Maturing technology makes 3T scanner a clinical must-have

"So what do you use your 3T system for?"

Five years ago, that question would have been a good opening for a detailed discussion on the strengths and weaknesses of high-field MRI.

Small joint imaging? Sure, just look at the exquisite spatial resolution.

Brain scans? Most definitely. Our neuroradiologists are seeing defects that you'd never see at 1.5T.

Abdominal MRI? Well, we'd like to, but there are still too many issues with artifacts right now.

Cardiac studies? Sorry, did you say 3T or CT?

Today, radiology department heads and practice managers are more likely to have a more succinct response: "What aren't we doing at 3T?"

Most, if not all, university-affiliated centers in the U.S. have at least one 3T system. Several outpatient imaging clinics have also invested in high-field MRI. Nowhere are the scanners being reserved for a small list of niche applications. Certain specialty scans may be directed toward 3T, but the scanners are earning their keep by tackling burgeoning clinical work lists.

Radiology staff at Fletcher Allen Healthcare at the University of Vermont in Burlington have been using 3T with their clinical caseloads for two years. The 500-bed teaching hospital has two 1.5T units and one 3T system for clinical referrals. A second 3T unit intended for research and a 1T open system for clinical work are being installed.

Patients referred for MRI at the Burlington hospital are typically allocated to whichever scanner has a free slot. Only patients with implants that have questionable 3T compatibility are excluded from the high-field magnet.

"We do not have a menu of things that we do at 3T and a menu of things that we do at 1.5T," said Dr. Steven Braff, a radiologist at Fletcher Allen. "But if clinicians specifically ask for 3T-neurologists typically do, and neurosurgeons might-then we would certainly accommodate them."

Patient scheduling is generally independent of field strength at the Duke University Medical Center as well. The center now has nine MR scanners devoted to clinical work, two of them 3T units. This generous MR provision follows the purchase of two additional 1.5T systems in 2006 to help cover outpatient imaging.

The choice of field strength for the new MR systems was guided by space considerations rather than expected performance, said Dr. Elmar Merkle, director of body MRI at Duke. Expansion of the center's MR capacity required the new magnets to be sited in trailers, but finding a freestanding 3T system proved impossible. Hence, the decision to grow the 1.5T base.

"We would probably have purchased a 3T scanner had there been a technical solution," he said.

Pregnant patients, individuals with ascites, and anyone with an implant that has not been 3T safety-checked will always undergo scanning on a 1.5T system. All other patients are scheduled on whichever scanner has free space.

Certain exams are definitely superior at 3T, Merkle said. Any applications reliant on contrast enhancement, MR angiography, or breast MRI, for example, gain significantly from the improved signal-to-noise ratio. Investigations that hinge on seeing very fine detail, such as musculoskeletal imaging, also benefit from the submillimeter resolution that 3T can offer.

"If I had a choice, I would do these patients at 3T. But I wouldn't reschedule or delay their exam because 3T wasn't available," he said.

The same philosophy applies at the University Hospital Michigan in Ann Arbor, where clinical MR imaging is distributed among five 1.5T systems and one 3T unit. Dr. Suresh Mukherji, director of neuroradiology, much prefers to schedule his brain MRA and cervical spine imaging referrals on the high-field scanner. But if the timing proves impossible, he would rather switch to 1.5T than ask a patient to wait.

"The radiology department sees about 3000 patients for MRI every month, so it is impossible to get every single study on the 3T that we would like. Logistically, that is almost impossible," he said.

Dr. Steven Mendelsohn, medical director for Zwanger-Pesiri Radiology on Long Island, also aims to do his MRA work at 3T unless scheduling prohibits it. His group runs two 3T and six 1.5T units from its six imaging centers, giving schedulers some play in juggling appointments. Priority for 3T time slots is given first to MRA, then to musculoskeletal applications after that. Remaining spaces on the scanners are then filled from the referrals list.

"Sometimes we will preferentially schedule body work on very, very thin patients to 1.5T. I hear other people say that, in general, they prefer to do abdominal work at 1.5T rather than 3T, but we have not found that," Mendelsohn said.

At Nevada Imaging Centers in the Las Vegas region, 3T is used preferentially to evaluate brain lesions and spinal cord disease. Certain specialty sequences, such as MR spectroscopy, diffusion tensor imaging with fiber tracking, and functional MRI, are done exclusively at 3T. Patients may be diagnosed with a brain abnormality after examination on one of the practice's four lower field strength magnets. Follow-up is then typically moved to 3T.

Preferential triaging of breast MRI to 3T will start later this year at New York University Medical Center. Patients referred for breast MRI are currently assigned fairly randomly to one of six 1.5T or two 3T scanners. This approach is set to change when NYU purchases a 3T system for its Clinical Cancer Center. The added 3T capacity will be filled by newly diagnosed breast cancer patients who are undergoing an extensive disease evaluation and subjects with a high risk of developing cancer.

"Our thinking is that 3T will give us more spatial resolution, so we are going to try and use it for those patients who are more likely to have an abnormality," said Dr. Linda Moy, an assistant professor of radiology at NYU.

RESEARCH TO ROUTINE

The case for shifting breast MR workup from 1.5T to 3T is not, strictly speaking, supported by hard scientific evidence. Such a case would need to demonstrate that the extra detail afforded by higher resolution MR scans did indeed alter patient outcome. NYU researchers are currently participating in a multicenter study to address this issue. They are also investigating whether the addition of MRS and perfusion and/or diffusion imaging may improve diagnostic specificity.

Moy is hopeful that as the 3T base increases, the case for high-field imaging will be demonstrated. But for now, protocols need to be refined and evidence gathered. Until this is done, efforts to reduce false-positive results by gathering spectroscopic or physiological data will not be reimbursed.

"Right now, none of these additional sequences are paid for. We are just using some magnet time to develop them," she said.

The time lag between research and routine use at 3T is undoubtedly shortening, according to Thoralf Niendorf, Ph.D., a professor of experimental MRI at Aachen University Hospital in Germany. Routine clinical studies now account for approximately 40% of the 3T workload there. Another 40% of 3T scanner time is devoted to clinical research, and the remaining 20% to basic research to advance MR methodology. This sort of split is common at European academic institutions, which have traditionally relied on research grants to fund investment in high-field technology.

The routine 3T clinical caseload will generally comprise a mix of structural and functional brain MRI, MRA, and cardiac MRI. Cardiac applications have conventionally been regarded as among the most challenging for 3T, because of motion artifacts and magnetic inhomogeneities, as well as localized tissue heating from radiofrequency deposition (specific absorption rate effects). But this is also an area where the increased signal could overcome some of the problems experienced at 1.5T. A higher magnetic field strength may, for example, make it easier to differentiate healthy and infarcted tissue during delayed-enhancement imaging to assess myocardial viability. Switching from 1.5T to 3T could also improve the visibility of tagging bands used to track myocardial wall motion and extend their longevity.

"The work we are doing with cardiac MRI is pioneering work, and we are trying to drive further progress. But our ultimate goal is to transfer all these results as fast as possible into the clinical environment," Niendorf said. "The delay between a new MR methodology and its first clinical application has shrunk from 18 months to just two or three months. Once an early feasibility study has been successfully completed, it is just another small step to widespread clinical usage."

Wider availability of multichannel high-field coils should overcome any remaining problems with 3T cardiac MRI, according to Braff. The results of cardiac perfusion studies performed at 3T at Fletcher Allen Healthcare are already better than those acquired at 1.5T. Viability imaging studies still have some way to go, however.

Iron these difficulties out, and 1.5T magnets could be destined for medical history museums. Braff predicts that future radiology departments and imaging centers will invest in 3T and high-field open scanners alone.

"Five years from now, this combination of magnets is going to supplant 80% of 1.5T systems," he said.

NEW APPLICATIONS

The rapid transition from research to clinical use can also be seen with MR cholangiopancreatography. Last spring, new sequences were becoming available that promised to rid 3T MRCP of the artifacts that have limited its diagnostic utility. Radiologists at Duke were among those investigating this development. Today, 3T MRCP is just another routine clinical application at Duke.

Radiologists at Clarian Health Methodist Hospital in Indianapolis are so impressed with the resolution of 3T MRCP that they no longer routinely perform pancreatic examinations at 1.5T, said Gerry Szkotnicki, director of Clarian's Neuroscience Center of Excellence.

"Three-T provides sufficient resolution to create a usable study that can preclude the need for more invasive investigations," he said.

Most of the 3T MR work at Clarian is brain imaging. The hospital is a large neurotrauma center and receives many requests for presurgical evaluation of neurovascular defects and brain tumors. Radiologists, clinicians, and surgeons are keen to implement up-and-coming neuroimaging protocols.

One area receiving considerable attention is the use of 3T perfusion/diffusion mismatch for stroke evaluation. Such a study could be used to guide neurovascular intervention and determine a patient's poststroke prognosis. Developments in functional neurosurgery, such as cortical and skull-based stimulators, have also focused attention on DTI with fiber tractography. Information derived from such a scan could provide critical pre- and postsurgical guidance for novel interventional treatments of functional conditions such as epilepsy.

To allow for these emerging neuroimaging applications, installation of a high-field MRI system in the neurosurgical operating room is planned. This could be a 3T unit, though no decision has been made yet. This move will require considerable technical expertise, not to mention financial commitment, Szkotnicki said. But, given the potential clinical gain, particularly for more complex cases, Clarian is prepared to make the investment.

"In the case of a borderline inoperable brain tumor, for example, or a case where surgery will have a high risk of neurological deficit, being able plan a case on a high-definition 3T system and then evaluate progress in the operating room would be of great value," he said.

A number of advanced 3T neuroimaging techniques have already entered routine clinical practice at University Hospital Michigan, Mukherji said. These include time-resolved MRA, fMRI, MRS, DTI, and perfusion imaging. The quality of physiological and functional imaging at 3T is substantially improved by the increased SNR, enabling the radiology team to perform these sequences confidently and routinely. This really demonstrates the true strength of higher field MRI, he said.

"The number one reason that people are investing in 3T is not necessarily because they can make the same diagnosis better," he said. "A brain tumor is still basically a brain tumor even when you can see it better. With these physiological and metabolic techniques, we are moving toward treatment monitoring."

Ms. Gould is a contributing editor of Diagnostic Imaging.