Towards Low Cost Biomedical Imaging

Medical imaging is one of the very best applications of technology — it allows us to peer inside of the human body without actually performing surgery. It’s non-destructive testing to the extreme, and one of the more interesting projects we’ve seen over the past year uses AC currents and an infinite grid of resistors to image the inside of a living organism. It’s called Spectra and it is the brainchild of [Jean Rintoul]. Her talk at the Hackaday Superconference is all about low cost and open source biomedical imaging.

We’ve seen some interesting medical imaging hacks in the Hackaday Prize over the years. There have been vein finders and even a CT scanner, but when it comes to biomedical imaging, the Spectra project is something different. Right now, it’s just good enough to image organs while they’re still inside your body, and there’s still a lot of potential to do more. Let’s take a closer look a how this works.

The idea behind Spectra is to use AC waves through a medium (a fruit, or a person, or a rat) and use tomographic techniques to image  the interior. Specifically, this is called Electrical Impedance Tomography, and you can surprisingly build a small version of this with very, very minimal hardware.

EIT electrodes on the chest of a 10-day-old baby

Electrical Impedance Tomography works due to the fact that different tissue types have different resistances. By sending an AC wave (around 10kHz or so) through a body, the inside can be reconstructed. Everything from lung volume measurements to muscle and fat mass to cancers can be detected with this technique. Yes, you’ll still need a tech or an MD to interpret the data, but this is a very inexpensive way to image the human body compared to current technology.

As for why this matters, this is a fantastically cheap way to look inside a body. Preventative care in medicine doesn’t have an affordable imaging solution; if you need an MRI or CT scan, odds are you’re only going to get that when you’re sick. A CT scanner will cost millions of dollars and MRI machines are even more expensive. [Jean]’s prototype electrical tomography setup can be build for well under a thousand dollars — and it looks she even has a crowd funding campaign in the work to get more test hardware out into the world. Making this both open source and low cost makes the technology more accessibility. We hope this ultimately leads to better medical care for everyone, and [Jean] is on the cutting edge of this advancement.

21 thoughts on “Towards Low Cost Biomedical Imaging

  1. First noted EIT in a piece of medical equipment about 15 years ago. And it did not seem to be well like by practitioners. Am guessing that its scope was severely limited by the FDA, and that the available embedded processing power was not yet up to the task.

    The value of the FOSS stuff would be directly proportional to FDA/NRCan/EU MDR approvals. And FWIW, many African and other third-world states have harmonized with EU directives. Much of the cost reduction would be a body of code and some example hardware where assessed per the ISO13485 and 82304 mess; and this was not addressed. Look at ‘software as a medical device’ in your favorite search engine.

  2. I would argue that it is not one of very best applications of technology, I can easily think of over a dozen better applications: cooking, sanitation/toilets, water purification, agriculture, housing, refrigeration, wheeled transportation, printing, electric lighting, currency, antibiotics, calendars/timekeeping and maps/navigation. Hell, I bet I could come up with a list where if items were as detailed and specific as “medical imaging” then medical imaging would not even crack the top hundred.

    Hell, even if you looked at technology that is almost exclusively applied to medicine, it might not even crack the top ten, for example: stethoscope, sphygnomonometer, scalpel, blood transfusions, IV drugs, antibiotics, gloves, masks, antiseptics, blood/urine testing, vaccines, casts/splints/braces, clean bandages, medical thermometers. I bet more lives would be lost each year if we lost almost any of the aforementioned technologies than if we lost medical imaging.

    It is infuriating how simple, widely used and old technologies are often considered irrelevant or insignificant when compared to advanced technologies when the opposite is often more true.

      1. Agreed there are lots of useful applications of technology. People have applied EIT to industrial pipes and oil and gas industry too. Medical Imaging is just a previously very expensive area not accessible to hackers. If it was made more accessible to people to improve and build upon, perhaps we could use it more!? 1/3 of the world has no access to any kind of medical imaging right now, and even in the US we use it fairly conservatively due to price. If there was no barrier to doing a scan, we could build much better datasets and let machine learning find anomalies in our yearly check up.

    1. Thanks for clearing that up I’ll bet many were confused before you helped out. The electrodes are going around the chest, not the abdomen, not the head.
      They need to encircle the chest to make the tomographic image. The chest is between the neck and abdomen. They are on the back and front of the chest. Neat, huh?

    1. Ha yeah, value of medical care in A CERTAIN COUNTRY AHEM is completely divorced from the value of the labor, equipment, and materials I’m afraid. Ever seen that graph of life expectancy versus healthcare costs? There’s one outlier so extreme it would be hilarious if it weren’t so sad.

      Still, great for developing countries now and probably great for medicine in general once the technology matures to its full potential. Pretty impressive stuff.

      1. Exactly the idea. Sometimes developing countries can even leapfrog first world ones when they take a different approach. It seems like the incumbents in US healthcare aren’t as aligned with patient needs in delivering quality healthcare. It would be great to explore alternate business models for innovating in health technology! I think it’s needed no where more than the US!

      2. You’re not paying for labor, materials, and equipment that much. The big cost is in R&D, testing, more testing, still more testing, paperwork, markups, insurance, and lawyers. Then more lawyers when one doesn’t work right, which increases the insurance, lawyers, testing, and paperwork.

        You just want a tomograph of something, no problem, I was doing that 20 years ago in rock and dirt.

  3. Great presentation, this young lady is sharp. Really got me excited… the hackaday hyperspectral imaging didn’t quite wake me up enough… or I hadn’t found a cleaner food source at the time yet…

    I enjoyed spectroscopy for sure… especially systems people couldn’t implement prior to me working on and being able to implement more valid holistically than most in the industry… if not anyone in the industry… have ever observed.

    There was a recent instructable where I wound up registering on Bruker to download their CT related software as well as download the instructable and other scanner info referenced in the comments section, to have mirrored.

    The 3D and 4D (4th being time in this note, i.e. dynamic) was the level I didn’t quite apply enough and wanted too. I was like teased with only demonstrating pilot Terahertz, Hyperspectral Imaging NIR, FTIR Microscopy and other Electromagnetic Spectrum (EMS) system ranges to image quality issues. Wasn’t only due to sabotage… though that was what I observed the most with my last Director acting like I had to learn to deal with failure of others to move up in that organization. Anywho…

    Tomography and Holography comprehension more clearly so I can visualize crystal clear mentally throughout the EMS applications is a goal of mine. I still have gaps in thorough imagining in my mind.

    This reminded me I need to download what was referenced in the presentation since I am not finding at the moment that I did.

    Here is the instructable and in the comments there is the Greek gentleman’s system too he has referenced you can find:
    https://www.instructables.com/id/Desktop-CT-and-3D-Scanner-With-Arduino/

    Good call on how simple practicing medicine is really. Reminds me of pattern recognition modeling and how easily that can be corrupted with the wrong operator or training set. I trained in first responding also… felt excellent how well I did compared to those who were licensed in health care. Really modern also in regards to technology and the history of intellectual thinking to comprehend Why what is being practiced effectively is What, Where, When and How… I guess the Who when unknown also. Technically, there are ancient medicinal protocols that are effective… though I’m not aware of other than in modern times being more aware and knowing Why in detail. With more thorough diagnostic methods… we will know with more certainty the variables to more valid lean sigma health care.

      1. Super interesting video. It’s so powerful to be able to see inside non-invasively. I agree, it would be excellent if we could easily scan any body part cheaply anytime – it would totally revolutionise medicine.

  4. Well, not sure how much “cheap” imaging would really change the cost, considering a huge portion of the costs are what the company pays for liability, regulatory issues, lawsuits, insurance, etc. This is where the bulk of the cost comes from.

  5. I honestly can’t see how this would be significantly cheaper to build than ultrasound.

    Hospitals pay >$5k for just a 16 probe EEG machine and $40k for an ultrasound machine that not only provides 4x better resolution than her fetal image, but can also show blood directional flow, produce 3D images, provide accurate measurement of organs and distinguish between important cell types. ETI could do one of these things.

    For ETI to improve, it requires higher frequencies and more contacts. So essentially, a few dozen of the ultrasound machine’s most expensive component. Sure, ETI wouldn’t need the ‘expensive’ $3k ultrasound probes, but it would need 10x as much ~$30k front end circuitry.

Leave a Reply to Ostracus Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.