To say that the process of installing a magnetic resonance imager in a hospital is a complex task is a serious understatement. Once the approval of regulators is obtained, a process that could take years, architects and engineers have to figure out where the massive machine can be installed. An MRI suite requires a sizable electrical service to be installed, reinforced floors to handle the massive weight of the magnet, and special shielding in the walls and ceiling. And once the millions have been spent and the whole thing is up and running, there are ongoing safety concerns when working around a gigantic magnet that can suck ferromagnetic objects into it at any time.
MRI studies can reveal details of diseases and injuries that no other imaging modality can match, which justifies the massive capital investments hospitals make to obtain them. But what if MRI scanners could be miniaturized? Is there something inherent in the technology that makes them so massive and so expensive that many institutions are priced out of the market? Or has technology advanced far enough that a truly portable MRI?
It turns out that yes, an inexpensive MRI scanner is not only possible, but can be made portable enough to wheel into a patient care room. It’s not without compromise, but such a device could make a huge impact on diagnostic medicine and extend MRI technologies into places far beyond the traditional hospital setting.
Align, Flip, Repeat
We’ve previously covered the basics of MRIs and why the machines are so loud. As a quick recap, recall that MRI scanners are essentially huge, powerful magnets that align the spin axes of all the protons in water molecules in the body tissues with the long axis of the body. Powerful radiofrequency pulses from antennas inside the bore of the machine perturb the spins, and sensitive receivers listen for the faint RF signals that result when the pulses end and the proton spin returns to the aligned states. Additional coils inside the bore, which are called gradient coils and make the characteristic noises of an MRI, shape the main magnetic field slightly and allow it to be rastered around the body, resulting in spatial data that, coupled with the RF signal, is turned into the detailed images we’ve all marveled at.
https://www.youtube.com/watch?v=exgr-ZNkIAM
The problem with miniaturizing an MRI scanner is that for these devices, size really does matter. The more powerful and even the magnetic field generated by the machine, the better the resolution of the instrument. The typical scanner in a hospital MRI suite has a magnet that operates at 1.5 Tesla, a field strength that requires superconducting coils cooled to nearly absolute zero with liquid helium. The magnet accounts for most of the bulk and weight of a large MRI scanner.
While superconducting magnets are standard in medical MRI scanners, they are finicky beasts. They require an uninterrupted power supply to ensure that the magnet doesn’t quench, which occurs when the coils exceed their superconducting temperature and increase in resistance. When this happens, the current flowing through the formerly zero-resistance coils creates excess heat, which boils off all that expensive liquid helium. Superconducting magnets have revolutionized MRI, but have also anchored scanners firmly in the MRI suite.
Permanent Magnets Work Too
Luckily, superconducting magnets are not strictly required for usable MRI scans. Clinically useful images can be obtained using any magnet that can create a strong, even field. Some large clinical machines still use permanent magnets or resistive electromagnets, but these have disadvantages. Permanent magnet scanners tend to have weaker fields and thus produce lower resolution images, while non-superconducting electromagnets have massive electrical needs and require constant cooling.
Still, permanent magnets are an attractive way to reduce the size of an MRI scanner, especially given advances of the last few decades in magnetics technology. The same rare earth elements that are being used to make the powerful magnets in everything from the brushless DC motors in drones to the powerplants of electric vehicles have found their way into permanent magnets powerful and compact enough that a bedside MRI scanner is possible.
Hyperfine Research, a company formed in 2014 by scientist and entrepreneur Jonathan Rothberg, recently won initial FDA approval for its Hyperfine point-of-care (POC) portable MRI system. Utilizing a permanent magnet of only 0.064 T, the scanner strips away everything not dedicated to the production of an image. There’s no patient handling system; rather, the machine is wheeled to a patient room, where the adjustable bed is used to position the patient. Power requirements are modest enough that the machine just plugs into a regular outlet. At 1,400 pounds (653 kg), it’s still a bulky beast, but it’s a far cry from the multiple ton behemoths that usually grace MRI suites.
Bring the Machine to the Patient
To be clear, Hyperfine is not marketing its POC machine as a replacement for standard MRI scanners. The POC scanner has a far lower resolution than a traditional MRI, and the images it produces would never take the place of those produced by a machine with a powerful superconducting magnet. And while Hyperfine’s marketing literature is quick to point out that the machine could bring MRI to underserved populations, like those in developing countries, chances are pretty good that the $50,000 price tag — peanuts compared to an MRI suite — will prove prohibitive.
The main niche that will likely be filled by this machine is revealed by the partner hospitals that are currently testing the scanner. Of the six hospitals, three have the machine in the neurology department, two are in the emergency department, and one is in pediatrics. The ability to quickly wheel the machine into a patient room and perform a survey MRI scan of the head will be invaluable to these specialties. ER and ICU patients, especially trauma patients, are often too sick to move to an MRI suite, so the ability to assess a traumatic brain injury or a stroke rapidly and without leaving the department could be critical to better outcomes.
It remains to be seen if the low-resolution images the Hyperfine POC system can produce will be worth the price of admission. But if the last decade has taught us anything, it’s that technological leaps, especially in magnetics and computers, come when you least expect them, and often lead to innovations that make what was once science fiction into everyday products. With advances in superconductors, we may even someday see POC MRI scanners that rival today’s big magnet scanners, and at a fraction of the cost.
[Featured images: Hyperfine]
Inevitable progress.
MRI was first known as NMR and used mostly in chemistry. Nuclear Magnetic Resonance. Sample size was just a test tube and the unit was half the size of a home refrigerator. No surprises here folks…. just progress by from dropping the habit of thinking bigger is better and looking back to the origins of the process back when such magnetic fields were just being developed.
Yup, “Nuclear” as in the nucleus of hydrogen atoms, the way it functions. Not as in “plutonium” and “Nagasaki”, but people are stupid, and rather than waste millions of doctor-hours explaining the difference, they just changed the name.
If you can fit your head into a test tube I’ll agree with you.
The volume of the (as much as possible) homogeneous magnetic field in this thing is about 4 orders of magnitude bigger then a regular test tube.
Is a magnetic field needed? Is there any nonlinear effects from the magnetic part of two different radio waves?
Yep, Nuclear Spin Resonance (NSR) fun. I guess not marketing as either NMR or NSR was “politically” correct as MRI was as a term bolstering for sales and consumers demand.
Technically, nuclear magnetic resonance isn’t limited to only hydrogen atoms nucleus. All atomic nuclei have nuclear and electronic spin though I forget the details of why some can be measure more than others.
Furthermore, I’m amazed at finally the advancements have been made to make bench top NMR’s that don’t require liquid gas cooling. Seems the same advancement has to be made with MRI’s. Seems even there is potential for zero-field (not as more likely) and earth field (way more likely) NMR/MRI too.
On another note… NIR is moving forward also with some awesome progress by the OpenWater Company team:
https://www.openwater.cc/
You forgot to mention sometime HaD contributor and all-round accomplished guy [Gregory L. Charvat] is a cofounder of Hyperfine.
That really ought to have been mentioned. But still, nice hack, Greg!
Yeah, we absolutely should have mentioned it!
(I didn’t know. Sorry.)
Understandable. They pivot so fast it’s hard to keep up.
Liquid nitrogen superconductors have been around since the 1980s. I thought by now some one would have marketed on that didn’t need 4K.
They did. There were rare earth magnet MRIs available back in the 90s at least (when I first came across them, Toshiba IIRC ) – used for patients where the normal small tube was restrictive or the high field caused problems with other equipment. But I don’t recall them being very portable.
Yes, the Japanese pioneered low cost, small MRIs some time ago. Don’t know the exact date. Japanese doctors complained that the big MRIs in hospitals were too expensive and hard to get appointments for their patients; they asked MRI companies to come up with something smaller and cheaper that they could use in their offices. The companies did so and the cost was about $50,000. That meant a GP could pay it off in a year or two. Having an MRI to “look” inside a patient is a transforming experience for most doctors. It’s true the resolution isn’t as good, but they didn’t care; just having that look inside was valuable.
All high temp superconductors we know today are ceramics. Very brittle ceramics. It’s one thing to make a fist-sized puck and a totally different one to make a 1m (3-ish feet) bore ring magnet (the forces from the magnetic field would break it).
There are teams working on ways of making large, high temperature superconducting magnets. Not just for MRIs, but if fusion power is ever to be a thing, those would help. A lot.
There are indeed people working on this a group in wellington New Zealand has built several MRI system with high temp superconductors not full body but for extremity’s like legs and arms. https://www.wgtn.ac.nz/robinson/research/magnet-systems
I happen to share on office with some of the people building theses systems.
In regard to the weight of current MRI machines.
When Mayo Clinic was planning their Gonda building, they were targeting for a final height of 20 stories.
But, they specified I-beams for a 65 story building to compensate for the added weight of the CAT and MRI scanners
the completed building would hold.
Heh, that’s probably why the MRI dept is in the basement in a lot of older hospitals.
Come to think of it, another one I’ve been to kind of cheated, they have a sort of extended corner to the new building where their MRI is. Like a tower corner, but not higher than rest of building, and the building is recentish, only a decade old. So I suspect that they put extra strength and structure just in that one corner for all the heavy stuff and the rest of the fairly large area floors on the same levels are standard strength instead of 3x.
Also in the basement to limit vibration, and make it easier to shield the room from RF interference.
Don’t forget “Extremity MRIs”; used for arms, legs, hands, and feet.
https://4rai.com/blog/3-types-of-mri-machines
Does anyone else remember the SciAm Amateur Scientist column for a home-built NMR device back in the late ’60’s? I built it. It was the coolest thing I ever made in my youth.
https://hackaday.io/project/1376-pyppm-a-proton-precession-magnetometer-for-all
Was it something like this?
found it. april, 1959.
Not quite the same, as it used a large magnet.
Probably one of the projects [here](http://www.ke5fx.com/stong.pdf)…
One of the projects is called “Transistorised drive for telescopes”.
Which is quite possibly the most 1950s amateur scientist thing I’ve ever heard.
Wow, that’s one of the best PDFs I’ve seen on HaD!
The only thing I could think of while looking at that picture was “get your ass to mars”
“Has Portable MRI got to the floor yet?”
“Uh, no, it looks like it’s stuck to the elevator.”
“You mean stuck _in_ the elevator?”
“No, to the elevator. And three compressed gas tanks, a Mayo stand, and a construction worker’s steel-toed boots.”
And the poor dude with the metal plate in his arm from the snowboarding accident…
In the early days of YUGE magnets, one getting ready to be shipped, pulled the forks off of a forklift and pinned a worker between a fork and the magnet, breaking his leg.
An EMT responding was surprised the scissors he was using to cut the victim’s pants, kept jumping out of his hands and sticking to the magnet.
No, I do not believe this is just an “urban legend”.
What is the wattage for a machine of this size? I mean, do you need 220v in order to power this thing?
Well, their blurb says it “plugs into a standard wall outlet”, and since these guys are plainly American, I’m sure they mean the usual NEMA 5-15 120V receptacle, from which they can suck 1.5 kW.
A full-size MRI I’m familiar with required one 50 kVA supply for the MRI systems and a second 50 kVA for the cooling equipment. The battery for its UPS filled a large room and used 48 (car battery size) lead-acid batteries. Peak output of the gradient amplifiers was 750 kW. The RF amplifier can hit 35 kW. Current in the main (superconducting) magnet was 350 amps (though cost nothing once charged). It cost over $100 per day in electricity to run.
This little machine is a tenth the size, does not require that much cooling, and won’t require that much power for either gradients or RF, but the peak power requirements are still very likely in the multikilowatt range. If there’s enough power supply capacitance, or a small battery, it’s certainly possible the average power could be low enough to come from a single 120V 12.5A socket.
Portable CT scanners already in the market use a small battery pack for exactly this reason, and run off a normal 120V socket, even though their peak power requirements are in the 30-40 kW range for a few seconds for a scan.
I’m just guessing here, but I think a horse-pistol room would have a 120 V 20A sockets available.
You find a way to make it so I don’t lose my mind when going inside one of these things, you’ll be my best friend.
Diazepam. Or Chloryl hydrate. Propofol works too, but probably overkill.
An edit button please, HaD. It’s “chloral hydrate”
… It really doesn’t matter at all if it’s portable or not if they don’t get rid of the ludicrous licensing fees.
“this is a life saving miracle of technology”
…. too bad nobody can afford it, and literally nobody can own one.
MRI machines are cheap, the literal wallet rape and the inability for anyone (except the patent holder) to actually own one is a travesty.
Interesting. What are the license fees you mention, and how much are they?
There are 13 000 MRI systems in the USA, and about that many again in the rest of the world. It’s clear that *somebody* owns at least one.
But, please, tell us: How much would a MRI system cost if you were to engineer, build and sell it? Be specific on the specifications so we can compare with current marketed systems.
Mitsubishi made what was supposed to be the first fixed magnet MRI that I worked on installing in an 18 wheeler van as an assistant engineer. The magnet weighed 60k pounds and unshielded would wipe out your credit cards at 60 paces or yank a tool out of your hands and take a block and tackle to remove it. But the shielding in both magnetic and emf was good enough to get good MRI scans driving in traffic. By moving around many small towns had access to an MRI that could never pay for one that was permanent. This was many years ago ~1990ish
Why does the magnet have to be single solid doughnut construct? Why can’t an electromagnet be used instead if what you need is a uniform magnetic field?
The magnet used in this portable scanner is not a ‘doughnut’ (solenoid) type: it’s a pair of large flat disk-shaped permanent magnets, one above and one below the scanning volume. No power required. Meaning it can’t be turned off either, nor does it have very high field.
The solenoid-type magnet used in more conventional MRI systems IS an electromagnet. Early systems did use ordinary copper wire and required prodigious amounts of electricity to work, and needed the corresponding cooling systems to remove the heat. They were limited to quite low fields (0.15T or so) because of this.
Superconductors permit much higher current and field, and require no power once charged, but do need cryogenic temperatures to work, which does require cooling — once cooled, a magnet needs about 0.5 watts of cooling to stay at 4.2K, and takes about 50 kilowatts from the wall to do it. The newer superconductors like MgB2 promise being able to run at 20K — not even close to liquid nitrogen’s 77K, but they would relax that cooling requirement quite a bit.
Damadian’s Fonar system used a huge (>100 ton) permanent magnet that didn’t require electricity or much cooling, but still was fairly low field but had siting and cost difficulty because of its weight.
Who cares about portability. Bring the cost of having one down to a reasonable $200 to $300. I had an MRI on my knee two weeks ago. Cost $6400.00!!!!! What. For one MRI!!!! No wonder medical insurance is so stupidly high in the US. Costs are totally out of control.