It has been 15 years since the first MRI scanners began arriving at well-heeled hospitals and imaging centers. Today there are more than 5,000 in use at fixed sites and in shared mobile vans, producing some 11 to 12 million clinical studies annually, according to data provided by Technology Marketing Group, Des Plaines, Ill.

The early high-field magnets were novelties that produced beautiful images. But as MRI became a standard clinical tool and original equipment manufacturers geared up to offer a variety of systems with varying field strengths to meet a seemingly insatiable demand, serious concerns about machine performance and image readability began to emerge. By the early 1990s, many influential members of the American College of Radiology (ACR) had become convinced that steps needed to be taken to protect patients from misdiagnoses engendered by substandard MRI equipment producing murky images unreadable even with night-vision goggles.

The first tentative step to address the problem was creation of an MRI standard in 1992, but it became evident that stronger measures were required, recalls William G. Bradley, Jr, MD, PhD, who headed the ACR committee on standards and education at that time. “So we decided to put together something more demanding, which was an accreditation program for MRI sites modeled after the Mammography Quality Accreditation Standard (MQAS).The MQAS has without question dramatically improved the quality of mammography and has since been adopted by the federal government [as the Mammography Quality Standards Act]. The ACR hopes MRI accreditation will be embraced by payors as a standard for reimbursement.”

The accreditation program was finally approved by the ACR Board of Chancellors in November 1996. “We had about 100 people involved and 100 sites were tested at all field strengths,” recalls Bradley, director of MRI at Long Beach Memorial Medical Center, Long Beach, Calif, and professor of radiology at the University of California-Irvine.

Still, MRI accreditation got off to a slow start, in part because many ACR members had reservations about the need for what was certain to be a time-consuming and expensive process.

“If everyone practiced the way we do, you would not need accreditation,” notes David Seidenwurm, MD, a neuroradiologist for the past 9 years with Radiological Associates of Sacramento (RAS), the largest private radiology practice in Northern California. “But I think that by setting a minimum standard, in some cases that is going to be what serves the patient best. Whether accreditation is necessary or not, I think it is coming.”


Initially, many applicants failed because they used the wrong pixel size, or they were not scanning their phantom properly, Bradley reports. But more explicit instructions were developed, and the move to obtain MRI accreditation is now in high gear, helped by some prodding from payors like Aetna/U.S. Healthcare, which has announced it will provide reimbursement only for MRI images made at accredited sites, effective January 1, 2001. Other managed care organizations and the Blue Cross/Blue Shield organizations in some states are reportedly ready to follow suit, according to Jeffrey C. Weinreb, MD, current chair of the ACR Committee on MRI Accreditation, and the ACR is working to get additional payors to sign on.

As of October 1999, 1,772 MRI sites have applied for accreditation, and 389 units have been accredited to date. Considering that many fixed sites have more than one magnet, and that mobile units may serve two or three hospitals, this number probably represents 35% to 40% of all magnets now in use. Applications continue to arrive, and while the initial backlog has been reduced by training additional radiologists and physicists to assess the images and data, it still typically takes 4 months from submission to accreditation. In some instances, however, accreditation has been delayed because the phantom data submitted have been so difficult to read that the ACR has had to send physicists to the site.

“Going into this, I was not absolutely sure that MRI accreditation was needed,” admits Weinreb, who is professor of radiology at New York University School of Medicine, and cochairman of radiology and director of MRI at the New York University Medical Center, New York City. Weinreb has had the opportunity to see most of the clinical images submitted for accreditation. “Keep in mind that sites have the freedom to submit their best images, and have selected what I see,” he notes. “Frankly, I have been astounded at some of what has been submitted. Literally there are images that people are making on good machines-machines that you know are capable of better work-that are simply inferior. So if I had any reservations before about the value of this process, and I know that it is not perfect, I am very, very confident now that it is a good development.”


The accreditation program is designed to assure that not only the MRI hardware meets specified criteria, but that the technologists and the interpreting physicians have appropriate training, qualifications, and continuing medical education. “What it does not do is test the capability of the interpreting physician to interpret accurately,” Weinreb said. “It does not credential the radiologist or any other doctor that is reading at the site. That is the responsibility of the American Board of Radiology, the American Board of Cardiology, and the American Board of Orthopedic Surgeons.”

At multi-magnet sites, all magnets must meet the criteria. Shared mobile units that use the same physicians, technologists, and protocols at each visited site count as a single site. “But that is a minority of them,” Weinreb notes.

Some radiologists have expressed the opinion that the whole purpose of accreditation seems to be to weed out substandard systems and casual practitioners, but Weinreb says the ACR is not on a search- and-destroy mission. “The accreditation process is meant to be educational and not punitive. Everybody has their own way of thinking about this,” he says. “The goal is to improve the quality of MRI, not to get rid of lousy equipment. It may mean upgrading it, or tweaking it, so it performs better.”

Sites receive report cards that inform them of their specific deficiencies. And in most cases, the ACR will recommend, if possible, what can be done to minimize or eliminate those deficiencies. “By making people aware that their equipment is not performing up to a minimum standard, we hope that the worst sites will do something to improve their imaging,” Weinreb says. “Reimbursement policies may make that happen. But you would like to think there are some people who would actually care about [improving]. Maybe that is idealistic or unrealistic.”

“Of course, there are going to be sites that do not receive accreditation,” he adds. “But the reason may not be the clinical images. There are phantom data as well as other information that are part of the equation.”

Still other sites may not be accredited because they never applied. “Some people know they are going to fail so they are not going to apply,” Bradley says. Thus, accreditation will not necessarily eliminate people with older systems producing poor images. But, he adds, “it will if all payors insist on ACR accreditation.”

The concept of tying accreditation to reimbursement, as it is for mammography, makes sense to Lawrence Muroff, MD, president of Tampa, Fla-based Educational Symposia, Inc, who has been involved in educating diagnostic imagers since the mid-1970s.

“By having accreditation as an accepted means of setting the bar for quality, you have protected the patient at least from having poor quality studies,” Muroff says. “If then you further link that to reimbursement, you have assured that people who are just casually in the field — the marginal players who should not have been there in the first place — will be forced out of the market. Accreditation also has a braking effect on the unnecessary proliferation of MRI equipment.”


An ongoing complaint of many radiologists is that the accreditation program puts too much emphasis on hardware and not enough on interpreting the images. “Everyone has to sign a sheet that says they are capable of reading images and have read a certain number. But the number [500] is not very high. And if you read [that many] badly, it does not mean you can really read them,” says Allen Elster, MD, professor of radiology, Wake Forest University, and director of the MRI Center at Wake Forest University Baptist Medical Center, Winston-Salem, NC. “The fundamental problem in MRI is not the equipment, it is the interpretation of the images. There are just too many people around who take an MRI course over a weekend and become an expert.”

Doing overreads at Wake Forest, Elster constantly finds misdiagnoses. “It is not a trivial problem,” he believes. “Anybody in an academic center will tell you this. Patients are referred in or have come in because they have problematic cases, and everyday we read at least a half dozen reports of scans from outside MRI centers. Invariably there are significant errors made on a large number of those outside reports. Either the radiologists do not recognize an abnormality or they interpret an abnormality they see in the wrong way.

“Unfortunately, the people who are not very qualified at reading scans are usually the same ones who would not want someone in an academic center to overread all of their scans. The really bad people do not want any type of control like that. The people who really need our overread services do not ask for them. The people who send us business are the good, conscientious radiologists who want what is best for the patient.”

“MRIs are easy to read if you are willing to accept mediocrity,” notes Alan D. Kaye, MD, chairman of the radiology department at Bridgeport Hospital in Connecticut. “If I am paying $550 to $1,000 for a test, I want to make sure that someone with expertise is interpreting it.

“Accreditation will set the floor, but we can do a lot better than that,” he adds. “Anyone can buy good images. But what is more challenging than producing good images is setting up the correct protocols and interpreting the subtleties. My sense is that accreditation is not going to be able to address those issues satisfactorily.”


While some physicians such as Muroff believe accreditation will slow somewhat the proliferation of MRI, a major stumbling block to rampant growth in many states remains the provision for a Certificate of Need (CON). Currently 23 states have CON laws that apply to MRI. Some apply only to hospitals. Others are more byzantine.

Connecticut’s CON law, for example, is one of the more complex and restrictive. It applies to major medical equipment purchases exceeding $400,000 by hospitals and health care facilities, but exempts private physician practices. Thus, an orthopedist can install a $1 million MRI in his office without getting a CON.

The Connecticut CON was originally promulgated to regulate costs and control the proliferation of technology. The state has just formed a work group, including a cross-section of independent radiologists, to reexamine the CON rules, particularly the cost threshold, which many feel is inconsistent with the idea of advanced quality since substandard equipment can sneak in under the regulatory radar.

“The requirement for a CON in Connecticut has definitely slowed the proliferation of MRI,” Kaye says. He heads Bridgeport Radiology Associates, which operates the MRI at Bridgeport Hospital as well as MRIs at two of its four offices. “[Submission and approval of a CON] is a long, drawn-out process, but it has its merits.”

One of the benefits, Kaye says, is that it restricts the marginal players. Another is that the process takes into account who is going to do the reading. If MRI is restricted to larger hospitals or groups that have the ability to provide subspecialty reading, the patient is going to be in the hands of an expert.

While CON may limit competition, it has not kept prices artificially high, Kaye says. “We are now in the $550 to $750 range, and have been there for the past 3 years. The managed care companies have not indicated they want to restrict imaging to one or two locations, but they seem to be trying to limit imaging in general by having onerous and sometimes irrational precertification rules.”

The first high-field MRI imaging systems cost in excess of $2 million, in addition to hefty site preparation expense — often an additional $500,000 — and ongoing operational costs. Today $500,000 will purchase a low-field open-bore MRI requiring substantially lower operating costs and far less in site preparation expenses. And prices may get lower still.

All this means that an on-site MRI is now within the financial sights of even the smallest hospital that wants one, Muroff says. “There are some scanners that can break even with just a couple of studies a day.”

Putting a low-field MRI in a small hospital where radiologists have minimal training in interpreting magnetic resonance images has positive and negative implications.

“Some lesions are going to be so obvious on the MRI scan that even an inexperienced person can see them,” Elster notes. “So to the extent that the modality can produce a superior quality image, that should help. To the extent that there are a number of artifacts and pitfalls when interpreting MRI scans, it could have a downside as well.”


It is not just rural and small town hospitals that are queuing up for the open-bore MRI. Radiological Associates of Sacramento, with more than 40 doctors and in excess of 500 employees, recently added two low-field open-bore MRI and one high-field MRI to the four high-field magnets it operates in freestanding imaging centers and local hospitals. It now has seven MRI systems at six sites in the Sacramento area.

Buying a new MRI involves a complex equation, Seidenwurm says. “You have to look at your practice needs,” he says. “The high-field magnet was for a very demanding hospital practice with active stroke, pediatrics, neurosurgery, and oncology services. We needed to satisfy those requirements so we went with a high-field echoplanar-capable system. One of the low-field systems went to a trauma hospital where we believed that an open-bore MRI would best serve critically ill patients who might have accompanying life support equipment.”

The other open-bore MRI went to an RAS imaging center to complement a high-field system and provide the proper tool for the special needs of larger patients, children, and persons with claustrophobia or metallic hardware other than pacemakers, according to Seidenwurm.

The economics of the low-field-strength open-bore MRIs are considered a threat to mobile MRIs, which as recently as 1997 served 1,485 sites, according to a survey made by the Technology Marketing Group. Some say mobile MRI will go the way of the passenger pigeon, but not everyone agrees. “There are some things you can do only with a high field,” says Bradley of Long Beach Memorial Medical Center. In his best-of-all-possible-worlds scenario for a small hospital, Bradley says he would “put in a low-field open magnet to take care of my routine scans, but once a week the mobile would come by and I could do my coronary magnetic resonance angiography, spectroscopy, and fusion imaging.”

In apparent agreement, two smaller Connecticut hospitals recently signed a new 5-year agreement to continue mobile MRI services. The proliferation of MRI has dropped the cost of a scan here from the $800 to $1,200 range typical in the early years to the current $500 to $650 price range.

However, in some states with an overabundance of MRI sites, such as Florida and California, competitive pressures have given rise to imaging brokers who contract with underutilized sites to deliver customers for discount scans, with reimbursement often pegged as low as $300.

“There are people who broker scans with altruistic goals and there are people who do it purely for crass profit with the worst equipment you could ever imagine,” Muroff says. Brokering reportedly has been around for 7 or 8 years. “Some networks were set up to provide for uniformly good quality studies, interpreted by credible, knowledgeable individuals, at competitive rates. Other networks, however, were set up to troll as many cases through at whatever cost they could get on whatever types of equipment were available. So scan brokering can have benefits and it can have bad effects. This is an area where accreditation can chop off the poor equipment, the shoddily done study.”

“I do not think MRI has peaked,” says RAS’s Seidenwurm. “But I do not think that the same explosive growth that occurred in the first 5 years of MRI is going to occur again. There will be a continuing desire to improve the quality and versatility of MRI, and to expand into new applications, but I do not think this will be as rapid as it was a few years ago.”


Richard B. Elsberry is a contributing writer for Decisions in Axis Imaging News.