IEEE TRANSACTIONS ON MAGNETICS, VOL. 25, NO. 2, MARCH 1989

I

1755

SUPERCONDUCTING MAGNETS FOR WHOLE BODY MAGNETIC RESONANCE IMAGING

Michael F. Murphy

Oxford Superconducting Technology 600 Milik Street

Carteret, NJ 07008

3. Abstract

Superconducting magnets have achieved preeminence in the magnetic resonance imaging (MRI) industry. Further growth in this market will depend on reducing system costs, extending medical applications, and easing the present siting problem. New magnet designs from Oxford address these issues. Compact magnets are economical to build and operate. Two 4 Tesla whole body magnets for research in magnetic resonance spectroscopy (MRS) are now in operation. Active-Shield magnets, by drastically reducing the magnetic fringe fields, will allow MRI systems with superconducting magnets to be located in previously inaccessible sites.

Introduction

To date, Oxford has manufactured in excess of 1,000 one-meter bore superconducting magnets for whole body magnetic resonance imaging scanners. Overall the great majority, about 90% of today's MRI systems, are built around superconducting magnets, with the balance utilizing resistive and permanent magnetic technologies. Superconducting magnets have achieved preeminence in this application because they offer the following unique capabilities:

1. Superconducting windings can have a nearly unlimited ampere-turn capability and can, therefore, provide strong magnetic fields over large volumes economically. Superconducting magnets in routine clinical use todayrange in field strength from 0.35 Tesla to 2.0 Tesla. Resistive and permanent magnets have been applied to MRI only up to the lower end of this range. Imaging at higher fields increases the signal-to-noise ratio which can translate into a reduction of scanning times and an increase in spatial resolution.

2. Superconducting magnets, in general, provide greater field uniformity, allowing imaging over a wider field of view. Weight and power consumption constraints lead to compromises in the useful imaging volume in permanent and resistive magnet designs. Improved field homogeneity has beneficial effects in terms of RF power deposition and in the terms of image quality.

Superconducting magnets provide magnetic fields of unsurpassed temporal stability. Unlike resistive and permanent magnets, superconducting magnets operate in the persistent mode and are isolated from thermal and electrical transients. Consequently, field stability of better than 0.1 ppm/hour is routinely achieved. Field instability would adversely affect image resolution.

4. Superconducting magnets for MRI are highly reliable and easy to operate. Because superconducting magnets are not electrically lossy, power supplies are only needed for initial energization, afterwhich the magnet will continue to operate in the persistent mode. Consequently, they are unaffected by power interruptions. Efficient cryogenic performance (specifications on helium consumption are now typically < 0.4 liter/hour) allows operation in most cases without refrigeration systems. The only maintenance requirement is the periodic topping up of the cryogens.

Reauirements for Further Growth of MR

The MR installed base in the U.S. is fast approaching 1,000 systems. Impressive as the growth rate of the MRI industry has been, the installed base of the competing technology of x-ray Computed Tomography (CT) has risen to nearly four times that of MR. Several advances are neccessary in order for the MR volume to approach that of CT.

The price of an MR system is currently about twice that of CT, which puts MRI at a relative disadvantage over CT. In areas where both modalities are capable of detecting lesions to a comparable extent, it becomes difficult to justify the added cost of an MRI exam. Consequently, the price of an MRI system must be brought down if the industry is to grow to its full potential. Although the magnet price is only about 20% of the selling price of an MR system, the increasingly cost conscious industry demands magnets that are both economical to build and economical to operate. Compact, economical magnets can open up applications for MRI that as yet are not practical.

0018-9464/89/0300-1755$01.00@1989 IEEE

1756

Specific medical applications for MRI are to date less widepspread than are applications for CT. Because MRI yields excellent soft tissue contrast and it is also unobstructed by bone, it was predicted early on that MRI would cut deeply into CT's head and spine caseload, and so it did. In fact, the average mix of patient MRI exa.ms emphasize central nervous system imaging with head and spine exams representing about 7O0I0 of the total exam volume. CT, however, is still preferable to MR in some clinical applications. Yet, with continued improvements in surface coil technologies, MRI is gaining acceptance in neck and extremity imaging. Likewise, rapid imaging techniques which minimize motion artifacts promise to extend applications of MRI into abdominal, pelvic and cardiac imaging. In fact, despite CT's apparent strongholds, it is generally believed that MRI could someday supplant CT entirely.

Investigators continue to search for biological and clinical applications in spectroscopy. Such applications derive clinical information from in vivo MR spectral data. Spectroscopic analysis relies on MR's capability to identify particular chemical compounds by their chemical shift. This unique information is not available from any other modality. Typically nuclei other than hydrogen are used including phosphorus, carbon, flourine and sodium. Likewise, techniques for imaging these various nuclei continues to be developed with specific clinical objectives in mind. However, Since these biological nuclei are less abundant within human tissue and inherently less sensitive to magnetic

fields than hydrogen nuclei, higher field strengths are needed in order to create enough magnetization to detect a signal. The ultimate success of multinuclear magnetic resonance spectroscopy in achieving routine clinical application would have an important effect on the total market requirements for superconducting MR mag nets.

A final impediment to the greater utilization of MRI equipment is the difficulty end users have in finding suitable sites for the magnets. The problem arises from the impact of the magnet's fringe field on the clinical environment. Because magnetic fields can adversely effect pacemakers and cerebral aneurysm clips, general public access to fields above five gauss must be restricted. For a 1 .O Tesla magnet, this means an area approximately 21 m x 17 m falls within the restriction zone and as many as five stories in a multistory hospital are affected. Reduction of the fringe field is possible by use of steel shielding, but the fact remains that MR installations often require special purpose buildings or extensive site renovations to accommodate the equipment. The cost and planning associated with such projects continue to slow the diffusion of the technology.

New Maanets from Oxford

Oxford Superconducting Technology in the US. and Oxford Magnet Technology in the U.K. have brought new magnet designs to the marketplace which address the issues raised above. . .

ComDact Maanets

Compact Magnet

In creating a second generation magnet that offers true cost advantages while not compromising features already accepted by the medical community, Oxford introduced the Compact magnet in 1986 which now is in full production. The Compact magnet design offers reductions in size and weight while also achieving reductions in manufacturing costs and operating costs. Compact magnets operate at fields from 0.35 Tesla to 1.5 Tesla. They were developed to cope with the toughest challenge, that of the mobile environment. Their lighter weight (under 3800 kg for 0.5 Tesla) and reduced size make them easier to install in mobile vans as well as in static sites. Their suitability for mobile use has been verified with tests on shaker tables. They can be ramped to 0.5 Tesla in under 10 minutes. Their helium consumption without refrigeration is the best in the industry, specified at 0.35 liters per hour. Actual performance is typically much better than specification. Addition of a Gifford-McMahon two stage refrigerator can reduce this further to 0.1 5 liters per hours, and eliminate nitrogen consumption. Although the outside dimensions of Compacts are significantly smaller than conventional magnets, the bore size of the Compact is a full one meter, for compatibility with existing gradient coil and RF coil designs and compatibility with possible future gradient coil enhancements. In summary, Compact magnets provide a number of economies. They serve the mob "? MRI

1757

market, Which allows costs to be spreaa among a number of hospitals. They are interchangeable between mobile and static installations, simplifying manufacturing planning. Their smaller size allows savings on material costs. Their efficient use of cryogens reduces operating costs. These economies are achieved without other performance compromises.

PCr \

4 Tesla Maanet

4 Tesla Whole Body Magnet

A new tool for Magnetic Resonance Spectroscopy (MRS) research, to supplement the 2 Tesla High Homogeneity magnets presently in service, is now available. Two 4 Tesla whole body superconducting magnets manufactured by Oxford are currently in operation. The design challenges included:

a) Maintaining the required dimensional accuracy of the windings in the presence of the very large i ntercoil forces.

b) Providing for the safe dissipation ofthe 35 megajoules of stored energy in event of a quench.

These magnets reached 4.2 Tesla without quenching. Intentionally induced quenches at 4 Tesla verified the protection scheme. A key specification for a magnet whose primary use in MRS is the homogeneity. Measured values for these magnets were 6 ppm on a

40 cm diameter spherical volume (dvs), and 0.2 pprn on a 15 cm dsv.

These initial units have a one meter diameter bore, however, Oxford since has designed a premium 1200 mm bore version and also a 850 mm bore version to save on size and cost.

1

Human 4 Tesla 31 P DRESS spectrum

in 10 sec

11- ATP i

4 Tesla Spectrum (Courtesy of General Electric Medical Systems)

Early clinical applications of MRS will most likely be done at 1.5 and 2.0 Tesla, where a significant installed base is available. Research at 4 Tesla should accelerate the developmevt of this potentially revolutionary medical technique.

Active-Shield Maanet

Active-Shield Magnet with refrigeration option

The magnet development having the largest impact on the MRI industry is the Active-Shield magnet. These magnets utilize superconducting shielding coils, at a larger radius than the primary coils, and in series with them, to drastically reduce the fringe field. This approach reduces the volume of space within the 5 gauss line by a factor of 20. The stray magnetic field is contained entirely within the imaging room.

1758

I 21’7’ 0 . 4 rn

1;. Active -Shield \ \

The 5 gauss line of the Active-Shield Magnet compared to a conventional magnet at two different fields: 0.5T and 1 .OT

Since the spring of 1987, Oxford has shipped in excess of 40 Active-Shield magnets at field strengths of 0.5 Tesla and 1 .O Tesla. The 0.5 Tesla Active-Shield magnet has been designed to be rugged enough for mobile use and it is being used as a mobile product. The 1.5 Tesla Active-Shield product is in the final stages of system development and assembly of the first production unit is eminent.

A sagittal image of the brain obtained using a 1 .OT Active-Shield Magnet. (Courtesy of Picker International)

Summary

Magnet technology has kept pace with other technical innovations in MRI, the first large scale commercial application of superconductivity. Further growth of the MRI industry will be aided by innovations in superconducting magnet technology.

Acknowledaements

The work described here was performed by the author’s colleagues at Oxford Superconducting Technology, Carteret, New Jersey and Oxford Magnet Technology, Eynsham, England.

References

1. D. E. Andrews, Magnetic Resonance Imaging in 1987, Advances in Cryogenic Engineering, vol. 33.

2. D. G. Hawksworth, I. L. McDougall, J. M. Bird, D. Black Considerations in the design of MRI magnets with reduced stray fields, presented at 1986 Applied Superconductivity Conference, Baltimore, MD.