New 3T MR scanners built for the clinical environment are barely out of their packing crates, so clinical experience is quite fresh. But imagers with a few hundred cases under their belts are finding that 3T is superior to 1.5T in bread-and-butter MR imaging in the brain, spine, and musculoskeletal system. And although acquisition time tends to be longer, radiologists don't have to confront some of the other issues that plagued users of the first-generation 3T units, such as siting logistics, reduced field-of-view with some coils, and prolonged T1 relaxation times.
The experience with 3T MR is so positive that many imagers, such as Dr. Mark Shapiro, head of neuroimaging at the Neuroskeletal Imaging Institutes in Winter Park, FL, claims they will never buy a 1.5T machine again.
Second-generation 3T magnets just started coming off the assembly lines in earnest in the last quarter of 2003, but they are rapidly entering the clinical sphere. Neuroskeletal Imaging Institutes, which installed its first clinical 3T scanner last February, is adding a high field strength magnet in each of its three other south Florida outpatient imaging centers. And vendors that installed 20 3T units in the U.S. last year are gearing up to install 150 this year in outpatient imaging centers and community hospitals.
The move to the clinic has been swift because manufacturers focused on end users. In addition to the technological accoutrements of high-field MR-parallel imaging and high-performance gradients-current 3T scanners have been designed to mimic their 1.5T counterparts. The machines are the same size as short-bore 1.5T units, so they don't have the siting or "patient-unfriendly" problems their heavier, long-tunnel 3T forebears did. The 3T units also have the same user interface and use the same software as 1.5T machines, so they can be used right out of the box in place of 1.5T systems.
And because the scanners come with a family of coils, there is no need to reposition patients or switch to a 1.5T magnet to counteract a contracted field-of-view. The restricted linearity of older head-only coils made it difficult to view the cervical spine and the craniocervical junction of a patient with a short, squat neck or broad shoulders who couldn't be positioned high enough in the coil, said Dr. Paul M. Ruggieri, head of MRI at the Cleveland Clinic. These patients had to be shifted over to a 1.5T MR machine to avoid obscuring the regions surrounding the foramen magnum and C1 and C2 vertebrae by signal loss.
Scan times are still longer at 3T than at 1.5T, but they are steadily declining as radiologists adjust to the learning curve and adapt to the peculiarities of imaging at high field strength. These include prolonged T1 relaxation times, which produce less T1 contrast in standard echo sequences.
Since the Cleveland Clinic began MR imaging at 3T three years ago, Ruggieri has been working to optimize parameters for conventional studies, one by one. He found that with standard or fast spin-echo for T2-weighted imaging, he could get the expected bump in signal to noise at 3T in about the same acquisition time as at 1.5T, because the T1 relaxation times didn't come into play.
But imaging time for conventional T1-weighted studies was at least 50% longer, as Ruggieri had to resort to fast inversion recovery to improve contrast. Fast inversion recovery prolongs acquisition because of the need for a series of concatenated acquisitions to cover the whole head: a concatenated study before and after gadolinium administration and a second postcontrast concatenated scan in a second plane for a typical brain tumor study.
To trim scan time associated with conventional T1 imaging, some institutions use magnetic prepared rapid acquisition GRE (MP-RAGE) rather than inversion recovery and cover the whole head in a single acquisition, Ruggieri said. Shapiro addresses the issue of loss of conspicuity of enhancing lesions by downplaying spin-echo imaging of the brain.
"There's no doubt that spin-echo T1 imaging at 3T will not achieve the same differentiation as 1.5T because of the long relaxation times. But with gradient-echo pulse sequences, the increased enhancement with paramagnetic contrast agents that comes with higher field strength, and the ability to trim slice thickness and capture images in all three planes with volume acquisition, we can still detect more lesions at 3T than we would with spin-echo imaging on 1.5T," he said.
On the whole, however, longer acquisition times often have less to do with 3T techniques or protocols than with user preference. Even though 3T MR's increased signal could be used to accelerate throughput, many imagers are taking more time in acquisition to obtain higher matrices such as 512 x 512 or 1024 x 1024 rather than 256 x 256.
And 3T acquisition time will eventually drop with the addition of transmit/receive phased-array coils, which will eliminate issues related to the specific absorption rate, Shapiro said.
Only a handful of sites have used 3T exclusively for clinical practice for more than a few months. Neuroskeletal Imaging Institutes, which is among the centers with the longest clinical experience, has been exclusively using 3T MR clinically to scan 22 patients a day and serving the same case mix it did at 1.5T: patients who need brain, spine, and musculoskeletal studies. But because of 3T, the facility can provide more imaging options. Every patient undergoing brain scanning has a perfusion and diffusion study, spectroscopy, and 3D fluid-attenuated inversion recovery (FLAIR), and every study can be done at a 1-mm slice thickness.
Even with older, research-oriented magnets, academic medical centers that have been sandwiching in clinical cases can demonstrate that the technology makes a diagnostic difference in brain imaging.
Dr. Gregory Sorensen, associate director of the Neurological MR Center at Massachusetts General Hospital reported at the 2003 RSNA meeting that in one-third of 50 patients, 3T MR revealed clinically relevant abnormalities in the brain that were missed at 1.5T. The technology also helped neuroradiologists follow patients who underwent surgery for obsessive-compulsive disorder, a destructive neurobiological condition that is not well understood.
Neurosurgeons at the Cleveland Clinic and MGH request 3T MR for functional studies, particularly for visualizing discrete differences in contrast between abnormal and normal areas of brain cortex and for locating minute cortical dysplasias in patients with nonlesion epilepsy. And high field strength MR has become the first choice at the University of Bonn in Germany for assessing the carotid arteries and cerebral vessels.
In body imaging, 3T MR is spotting tumors that are notoriously difficult to detect at potentially curable stages, counteracting distortion that complicates cardiac studies, and opening the door to physiologic imaging and spectroscopy.
As a result of 3T's additional signal to noise, spectroscopy is being done in the abdomen in a series of single breath-holds, said Dr. Robert Lenkinski, director of the 3T MRI program at Beth Israel Deaconess Medical Center. Spectroscopic analysis of a renal mass has been completed in eight 16-second breath-holds. A mediastinal mass has been evaluated spectroscopically in four breath-holds. Multinuclear spectroscopy of phosphorus metabolism to test the viability of muscle in the foot of a diabetic patient has been done in four minutes of signal acquisition.
In the U.S. and across Europe, 3T magnets are just about to start imaging patients on a routine basis.
"We are at an early stage," said Dr. Christiane Kuhl, an associate professor of radiology and head of MRI at the University of Bonn. "But if a center sees many neuroradiology and some non-neuroradiology patients, it can safely use a 3T scanner for both."
MS. SANDRICK is a freelance writer in Chicago.