Superconductors lower MRI costs

Magnetic resonance imaging (MRI) provides detailed images of organs and structures within the human body without invasive procedures or exposure to radiation. MRI scanners use magnetic fields to line up the spins of hydrogen atoms in the body. Expensive helium-cooled superconducting magnets generate the required field strengths of 1.0 to 1.5 Teslas, resulting in machines that cost $1 to $2 million. The high cost has limited the use of MRI in the current cost-constrained healthcare environment.

With increasing pressure to control costs, manufacturers have started offering less expensive low-magnetic-field-strength (0.2 Tesla and below) MRI scanners. Most of the cost savings comes from using an electromagnet instead of the superconducting one. But apart from the lower cost, these systems give doctors physical access to patients during imaging due to the smaller magnets. These "open" systems also benefit very large or claustrophobic patients, who cannot be scanned in a traditional closed "doughnut" system.

A superconducting coil on a wafer replaces a conventional copper coil int this MRI probe assembly. Liquid nitrogen was used to cool the wafer to 77K in the first prototypes. Subsequent units use a mechanical, closed-loop heat exchanger.

Low-field machines are easier to maintain because technicians don't have to deal with liquid helium. And with lower costs, MRI machines dedicated to imaging extremities such as the knee, ankle, elbow, foot, and hand are becoming more prevalent. However, the lower field strength of these machines results in poorer quality images and longer scan times compared with million-dollar machines.

To address this drawback, Conductus Inc., Sunnyvale, CA, has been working with Siemens AG Medical Engineering Group to develop superconducting receivers for the Siemens Open 0.2T MRI system.

Superconductors are materials that undergo a transformation at reduced temperature that gives them the ability to carry electrical currents without any resistance. Without resistance, electrical signals are not dissipated in the form of heat, so all manner of electrical and electronic devices and components become far more efficient.

An MRI machine's radio-frequency receiver uses specialized copper coils to sense the signals generated by applying a magnetic field to the human body. Low-loss superconducting receiver coils in MRI machines can significantly enhance performance by improving the signal-to-noise ratio (SNR). Higher SNR leads to either increased image quality or faster imaging time.

Using processing techniques borrowed from semiconductor manufacturing, Conductus deposts superconducting material on a sapphire wafer and etches it into a coil pattern using photolithography.

In fact, superconducting coils could improve SNR to the point where low-field machines could produce images similar in quality to those of the more expensive machines. Achieving higher performance in low-field machines could also lead to the use of MRI in a wider range of diagnostic imaging applications, such as routine screening for breast cancer.

"There are a number of ways radiologists can use better signal-to-noise ratios," says Laurie Mann, senior engineer at Conductus. "The most obvious one is to get a better quality image with better resolution. They can also use it to achieve the same-quality image but in a shorter time."

Shortening imaging time is a worthwhile goal. MRIs take from 1 to 30 minutes, depending on the body part, and the patient has to remain still for the entire process. Decreasing imaging time also benefits the bottom line because billing is usually done on a time basis.

"The coil is like a radio receiver," explains Mann. "Hydrogen nuclei are excited by the magnetic field, which lines up their spins. A pulse at a specific frequency causes the spins to persist. As spins decay, or precess, they emit an echo, which the coil receives, as the nuclei relax back to their normal state." Encoding and interpreting the echo creates the image.

Conductus makes its superconducting components from a material called YBCO, mercifully short for yttrium barium copper oxide (Y₁Ba₂Cu₃O₇), which was the first superconductor discovered with a critical temperature above 77K--the boiling point of liquid nitrogen. Previous superconductors had required temperatures of a few degrees Kelvin, which could only be achieved with liquid helium--a more expensive fluid.

Starting with a sapphire wafer as a substrate, Conductus technicians make the coils using the thin-film deposition and photolithographic processing techniques pioneered in the semiconductor industry. Sapphire is a good heat conductor and is structurally compatible with the YBCO, notes Mann.

"We operate the superconductors at the temperature of liquid nitrogen (77K) using a closed-cycle system," says Mann. "There's a working fluid--high-pressure helium--that goes through a compressor and is sent through heat exchangers to remove the heat from the wafer." This technique is more beneficial than an open-loop system from a user standpoint, according to Mann, because it's self contained and the operator doesn't have to replace cryogenic fluid.

Mann and Conductus are now working out the packaging issues. The subassembly has to be insulated from the patient, and needs to maintain a vacuum and a very low temperature. "Cryopackaging is where the design challenges are," says Mann. "We want to make it easy to use, so all a radiology technician has to do is turn the MRI machine on."

A team led by Dr. Jonathan Lewin, director of MRI at University Hospitals of Cleveland, which is affiliated with Case Western Reserve University, is doing preclinical testing of the Siemens Open system with Conductus' superconducting coils.

The open MRI scanner on the left uses a C-shaped electromagnet that allows doctors access to a patient--unlike the traditional closed 'doughnut' configuration.

"We're using our scanner primarily to take advantage of access to the patient for interventional procedures," says Lewin. "We've done about 100 diagnostic procedures using needles to sample or inject or to biopsy areas of the body while somebody is being imaged--an MR image is updated every second."

"In the head and neck," continues Lewin, "it's very difficult to position a needle, and it's difficult to see the detail you need on any other imaging modality."

In addition to doing preclinical testing, Lewin's team is helping develop the superconducting coils in collaboration with Conductus. So far, they've been imaging phantoms--non-human objects such as fluid-filled beakers--to get the system calibrated to fully take advantage of the signal-to-noise capabilities of the coils.

Because of the lack of resistance within the coil windings, you can theoretically have an SNR up to 10 times that of conventional copper coils, says Lewin. "In our setting, it's probably closer to two or three times copper," he notes. "But that still means you can cut your scan time in half or increase your resolution by a factor of two. A small SNR advantage can make a very big difference in imaging--especially in low-field machines."

Imaging phantom subjects using a Siemens Open MRI machine with prototype superconducting receiver coils has resulted in a 100% increase in the signal-to-noise ratio of the system, says Lewin.

"You don't want imaging to take a long time during an interventional procedure," stresses Lewin. "Conventional coil engineers are very happy if they can get a 10 to 15% increase in SNR by a major coil modification. Doubling it is really a huge step. Our hope is that the increased sensitivity will allow us to improve interventional procedures."

Says Conductus' Mann: "We've improved the low-field systems almost to the point where the resolution is as good as that of a high-field system--and we should achieve that goal by the end of the year." Look for low-field MRI scanners with superconducting receivers to be available within a year.

Hurdles to overcome

How superconductors work

H. Kammerlingh Onnes discovered the first superconducting material--mercury--in 1911 when he found that its electrical resistance dropped sharply to an infinitesimal value at 4.2K. The first theory to explain the phenomenon came 50 years later. John Bardeen, Leon Cooper, and Robert Schrieffer proposed a state of matter in which electrons in a superconductor form "Cooper pairs" that cooperate to avoid the collisions and interactions within a metal that lead to ordinary resistance. This pairing up can only occur when the temperature is low enough that the ordinary thermally induced motions of the electrons are sufficiently reduced.

However, this theory doesn't appear to apply to so-called high-temperature superconductors--materials that exhibit superconductivity at temperatures higher than 77K, the boiling point of liquid nitrogen.

In fact, researchers don't know for sure how high-temperature superconductors work, but they do have theories.

In 1988, a team of American physicists led by William A. Goddard, III suggested that when a material becomes superconductive, a few electrons are displaced from the oxygen atoms, which causes the atoms to become magnetic. Adjacent copper atoms are pulled into line by the magnetism and lose electrons into the quantum pockets, or "holes," created by the displaced oxygen atoms. This process creates new "holes," and electrons continue to flow through the material indefinitely as long as the transition temperature stays constant.

This theory, which has so far proved accurate, predicts a transition-temperature limit of about 200K for the copper oxide group of superconductors. A material could be cooled to that temperature by dry ice--which is even cheaper and easier to maintain than liquid nitrogen.

Other applications