High-Field MRI: Race to the Top?

Last week, exactly 25 years after the first "high-field" (1.5 T) MRI scanner was introduced, radiologists gathered at an ECR 2011 session in Vienna, Austria, to ponder what its chairman called "higher and higher field magnets."

"Where," asked Olivier Clement, director of imaging research at Descartes University in Paris, asked, "is it going to stop?"

Only eight years ago, 3 T scanners were introduced to address the vexing problem of low resolution, and now there are between 40 and 50 7 T clinical scanners worldwide, either in operation or "on the brink," said medical physicist Jürgen Hennig, scientific director at the University Hospital of Freiburg. One of them is practically around the corner from the conference, at Vienna's Center of Excellence for High Field MR.

It's located an easy walk from the hospital to encourage referrals, director Siegfried Trattnig said at the opening press conference for the congress, where he showcased some of its uses. Doctors use sodium MRI imaging to visualize glycosaminoglycans as a marker for damage to cartilage transplants in knees. "You can see ultrastructures in cartilage in vivo," he said, "where before you could only see it in specimens."

The Vienna center is the first worldwide installation for CEST (chemical exchange saturation transfer) imaging, which is being used for early diagnosis of osteoarthritis and disc degeneration. "Shorter scan times at 7K make it really possible for sports medicine, muscular diseases, and oncology," he said, adding that it also has high potential in multiple sclerosis and epilepsy.
Although the new scanner is only a research facility, Trattnig added, "neurosurgeons here realize that 7 T is more accurate," and they're willing now to send their patients for 7 T rather than 3 T."

The motivation to surpass 1.5 T in the first place was that the higher signal-to-noise ratio at higher field strengths offers both faster imaging and better resolution, which allows scanning of ever-smaller volumes. In a standard 1.5 T clinical scanner which can take one image of a patient in a minute, said Luc Derasse of the University of Paris South, scanning a small animal could take up to two years. Studies of rodents are now numerous at 4 T and 7 T. Functional MRI would not be possible without the higher accuracy of 7 T, nor would the molecular or cellular targeting that is developing apace.

All of this comes at a cost (and not just in dollars or euros). The chief challenges are the risk of radiofrequency damage to tissue-the so-called "SAR (specific absorption rate) problem” - and the need to create and fine-tune multiple arrays of small coils to create a more manageable signal/noise ratio than a single huge coil magnet would generate. This works reasonably well, but leads to non-homogeneous signals, said Hennig. The problem is "intensely researched and largely unsolved."

Another challenge is the fact that contrast media which have proven their worth at 1.5 or 3 T may respond in a completely different way or even prove useless at higher field strengths. An informed choice based on a solid understanding of contrast material specificity will be ever more crucial in the future, predicted Derasse.

The race to the top continues, with ultrahigh-field magnets already in testing. Derasse showed an image of the 9.4 T magnet in testing at the University of Minnesota's Center for MR Research. Gradients and magnetization are stronger, naturally, and it's more effective with targeted contrast media, he said. But this too comes at a price: increases in T1, decreases in T2 and T2*, shorter echo time, and longer acquisition bandwidth. A move to coil arrays with 16 to 32 channels allows localization of the signal, but this increases the problem of noise arising from the sample in addition to the magnet. Supercooling is used to reduce this problem, he said, and new superconducting materials are on the horizon.

The developing field of 7 T MRI has been "largely built by neuroscience," according to Hennig, and has not provided what neurologists really want, which is comparisons between people's brains, healthy and otherwise. Few studies to date have analyzed activation in different layers of the cortex, for instance. Meanwhile a vast amount of revelations at 7 T have come from flow studies in the brain, "an enormous wealth of information which will only slowly be introduced into application," he said.

He showed several titillating slides, such as susceptibility-weighted images from the German Cancer Research Center at Heidelberg that show the introduction of microbleeds in the brain after radiation treatment. There are numerous applications outside the brain as well, Hennig added. Spinal imaging is "doable," and cardiac imaging at 7 T produces scans of very high quality. But there's "no killer app yet."

"I have put my bets on oncology," he went on, "because I think the cost of the exam will be negative compared to the cost of the therapeutic solution."

Having seen such images, Henning asked the audience, would you buy a 7 T scanner? It's a very expensive machine, he cautioned, and you'll need to count the capital costs as well: the magnet is huge and weighs 36 tons. Appreciation of 7 T MRI will increase, he predicted, when it fits on the footprint of a 3 T scanner.