Best Product Entry: A HSDK For Ultrasound Imaging

As an entry into this year’s Best Product portion of the Hackaday Prize, [kelu124] is developing a hardware and software development kit for ultrasound imaging.

Ultrasound is one of the primary tools used in modern diagnostic medicine. Head to the doctor with abdominal pain, and you can bet you’ll be seeing the business end of an ultrasound system. While Ultrasound systems have gotten cheaper, they aren’t something everyone has in the home yet.  [kelu124] is working to change that by building a hardware and software development kit which can be used to explore ultrasound systems. This isn’t [kleu124’s] first rodeo. HSDK builds upon and simplifies Murgen, his first open source ultrasound, and an entry in the 2016 Hackaday prize. [kelu124’s] goal is to “simplify everything, making it more robust and more user-friendly”.

The system is driven by a Raspberry Pi Zero W. A custom carrier board connects the Pi to the pulser block, which sends out the ultrasonic pings, and the analog front end, which receives the reflected signals. The receiver is called Goblin, and is a custom PCB designed [kelu124] designed himself. It uses a variable gain amplifier to bring reflected ultrasound signals up out of the noise.

A system like this would be a boon both to hackers and medical professionals working in the field. Ultrasonics can do more than just imaging. You can decrease healing time with ultrasonics, or even levitate things!

33 thoughts on “Best Product Entry: A HSDK For Ultrasound Imaging

  1. It would NOT be a boon to medical professionals, they would not be allowed to use it, even if they were the sort of specialist that had training and qualifications related to the use of the ultrasound. Sorry but that is a fact, nothing to do with the project, that is just how that highly regulated and “closed shop” profession operates. I see more use for it in science labs actually.

    1. Actually it could be used. Some years ago I developed an improved portable ECG data acquisition platform for a Cardiology clinic. After the final product came out, we had to get all sorts of certificates, and it was tested by the docs for half a year, but they accepted it. I think the secret here is to create something at least partially consulting with a medic.

      1. Medical Equipment manufacturing requires a few things.
        1.) All components have a documentation trail compliant with local medical standards, and chip supplier licensing restrictions.
        e.x. Power supply modules are expensive, but it is where 99% of consumer product people fail lab test.

        2.) The plastics manufacturer has medical device certification. Many will tell you anything to get your business, but you will only find 1 or 2 local factories that can legally create sterile packaging.

        3.) Clinical therapeutic use is very different from medical use, and there are very good reasons hospital equipment looks like it is from the 90’s. Not all turn-key assembly houses are ISO certified for medical devices, and not all certification labs can handle this device class (it is a legal nightmare if something goes wrong).

        4.) Many engineers think they are qualified to do anything, but they would be wrong. This area is a specialization, and people can actually die if you get it wrong (in reality, given enough time some sick people will die anyway). Thus, often it is unpopular given a design process may take several years to complete.

        5.) Having clinical trials by actual medical Doctors is not just a good idea, it is a legal requirement in some areas.

        This particular area of study is not fun, and this project is nowhere near legal for a clinic.

        1. “…This particular area of study is not fun, and this project is nowhere near legal for a clinic…” .. agreed.. however, neither is a ballpoint pen legal for use in an emergency tracheotomy.

          Given the choice, I’d skip the legal formalities if the alternative is death. I can see this project inspiring others to explore ultrasound systems, whether it leads directly to a “legal” medical device or not, time will tell. This is, after all hackaday, not Philadelphialawyeraday.

          1. Although I may appear as an “Idiot” to the uninitiated… the advice given will improve the probability of their projects success. I admit I’ve only launched 14 products in my career, but have witnessed countless others fail hard for simple mistakes.

            Did anyone warn the amateur about deep tissue burns from high powered ultrasonics yet…?
            You may be mistaken if you think this is just a legal concern.

        2. Medical devices in the US (and the parts of the world which use the FDA regs as their standard) require a lot more than that. They are regulated by 21 CFR part 11. You have to prove it’s safe and effective, do detailed risk analysis, thorough verification and validation, careful control of your entire supply chain, and mountains of documents. People ask why medical devices cost so much compared to the cost of the stuff that goes into making them, and that’s why. It’s also why the stuff that gets to market usually works and why it makes the news when they actually harm someone in a serious way.

          Does it hinder the development of new stuff? Sure. Do I want it to happen any way? Oh, hell yes.

          – former medical device guy

          1. Not to mention there is a whole world out there that doesn’t have access to pricey medical equipment. I can see medical charities operating in third world countries seeing something like this as a viable alternative to “tell me where it hurts” diagnostic techniques. While not legal for use in the US, the US is NOT the whole world. Innovations like this can also serve to apply downward pressure on the cost of legal devices when doctors and clinics start questioning the heretofor unquestionable subject of price.

    2. Smart phones isn’t approved for medical use and a doctor used it with cardboard shroud to examine a 3D image before a major operation.

      3D printing is still mostly used everywhere but medical and once in a while I hear of 3D printed part being used.

    3. As a medical professional travelling the world, and at one time involved in one of Kelu’s projects (Thanks for the well wishes, amigo! Things are still rough, but we’re living life to the fullest. Also, greetings from Azerbaijan and Doha!) I can’t tell you how much of a boon this would be for any of us that get out into the field.

      Out in the mud and muck, we’re literally stuck with what we can bring with us. It isn’t a matter of getting a patient to a higher level of care. In many cases there isn’t a higher level of care available. And there is simply no way that I can afford to bring a $4,000 pocket sono-imager into the field (forget the $40,000 unit I tested a couple of years ago).

      But give me an inexpensive one? Something portable, that I can plug in to my phone? Hell yes.
      Not certified by the FDA? OK. I promise not to use it on a patient that will complain to the FDA.

      The established medical infrastructure is insanely difficult to work in. Yet the vast majority of the population on this planet will never benefit from an established medical infrastructure (except through donations with a lack of technical support or through back-channel/black market purchases of dubious quality).

      This is a stupidly simple proposition: Take an overly expensive technology, make it work, and get it in to the hands of people who can help the under-served.

      Better yet, open source it. Get it out into the public, and look for those constructive comments to make it better.

      Kelu124, again, my hat is off to you. My apologies for not helping more, but life really did get very crazy.

      Continue to do great things, my friend!

      1. Even then, as a person who has observed what happens in places like Central America, I can tell you that the “establishment” reaches far out into most places and you really do have to “go bush” before you are not under their control and scrutiny. Even then before you leave to go into the field you may find yourself having to answer questions about what exactly you are doing. I’m not sure if it is paranoid security services or petty public servants but they do place control ahead of the welfare of their most needy.

      2. As long as you are okay with accepting full responsibility if anything bad happens to the patient. Who is going to ensure calibration and accuracy? Do you trust this instrument such that you don’t have to verify it at least on a week to week basis? Who will be blamed if it leads to a misdiagnosis? Who will ensure bugs get fixed if a fault is found? Who will be tasked to propagate the changes and ensure compliance?

        My point is that I am all for a cheaper and easier to use/obtain medical device but I need to protect my patients above all else. I am not going to subject my patients to something that does not have traceability and accountability. That is just immoral. There must be a clear and orderly entity to assign blame and obtain compensation from that will not arbitrarily go away on a whim.

        Projects of this class are very interesting and if they want to go the route of FDA approval and testing in clinical trials then I am willing to offer help. Just don’t assign the label that it is a cheaper medical device until it has the same certifications and proven to be non-inferior.*

        * As an aside, I find lots of projects on Hackaday touting themselves as cheaper medical devices. I think this is just plain wrong. You need to have an equivalent comparison as in you need to go through the process they have gone through. Besides certification and testing, one often overlooked aspect is component selection as a large amount of off the shelf components are not approved for medical applications in the first place. You often have to specifically look for medical grade components which cost many times what a normal component will cost or at least get approval from the manufacturer in writing as well as their traceability/compliance documentation.

        1. I guess I’ve just sat by the bedside too often, wishing I had a piece of gear that could give me better information. That better information, in the context of the totality of the patient, may make a treatment decision easier/more accurate/less painful.
          If you’re a medical professional of any experience, you’ve had to confront the possibility that the data in front of you is not the data you think it is (wrong patient, improperly calibrated systems, circumstantial perturbations, etc).
          The point being: You shouldn’t be an idiot and take any data in isolation. We don’t cath every patient with chest pain.
          Using an ultrasound is a great example. It’s minimally invasive with no documented lasting effects that I’m aware of. It can make the difference between blindly going after a foreign body or tapping a fluid space.
          Do I need an ultrasound to do a pericardiocentesis? No. Is it useful? Hell yes. Can I look at the image in front of me and decide “This piece of equipment is not helpful in this case”? Yes

          Who is responsible? I am. I am just as much responsible using this ultrasound as I am the majority of the patient care equipment I have encountered in the field. The only time I’ve seen up to date calibrations and inspection certificates on them is when they came fresh off the boat from whoever donated them. I know that this equipment may have issues, but it is what I have on hand. And the logistics train behind keeping it running is too long and too wide.

          Flip it around: You’re in a hospital in your favorite technologically competent country. You’re using a completely certified, freshly inspected name-brand ultrasound imager… Except, it isn’t working right. It could be any number of things: improper calibration, wrong settings, incorrect training, etc. But the bottomline is this: You know you can’t trust the data you are getting because of your experience. You morally can’t use this piece of equipment. It literally has the return to service sheet on it, someone just looked at it and said “This is safe to use”, and you can’t use it because your judgment says this is suspect.

      3. First, I respect the project. It’s impressive, and a great first step. But when you say medical devices are overpriced, perhaps you’re not aware of what’s happening on the back end. Just a few examples:

        1) Can you prove the device is not mixing up data from one patient to another? Which mitigations prevent it?

        2) What mitigations do you have in place to prevent counterfeit parts from slipping into your supply chain? Are the controls on your supplier adequate? Can you prove it?

        3) Are all the parts rated for the environment they will be used in? The Pi Zero doesn’t have a published range of operating temps, I looked. So how do you know it will continue to work and not produce errors in the device’s output? Your gut feel isn’t enough.

        3) Have you considered what happens to the patient, operator, and bystanders if parts of your device fail? What mitigations do you have in place for imaging artifacts that can cause misdiagnosis?

        4) In the event of a problem in the field, do you have tracking in place to make sure you can contact your customers and get out fixes, or a recall?

        5) Formal change control? Packaging? Labeling? All covered in the regs.

        These are the sorts of things that make it so airplanes don’t regularly fall out of the sky, and why medical devices work so routinely that you just assume they will. It’s expensive, and it’s hard work.

        I have the utmost respect for medical professionals and the work they do, but don’t diminish the value of the controls that are in place around medical devices. They are an utter pain in the butt, but they reduce the odds of inadequate tools and snake oil making it to market. That costs, and it has value.

  2. For all you medical device professionals on here that keep talking about clinical trials, What country are you in that requires them? Most ultrasounds (for diagnostic purposes) in the US are Class II devices. I believe that the MDD classify them as Class IIa. If clinical trials are encouraged in your country, what is proven in them that couldn’t be proven through benchtop testing?

    Here is someone trying to do good. Either help him (e.g. be aware that ultrasound can cause deep tissue burns, please be careful with self experimentation, also traceability in the medical device industry is important), point him to something useful (The FDA has a guidance document on the parameters they find important. see “Guidance for Industry and FDA Staff – Information for Manufacturers Seeking Marketing Clearance of Diagnostic Ultrasound Systems and Transducers”), or shut up. This isn’t another fidget spinner. It has the potential to do real good.

    Good luck.

    PS. The FDA waves fees for certain devices when the fees would be a “Barrier to Innovation.” Also, your goals are lofty enough and there are enough phantoms laying around that you could probably get a clinician to let you borrow them after hours for your substantial equivalence testing.

    1. The trials are to prove that they are not inferior to other products that are designed to do the same thing using the same modality. You have to be able to demonstrate equivalency (parametric output for sensitivity and specificity analysis) or label it differently so you don’t confuse people who use it. Bench top testing hardly replaces real world testing. Think about what you are saying. If that is true then why would we have beta testing or usage feedback? Just test in-house and release the product into the world. Surely no major problems will occur. Pfft.

      The only major criticizism I am pointing out is people in general calling these projects an equivalent medical product when it isn’t. There is nothing about saying they can’t develop it into a medical device. It just isn’t there yet. It’s the same as calling a pile of wood a table. You can make it into a table. No one is disputing that. You just can’t call it a table in its current form.

      If the creator wants specific feedback on something then he or she can ask for it. No one is going to write a tutorial on medical device development in the comments. No one has that much free time. If there are specific questions then fine otherwise go hire someone to teach you as it requires a big commitment to hold someone’s hand.

      Just like you are able to freely criticize people, other people are free to criticize others so no… Telling others to shut up when you are criticizing is just being a pot calling the kettle black. In addition, no one says the creator has to abide by whatever is said in the comments. Just read them, think about it, and if you don’t agree then you can ignore. It’s part of being an adult. Not everything has to be said in a positive way. Being straightforward is acceptable. Try going through medicine (med school and residency) if you want a good life example. You get something wrong, misinterpret it, or be ignorant about it your feet will be held over the fire by your attendings and peers. Some even make fun of you for it. You end up growing some thick skin and learn to not take criticizism personally.

      1. Hey Toby – thanks for the comment. A lot is being said about medical devices – which are a hell to make, ensure quality, and get certification for, and even harder that was is said! That’s the main reason why _this does not aim at being a medical device_. It’s just a dev kit for hacking / academic / r&d purposes =)

      2. The text of the body was directed toward LOL. I just couldn’t move it once it was added.

        I think that the medical community (clinicians, engineers, and regulatory departments) has a huge chip on our shoulders in that we think what we do can’t be done by mere mortals. As a result, we make things more complicated than they need to be. From a regulatory prospective, clinical trials are not required. According to both US and EU regulatory bodies, both safety and efficacy can and should be well established prior to it being used in a clinical setting. If you think about it, if you don’t know how stable your power is, or if you can’t determine what your looking at in a phantom, you really shouldn’t use it in or on a person.

        Validation in a clinical setting still needs to happen. But these don’t need to rise to the level of clinical trials. This is where you realize that the cord should have been a foot longer, because the counters in most rooms are higher than expected. Or it needs a bigger battery because the backwoods town in Appalachia only has power 6 hours of the day.

        My point wasn’t don’t criticize. It was don’t crush his soul. This isn’t good for medical residents and their patients (see Leisy and Ahmad 2015). It probably isn’t good for someone volunteering his time.

  3. There is a functional safety standard IEC 60601 that applies to most medical devices of this nature. It’s a derivative from the parent general functional safety standard IEC 61508.

    It’s not impossible to achieve compliance by any means, but you would struggle to do retrospectively – the chances you did everything you needed to and in the way required are basically zero if you don’t set with compliance in mind.

    So you would basically end up redesigning your device, but if you had something good it would still be well worth making a backyard proof of concept, then look to partner with a group that supported the proper development process for the to market version.

Leave a Reply to AndyCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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