Some initial thoughts as a practicing radiologist:
- This looks really cool and I hope they keep innovating on this. I love seeing new modalities develop and despite my (many) reservations and criticisms, if even one good use case comes out of it that truly helps people, it's tech money well spent imo.
- They show the reconstructed images as though they are a low resolution CT, and promise that quality will improve as they iterate. This is cool, but ultrasound is not CT. Ultrasound cannot image the lungs, as they are filled with air. You cannot find bone lesions, as the sound waves do not penetrate the cortex. You cannot image many structures in the abdomen if they are surrounded by gas-filled bowel. The brain is encased in bone, so you might get some penetration but it will be very limited. Even with theoretically perfect AI reconstruction, these scans will not be true "full body" in that there will be structures that are not reliably imaged. Imagine paying for weekly full body scans for years, everything looks fine, then its the lung cancer surrounded by air and invisible to ultrasound that kills you (that's why we use CT for lung screening!)
- The images they show are very cool, and do appear to show the correct structures. I realize this is early, but fuzzy shapes of organs is very, very far from medically useful. The whole point of screening is to identify problems early, often by definition, small. This technology looks like it will be best for seeing large, superficial (close to the skin) structures, whereas for effective screening, you want the opposite - small, deep structures.
- "Incidentalomas" or unexpected, probably benign, findings are annoying to physicians, but I in general have no problem with people collecting data on themselves where they can. To me it's similar to heart rate monitors or home blood pressure cuffs. The main issue here is education, so that patients know what the data is and is not telling them. The more complex the data, the more difficult that is.
- Many people mistakenly believe that early diagnosis is the final boss in medicine, that if only we could find every cancer early we could prevent all those deaths. There are, in fact, many, many other hurdles and bottlenecks. Many chronic, expensive diseases do not have clear imaging manifestations. The claim that "it's completely possible that with enough early imaging in the future, the world could avoid 30% of all deaths and 50% of all healthcare costs", I think, to any practicing physician, would sound completely divorced from reality.
> The brain is encased in bone, so you might get some penetration but it will be very limited.
Radiologist as well. Remember this is full wave inversion not pulsed wave B mode. You can get much more useful information from both high low frequency and capture transmitted waves.
There is promise with this and we use it for example with MRgFUS. With advanced computational models or patient specific CT/ZTE MR aberration correction it is theoretically very feasible to image the brain with ultrasound, whether that’s more useful than say portable low field strength MR is a different question altogether.
> This is cool, but ultrasound is not CT.
Not to be pedantic but since this is a tech forum I would clarify that FWI US is computed tomography by definition (at least in this and many applications). Gas degrades conventional CT too, it’s just worse with US as you have little to no forward propagation and of course innumerable interfaces in the lungs to reflect and scatter.
I'm not a medical doctor at all, just an engineer who works in medical devices and I'm definitely sceptical about this.
I'm not totally sure of the value of an imaging system that only gives you very low resolution images if they're not accurate enough to determine anything from. You'd need a secondary CT or MRI anyway so why not skip to that?
My real concern is the dependence on external servers to reconstruct the images
Edit: From reading other people's comments, people are acting as if this is the first device trying to sell itself as improving pre-diagnosis imaging and this is totally revolutionary. This is not, and if any of the other products have convinced the entire medical industry that frequent imaging is beneficial then neither will this
"why not skip to that?" MRIs and CT scans are expensive, require referrals, and you usually can't get them without believing you already have an issue. If this technology can get to a point where it's high enough resolution, cheap enough to just have at spas, and shared across the world then people will be able to know if/when they should get that secondary scan before symptoms start.
Probably cheaper and substantially better contrast resolution to use low field strength perma magnet MRIs with advanced computation to be honest.
I do a lot of MRI analysis including segmentation of small structures in the hippocampus called the hippocampal subfields. To collect these segmentations, we collect partial-field-of-view high in-plane resolution T2-weighted images on a 3 or 7 Tesla magnet. These sequences are generally only included in research protocols if the research specifically cares about hippocampal subfields...therefore they are rarely collected. There have been attempts to enable segmentation of these small structures using lower resolution T1-weighted scans, leveraging deep-learning or other models trained on concurrent T2w high resolution scans and the lower resolution scans within the same subject, allowing the model to predict the higher resolution information from the lower resolution inputs. This produces spectacularly beautiful segmentation on shitty data. Data whose resolution is about the same as the thickness of the structures you are segmenting or less. The problem is this: 1. The lower resolution image barely has any information in it on these smaller structures 2. The accuracy of the resulting segmentation depends entirely on how much the person fits the training distribution. But much research is on specific populations: children, autism, etc. 3. Some big names in imaging analysis tools have published these tools, lending their credibility to them. 4. The beautiful segmentations and (3) tend to convince non hippocampal experts that the resulting data is trustworthy, especially to an eager beaver researcher trying to maximize the impact of their already collected datasets.
I've rejected a number of papers for this.
But my point is this. Midjourney Medical might train a model to produce pretty images with this technique, but the more they need to depend on deep-learning models to get usable data, the more that the match between the training distribution and patient will matter.
this is why i always come to hacker news for the expert opinions. thank you for being critical yet optimistic.
You don't know if the opinion is expert. You don't know who that person is or what credentials they even have. Blindly trusting something that sounds right is a terrible way to inform yourself.
I'd agree more if MidJourney had decided to announce their product plans with a white paper instead of glossy 'product vision' marketing spin and virtually no information as to how they hope to solve the vast technical leaps necessary to convert the transducer chip they licensed from Butterfly's low-cost, handheld, pocket-sized USB ultrasound device into a contactless, 360 degree, 60 second full body scanner.
Given MJ's extraordinary claims and lack of detail, I thought the GP's response was well-calibrated, especially given MJ's unfortunate choice to lean into vaguely implying this has 'medical' utility, despite providing zero evidence (or even plausible theory) their approach could ever have diagnostic value greater than Butterfly's FDA-approved, handheld, full contact USB pocket scanner which is available now and plugs into a mobile phone. They are using 40 of the exact same transducer chip (designed for full contact use) from 200-400 times farther away. You can use the existing full contact Butterfly scanner today and just move it to 40 different angles. It would take a couple minutes longer, provide vastly greater resolution and is proven to have diagnostic value.
Bayes Theorem...the chances this rather milquetoast and balanced analysis was written by someone with no knowledge is vanishingly low IMO.
With LLM tools being widely available, it's extremely high, imo.
Idea for a website/documentary -- have experts respond to a piece of news, or provide commentary. Put a few expert pieces alongside a few LLM outputs, have people guess/work out which is which. Have the same people tell you why.
If on a website, rank the results; present the 'how I worked it out' info for the best spotters (and you could interview them). Keep the answers secret for a few weeks, then reveal them in a way that the game is still playable.
It's repeatable, every few months you could interview new experts (or the old ones again), get new models.
Kinda like the critical thinking version of images of a pelican on a bike.
I love your idea and would enjoy seeing the results of that controlled experiment.
I'm also interested in the broader impact of using LLMs in place of web search for general Q&A when we want 'to know things'. It's pretty clear the way LLMs are being used for knowledge acquisition now is often less accurate while 'feeling' more certain. Even if we set aside explicit hallucinations, I suspect it's still less accurate.
Hah, I'm neither a bot nor written with any help from an LLM, but I'll take the fact that you can't tell the difference as a compliment :)
It's not meant as a slight against you, but an observation on people blindly trusting Internet commenters in a time where our trust is (or should be?) at an all-time low for such content, so we should check our priors to ensure what we're consuming can be trusted and is verifiably true as LLMs exude confidently incorrect behavior.
I want to point out that you are posting from a 5-months old account (squarely within a time frame where LLM-powered accounts would be created), with an UUID-sounding username, claiming a 10-year-old professional history in the field, and using those credentials to bring up the possibility that a 14-year-old account claiming to be an MD with `md` in the username and _lots_ of comments is LLM-generated and asking for skepticism.
It's not particularly helpful; you could easily have done the 5 minutes of work.
Congrats, you've discovered people are always joining an anonymous platform, and some want pseudonymity? If I put "md" in my name, am I suddenly a doctor then? Your reading comprehension needs some work, I've been working for 16 years and it isn't being used as a cudgel to appeal to authority, but sharing some context about who I am and what my perspectives are.
Someone using md in their account name for 14 years is really committing to the bit if they aren’t a doctor.
I went back in their comment history before LLMs existed and found comments where they claim to be a doctor and sound like they know what they are taking about. I’m not a doctor but my wife and many of our friends are, so I know what they sound like.
I appreciate your feedback, I did not dig through their comment history to make that discovery, but saw someone showing full trust in a random user on an anonymous forum and wanted to encourage more critical behavior.
I get the intention, but you made an assumption that the person you replied to didn’t spend 30 seconds find some more supporting evidence before they made their post.
But as far as trust goes, Hacker News has historically been a fairly high trust community. LLMs have the potential to change this dynamic, but I don’t think encouraging people to assume that every post is an LLM is helpful. I don’t think a community with that level of distrust is possible, and at that point we should just all walk away.
The parent comment is demonstrating the exact type of internet savvy critical thinking that you’re ostensibly arguing for, but you seem to be reflexively defensive.
Making an ad hominem attack on a random poster asking for critical thinking on who to trust and then making false claims is not "internet savvy critical thinking" since they can't string two thoughts together coherently (a bio blurb vs. an appeal to authority in a random comment).
Not taken negatively, and you're right. This is even more difficult with things that are opinion, and not clearly verifiable.
We don't know if any of the opinions on HN are expert for that matter. Yet somehow, it's survived and grown for 20 years.
"Blindly trusting" is not the same as "learning some new questions to ask about the validity of the given claims". The Midjourney announcement is not providing detailed medical credentials either.
they have md in the username
True.
https://en.wikipedia.org/wiki/On_the_Internet,_nobody_knows_...
> ultrasound is not CT
They're using "CT" in its literal sense: tomography*, using computers. In this case, ultrasound is the penetrating wave rather than x-ray. It is of course a very different thing than what the medical world knows as "CT" today.
*https://en.wikipedia.org/wiki/Tomography
we can do tomography on any round-robin-rectify multi-pov source, doesn't have to be x-ray is just de facto use in medicine, closer to at min marketing ed problem
Not a physician. Some observations on these statements.
The predicate is "given how we practice medicine and the limits of humans ability to interpret the imaging modalities we have."
The more specific predicate is "for my specialty would this replace or prove superior to the tools that I have?"
Both of these are totally reasonable, however the history of medicine, and science in general, is that creating new ways to look at things has a tendency to reveal information that we never knew we needed.
For example, for years I thought of blood sugar as something that was either in a good or bad range. Then I tried a continuous blood sugar monitor. The full picture of the body's response to specific foods that I ate was eye opening. There's so much more to learn when you get a higher resolution (temporal in that case) view into your body.
Another wonderfully hopeful example is the retinal imaging ML work done by google. A completely non-invasive image of the retina for diabetic issues, that also happened to be able to predict things like age, sex, smoking status, previous cardiac events and more! Just take high-res pictures of things! The body is interconnected in ways that you can infer from one system so much about others.
So while I don't think anything the Dr. said is "wrong", I think it represents a very common blinkered mindset of pragmatic practitioners who need to deliver reliable performance daily.
Your points are well taken, and I think this is the fundamental struggle of anyone who works in a narrow and deep field. It's truly difficult to see things from a different point of view sometimes. It certainly could turn out that this ultrasound setup gives truly new information, but, it isn't really a new way of generating an image, it's the same physics we've used to generate images from sound waves for decades, and that modality comes with some pretty hard physical limitations that this demo does not directly address. Time will tell, if they don't run out of money. I'm hopeful!
It's true that sometimes deeply experienced professionals are less likely to discover or accept novel methods based on new approaches or technologies. Unfortunately, in this case MJ's proposed product is being deceptively presented to appear as if it's the kind of bold new approach we all hope to see in medicine - but it's not. It's a repackaging of an existing product that's already on the market.
MJ is buying the transducer chips used in Butterfly's low-cost, handheld, pocket-sized USB ultrasound device (it's not an R&D license, they're literally buying the same chip). The repackaging is to turn it into a contactless, 360 degree-at-once, 60 second full body scanner. Every aspect of the repackaging provides the same singular benefit over the Butterfly device: convenience. Unfortunately, every aspect of the repackaging has the same two downsides: lower resolution (meaning lower diagnostic value) and higher cost.
Spoiler alert: moving the imaging transducers 200-400 times farther away from your organs and introducing a large volume of water between the transducer and your skin in no way improves resolution or diagnostic value (quite the opposite (exponentially!)). Having 40 transducers on a hula hoop ring that far away offers no value over having one transducer much closer and moving to as many angles as necessary to image the volume of interest - except it might be a few minutes faster.
So, this isn't an "exciting new approach to medical imaging." It's a marketing repackage of an existing medical product into a non-medical, higher-cost, 'spa experience' with trendy, tech-adjacent appeal and vaguely medical-ish window dressing (it's carefully disclaimed has having no medical value in this form). Since the exact same chip is already available in a much less expensive, far more ideal form that's fully repositionable to any angle, is closer (and can deform skin to get closer still), the real question is how much medically-relevant diagnostic value could MJ's repackage of the same chip offer over the existing better, cheaper implementation? Butterfly's version is already FDA approved with proven diagnostic value. And all of MJ's hype around 'AI-powered' isn't about improving diagnostics, it's only necessary to recover at least some signal from the hurricane of noise and multi-path issues created by the adding MJ's cool-looking convenience features. But slowly being lowered into that tank with the neat ambient light ring sure looks sci-fi, right?
Exponentially would make it computationally intractable by definition, and is the proposal suggestion that having simultaneous sensors can be used to reconstruct more information than the conventional way of using a handheld single sensor? Is that plausible?
> The full picture of the body's response to specific foods that I ate was eye opening.
you were surprised to find out that stress and carbs raise blood sugar?
I'm doing this now, and the thing I find most surprising is that there seems to be some invisible persistent state that gives high or low sugar a sort of momentum - so if I've been doing a fair amount of physical activity for a month, I can lay slugabed for two days and still drive to have a slice of pizza in the afternoon without trouble; but if I've been slacking on the activity piece, or arguing with my spouse, or travelling and eating a lot of dubious things, I can walk five kms to a pizza place, eat the slice, walk the 5 kms back, and it will still spike. Also I have issues with the CGM being higher than the prick-blood test, like 40 points higher rather consistently. A1C is still dropping, but the CGM numbers are more directionally accurate than numerically accurate.
I'm far more willing to trust someone in the field every day (you), over keyboard warriors who want to validate their new toy is alive.
They spent a lot of time selling the spa, and not a lot of time showing us the data.
That's because there isn't any data yet, at least not enough from real patients to be meaningful. I would love to see some of the raw imaging data they have generated though, if that's what you mean.
> I would love to see some of the raw imaging data
Just look at images from the Butterfly IQ3 handheld ultrasound device which has been on the market a while (https://www.butterflynetwork.com/iq3). Midjourney is repackaging 40 of the exact same chip around a big, non-contact ring. Since MJ is placing the devices 200 to 400 times farther away from your organs and sending sound waves through a large volume of water before contacting your skin (instead of a thin smear of gel) the images will be much lower fidelity.
Given that current ultrasound probe technology (including butterfly) relies on the probe being essentially in contact with the tissue being imaged, it’s hard to imagine how this set up can be effective with the imaged volume so far from the transducers, since there will be a huge amount of dissipation in the water bath, but maybe they have found a way to solve that? Also, I imagine that the quality of the images, such as they are, will fall off very quickly in larger patients. Will be interesting to see.
Exactly my concern too. There are techniques like synthetic aperture focusing which can correct some of these errors to some degree but they're complex and have harsh limits of their own. It's always better to not have the errors introduced by the distance and water volume in the first place. The thing which makes no sense about this entire approach is we already can not have those errors.
I've been looking up relevant data and reading some papers to determine if I'm missing something there but, so far, the approach looks pretty much 'all downside' with the few upsides being: 1. Faster to image full body, 2. Don't have to have some technician poking you with an ultrasound wand, 3. Looks cool?
But I'm just an imaging and DSP guy, you're the actual radiologist. If you don't mind there's one question I'm not sure about. Trying to 'strong-man' the product concept, the only potential benefit of the approach I haven't crossed out is if there's any meaningful value from having additional simultaneous receivers off-axis from the emitter? I mean value which can't be gained from just moving a single emitter to another axis, grabbing more images and then cross-registering those. Even then, the off-axis receivers are always co-planar with the emitter, which seems like it would greatly limit any utility.
The downside column I've got so far is vast... and it's not just distance, there's also the turbulance in the water, micro-bubbles from the ongoing submersion of body and platform into the tank, the thermal disruption at the boundary layer, the fact the human is freestanding with no support while being submerged means they'll be far less stationary than a human comfortably reclined on a ultrasound table, it goes on and on.
There also isn’t a spa yet.
I have a friend who may be dying. It's not cancer but equally dangerous. They caught it early but they're powerless to do anything. We're just watching, waiting, and hoping.
> "Incidentalomas"
Good word for it. We were a bit worried about something that showed up my my mum's scans, as if an inoperable tangerine-sized lung tumour wasn't a big enough problem. It was a round dot about the size of a pea on her adrenal gland that lit up like crazy in contrast dye. Now, that as you probably know was a worry because oh shit, lymphatic system involvement, that's going to spread like crazy.
But after two years of immunotherapy, and about six years after being detected while the lung cancer is gone, the pea is still there, still as bright as ever, neither getting bigger nor smaller.
No-one is in a hurry to poke at it and see what it is, just in case that turns out to be a mistake.
89 this year and at least got to see her grandchildren start primary school, so good work from your lot and the oncologists!
Now I know it's an "incidentaloma" :-)
> "Incidentalomas" or unexpected, probably benign, findings are annoying to physicians
For a lot of these things I wonder why they don't just do multiple scans just to see how things develop. Is it a cost issue?
That's generally exactly what we do, which if we need to follow 2x or 10x incidental lesions in the population, leads to cost and availability problems. A lymphoma patient in remission needs follow up scans too, and I don't want them to have to wait 3 months because thousands of people are now following up their benign adrenal adenomas.
If you could dramatically reduce cost and improve availability, would this still be a problem?
What's the limiting factor that prevents medical imaging from getting cheaper and more available?
The machines are expensive (millions range per MRI scanner), staffing the machines nearly around the clock with highly educated technologists, repair/maintenance of expensive specialized machinery, radiologists to read each scan (esp with a current shortage), means it’s very expensive to set up and run an imaging center. Opening and owning an imaging center used to be seen as fairly lucrative. and many radiology private practices did just that, however, the economics have changed over the years with dropping reimbursements, staffing shortages, etc and now often these imaging centers are seen as a liability rather than an asset.
Excluding the cost of X-ray/CT/MRI machines, operating them, getting people to them and through them, sometimes injecting contrast, and sometimes dealing with side effects of said contrast, radiologists, I think. You can scale all of the above except interpretation. AI is the natural next thought for how to scale that part, but it's been thought that this would happen any moment for over a decade.
This is definitely a part of standard follow up for small findings and part of the guidance for incidentally detected lung nodules smaller than <8mm.
I think mammography is a great example. Many people are quite surprised to hear that the Positive Predictive Value of a screening mammography is only in 10-15% range. This despite mammography being a pretty sensitive test. This is because despite good test performance characteristics, applied across a large population of relatively health people, the 2-5% false positive rate is a large number of people.
Statistically overdetection leads to poorer outcomes because interventions have a risk as well. That's why everyone doesn't get a yearly full body CT scan, for example. The current guidelines are based on optimizing for maximum good, and believe it or not some things are best not known about because the risk of dying from it is about the same as the risk of the treatment.
Because people discovering "Incidentalomas" will be too freaked out to wait "just to see how things develop".
Cost, time, and for things like CT's, trying to limit your radiation exposure
Couldn't have said it better.