After many years of being considered a promising new technology, cardiac magnetic resonance imaging (CMR) is likely to assume a major clinical role in the near future. While CMR already has established itself as clinically useful for high-quality anatomic imaging of the heart in such selected conditions as congenital heart disease and tumors involving the heart, these conditions represent a minority of patients with cardiac disease. MRI is used generally only as a second-line method of imaging the heart, such as, for example, when an echocardiogram is ambiguous or technically difficult to obtain.

What has changed recently is the development of new imaging hardware and methods that promise to provide new cardiac functional and perfusion information, equivalent or superior to that available from conventional methods. This should permit CMR to play a more important role in the evaluation and management of patients with ischemic heart disease, a major public health problem. There additionally is much progress in and active ongoing research on the development of MR coronary angiography and even MR-guided interventional methods. With the prospect of performing a much larger number of patient examinations, along with eliminating other examinations currently being performed, the stakes involved in the question of who will perform CMR have been raised. The major manufacturers of MR equipment are developing products specifically directed at a burgeoning CMR equipment market, in whatever departments it may be based, and both radiology and cardiology practices are deciding how aggressively to pursue this new clinical imaging specialty.


Cardiac applications of MR share the advantages of MR in other parts of the body, such as the ability to obtain high-resolution tomographic imaging in arbitrary planes of orientation with good intrinsic tissue contrast, without the use of potentially harmful ionizing radiation. However, the heart has an obvious imaging problem in that it moves, with both the heartbeat and respiration, on time scales that are short in comparison with conventional imaging times. This movement can result in image motion blurring and artifacts. Synchronizing image data acquisition with the heartbeat permits acquiring data at consistent phases of the cardiac cycle (gating) over multiple heartbeats (assuming the patient has a stable cardiac rhythm), effectively freezing the cardiac motion, at the cost of only being able to acquire data for each image during a corresponding small fraction of the imaging time. The remaining available imaging time can be used either to acquire multiple images at each location, which can be played back as a movie loop to show motion, or to acquire images at multiple locations, each of which may be acquired at a different phase of the cardiac cycle. Similarly, the respiratory cycle can be monitored and used to gate data acquisition to reduce respiratory blurring, but at the cost of a further reduction in the time available for image acquisition. Higher speed imaging can be used to acquire images during suspended respiration, or even in a fraction of a heartbeat, although generally with some loss in image signal-to-noise ratio. Ongoing advances in the capabilities of MR image acquisition hardware, such as magnetic field gradients and other electronic and computer components, have led to a steady increase in the speed of CMR, permitting high-quality anatomic imaging evaluation of the heart in times comparable to those for other parts of the body.

New functional imaging capabilities of CMR are being developed that may significantly improve its clinical utility for the evaluation and management of ischemic heart disease. CMR is unique in its ability to acquire high-quality tomographic images of the heart in true 3-D registration to each other at multiple phases of the cardiac cycle (although research continues on improved high-speed and gated radiographic computed tomography methods). This has led to its acceptance as the current gold standard for the assessment of global cardiac function measures, such as stroke volume and ejection fraction, at least for research purposes. As the superiority of CMR to conventional cardiac imaging methods such as echocardiography and radionuclide blood pool imaging becomes further established, this may lead to more CMR examinations to evaluate cardiac function. Furthermore, the unique motion sensitivity of CMR can be exploited with novel methods, such as noninvasive tissue tagging with spatial modulation of magnetization (SPAMM), to provide unique kinds of information on regional wall motion. This includes the transmural distribution of contraction and measurement of nonradial components of motion, such as torsional motion and circumferential and longitudinal shortening. Much information can be gained from a simple visual inspection of these tagged images; quantitative analysis of regional wall motion may prove to be even more valuable. While initial experience with these new capabilities is still very limited, they should significantly advance the evaluation of the functional consequences of ischemia and other heart diseases. This, coupled with its other advantages, could lead to CMR becoming a dominant method to evaluate cardiac function..

Another new area of development in CMR with great potential clinical utility is the evaluation of regional perfusion, particularly with the use of very rapid imaging (at multiple slice locations per heartbeat) to follow the passage of a bolus of MR contrast agent through the heart. Again, while visual inspection of such imaging sequences readily reveals areas of poorer flow through their delayed and reduced contrast enhancement, quantitative analysis may provide additional useful information. Coupled with its superior spatial resolution, CMR could thus potentially replace most current cardiac nuclear imaging examinations for perfusion evaluation.

As in other parts of the body, magnetic resonance angiography (MRA) is continuing to improve, with coronary MRA starting to approach diagnostic quality, at least for screening purposes. Coupled with promising early developments in interventional methods using CMR, this could eventually lead to CMR replacing a significant number of current fluoroscopically guided catheter-based studies.

Despite all the promise of CMR as a one-stop shop for cardiac imaging, it must be kept in mind that it is still evolving, albeit rapidly, and little is ready yet for routine clinical use. In addition, few efficacy studies have been performed, limiting potential reimbursement. However, the prospect of improving the accuracy of cardiac evaluation with CMR while reducing costs by eliminating the need for other examinations is very appealing and is certain to evoke much interest.


To help us anticipate the future course of the involvement of radiology in CMR, we can examine the history of the three currently dominant conventional cardiac imaging methods: cardiac catheterization, radionuclide imaging, and echocardiography. In all three cases, initial technical and clinical development was carried out largely by radiologists and the physicists and engineers working with them. As the clinical utility of these imaging methods became clearer, there was progressively greater involvement by cardiologists and progressively less involvement by radiologists, so that now the great majority of such studies are carried out by cardiologists, with radiologists involved in only a relatively minor way.

There have been many factors involved in this loss of the other cardiac imaging methods from radiology. An obviously important factor has been the issue of who controls the patients. It clearly is advantageous financially for cardiologists to have imaging studies on their patients performed by themselves or an imaging specialist in their group, rather than refer them to a radiologist elsewhere. Another factor has been clinical expertise. Although trained in cardiac disease as part of their residency, radiologists no longer have to face a specific cardiac-directed portion of their board examinations; in a version of the chicken-and-egg question, they have stopped taking cardiac imaging as seriously as imaging of other parts of the body, in anticipation of it playing only a minor role in their practice. Yet another issue could be the potential risks to the patient of stress examinations performed for evaluation of ischemic heart disease, which are different from those involved in other parts of radiologic practice and perhaps are somewhat intimidating.

What has happened typically in the evolution of other areas of cardiac imaging is a progression from early imaging being performed largely by radiologists, to a joint performance of imaging by radiologists and cardiologists with interpretation by radiologists, to performance largely by cardiologists with later radiologist review, to a final effective elimination of radiologist participation altogether. This has generally happened in association with local turf battles between radiology and cardiology over who controls the imaging equipment and the revenue generated by the examinations; cardiologists generally have ultimately acquired their own equipment, dedicated for performing their own examinations. The outcome of these battles has been determined only partly by issues of relative clinical competence (radiologists know imaging better, while cardiologists know the diseases better, but each can fairly readily acquire the additional knowledge and skills required to do the clinical imaging well), but also by local factors such as control of expensive imaging resources, control of patient referrals and hospital admissions, and sheer force of personality.

That such an outcome is not inevitable for CMR is shown by the importance of radiologists in the imaging of other parts of the body, at least in the United States, although there are continuing skirmishes in areas such as neuroimaging. There also is continuing evolution of other aspects of medicine that could have an impact, such as the growing influence of managed care that could seek to diminish the relative importance of cardiology specialists and reduce the amount of reimbursement for imaging studies to the point that there may be little incentive to fight over who gets to do them.


With the increasing likelihood of CMR playing a major role in the evaluation of cardiac patients in the relatively near future, there is growing interest in what has been, up to now, a relatively narrow market. After early attempts to develop a dedicated CMR system by one vendor and, later, a joint venture between two others were abandoned as being premature, the major equipment manufacturers have primarily focused on developing add-on cardiac imaging packages for their general purpose MRI systems. Recently encouraged by ongoing technical advances and increasing clinical interest, the major manufacturers have begun developing MRI systems specifically tailored for CMR applications (although still able to be used for other applications). Such CMR-specific features as high-performance gradient systems, specialized imaging sequences and user interfaces, and specialized image analysis and display software are bundled together and are being marketed as CMR systems to a specifically cardiac-oriented audience. While many areas of CMR are still very much under development and their real clinical efficacy still remains unproven, the large amount of money potentially involved makes this an area of great interest to both radiology and cardiology. As far as the manufacturers are concerned, their business is to sell equipment; while they will not wish to alienate current customers, they will seek to meet the needs of any new market, wherever it may be based.

While the relative local influence of different disciplines will certainly be important in determining the outcome of turf battles, there must also be clinical and technical competence in CMR behind such maneuvering. Although economic and political forces are obviously important in the real world, medicine is ultimately about taking care of patients, not just fighting for market share. Radiologists generally have a much better background than cardiologists in the complex factors involved in producing and interpreting good-quality general MRI examinations. However, radiologists will have to learn the additional specific technical aspects involved in performing CMR and the clinical aspects of interpreting it well, or all the political muscle available will ultimately be of no avail in holding onto CMR as a primarily radiologic procedure. Cardiologists are also capable of learning these aspects of CMR, so if radiologists decide to wait until the manufacturers deliver simple turnkey systems for CMR prior to acquiring the additional necessary knowledge, they will have lost the potential edge they now have by virtue of their current MRI knowledge and installed systems.

As part of radiology’s demonstration of CMR competence, it is necessary to incorporate additional specific training requirements in the radiologic training program; for radiologists already in practice, additional formal training opportunities must be provided, for example, in association with annual radiology society meetings. A new intersocietal-based ongoing series of training courses in cardiac imaging for radiologists is being planned under the aegis of the American College of Radiology. In academic centers, cardiac imaging has become almost a vestigial part of radiology departments; it will have to be rebuilt, with active involvement in both more basic and clinical research. As the CMR field evolves further, it may be appropriate to develop formal subspecialty training and certification. In those academic centers where suitable relationships can be worked out, a collaboration between radiology and cardiology would theoretically lead CMR to its most effective development. In private practice this would be less likely to be readily worked out, although some cardiology groups may even hire a radiologist to run a CMR system for them. While fee splitting would be unethical, some sort of arrangement whereby cardiology gets to share in the revenue generated by cardiac MRI cases done in radiology may be necessary to help motivate a significant flow of case referrals. In any case, radiologists will, of course, have to provide good-quality CMR studies in order to be entrusted with these patients.


While the outcome of local turf battles will depend in part on the relative political strength of the local players, hardball attempts to monopolize CMR by either side are likely to lead to polarization of the departments and impede the kind of collaborative effort that will be required to develop optimally this promising new field for the benefit of patients. The relative strengths of radiology in imaging and cardiology in patient care can be deployed best through a collaborative effort to develop and apply CMR. Just as no specialty can really claim to own an organ system, no specialty can really claim to own imaging. While it will be appropriate for radiology to use whatever legitimate means it has available to avoid being squeezed out of CMR, it is not going to be possible, and would not be appropriate, for radiology to keep CMR entirely to itself. One potential outcome is that while some high-powered private and academic cardiology centers will likely purchase and use dedicated CMR systems (hopefully just for CMR cases), many of these systems also will be sited in academic and larger private radiology practices. Most of the technical innovations developed for these specialized systems probably will find their way ultimately onto the more general-purpose MRI systems used widely by radiology practices. If these radiologists have the necessary training to provide good-quality CMR examinations, they will be able to study cardiac patients referred to them from cardiologists with smaller practices that would not justify the investment of money and time needed to acquire and implement the technology necessary to do CMR, as well as cardiac patients directly referred from other physicians.


Leon Axel, PhD, MD, is professor, Department of Radiology, Pendergrass Diagnostic Research Laboratory, University of Pennsylvania Medical Center, Philadelphia.