect every single minor lesion in small coronary side branches. Also, with the advent of the third-generation coronary MRA techniques, improved postprocessing, greater interest from clinical users, the availability of new dedicated “cardiac” MR scanners, and new MR blood pool agents (the latter two indicating industry's interest in this issue), the use of coronary MRA in daily clinical practice will soon become reality. Coronary MRA for lesion detection is now available to clinical users and not only investigators at major academic centers or clinics. Coronary MRA will have an important role in the workup of patients with ischemic heart disease in the very near future.
CORONARY ARTERY LESION DETECTION BY MRI: HOW DOES IT DIFFER FROM OTHER VASCULAR TERRITORIES?
The physiology of blood flow in coronary arteries is very different from the physiology of blood flow in large vessels such as the carotid arteries and the abdominal aorta. In the coronary arteries, flow velocities are lower than in other large vessels; there is less or no turbulent flow and there is potential for collateral blood flow. Normal coronary arteries and veins appear as high-signal linear structures on coronary MRAs. Distal to a total occlusion, vessels with no collateral flow have decreased or background (fat-suppressed) signal intensity. However, occasionally a vessel shadow is visible, possibly corresponding to the wall of the occluded vessel, and this could be misconstrued as representing reduced flow beyond a subtotal occlusion. Reverse flow distal to a total occlusion cannot be distinguished from forward flow past a subtotal occlusion without determining the direction of flow.
Figure 1 shows a long hemodynamically significant proximal right coronary artery lesion, as seen on both the conventional angiogram and on the coronary MRA.
Discrete coronary lesions, both total and subtotal occlusions, may appear as either areas of signal void, decreased flow signal, or vessel wall irregularity. It is thus very important to be aware of these important differences between coronary MRA and MRA of large vessels elsewhere in the body. Also, because of collateral flow, a total occlusion may have slow-moving fluid present in the vessel distal to the occlusion and show a normal vessel signal.
Diffusely diseased coronary arteries cannot be distinguished from small-caliber normal vessels by coronary MRA. A dominant left coronary system with a small-caliber RCA can simulate diffuse disease in the RCA, and vice versa. Women with small-caliber coronary vessels are also more difficult to study with coronary MRA.
PIONEERING WORK WITH FIRST-GENERATION CORONARY MR ANGIOGRAPHY: THE 1993–94 PERIOD.
A historical review of the initial reports on the use of coronary MRA for coronary artery lesion detection can be very educational. It illustrates a typical cycle: the initial enthusiasm of the first very positive and promising preclinical reports and then the more negative follow-up reports when other investigators try to duplicate the initial work. This seems to be a very typical scenario in the early stages of new imaging technology implementations in medicine. Similar scenarios can be found in other areas of thoracic imaging, such as the use of computed tomography (CT) for pulmonary embolism detection and the use of electron beam CT (EB-CT) for three-dimensional (3D) coronary artery visualization.
WORK BY MANNING ET AL., REPORTED IN 1993
Initial clinical testing of the first-generation coronary MRA technique was performed at the Beth Israel Hospital in Boston and reported in the March 1993 issue of the New England Journal of Medicine. In this historical article, Manning et al. reported for the first time on the ability to detect coronary occlusion and stenosis by using MRA in a series of 39 patients referred for elective coronary angiography. This paper brought hope to the field of cardiology and medicine that a noninvasive coronary imaging technique soon would become available. It created great interest by industry and MRA vendors for this new and extremely promising application of cardiac MRA. More than anything else, it helped increase the research efforts in new uses of cardiac MRA techniques. The national and international press got involved, and this report went beyond just the medical media.
In this first clinical study, coronary MRA, as compared with conventional angiography, had 90% sensitivity and 92% specificity for correctly identifying individual vessels with 50% or greater angiographic stenoses. The corresponding positive and negative predictive values were 0.85 and 0.95, respectively. The overall sensitivity and specificity for correctly classifying individual patients as having or not having serious coronary disease were 97% and 70%, respectively, with corresponding positive and negative predictive value of 0.90 and 0.88, respectively. Lesion detection was highest in the left main coronary artery. Coronary MRA showed the lowest sensitivity for lesion detection in left circumflex artery (71%) and the lowest specificity in the right coronary (78%). In a follow-up study of 72 patients with 81 significant lesions in a total of 271 visualized coronary arteries (and 39% of patient without significant lesions), the sensitivity for an individual vessel with stenosis remained 90%.
WORK BY DUERINCKX ET AL., REPORTED IN 1994
Subsequent attempts by other investigators to reproduce these results based on blinded studies have not been as successful. Duerinckx et al. were the first to report in December 1993 that the 90% sensitivity for lesion detection by using first-generation coronary MRA was optimistic and could not be reproduced in a double-blinded prospective study. Their study was published in the December 1994 issue of Radiology. The publication of the article by Duerinckx et al. confirmed what several other investigators and industry people were starting to discover and had to admit; that although very promising, and in some patients spectacular, the images routinely produced with first-generation coronary MRA were not consistent enough to reproduce the results first reported by Manning et al. This paper has historical importance, because it set a more realistic tone for what first-generation coronary MRA can accomplish.
As can be expected, great controversy erupted over these discrepancies in the performance of first-generation coronary MRA techniques. The original correspondence between Manning and Duerinckx was published in the June 1995 issue of Radiology. Duerinckx et al. were not alone in disagreeing with Manning et al. Indeed, from 1994 to 1995, the reported sensitivity for detection of significant (≥50%) lesions was highly variable: 36% in an initial study by Post et al. of 14 patients with 14 significant lesions; 33% to 75% in a follow-up study by Post et al. of 35 patients with 35 stenoses; 63% average sensitivity (range, 0%–75%) in a study by Duerinckx et al. of 20 patients with 27 proximal lesions; 65% in an initial study by Pennell et al. of 17 patients with 23 lesions; 88% in a follow-up study by Pennell et al., using a nonblinded analysis of 31 patients with 41 lesions; 83% to 100% in a study by Nitatori et al. in 50 patients; and 56% in a study by Mohiaddin et al. in 16 heart transplant patients with nine lesions.
WHY THE DISCREPANCY IN RESULTS BETWEEN MANNING AND DUERINCKX?
The failure of other investigators to achieve a 90% sensitivity and 92% specificity for coronary lesion detection was not caused by differences in techniques relating to relative poor temporal resolution, shorter echo times, or worse patient cooperation, unlike some statements to the contrary. In fact, both Manning and Duerinckx used the same MR pulse sequences designed and provided by the same company (Siemens Medical Systems, Inc, NJ). Both investigators took great care in patient preparation and prescan breath-hold practice. In retrospect, the discrepancy in results can most likely be explained by interobserver differences in interpretation of a similar set of MR images. The most important difference between the Manning and Duerinckx studies may well be the “human factor,” that is, the level of confidence chosen by each investigator to call an anomaly on an MR image of a lesion.
An important practical problem with first-generation 2D coronary MRA is the potential for overreading or underreading of coronary lesions. Because only one slice is viewed at a time, and spatial registration between slices in never guaranteed, it can be difficult to distinguish between an apparent lesion caused by an out-of-plane vessel course and a true lesion. This is compounded by the limited spatial and temporal resolution (typically 1 mm in-plane spatial and 80 to 100 ms temporal resolution). It is important for users of this technique to understand these interpretative difficulties. With second-and third-generation coronary MRA techniques, many of these problems disappear. One has to adhere to strict radiologic interpretation criteria or else one would have false-positive readings in almost 100% of normal patients.
Several explanations are possible for misinterpretation of a coronary lesion on 2D coronary MRA. These include spatial and temporal resolution, timing of the acquisition (within the cardiac cycle), irregular breathing and misregistration between breath-holds, partial volume averaging of tortuous vessels and overlapping adjacent anatomic structures, reverse flow past an area of stenosis caused by collateral flow, and the unknown signal and contrast characteristics of vessel wall and plaque on coronary MRA sequences.
LESION DETECTION USING FIRST-, SECOND-, AND THIRD-GENERATION CORONARY MRA: THE 1995–2000 PERIOD
It was originally hoped that coronary MRA could become a noninvasive screening tool for the detection of coronary artery lesions. The results with first-generation coronary MRA in the 1993–1994 period were very promising but somewhat variable. The first-generation 2-D breath-hold MR imaging technique was tested by several investigators on a total of approximately 224 patients and was able to detect significant coronary artery lesions (greater than 50% diameter narrowing) with a sensitivity ranging from 33% to 100% (see Table 2, top). The reported sensitivity also varied depending on the vessel examined. Because all the early studies were done with a surface coil positioned on the anterior chest wall, the more posteriorly located circumflex coronary artery was the more difficult vessel to visualize. More recent studies have used phased-array thoracic coils, with better visualization of the circumflex coronary artery. The early 3D techniques did not make a significant difference. The initial clinical results were very disappointing, with a sensitivity of 0% to 38%, as shown in Table 2. Then in 1994 to 1995, second-generation 3D coronary MRA trials were performed, using navigator-pulse feedback. With the original implementation of these techniques, a few investigators were able to show a sensitivity of 65% to 87% for lesion detection, as shown in Table 2. Many modifications and improvements for this second-generation coronary MRA approach have been proposed and implemented. These improved second-generation coronary MRA techniques show great promise. However, even these improved navigator techniques still require long acquisition times, with all the inherent risks of erratic breathing or motion. Thus, there is interest in developing newer techniques, referred to as the third-generation coronary MRA, that allow multiple slice acquisition in a single breath-hold. Only very preliminary clinical results are available for the third-generation coronary MRA techniques.
The results of all the preclinical trials of coronary MRA for coronary artery lesion detection are summarized in Table 2. After 7 years of preclinical trials using many different coronary MRA techniques, no technique has yet emerged as superior and can provide a sensitivity and specificity for coronary lesion detection that compares with traditional contrast coronary angiography.
The performance of first-generation coronary MRA is still relatively time-consuming and requires the presence of a trained cardiac imager. With a moderate amount of training and practice, most radiologists and cardiologists should be able to perform these studies. Several excellent review articles have provided guidance for the time-efficient performance of these examinations.
UNRESOLVED ISSUE: WILL THE SENSITIVITY FOR LESION DETECTION BE GOOD ENOUGH FOR CLINICAL PRACTICE?
The most important limitation to the clinical implementation of coronary MRA is the great variation in the appearance of significant (≥50%) coronary lesions and the large number of image artifacts that can be misinterpreted as representing lesions. Although this has been best documented for first-generation 2D coronary MRA techniques, preliminary evidence in normals seems to indicate that the problem may be worse with the initial implementations of the second-generation techniques by use of navigator pulses. In a study of normal volunteers imaged with a 3D coronary MRA technique and retrospectively gated navigator feedback, Stehling found that approximately 50% of the examinations were of unacceptable quality, making interpretation of the images difficult or impossible. Nevertheless, several investigators, using this early implementation of the navigator technique (and using the same MR scanner and software from the same vendor), reported much better results in patients.
It is obvious that the coronary MRA techniques available today have limitations when it comes to lesion detection. Clinical studies are needed to determine whether these techniques, with their limitations, would still have an impact on selecting clinical pathways for patient care and intervention. The ultimate clinical utility of a coronary MRA screening technique should be compared with other noninvasive alternatives such as coronary calcium screening by using EB-CT, EB-CT visualization of the coronary vessels, echocardiography, nuclear medicine studies, and even treadmill stress testing. None of the existing noninvasive techniques have a 100% sensitivity. Each technique has its own limitations, and the ultimate relative value of coronary MRA with respect to these techniques remains to be determined in future studies.
Another problem is that the expectations for coronary MRA were somewhat unrealistic to begin with. MRA in peripheral vessels and even in larger vessels such as the carotid arteries has well-known limitations and image artifacts. Nevertheless, MRA of the carotid arteries is well accepted and clinically used. Expecting that MR techniques applied to the small vessels on the heart that are subject to cardiac and respiratory motion would be able to perform better than in relatively stationary vessels such as the peripheral vasculature or the carotid arteries was perhaps unrealistic. In addition, one expects a very high sensitivity for lesion detection in the coronary territory, whereas a lesser sensitivity in the peripheral vasculature would be acceptable.
Lesion appearance and the presence of collateral flow also complicate image interpretation, as described earlier. Discrete coronary lesions (total and subtotal occlusions) may appear as areas of signal void, decreased flow signal, or vessel wall irregularity. A subacute thrombus can mimic flowing blood. This is sometimes different from the more traditional appearance of lesions in larger vessels, as seen on MRA. Also, because of collateral flow, a total occlusion may have slow-moving fluid present in a vessel distal to the occlusion and show a normal vessel flow signal. A correlate of this is the fact that the significance of a coronary lesion cannot be quantitated on coronary MRA. The typical spatial resolution is 1.5 × 1 × 4 mm for coronary vessels measuring 2 to 4 mm in diameter. The problems with measuring luminal diameter narrowing on coronary MRA have been clearly outlined in two recent studies. Pennell et al. classified the degree of MR signal loss in the region of a known coronary artery lesion as severe or partial and in addition noted when an irregular vessel outline was present. The main percentage of stenosis was significantly different for each type of signal loss, suggesting that visual assessment of signal loss can be a qualitative guide to stenotic severity. They also measured directly on the viewing screen the relation between the proportion of magnetic resonance signal loss at the site of stenosis and compared this with the diameter of the stenosis as assessed by conventional angiography. They found that there was a significant relationship between stenosis severity and magnetic resonance signal loss, although with a considerable scatter in individual measurements. They were never able to measure “luminal diameter reduction.” Shimamoto et al. provided a detailed analysis of the use of spatial profile curves to better measure the diameter of coronary arteries on MRA. This problem in quantitating the severity of stenosis was discussed early on by Duerinckx and Urman and Pennell et al.. Duerinckx et al., have suggested that the problem in quantitating the severity of stenosis can be circumvented by using “confidence levels” for the presence of significant disease and by using receiver operating curve (ROC) analysis to interpret image data.
IS THERE A GOLD STANDARD TO EVALUATE NEW TECHNOLOGIES?
A detailed protocol for the interpretation and reading of coronary MRA has been described by Duerinckx et al.. This protocol includes, for each segment of vessel (for the visualized portion of the coronary anatomy), the following: (1) length and width of vessel; (2) grading of image quality; and (3) likelihood of presence of significant lesion or any (mild or moderate) disease (i.e., confidence levels for ROC analysis). This protocol, together with an estimate of intraobserver and interobserver variability in interpretation, could form the basis for multicenter evaluations of any of these new coronary MRA techniques. However, this presumes the existence of a “gold standard” to assess the MRA findings. Does such a standard really exist? We discuss the problems with the generally accepted standard of iodinated contrast–based angiography.
The presence or absence of coronary artery lesions is traditionally evaluated with coronary angiography. However, great interobserver variability in the interpretation of coronary angiograms has been well documented. Moreover, the assessment of significance of coronary stenosis based on interpretation of the coronary angiogram only is questionable. We briefly address both of these issues.
One older study by Zir et al. from 1976 showed that in only 65% of coronary angiograms did four observers agree about the significance of a stenosis (defined as greater than 50% of luminal narrowing) in the proximal or mid-left anterior descending coronary artery. In this study, four experienced coronary angiographers (two radiologists and two cardiologists) independently assessed the location and degree of coronary artery stenosis in 20 coronary angiograms and showed a substantial degree of variability in the interpretation of coronary angiography. This was even more surprising in the sense that all four angiographers were from the same institutions and had frequently participated as a group in the interpretation of coronary angiograms. The 20 angiograms in their study included the typical spectrum of coronary angiographic findings with very few normal coronary arteries or cases with single-vessel disease.
The concept of what is a “gold” standard is constantly being reevaluated and questioned. The problems of interobserver variability are not unique to the interpretation of coronary angiograms but have also been reported in the interpretation of other vascular imaging studies. The reasons for the observed variability in coronary angiogram interpretation in the older studies such as that of Zir et al. can be related to several factors: a limited number of projections; partial visualization of the vessel segments with lesions; the quality of the coronary angiograms; and differences in perception. Some of these problems no longer exist with the newer cardiac cathetherization laboratories. However, many important problems remain. One of the most important factors contributing to the variability in interpretation was disagreement as to which portion of the vessel to use as a normal reference when calculating the percentage of coronary stenosis. Frequently large segments of the vessels are abnormal, making such determinations difficult. There may be vessel dilation, lesions may be multiple, or lesions may be smooth or tapering. Several approaches may reduce the error attributable to interobserver variability. One would be to combine independent readings. Another would be to use a consensus opinion based on joint reading of the coronary angiogram.
These difficulties in determining the level of stenosis of a coronary lesion could partially explain the problems with the poor correlation between coronary stenosis and functional significance of a coronary lesion. The variations in cross-sectional shape of the lesions also account for some of the discrepancies. This well-known phenomenon can be solved by obtaining multiple projections of single lesions. Unfortunately, in the case of a left main coronary artery or proximal left anterior descending coronary artery, often many overlapping vessels are difficult to separate even with several views.
One has to keep these discrepancies in coronary angiogram interpretation in mind when designing and interpreting studies comparing new technologies such as coronary MRA or EB-CT angiography with conventional coronary angiography. If one presumes that the conventional contrast-based coronary angiography is the gold standard, and does not take into account the great variability in interpretation, one may draw erroneous conclusions about the performance of either MR-or EB-CT–based coronary angiography. Thus, the results of prospective blind studies comparing both technologies should be interpreted with great caution. Each study should include a retrospective analysis of the data, because in some instances coronary MRA or EB-CT angiography may provide additional information that is difficult to gain from a traditional conventional coronary angiogram that provides a limited number of views of multiple superimposed vessels.
CORONARY LESION DETECTION BY MR IN THE FUTURE
Coronary MRA technique development has been proceeding at a very fast pace, indeed so fast that most techniques cannnot be evaluated by more than one or two investigators before everybody moves on to the next technique. Clinical users should not be disappointed by this or view this as an indication that these techniques are not working. First-generation coronary MRA techniques are more than adequate for a large number of specific clinical applications, and are even capable of detecting significant coronary lesions with a sensitivity of 65% to 80% in many centers. The early implementation of the second-generation coronary MRA techniques (using retrospective respiratory gating) required long acquisition times, which explains their unpredictable behavior. When images were good they were very good, but equally often they were very bad and uninterpretable. The improved second-generation coronary MRA techniques using navigator pulse feedback and adaptive prospective (or retrospective or interactive) correction of slice position or signal averaging dramatically improve the reliability and image quality and thus increase the sensitivity of the technique. The newer third-generation techniques appear very promising and easy to use. The advent of blood pool agents also will impact our capability to detect lesions. Hopefully, we will soon have a useful array of complementary noninvasive MRA approaches to the evaluation of coronary lesions that will be easy to use and will provide reproducible results.
Borelli C, Berthezene Y, Olteanu B, et al. Subacute coronary artery thrombus: MRI findings. J Comput Assist Tomogr 1997; 21:962–4. Bunce NH, Pennell DJ. Coronary MRA: a clinical experience in Europe. J Magn Reson Imaging 1999; 10:721–7. Caputo GC. Coronary arteries: potential for MR imaging [Editorial]. Radiology 1991; 181: (3):629–30. Caputo RC. Coronary MR angiography: a clinical perspective [Editorial]. Radiology 1994; 193:596–8. Chernoff DM, Ritchie CJ, Higgins CB. Electron-beam CT coronary angiography: imaging characteristics in normal coronary arteries [Abstract]. Radiology 1996; 201:274. Danias PG, Edelman RR, Manning WJ. Coronary MR angiography. Cardiol Clin 1998; 16:207–25.Danias PG, McConnell MV, Khasgiwala VC, et al. Adapative navigator correction of slice position in real-time for MR coronary angiography [Abstract]. In: Book of Abstracts of the 4th Meeting of the International Society of Magnetic Resonance Medicine, Vol 1. New York, NY, April 27-May 3, 1996:177. Danias PG, Stuber M, Edelman RR, et al. Coronary MRA: a clinical experience in the United States. J Magn Reson Imaging 1999; 10:713–20. De Cobelli F, Cappio S, Vanzulli A, et al. MRI assessment of coronary stents. Rays 1999; 24:140–8. De Man F, De Scheerder I, Herregods MC, et al. Role of intravascular ultrasound in coronary artery disease: a new gold standard? An overview. Acta Cardiol 1994; 49:223–31. Dinsmore RE. Noninvasive coronary arteriography: here at last? [comment]. Circulation 1995; 91:1607–8. Doriot PA, Dorsaz PA, Dorsaz L, et al. Towards a gold standard for quantitative coronary arteriography. Phys Med Biol 1997; 42:2449–62. Duerinckx AJ. Coronary MR angiography (invited article). Radiol Clin North Am 1999; 37:273–318. Duerinckx AJ. Coronary MR angiography [Response-to-Letter-to-the-Editor]. Radiology 1995; 195:876. Duerinckx AJ. Coronary MR angiography. MRI Clin North Am 1996; 4:361–418. Duerinckx AJ. Coronary MR angiography. Radiol Clin North Am 1999; 37:273–318. Duerinckx AJ. Review: MR angiography of the coronary arteries. Top Magn Reson Imaging 1995; 7:267–85. Duerinckx AJ, Atkinson D, Hurwitz R, et al. Coronary MR angiography after coronary stent placement (case report). Am J Roentgenol 1995; 165:662–4. Duerinckx AJ, Atkinson D, Hurwitz R. Assessment of coronary artery patency after stent placement using magnetic resonance angiography. JMRI. J Magn Reson Imaging 1998; 8:896–902. Duerinckx AJ, Atkinson DP, Mintorovitch J, et al. Two-dimensional coronary MR angiography: Limitations and artifacts. Eur Radiol 1996; 6:312–25. Duerinckx AJ, Urman M. Two-dimensional coronary MR angiography: analysis of initial clinical results. Radiology 1994; 193:731–8. Duerinckx AJ, Urman M, Sinha U, et al. Evaluation of gadolinium-enhanced MR coronary angiography [Abstract]. Radiology 1993; 189:278. Duerinckx AJ, Urman MK. Hemodynamically significant coronary artery lesions missed with MR coronary angiography [Abstract]. JMRI 1994; 4:80.Duerinckx AJ, Urman MK, Atkinson DJ, et al. Limitations of MR coronary angiography [Abstract]. In: Printed Program of the First Meeting of the Society of Magnetic Resonance (SMR). Dallas, Texas, March 5–9: 1994; JMRI 1994;4:81. Feldman T, Laskey WK. Alchemy in the cath lab: creating a gold standard [editorial; comment]. Cathet Cardiovasc Diagn 1998; 44:14–5. Galbraith JE, Murphy ML, Desoyza N. Coronary angiogram interpretation: interobserver variability. JAMA 1981; 240:2053–9. Goldsmith MF. Realizing potential of MR coronary angiography may ease patients' test load and diagnosis costs (Editorial). JAMA 1994; 271:256. Hartnell GG, Hughes LA, Finn JP, et al. Magnetic resonance angiography of the central chest veins: a new gold standard? Chest 1995; 107:1053–7.Hastings MM, Andes S, Hsu A. The promises of autopsy: still the “gold standard” of quality? [Abstract]. Abstr Book Assoc Health Serv Res 1997;14.Hu BS, Meyer CH, Macovski A, et al. Multi-slice spiral magnetic resonance coronary angiography [Abstract[. In: Book of Abstracts of the 4th Meeting of the International Society of Magn Reson Med (ISMRM) Vol 1. New York, NY, April 27–May 3, 1996:671. Huber A, Nikolaou K, Gonschior P, et al. Navigator echo-based respiratory gating for three-dimensional MR coronary angiography: results from healthy volunteers and patients with proximal coronary artery stenoses. AJR 1999; 173:95–101. Johansson LO, Nolan NM, Taniuchi M, et al. High resolution magnetic resonance coronary angiography of the entire heart using a new blood-pool agent, NC100150 injection: comparison with invasive X-Ray angiography in pigs. Journal of Cardiovascular Magnetic Resonance 1999; 2:139–44. Karkow WS, Cranley JJ. Variations in interpretation of arterial stenosis. J Cardiovasc Surg 1989; 30:826–32. Kessler W, Achenbach S, Moshage W, et al. Usefulness of respiratory gated magnetic resonance coronary angiography in assessing narrowings >50% in diameter in native coronary arteries and in aortocoronary bypass conduits. Am J Cardiol 1997; 80:989–93. Kotsakis A, Tan KH, Jackson G. Is MRI a safe procedure in patients with coronary stents in situ? (editorial). Int J Clin Pract 1997; 51:349.Kramer CM, Rogers WJ, Mankad SV, et al. Short and long-term safety of magnetic resonance imaging soon after stenting for acute myocardial infarction [Abstract]. Presented at the Society for Cardiovascular Magnetic Resonance, Second Meeting, January 22–24, 1999, Atlanta, GA, 1999:99. Lethimonnier F, Furber A, Morel O, et al. Three-dimensional coronary artery MR imaging using prospective real-time respiratory navigator and linear phase shift processing: comparison with conventional coronary angiography. Magn Reson Imaging 1999; 17:1111–20. Manning W, Edelman R. Magnetic resonance coronary angiography. Magn Reson Q 1993; 9:131–51. Manning WJ, Edelman RR. Coronary MR angiography [Letter to the Editor]. Radiology 1995; 195:875. Manning WJ, Li W, Edelman RR. A preliminary report comparing magnetic resonance coronary angiography with conventional angiography. N Engl J Med 1993; 328:828–32.Manning WJ, Li W, Wielopolski P, et al. Magnetic resonance coronary angiography: comparison with contrast angiography [Abstract]. In: Book of Abstracts and Proceedings of the 2nd Meeting of the the Society of Magnetic Resonance (SMR), Vol 1. San Francisco, California, August 6–12: 1994:368. Marcus M, Skorton D, Johnson M, et al. Visual estimates of percent diameter coronary stenosis: “a battered gold standard.” JACC 1988; 11:882–5.Mohiaddin RH, Bogren HG, Firmin DN, et al. MR angiography of the coronary arteries in heart transplants [Abstract]. In: Book of Abstracts of VII International Workshop on Magnetic Resonance Angiography. Matsuyama, Japan, Oct 12–14, 1995.Mohiaddin RH, Bogren HG, Lazim F, et al. Magnetic resonance coronary angiography in heart transplant recipients [Abstract]. In: Book of Abstracts of the 4th Meeting of the International Society of Magn Reson Med, Vol 2. New York, NY, April 27-May 3, 1996:671. Mohiaddin RH, Bogren HG, Lazim F, et al. Magnetic resonance coronary angiography in heart transplant recipients. Coron Artery Dis 1996; 7:591–7. Moshage WEL, Achenbach S, Seese B, et al. Coronary artery stenosis: three-dimensional imaging with electrocardiographically triggered, contrast agent-enhanced, electron-beam CT. Radiology 1995; 196:707–14. Müller MF, Fleisch M, Kroeker R. Coronary arteries: three-dimensional MR imaging with fat saturation and navigator echo. Radiology 1996; 201:274. Müller MF, Fleisch M, Kroeker R, Chatterjee T, Meier B, Vock P. Proximal coronary artery stenosis: three-dimensional MRI with fat saturation and navigator echo. J Magn Reson Imaging 1997; 7:644–51. Nissen SE, Gurley JC. Assessment of the functional significance of coronary stenoses. Is digital angiography the answer? [Editorial Comment]. Circulation 1990; 81:1431–5.Nitatori T, Hachiya J, Korenaga T, et al. Clinical application of coronary MR angiography: studies on shortening of time for examination [Abstract]. In: Book of Abstracts of the 3rd Meeting of the Society of Magnetic Resonance and the 12th Annual Scientific Meeting of the European Society for Magn Reson Med and Biology. Nice, France, August 20–25, 1995. Nitatori T, Hanaoka H, Yoshino A, et al. Clinical application of magnetic resonance angiography for coronary arteries: correlation with conventional angiography and evaluation of imaging time. Nippon Acta Radiol 1995; 55:670–6. Nitatori T, Yoshino H, Yokoyama K, et al. Coronary MR angiography: a clinical experience in Japan. J Magn Reson Imaging 1999; 10:709–12. Oshinski JN, Hofland L, Mukundan S, et al. Two-dimensional MR coronary angiography without breath holding. Radiology 1996; 201:737–43. Patel MR, Klufas RA, Kim D, et al. MR angiography of the carotid bifurcation: artifacts and limitations. AJR 1994; 162:1431–7. Pennell DJ, Bogren HG, Keegan J, et al. Assessment of coronary artery stenosis by magnetic resonance imaging. Heart 1996; 75:127–33.Pennell DJ, Bogren HG, Keegan J, et al. Coronary artery stenosis: assessment by magnetic resonance imaging [Abstract]. In: Book of Abstracts of the 11th Annual Scientific Meeting of the European Society for Magn Reson Med and Biology. Vienna, Austria, April 20–24, 1994;374.Pennell DJ, Bogren HG, Keegan J, et al. Detection, localization and assessment of coronary artery stenosis by magnetic resonance imaging [Abstract]. In: Book of Abstracts and Proceedings of the 2nd Meeting of the the Society of Magnetic Resonance, Vol 1. San Francisco, California, August 6–12, 1994:369. Pennell DJ, Keegan J, Firmin DN, et al. Magnetic resonance imaging of coronary arteries: technique and preliminary results. Br Heart J 1993; 70:315–26. Post JC, vanRossum AC, Bronzwaer JGF, et al. Magnetic resonance angiography of anomalous coronary arteries. A new gold standard for delineating the proximal course? Circulation 1995; 92:3163–71. Post JC, van Rossum AC, Hofman MB, et al. Clinical utility of two-dimensional magnetic resonance angiography in detecting coronary artery disease. Eur Heart J 1997; 18:426–33. Post JC, vanRossum AC, Hofman MB, et al. Three-dimensional repiratory-gated MR angiography of coronary arteries: comparison with conventional coronary angiography. American Journal of Roentgenology 1996; 166:1399–1404.Post JC, vanRossum AC, Hofman MBM, et al. Clinical utility of two dimensional breathhold MR angiography in coronary artery disease [Abstract]. In: Book of Abstracts of the 3rd Meeting of the Society of Magnetic Resonance and the 12th Annual Scientific Meeting of the European Society for Magn Reson Med and Biology. Nice, France, August 20–25, 1995.Post JC, vanRossum AC, Hofman MBM, et al. Current limitations of two dimensional breathhold MR angiography in coronary artery disease [Abstract]. In: Book of Abstracts and Proceedings of the 2nd Meeting of the the Society of Magnetic Resonance, Vol 1. San Francisco, California, August 7–12: 1994:508.Post JC, vanRossum AC, Hofman MBM, et al. Respiratory-gated three dimensional MR angiography of coronary arteries and comparison with X-Ray contrast angiography [Abstract]. In: Book of Abstracts and Proceedings of the 2nd Meeting of the the Society of Magnetic Resonance, Vol 1. San Francisco, California, August 7–12, 1994:509.Regenfus M, Ropers D, Achenbach S, et al. Gadolinium enhanced 3D breathhold magntic resonance angiography for detection of coronary artery stenosis in oblique projection angiograms [Abstract]. In: Book of Abstracts and Proceedings of the 7 th Meeting of the International Society of Magnetic Resonance Medicine, Vol 2. Philadelphia, PA, May 22–28, 1999:1262. Rogers LF. The heart of the matter: noninvasive coronary artery imaging [Editorial]. AJR 1998; 170:841.Rogers WJ, Kramer CM, Simonetti OP, Reichek N. Quantification of human coronary stenoses by magnetic resonance angiography [Abstract]. In: Printed program of the second meeting of the Society of Magnetic Resonance (SMR), Vol 1. San Francisco, California, August 6–12: 1994:370. Rushing L, Joste N. The surgical pathology report: standardizing the “gold standard” [editorial]. J Surg Oncol 1997; 65:1–2. Sacerdoti D, Gaiani S, Buonamico P, et al. Interobserver and interequipment variability of hepatic, splenic, and renal arterial Doppler resistance indices in normal subjects and patients with cirrhosis. J Hepatol 1997; 27:986–92. Sakuma H, Kawada N, Takeda K, Higgins CB. MR measurement of coronary blood flow. J Magn Reson Imaging 1999; 10:728–33. Sandstede JJ, Pabst T, Beer M, et al. Three-dimensional MR coronary angiography using the navigator technique compared with conventional coronary angiography. AJR 1999; 172:135–9. Shimamota R, Suzuki J-i, Nishikawa J-i, et al. Measuring the diameter of coronary arteries on MR angiograms using spatial profile curves. AJR 1998; 170:889–93. Singh K, Bonaa KH, Solberg S, et al. Intra-and interobserver variability in ultrasound measurements of abdominal aortic diameter: the Tromso Study. Eur J Vasc Endovasc Surg 1998; 15:497–504.Stehling M. Coronary arteries: experience with navigator echoes. [Abstract]. In: VIII International Workshop on Magnetic Resonance Angiography. Rome, Italy, October 16–19, 1996. Taylor AM, Panting JR, Keegan J, et al. Safety and preliminary findings with the intravascular contrast agent NC100150 injection for MR coronary angiography. J Magn Reson Imaging 1999; 9:220–7. van Beek, EJ, Bakker AJ, Reekers JA. Pulmonary embolism: interobserver agreement in the interpretation of conventional angiographic and DSA images in patients with nondiagnostic lung scan results. Radiology 1996; 198:721–4. van Geuns RJ, Wielopolski PA, de Bruin HG, et al. Magnetic resonance imaging of the coronary arteries: clinical results from three dimensional evaluation of a respiratory gated technique. Heart 1999; 82:515–9.vanGeuns RJM, Wielopolski PA, Rensing B, et al. Clinical evaluation of breath-hold MR coroanry angiography using targeted volumes [Abstract]. In: Proceedings of the Sixth Scientific Meeting of the International Society for Magnetic Resonance in Medicine (ISMRM), Vol 1. Sydney, Australia,: April 18–24, 1998:324. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications. A statement for health professionals from the American Heart Association. Writing Group. Circulation 1996; 94:1175–92. White CS, Laskey WK, Stafford JL, et al. Coronary MRA: use in assessing anomalies of coronary artery origin. J Comput Assist Tomogr 1999; 23:203–7. Wielopolski P, vanGeuns RJ, deFeyter PJ, et al. Breath-hold coronary MR angiography with volume targeted imaging. Radiology 1998; 209:209–220. Wielopolski PA, van Geuns RJ, de Feyter PJ, et al. Coronary arteries. Eur Radiol 1998; 8:873–85. Wielopolski PA, van Geuns RJ, de Feyter PJ, et al. Coronary arteries. Eur Radiol 2000; 10:12–35.Woodard PK, Li D, Dhawale P, et al. Identification of coronary artery stenosis with 3D retrospective respiratory gating [Abstract]. In: Ninety-sixth Annual Meeting of the American Roentgen Ray Society, Supplement to AJR. San Diego, May 5–10, 1996:166. Woodard PK, Li D, Haacke EM, et al. Detection of coronary stenosis on source and projection images using three-dimensional MR angiography with retrospective respiratory gating: preliminary experience. AJR 1998; 170:883–8. Woodard PK, Li D, Zheng J, et al. Coronary MR angiography. Magn Reson Imaging Clin North Am 1999; 7:365–78. Yoshino H, Nitatori T, Kachi E, et al. Directed proximal magnetic resonance coronary angiography compared with conventional contrast coronary angiography. Am J Cardiol 1997; 80:514–8. Zir LM, Miller SW, Dinsomore RE, et al. Interobserver variability in coronary angiography. Circulation 1976; 53:627–623.
© 2001 Lippincott Williams & Wilkins, Inc.