MRI Sheds Its Shielding and Superconducting Magnets (2024)

Magnetic resonance imaging (MRI) has revolutionized healthcare by providing radiation-free, non-invasive 3-D medical images. However, MRI scanners often consume 25 kilowatts or more to power magnets producing magnetic fields up to 1.5 tesla. These requirements typically limits scanners’ use to specialized centers and departments in hospitals.

A University of Hong Kong team has now unveiled a low-power, highly simplified, full-body MRI device. With the help of artificial intelligence, the new scanner only requires a compact 0.05 T magnet and can run off a standard wall power outlet, requiring only 1,800 watts during operation. The researchers say their new AI-enabled machine can produce clear, detailed images on par with those from high-power MRI scanners currently used in clinics, and may one day help greatly improve access to MRI worldwide.

To generate images, MRI applies a magnetic field to align the poles of the body’s protons in the same direction. An MRI scanner then probes the body with radio waves, knocking the protons askew. When the radio waves turn off, the protons return to their original alignment, transmitting radio signals as they do so. MRI scanners receive these signals, converting them into images.

More than 150 million MRI scans are conducted worldwide annually, according to the Organization for Economic Cooperation and Development. However, despite five decades of development, clinical MRI procedures remain out of reach for more than two-thirds of the world’s population, especially in low- and middle-income countries. For instance, whereas the United States has 40 scanners per million inhabitants, in 2016 there were only 84 MRI units serving West Africa’s population of more than 370 million.

This disparity largely stems from the high costs and specialized settings required for standard MRI scanners. They use powerful superconducting magnets that require a lot of space, power, and specialized infrastructure. They also need rooms shielded from radio interference, further adding to hardware costs, restricting their mobility, and hampering their availability in other medical settings.

Scientists around the globe have already been exploring low-cost MRI scanners that operate at ultra-low-field (ULF) strengths of less than 0.1 T. These devices may consume much less power and prove potentially portable enough for bedside use. Indeed, as the Hong Kong team notes, MRI development initially focused on low fields of about 0.05 T, until the introduction of the first whole-body 1.5 T superconducting scanner by General Electric in 1983.

The new MRI scanner (top left) is smaller than conventional scanners, and does away with bulky RF shielding and superconducting magnetics. The new scanner’s imaging resolution is on par with conventional scanners (bottom).Ed X. Wu/The University of Hong Kong

Current ULF MRI scanners often rely on AI to help reconstruct images from what signals they gather using relatively weak magnetic fields. However, until now, these devices were limited to solely imaging the brain, extremities, or single organs, Udunna Anazodo, an assistant professor of neurology and neurosurgery at McGill University in Montreal who did not take part in the work, notes in a review of the new study.

The Hong Kong team have now developed a whole-body ULF MRI scanner in which patients are placed between two permanent neodymium ferrite boron magnet plates—one above the body and the other below. Although these permanent magnets are far weaker than superconductive magnets, they are low-cost, readily available, and don’t require liquid helium or to be cooled to superconducting temperatures. In addition, the amount of energy ULF MRI scanners deposit into the body is roughly one-thousandth that from conventional scanners, making heat generation during imaging much less of a concern, Anazodo notes in her review. ULF MRI is also much quieter than regular MRI, which may help with pediatric scanning, she adds.

The new machine consists of two units, each roughly the size of a hospital gurney. One unit houses the MRI device, while the other supports the patient’s body as it slides into the scanner.

To account for radio interference from both the outside environment and the ULF MRI’s own electronics, the scientists deployed 10 small sensor coils around the scanner and inside the electronics cabinet to help the machine detect potentially disruptive radio signals. They also employed deep learning AI methods to help reconstruct images even in the presence of strong noise. They say this eliminates the need for shielding against radio waves, making the new device far more portable than conventional MRI.

In tests on 30 healthy volunteers, the device captured detailed images of the brain, spine, abdomen, heart, lung, and extremities. Scanning each of these targets took eight minutes or less for image resolutions of roughly 2 by 2 by 8 cubic millimeters. In Anazodo’s review, she notes the new machine produced image qualities comparable to those of conventional MRI scanners.

“It’s the beginning of a multidisciplinary endeavor to advance an entirely new class of simple, patient-centric and computing-powered point-of-care diagnostic imaging device,” says Ed Wu, a professor and chair of biomedical engineering at the University of Hong Kong.

The researchers used standard off-the-shelf electronics. All in all, they estimate hardware costs at about US $22,000. (According to imaging equipment company Block Imaging in Holt, Michigan, entry-level MRI scanners start at $225,000, and advanced premium machines can cost $500,000 or more.)

The prototype scanner’s magnet assembly is relatively heavy, weighing about 1,300 kilograms. (This is still lightweight compared to a typical clinical MRI scanner, which can weigh up to 17 tons, according to New York University’s Langone Health center.) The scientists note that optimizing the hardware could reduce the magnet assembly’s weight to about 600 kilograms, which would make the entire scanner mobile.

The researchers note their new device is not meant to replace conventional high-magnetic-field MRI. For instance, a 2023 study notes that next-generation MRI scanners using powerful 7 T magnets could yield a resolution of just 0.35 millimeters. Instead, ULF MRI can complement existing MRI by going to places that can’t host standard MRI devices, such as intensive care units and community clinics.

In an email, Anazodo adds this new Hong Kong work is just one of a number of exciting ULF MRI scanners under development. For instance, she notes that Gordon Sarty at the University of Saskatchewan and his colleagues are developing that device that is potentially even lighter, cheaper and more portable than the Hong Kong machine, which they are researching for use in whole-body imaging on the International Space Station.

Wu and his colleagues detailed their findings online 10 May in the journal Science.