Butterfly iQ – A cheap handheld ultrasound tool with AI smarts inside | Hacker News Reader

Butterfly iQ – A cheap handheld ultrasound tool with AI smarts inside

spectrum.ieee.org

238 pointstaion147 comments

Perspective from an emergency doctor. Many comments ask if non-medical persons would use ultrasound should it become affordable. I think not.

There aren’t enough common medical conditions that you would monitor with a personal ultrasound[1]. Also it’s a tough skill to learn without a solid background in anatomy. This would deter many.

Primary care doctors (eg family, internal, and emergency medicine) will benefit most from affordable ultrasound. We are learning that it’s a powerful diagnostic tool when used along side the physical exam. Some zealots have equated bedside ultrasound to be the biggest advancement to medicine since antibiotics. This notion I feel is exaggerated, but— it taps into the underlying excitement in the medical community for bedside ultrasound.

Here are some sample cases.

Patient came in with all the symptoms and findings of a stroke— altered mental status, inability to move their left arm. Before giving the treatment for a stroke, a potent blood thinner called tPA, the doctor decided to do an informal ultrasound of the patients heart. He found the patient had a massive dissection of their aorta. The patient wasn’t getting adequate blood flow to their arm or brain. Had the patient been given tPA they most likely would have died. A quick bedside ultrasound revealed a difficult diagnosis and saved the patients life.

Another case- a young female came in unresponsive and without a pulse. Her husband said she had complained of belly pain for the past several days but nothing else. While doing cpr we placed an ultrasound on her abdomen and saw a massive amount of blood in her belly. A pelvic binder was applied and we immediately started massive transfusion for presumed ruptured ectopic pregnancy. In this case the easy access to ultrasound gave us a diagnosis in under a minute.

Both of these cases occurred in the emergency department. Of all specialties I think emergency medicine has been quickest to adopt bedside ultrasound. Other fields however could benefit too. Ultrasound is underutilized in family medicine and internal medicine clinics. Making it affordable, say close in price to a stethoscope, should encourage its use in these specialties and lead to more discoveries of its use as a diagnostic aide.

1. patients with an abdominal aneurysm may wish to monitor its size with a personal ultrasound. This seems extreme.

Where do you see ultrasound being useful in the routine assessment? As a tool to rule in or out certain differentials, sure (like the examples you gave above), but I don't see the case for routine use in the GP's office.

I think "pocket" ultrasound will be (and is being) adopted by EMS well before it makes inroads in primary care.

Agreed, it’s better for acute complaints instead of routine physical exams. If a patient with COPD or CHF comes into your primary care office with a little cough, then you could enhance your bedside workup with a quick cardiac and pulmonary ultrasound. You could see if they have pleural effusions, pulmonary edema, consolidations, and make an informal assessment of their ejection fraction. This info takes about two minutes to acquire, and helps you rule-out dangerous diagnosis.

I agree with your assessment.

In my experience, ultrasound has revolutionized trauma care - basically getting rid of the invasive Diagnostic Peritoneal Lavage (where they put a needle in the abdomen, inject liquid, suck it out, and then see if it has blood in it).

Other examples where it can be used in a primary care setting:

1. Carotid stenosis (narrowing) - we know that this can be a cause of stroke and listening by stethoscope is likely useless (the really narrowed arteries don't produce much noise). Having ultrasound that is cheap and accessible might help catch more of these cases in a primary care setting

2. Heart defects resulting in athlete sudden deaths. Having a primary care doctor evaluate an athlete's heart during a sports physical with ultrasound might help catch some of these cases.

3. Outpatient evaluation of deep vein thrombosis in sedentary populations. This is important because this can help stop blood clots to the lungs.

However, as with your examples, these use cases all involve a medically trained operator. Like you, I am not seeing a lot of potential consumer use cases.

Counter point to your whole argument: medical professionals without the funding to acquire "proper" hardware. This isn't a DIY solution for hypochondriacs, it's access to technology that might not otherwise be able to afford it. Maybe veterinarians too.

That's not a counter point at all. It said "Many comments ask if non-medical persons would use ultrasound should it become affordable. I think not."

Yup! I'm halfway through the comments, and yet to see a mention of how cheap ultrasound machines would be a massive boon for healthcare in developing countries.

This is the first thing that came to mind. The second is giving med students more hands on time with ultrasounds. The price is low enough that students could practice on each other in class.

I saw ultrasounds live for the first time last year due to a worry about my son's nose bone. My thought was that sonogram interpretation is an impressive skill to develop. I've done remote sensed imagery interpretation for a few years in university and it felt like just that. To a normal person it's just blurry blobs but to an expert it's a map.

How about using an ultrasound for body composition? http://intelametrix.com/FullSite/ BodyMetrix has been around for a while and has about the same price of $2000

Great point - the difference between portable ultrasound machines like the one presented here and the Butterfly Network one is the mode of operation.

The BodyMetrix system uses A-mode (Amplitude Modulation) which essentially uses one signal or pulse and plots the amplitude change over depth. The changes in amplitude correspond to differences in tissue. [I conclude they use A-mode from the abstracts of the papers linked in their Science and Validation section]

The Butterfly Network system is a B-mode (Brightness Modulation) one. These have traditionally had many signals and are used to create the greyscale images commonly seen when referring to ultrasound. B-mode imaging has traditionally required a fair amount more in terms of hardware, software, and thus cost.

> Many comments ask if non-medical persons would use ultrasound should it become affordable. I think not.

I immediately thought of a non-medical use case.

A pregnancy and birth in Singapore costs $5-20k, so an extra $2k is not that much to be able to see your baby move and grow whenever you like as opposed to having to wait for your monthly appointment (and perhaps post photos on Instagram). It's a bit much for me but I might have bought one for $100.

Some redditors working in US ER departments complained of mothers turning up with "complications" mid-pregnancy because of the mandatory ultrasound that follows, allowing them to see their child, at which point their symptoms mysteriously vanish and they head home smiling.

> with "complications" mid-pregnancy because...

That's terrible :-( I know many GPs in Australia would love to do the ultrasound for them + bill for it, but it would be pretty affordable. This case is literally the reason mentioned at GP conferences for why you could buy an ultrasound - to make pregnant women happy because they see the pictures.

Luxury ultrasound is a growing business in the US. Wanna know the gender at 14 weeks? No problem, $40 in the mall. Want a DVD with a recording of the baby's heartbeat and a 20 minute video of it squiring around in 3D? $100. Want to stick that baby heartbeat into a teddy bear? Dude, you are kind of weird....$30. Sign up for our monthly "checking in on baby" package ultrasound discount.

I'm not kidding, it's very real: http://forever-yours.us/3d-4d-ultrasounds-2/

Thats actually really affordable considering how often people have children. The teddy bear thing is a really cool idea.

Where there's money to be had, there's someone willing to exploit it.

http://www.ctvnews.ca/health/ultrasound-clinic-blames-identi...

And when you get caught, shout "IT WAS A VIRUS!!!! CLONING IMAGES!!"

I dont think I have ever encountered a virus that cloned images. The only way I can even imagine it true is if a malfunctioning root kit was somehow installed and affected file writes....but in 30 years..haven't seen it. Likely they were just lazy so they kept handing out the same image figuring no one would know.

@davycro just said "it’s a tough skill to learn without a solid background in anatomy". If you have your home ultrasound and cannot accurately locate / track your baby, you may identify the wrong thing, and worry "it's not moving!"

Our first child came very late, and in those last weeks we downloaded a "baby heart monitor" on our iphone - it simply used the microphone with some isolation / amplification. We pressed the microphone end of the phone into the abdomen wall, and when it works we heard an amplified heartbeat out the speakers - very reassuring. But when it didn't work - when we couldn't find the heartbeat - we worried. Is it a false negative? Or should we rush to the hospital?

Or even worse you put the microphone on the epigastric vessels and hear moms heartbeat instead of the babies.

Why is this even worse? I assume you're suggesting it would potentially give a false sense that everything is ok when it isn't. But since such apps aren't intended as diagnostic tools, is it really a problem?

My wife had risky pregnancies. There is cheap professional but portable ultrasound you can rent. We had one for like 6 months each time (3x). It is a lifesaver when you know baby is OK and you don't have to rush to emergency for every little thing. Otherwise we would go crazy. Since we monitored at home blood pressure and sugar level, we were really good and not as much in the dark as we would be.

You could hear baby heart and also blood rushing to placenta. And those have distinct sounds and places you could check.

Can you link to any additional resources and/or the specific model you can recommend from personal use? Either way - thanks for sharing the info!

I managed to find it. I was waiting for my wife to come home, but I found it, this site you just can't miss :)

http://www.heartbeatsathome.com/

This is what we used and it is more than enough and you can carry it with yourself. Their customer service is also really nice, but you will not need it. Buy extra gel for doppler and you will fine.

I recall some unease over Tom Cruise using an ultrasound machine at home when his then wife Katie Holmes was pregnant [1]. Do the concerns over the home use of such machines still hold in this case?

[1] http://www.nbcnews.com/id/10309963/ns/health-womens_health/t...

The possibilities with deep learning are ridiculous. A doctor could point it at someone's heart or stomach and it could tell if it was normal heart or there was a problem with it and possibly what kind of problem.

In reality, we can't even get good results from ECG readings, and that's "pure data". Ultrasound results would be way harder (depending on position, orientation, person's body)

I wonder what about industrial use? Wouldn't USG be useful e.g. as a cheaper alternative to x-rays in searching for hidden structural damages in materials?

The reason you need a gel between the USG head and the body is to bridge the gap with something of a similar speed of sound. Barriers between materials of significantly different speed of sound reflect large part of sound waves; having such a strong barrier at the skin would make the SNR much worse.

Speed of sound in metals is much larger than speed of sound in water or air. I'm unaware of liquids that could be used in place of USG gel for examining metal objects (maybe mercury? but that would be bad from the environmental and health/safety POV).

> Before giving the treatment for a stroke, a potent blood thinner called tPA, the doctor decided to do an informal ultrasound of the patients heart.

Out of curiosity's sake: IIUC blood thinners help a patient with an ischaemic stroke, but are harmful for patients with hemorrhagic stroke. Am I right? If so, how do people usually quickly rule out hemorrhagic stroke?

Head scan. CT or MRI.

Correct. We do a non-contrast head ct and ct angiogram of the head/neck before giving tPA. In this case we may have found the carotid dissection on the ct angiogram.

I have chronic kidney stones (over 100 at age 30) and would absolutely pay $2000 for a personal ultrasound. Often it is difficult for me to tell whether it is just a rough passage (extremely painful but not a medical emergency), or a large stone obstructing my ureter causing hydronephrosis (medical emergency). With a personal ultrasound, I would be able to look at my bladder and see urine flowing from the kidneys to rule out blockage and prevent a trip to the ER. Granted, I'd probably still have to go to the ER because it is difficult in the states (in my experience) to get access to adequate pain control in a non-emergency setting.

Hi, off topic but-- my father had these, several a year for two decades. Some mornings we would wake to a note on the counter that said something like "Kidney stone, at the hospital."

That being said-- the trick for him turned out to be stopping soda, and keeping a water bottle by his bed and drinking a gulp of water everytime he woke up-- basically staying hydrated during the evening, and at all other times.

Just wanted to mention this in case it was of any use. Also, keep an eye on RCC. :-/

Appreciate the reply, thanks!

While on the subject: https://www.webmd.com/kidney-stones/extracorporeal-shock-wav...

I'm glad you responded. Ive had ultrasounds, and CAT, PET, and MRI before. Never knew you could glean so much information off of an ultrasound.

Now, for the hacker/maker community - maybe you won't know this offhand. But are there applications for ultrasound outside of medical? Would it be good for 3d scanning internal structures of a certain size? Can it find/detect cracks inside substrates, like how Xrays is used in weld defect detection?

Im just as curious at the non-medical as the medical. And it'd be damn nice to be able to ultrasound myself and go, "that dont look right". Im not a doc, but I live with my body 24/7.

Ultrasound is used to detect faults in materials (cracks in metal etc), yes. I think also for some measurements in fluids.

I’m sure doctors are looking forward to ultrasound scanners being common household devices, and getting 50 emails a day from panicked people convinced that their gallbladder is a tumor.

I'm looking forward to having an inexpensive portable USS to help find veins. I doubt consumers will pay for one of these, except maybe for pregnancy.

Yeah, this sort of thing is making quick inroads in prehospital care (this is far from the first portable ultrasound). Finding veins, FAST[1] scans, assessing the heart more directly during cardiac arrest, etc.

[1]: https://en.wikipedia.org/wiki/Focused_assessment_with_sonogr...

We've been using portable USS machines for years but I suppose what I meant is one I can keep in my pocket and one that doesn't cost $150000.

The GE Vscan has been around for years.

That's interesting. I've not seen that before. Still seems to be far more expensive than what I'm thinking of, though.

I'm looking forward to the day when every junior doctor can have one around their neck like every doctor has a stethoscope.

Why? What gap do you think it fills in the routine assessment?

I'm all for expanding access to US, but I'm very wary of throwing gadgets for the sake of gadgets at doctors...

Patients who need an IVC and have brittle veins and peripheral oedema.

It comes up a lot.

I never said anything about routine assessment.

Sure, I would love a portable ultrasound in the back of my ambulance for that very reason. I still use my stethoscope _way_ more often than I start an IV where US would be handy.

I don't know why you keep putting words into my mouth. I never suggested it would replace a stethoscope.

You said every doctor should have one with them "like a stethoscope". I don't know where you got "replace" from...

My point is that a stethoscope is an essential tool for assessing every patient. An ultrasound is a specialized tool for assessing a very small subset of patients.

I said "can". I don't know where you got "should" from.

The benefits of portable units at a non-devastating price have more implications for democratization of quality of care than anything else. WebMD and Google have already let the other genie well and truly out of the bottle.

Edit: By the way years ago I read an article which talked about the price of ultrasound devices, and the bottleneck was the price of the transducers. Obviously that’s changed, but when did that change? Does anyone here know?

I work for a competitor to Butterfly, and the main difference between this device and other ultrasound transducers is the use of a CMUT array instead of the typical piezoelectric crystal. CMUT arrays are MEMS devices, made in a semiconductor fab using the same processes as other ICs. This makes them cheaper to produce than traditional arrays where the elements have to individually sliced and wired to the electronics. CMUT isn't new, it's been being worked on for 20 years, but it isn't used because it is much less sensitive than the piezoelectric arrays. Either Butterfly has made a breakthrough and are choosing growth over profit margins, or they are trying to make up for worse performance with a much cheaper price. We won't know which until their device is actually on sale.

Incredibly informative answer, thanks so much. One last thing, do you personally believe “breakthrough” or “discount” in this case?

20 years of CMUT being the next big thing in ultrasound has left me a sceptic, but I know some really smart people at Butterfly. I don't think they would bring a device to market that they didn't think was competitive. My guess is that it'll be okay, not the best imaging handheld device on the market but better than several of our competitors.

As an internal medicine phsysician who has been using a handheld daily for the past few years I would only add that high image quality is not always needed for many of our most important applications: Lung, IVC, bladder, DVT, abscess. US is dramatically more accurate than physical exam and stethoscope In these cases.

Agreed. I was talking specifically about CMUT vs Piezoelectric. If high image quality was always a requirement, the handheld market probably wouldn't exist at all.

Butterfly is right to focus on getting an answer over image quality. As a startup they will likely be able to iterate quickly on machine learning, which is critically important to growing the ultrasound market. Right now the biggest barrier to adoption is the difficulty of reading ultrasound images, and machine learning is the most promising solution to that. I think that is more likely to be their competitive differentiator than their use of CMUT.

It changed this week, this is the first commercially available device that uses silicon chip rather than piezoelectric crystals, thus the buzz.

For now, this isn't a consumer device [0]

0. https://twitter.com/matthewherper/status/923897406776889344

This is a non issue IMO, it would probably self correct in a few months. Also you could still take advantage of that.

The patient who made a few ultrasound videos along a period of time and still has pain worth seeing a doctor, can now spare the doctor some time by showing him a bit more data. Could be a data hell too I must admit, but somewhere somehow I'd hope people to learn a bit about biology and physics in the process and 1) know more 2) know when not to panic 3) know how to take care of their bodies

Thats technology. At some point you can see remote doctor sessions being a thing, and robotic doctors carrying out some basic surgeries.

I’m not saying that wouldn’t happen, but we need to get away from the idea that non-professionals shouldn’t have access to their own medical information. More education and empowerment is the answer, not less.

> More education and empowerment is the answer, not less.

The crux lies in determining the boundary between education and quackery. The resurgence of antivaxxers, homeopathy believers and friends doesn't exactly help the case for deregulation here.

Is there any evidence anti-vaccine and homeopathy movements are bigger than they were in the past?

Google Trends suggests homeopathy is up a bit recently, but way down from its pre-internet peak.

https://trends.google.com/trends/explore?date=all&q=homeopat...

> Is there any evidence anti-vaccine and homeopathy movements are bigger than they were in the past?

I'm talking about Germany here as I can't really speak for the US, where from 2011 to 2013 prescriptions rose 3.8% and OTC sales went up 23% (https://de.wikipedia.org/wiki/Hom%C3%B6opathie#Bundesrepubli...).

As for antivaxers, Europe as a whole has a huge measles problem which has risen in the last years (http://theconversation.com/whats-behind-the-sudden-rise-in-m...).

Are you saying proponents of quackery might misread somebody's home ultrasound or give misleading advice on how to read it? That doesn't sound any worse than what could happen with thermometers, stethoscopes, blood pressure meters, athlete wristbands, and inspecting the color of phlegm, poo, urine, skin, etc. Nobody would want to hide those things from people for their own safety.

With most of what you mentioned, the local quacker is not able to say "OMG YOU GOT CANCER, HERE TAKE THIS $1000/pill HOMEOPATHIC CRAP"... with an ultrasound image he is very well able to do so. And people will do it, thanks to human greed - if you need a confirmation google for "miracle mineral supplement" aka chlorine dioxide and the "industry" surrounding this.

This is why we can't have nice things, the potential of abuse is just too high :(

   > More education and 
   > empowerment is the answer,
   > not less

More empowerment and more choice is why the US healthcare system is so damn expensive compared to the rest of the world.

Explain?

He might be comparing with France where rich people can not pay for better cure than poor ones.

No that would be regulatory capture.

Maybe we can use machine learning to train a computer to identify these diseases. That would lot cheaper than a doctor and with enough data it could do a pretty good job.

ultrasound is pretty much the worst imaging modality to do this in all of medical imaging.

I'm not convinced that the subset that would do that with an ultrasound device is really any different from the subset that does this kind of thing already.

easy solution: charge per review of ultrasounds.

If I can go get a device, and somebody qualified to read the images I produce can also inform me how and where to take pictures - I'd pay for that kind of a service. Seems like a business model to me :)

Problem is that if you've ever had an ultrasound, it's not really something you can just tell someone how to do. They manipulate the transducer in real time, angling it just right, giving it just the right amount of pressure on the skin. Even if you did it in real time I can't imagine the frustration of a physician over the phone going "No move it to your right, now up, now back further, ok less pressure, now more pressure."

That'll be a nice worry for doctors to have in reaction to the spread of consumer-available precision imagery equipment.

Right now I think it's more valid to worry about the patients who can't receive care due to whatever host of valid economic/social/cultural reasons prevent them from doing so -- and one of the best ways to help that struggle seems to be the creation of cheaper and more available diagnostics and therapies.

I don't know about this specific device -- but any competition in the medical diagnostics market that may drive device price down seems to be a benefit.

Engineer at Butterfly Network here, very excited to see interest on HN. As you can imagine, we are hiring, there are a lot of very interesting open opportunities https://careers.smartrecruiters.com/4Catalyzer

As someone who just had an ultrasound, I think this is the right direction. I had to goto a diagnostics center where they specialize in scanned imagery. I go in, paperwork, wait, wait some more, get scanned, and leave. The technician just had me hold my breath while she put some jelly on my abdomen and moved the device around. Once in a while she would take ultrasound pictures. Snap, snap, move, breath, snap.

My GP had me go in order to troubleshoot a blood test. Now, if the GP had this device it would have saved me some time and effort, not to mention the external factors such as gas, taking up space on the road, and an appointment.

So yeah, please, GPs everywhere, load yourself with these devices as they become cheaper and cheaper.

edit: GP's to GPs

If the GP had the device, it would have taken them as long as the sonographer (or longer) to acquire the same high-quality images. In fact, it's so impractical for physicians in the US to routinely do their own ultrasounds that they hire sonographers to acquire those images. I don't think this device would have changed your experience, I'm sorry to say.

It appears there are courses just for physicians. [0] Also, the device itself may include a course of some sort. Without evidence, I find it hard to believe GPs aren't capable enough to use an ultrasound device.

[0] https://sonosim.com/who-we-train/

In my reply to you, I'm assuming that your GP is 100% capable of performing an ultrasound. Ultrasounds take a well-trained human being a long time to acquire.

My wife is in med school so I have a bit more familiarity with the subject than the average person.

GP's being capable enough isn't the issue. It's much like blood draws.

GP's are more than capable of drawing blood an sending it off to a lab, but that process requires significant support of outsides parties. It simply doesn't make sense for the GP to take care of it in house when they'll likely need to send it out to someone else.

In the care of ultrasound, it also introduces an issue of "certifying" results. Taking an ultrasound isn't hard, but reading it can be difficult. Even if a GP can read an ultrasound with 90% accuracy, it's still much lower than someone who specializes in reading digital imagery. Of the 90% that a GP can read accurately, how many of those can a GP read confidently? Certainly not 90%. At that point, they might as well send you to someone who will get it right the first time.

In house labs are advertised as profit centers for doctor practices:

http://www.physicianspractice.com/blog/three-good-reasons-of...

By consulting companies trying to sell their services. Take that article with a grain of salt.

The funding model in the US means that providers have all sorts of incentives to pay people other than doctors lots of money to help them maximize patient volume.

(I say providers above because doctors are often just part of the machine, but the machine doesn't want to waste the doctor's valuable time any more than an independent doctor does)

I'd love a push for more monitoring capacity in the form of cheaper non invasive medical tools.

One of the main health issue is waiting to fix an issue. Cost and exam delay are both factors into letting things sink deeper.

Also, lots of medical tech is IMO (feel free to take me down on this one) stuck into analog electronics and physics of the past[1]. Namely one way information live flow. Today we may be able to compensate through software logic. Think super resolution algorithms used in space exploration. Now just think how neat it would be to make use of these smartphone GPUs for once .

[1] don't get me wrong, they did breathtaking thinking most of the time, but it's a difficult art and not necessarily the best solution today.

So just as I needed reassurance this pops up https://news.ycombinator.com/item?id=15595269

timely :)

>The technician just had me hold my breath while she put some jelly on my abdomen and moved the device around. Once in a while she would take ultrasound pictures. Snap, snap, move, breath, snap.

Well, yes, and she probably went to school for quite some time to know exactly what gets a picture and what doesn't.

This is like saying "I had an appendectomy, and all the surgeon did was make some cuts an sew me up." You're not going to a surgeon because he owns a scalpel.

This is not a good comparison. The GP has the technician do the imaging, but the real problem is not so much getting the image, but interpreting what's on it. This is a skill that the GP has to have either way.

The GP surely has the skills to take and read the scan.

But it's not efficient use of their time.

Ultrasound techs do the imaging/take the photos, and the physician reviews them. Ultrasound techs make less money than physicians, that's why physicians are taken out of that part of the process.

It's like saying "The dentist should also clean my teeth". Well their time is better spent on procedures, leave the cleaning to the hygienist.

In the outpatient setting it might take days or a week or more to get an appointment for a formal US, then more time for the radiologist to read it, then more time for the GP to get the results and consider treatment changes then more time to get in touch With the patient. But if I as a physician perform and interpret the ultrasound in real time I can immediately begin appropriate therapy. This is the major benefit of point-of-care US as opposed to conventional US.

The issue of specialist departments seems to be a global problem. A friend of mine needed to get an image taken for a diagnosis. He was referred to another hospital because our local one didn't have the capability. In the end they used an SLR to take a photo and sent it to where it needed to go. Why couldn't they do that in the first hospital? Why can't the GP do that and send the image? I have no idea.

We're constantly led to believe that the revolution is coming where you can snap a photo of your arm, send it to the health service and get a diganosis. That's not happening for a long time, I think.

> “And saving those images is required to bill for the scan,” he notes.

If that throwaway comment doesn't worry people, then I don't know what does. I can understand low resolution, health risks, but a principle barrier to the hospital is that they can't charge you for using it? Fuck that system, quite frankly.

Yeah, that's a really dispiriting sentiment to read.

But the good news is that it's rooted in the hopefully-soon-to-be-obsolete pricing model. If ultrasound devices really can get down into the $2000 price range, then the good doctors will buy them as a cost of doing business, and just not bill for the ultrasound readings. Wait a decade or so, and that'll percolate from the good doctors to the rest of the field.

Yes, there are perverse incentives etc. etc., but I know a number of doctors who are in it for the love of medicine and helping people. The bureaucracy changes course slowly, but those leaders can and do make a difference over time.

One voice does not equal the system, but in any the push by some to bill for clinician performed US is largely driven by a need to generate funds to justify to administrators the purchase of a $70k device.

From a provider standpoint, however: If they are not technically adept at operating an ultrasound machine, and do not have the knowledge necessary to read an ultrasound, why spend $2,000? And on that note, is this device able to be billed for in order to recoup that cost? We have that problem right now with OCT-A in ophthalmology practices, where the technology exists and is FDA approved but cannot be billed for to recoup the costs of the hardware.

Yup, doctors specialize. When a GP gets orders a scan, such as a CT or x-rays, it's typically another doctor that specializes in the imaging area who examines it and writes a report. I don't think many GPs would have much of a use for an ultrasound because they would prefer a specialist perform it.

A few caveats. One, telepresence. You can have the specialist basically be there with you as you operate the probe. Everything gets done in office and the patient doesn't have to travel which can be a big deal in rural communities. I find most urban GPs don't even have a second thought about sending patients off for a scan at some other office.

Following that would be very remote doctors, those that might be labeled as a GP, but basically have to have a ton of specialist knowledge because of the wide range of things they might need to do.

And of course vets, especially large animal vets. Who are already using cool devices such as head mounted displays to view the ultrasound as they generally have their arm all the all the way up an animals rear end.

Hadn't thought about rural physicians, very good point. I come from a relatively rural community, but with a large hospital and specialist presence, so it kind of slipped my mind that there are even smaller communities that would benefit from this.

Telepresence sounds great in theory, but in practice telling anyone, even a physician with a great deal of anatomical knowledge, how to use an ultrasound probe can be frustrating at best.

Also did not know about vets using these new devices, very interesting!

>He also ran into difficulties with transmitting the images to the hospital’s database for storage. “And saving those images is required to bill for the scan,” he notes.

I think this guy is missing the point. Seems like the goal is to make giving an ultrasound so easy that it would be silly to charge for it. Just like the doctor doesn't (or at least shouldn't) have a special charge to listening to your chest with a stethoscope.

> Seems like the goal is to make giving an ultrasound so easy that it would be silly to charge for it.

A literal ape can do an ultrasound scan if trained well enough - you still need years of training to actually interpret the data you gather with an ultrasound image or by listening to a chest. This is what you pay for.

Can someone with a better understanding of the medical industry comment about the quality of this device? It seems like a small portable ultrasound device would be a great thing for diagnosis. But I understand there are all kinds of issues that might not be obvious to people without a strong medical background.

Is this as good as it seems, or are there features this lacks which prevents it from being that useful in the hospital setting?

In a hospital setting this isn't terribly useful. An ultrasound is not that big, and easily moved around on a small cart.

Small portable ultrasounds like this are gaining popularity in prehospital care (in ambulances, etc).

This isn't really a new concept though. Things like the GE Vscan have been around for years.

Certainly in a well-funded hospital, but might $2K also not look more appealing than $20K-$200K for a small clinic in a one-horse town?

Caveat, I have no idea how much a commercial ultrasound machine costs. But it's one machine Michael, what could it cost? Ten thousand dollars?

Even for a well funded hospital this should mean that the procedure now becomes less expensive. Especially important for pets where you can be out of pocket for half the price of one of these.

Each of the machines at my (large academic) hospital is used for 8-12 transthoracic echocardiograms every weekday, or ~2,500 times per year. An average ultrasound machine cost $140,000 in 2016 (per [1]). If we assume that the machines last 3 years (so far I can tell you that they last at least 5 years, but let's be conservative), then the hospital gets ~7,500 ultrasounds for $140,000, so just under $20 of the cost is attributable to the machine. According to Choosing Wisely[2], the average cost of a TTE is ~$1,000. So, only 2% of the bill seems to be attributable to the price of the ultrasound machine. So, at least at major academic medical centers, the device itself probably won't make a huge change.

I could imagine the standard of care changing, however, and that might make a difference. If cheap ultrasounds become ubiquitous and more physicians are trained in their use, the use of ultrasound among generalists might increase, freeing up specialists to take a more narrow focus. I could see this technology being particularly useful in emergency and critical care settings.

1 = http://www.modernhealthcare.com/article/20141216/blog/312169...

2 = http://www.choosingwisely.org/patient-resources/echocardiogr...

The mark-up in western hospitals notwithstanding I can see a large market for cheaper devices in countries that are not so rich.

Absolutely. This price point -- apparently 70x cheaper than the prevailing standard -- could definitely bring the technology to places where $140,000 is prohibitive.

$20K-$40K (about the price for ultrasound machines nowadays) just isn't that expensive for a piece of medical equipment. Your consumables are probably more expensive than that over a year.

I suspect this would be more useful for veterinarians--especially in one-horse towns. :)

I'm somewhat skeptical of an EMT's skill level when it comes to ultrasounds.

Where it's being used it's a tool for paramedics, not EMTs.

I believe the expensive part is the array of piezo sensors.

Resolution is driven by the number of piezoes. Previous low cost ultrasound sensors skimped on the number in the array, which limited effective resolution. Aside from the piezos themselves, you need enough real time inputs to process them.

It seems like this one does better, though details are scant.

There are no piezo sensors. These are MEMS transducers. Fabricated in processes similar to IC chips so cost scales well.

A MEMS piezoelectric transducer. Or "piezo sensor".

A big thing in the medical industry is proof of efficacy. Hospitals and medicine in general are extremely risk adverse. Health systems will often go to great lengths and cost to demonstrate that the equipment they use minimizes risk.

$150k machines come with a high price tag because they also come with a reputation (and even financial guarantees) of accuracy. This minimizes risk to the hospital.

This is a new device, from a new manufacturer, based on a new technology. Health systems will be extremely hesitant about adopting this until it has some real world experience. They'll want to make sure that they if their are risks with the device and if the manufacturer looks to be around in the long run.

I could see this being somewhat popular when the decision is the alternative is no ultrasound at all. If the health system has a proven system, they'll like look to use that before this device.

So how to stop people from taking this discreet, inexpensive ultrasound device to China where parents will pay a lot of money for a black-market ultrasound to be told the gender of their fetus?

http://shanghaiist.com/2017/07/12/ultrasound-van.php

http://www.havocscope.com/cost-for-illegal-ultrasound-and-ab...

Aborting girls in China is exaggerated and you're certainly over playing it.

What about the girls lives saved by cheap medical technology?

Certainly one theory on the gender imbalance is girls tend to get less medical care due to costs and as such die at higher rates.

Is 100 million "killed, aborted or neglected" girls exaggerated? Or do you have better sources than the Economist: http://www.economist.com/node/15606229 (Note that article is from 2010, there have been many more since.)

Dime a dozen that dispute that figure but my biggest issue is with this comment is -

> to China where parents will pay a lot of money for a black-market ultrasound

No, normal parents don't abort girls, normal parents don't infanticide girls. The Chinese are not some sort of subhuman race that on mass do awful things to their children.

Not a fan of the New Republic but it talks about the issue (and the racism behind it)

https://newrepublic.com/article/133845/truth-chinas-missing-...

I don't understand the argument. Should the company not create this largely beneficial technology lest it be used for bad things in China? Should they be responsible for every case of people abusing the technology?

Chinese companies hardly care about the fate of western blue-collar workers when they produce cheaper products for export. Nor should they, because this is not their responsibility. Similarly, the issue of illegal ultrasounds/abortions is the responsibility of Chinese society and the Chinese government, it has little to do with this product or the company developing this product.

Ultrasounds are already pretty portable, here's a common one used in the hospital I work in [0]. Used models are probably pretty easy to come by. Reducing the size from a laptop to a smartphone isn't going to change/create any industry like the one you described.

0 - https://www.absolutemed.com/picts/products/tnw800-GE-Voluson...

This is absolutely interesting to veterinarians. Many veterinarians would use this as an adjunct to the physical exam. In human med, the follow up of conditions like mitral valve insufficiency is covered by insurance but in vet med, it gets expensive going to the recommended cardiology visits paying out of pocket. Plus, we can use ultrasound for a lot of routine and emergency procedures. I signed up right away.

There are already a number of mobile ultrasound devices, including ones that plug into phones and tablets.

As someone who's actually done a few ultrasound courses, depending upon what you're trying to image it's not easy. It's not just about having the device, years of medical training come into play.

Also, the mobile devices have significantly lower resolution and don't seem to be super important in the hospital.

Yep, I've been playing with the Lumify (https://www.lumify.philips.com/web/) for a few years. I think my hospital has had it for about 5 years.

From what I can gather, this new device's selling point is the price. I'm not sure how much the Lumify is, but I had the impression that it was a few thousand dollars. And there's some talk about new AI technology packed inside the new device, what that is I am not sure.

Given the regulatory process, I would envision AI along the lines of "region proposal". That is software to highlight or draw one's attention to a certain portion of the image that "may" have X or something interesting. That's in contrast to diagnosis software where it'd say this portion "is" X.

I should add there are several hand-held portable ultrasound machines, like from SonoSite and Clarius, and would agree that the selling point here is the price point. To the best of my knowledge, the cost of ultrasound machines has gone down by two orders of magnitude with somewhat comparable resolutions. I believe the top-tier machines with the best image quality or resolution are still on the order of >$100k (Philips Epiq for example).

The cloud software offered by Butterfly should also be interesting if done well.

Watch the video, this thing is mobile, but has the resolution of a full size device. But yes, thinking it will be an at home device is a bit of pipe dream.

The interesting question is who/what does the diagnosis. I've had my kids fall over a bunch of times (with a couple broken bones in there - never ambiguous, sadly) and broken a couple bones in my life too (as well as done stuff that convinced me that something must be broken in there, when nothing was).

So I think there were essentially zero times when a cheap-hand-held ultrasound in that home would have added information. The question is really whether this could be enormously helpful to a GP? Can they be trained usefully to take over some of the role of an ultrasound technician and do first-line diagnosis? It'd be great to not have to do the standard schlep over to the diagnosis place, wait around, get the thing done, and wait to get the results back to the GP or specialist.

I imagine there's also a fair bit of computer "vision" work on this stuff too - I would presume people are working feverishly away on ways of reconstructing 3D imagery and actual diagnoses of an image via computer, as opposed to squinting at these mysterious pictures.

Ultrasound is not terribly useful in the diagnosis of fractures. It can detect certain types of fractures in kids under 12 or so, but can't rule out a fracture.

There are absolutely things a GP could use ultrasound for to rule in or out various diagnosis. A GP doesn't really need a "pocket" ultrasound though. A "normal" ultrasound would work just fine.

I'd be curious as to what the payment model for a GP to own a conventional ultrasound machine would look like in Canada, the US, and the UK. It's my understanding that the pocket ultrasound at it's price point would essentially pay itself earlier on while serving a purpose of being a triage tool.

Given the size and cost of a conventional cart-based machine, I think the pocket one is more well suited.

> The question is really whether this could be enormously helpful to a GP? Can they be trained usefully to take over some of the role of an ultrasound technician and do first-line diagnosis?

This is likely healthcare system / country / training dependent, but yes, many local clinics will have ultrasound equipment. GPs do use it for things other than emergencies (simple pregnancy checkup, ultrasound guided injections, etc.)

I know some that would find the "Can they be trained usefully" a silly/offensive question to be honest.

Cheaper, more portable, battery powered ultrasound would make the technology more accessible to developing countries. This will save lives eg from childbirth complications.

I saw no mention if they removed the ability for it to sex a fetus, if they haven't it will be a tool for aborting female babies.

That seems like an impossible feature for an imaging device. I guess you could train a neural network to recognize fetus gender and blur out the crotch area on the image.

They rely heavily on AI for processing the data, they absolutely could remove the ability for gender identification.

That depends on how they're applying their "AI technology." Maybe it's to improve the SNR for a given pixel? Maybe it's to interpolate additional resolution between pixels / lines? It seems possible, but not necessarily likely, that they're doing whole-image post-processing / analysis.

Not sure why you are getting downvoted but if this device becomes popular in developing countries like India there will be a lot of aborted female babies.

I never understood the logic of this issue. If a fetus is human, then both male and female fetuses should be protected. If a fetus is not human, then this can't be a women's rights issue because women are human. If abortion is acceptable, parents should be able to choose to abort for whatever reason they want.

Also if you withhold the sex of the baby in case some parents may abort due to gender, why allow screening for down syndrome since this is a much more common reason to abort?

> parents should be able to choose to abort for whatever reason they want.

http://www.nytimes.com/2005/11/25/opinion/missing-50-million...

They should sell these at a huge discount in exchange for an agreement to share imaging data and eventual diagnosis back to the company, as training data for their AI. An AI that could identify medical issues on an ultrasound with greater accuracy than a doctor would be worth a fortune -- and genuinely make the world a better place.

I know it's been studied extensively, but doesn't fda and others still advise caution about frivolous use?

ultrasound is pretty harmless. Consequences about frivolous use would mostly be due to unintended discoveries that might require further investigation, which can be dangerous.

Another great medical device that piggy-backs on an iPhone is the FDA-approved spirometer Wing:

Http://mywing.io

There is a lot of hate in these comments for this being a surprisingly cheap and potentially disruptive technology.

I was just having a lengthy discussion with an ultrasound technician last week. He himself admitted that:

1) The classes he found for free on YouTube were just as good as his technical training and certification.

2) These new smaller devices are actually superior to the older/larger machines because the signal is processed locally rather than remotely (on the machine through the cable). He said he regularly has to deal with interference/noise from the cable itself.

The $2K price point is half anything else I have seen on the market. But I also suspect they'll be able to get this 10X cheaper within 4 years.

If nobody bothered doing this, this was going to be on my bucketlist of companies to start. It has massive potential, I'm glad somebody else is paving the way though.

Finally, medical imaging like this has already been proven to be a sweet spot and a hotbed for the modern day AI wave (even if we have another winter, it is the right tech for the right time). Similar fields, such as skin conditions, etc. has seen AI expertise trained and accurate well beyond even first world doctors.

This is a game changer for sure.

Very good! I'm looking forward to this high-end gadget future. Also stumbled today over Lumu Power[1] which I will probably acquire.

[1] https://lu.mu

What is the point of this, at all?

It won’t do a damn thing in the USA since it “can’t” be used for diagnosis. The incumbent industry will make sure it is not approved.

That's not how FDA approval works. It's actually easier for device approval when others entered the space first thanks to the "substantially equivalent" 510(k) process. Here's a link to Butterfly's premarket approval: https://www.accessdata.fda.gov/cdrh_docs/pdf16/K163510.pdf

Reply on news.ycombinator.com