A decent portion of my family are MDs. This is accurate.
However, this is also one of the places where it's quite disappointing to see human squishiness in action. When we opt to not receive data because we know there'll likely be non-malignant anomalies, we are also depriving ourselves of the raw data that allows us to discern non-malignant from malignant, data which would help very much in deciding on the cases we do receive.
A stream of known false positives is worth its weight in gold in any detection system, it's just that this specific system happens to contain squishy feely human doctors that are susceptible to short-term human 'Please don't spend more money than necessary, we're already in the red for this quarter' kind of pressures from their superiors.
Had we enforced a full-body MRI for every patient, we would quickly amass enough imaging data to know which anomalies are malignant and which are not to a very high degree, which would counteract the temporary cost increase in chasing false positives. The fact that we don't want to eat the short term initial cost means some of those things that could be uncovered by the scan, things that could save a person from a malignant deviation, is left unseen, untreated, to its natural end.
Determining whether something is (non) malignant isn't trivial. It often requires a more invasive test (e.g., biopsy, exploratory surgery). We try to avoid those procedures not just because we're cheap or afraid, but because they have a real risk of harming us 'squishy humans'.
For example, detecting prostate cancer a few months earlier is often not a huge win on its own. It is even less of a success if doing so requires that four people have allergic reactions to the anesthetic and two others acquire some kind of infection at the biopsy site.
It's certainly possible that more data would eventually let us avoid most invasive followups, that's nowhere near certain. Even if it were, the ethical calculus is still pretty tricky.
”Determining whether something is (non) malignant isn't trivial. It often requires a more invasive test (e.g., biopsy, exploratory surgery)”
Not necessarily, waiting can in many cases tell you whether something is malignant. If we had, say, a century of monthly full-body MRIs of a million persons, together with their history and cause of death, the technology to align those scans across time, and the technology to analyze such a data set, a retrospective cohort study (https://en.wikipedia.org/wiki/Retrospective_cohort_study) probably could uncover quite some interesting and useful information.
If, at every step, you pick the option with the highest expected QALYs (https://en.wikipedia.org/wiki/Quality-adjusted_life_year), and you manage to make doing the MRI very efficient, I think doing that experiment even might pass an ethics committee (at the cost of making the analysis harder)
(might because one could argue that the patient would be better of if the money spent doing those MRIs were spent on something else)
Unfortunately, doing such an experiment isn’t practical (¿yet?).
First, the cost would be insane. A century of monthly MRIs is 1200 scans per person, or 1.2B scans to finish your hypothetical dataset. We pay about $US 550/hr for scanner time, using a research scanner that's subsidized (i.e., we're just covering costs, not making a profit). The article doesn't say how long the scan is. You can burn as much scanner time as you want chasing resolution/quality, but an hour seems reasonable. That comes out to $660B, which is...a lot, and we haven't even paid staff yet!
Second, we have sorta tried this already. There's a massive neuroscience project to scan tons of brains called the Human Connectome Project. They have 1200 subjects, some scanned multiple times, behavioral measures, health outcomes, the works....
People have certainly found stuff in the data (I'm using some of it right now), but it hasn't lead to wild breakthroughs. There's a ongoing debate about whether this money would have been better spent on hypothesis-driven research instead.
For comparison, the entire NSF budget for next year is around $7B. That would just about cover the imaging component of building a speculative and ethically-questionable data set, assuming someone else pays for the biopsies, analysis, and staff.
(The NIH does have more money, but also funds trials, vaccines, and other stuff that we probably don't want to cancel for a century).
ML & better imaging algorithms should help a lot. Times could be reduced from 45min to 15min [1]. Scanning a beating heart can be reduced from 4min+ to 25sec [2]. Also ML can aid in comparing past images to current ones [3], which would give you quick insights into what changed.
Two ideas for time reduction I haven't seen discussed but perhaps might also help:
1. Don't scan at the same resolution across the whole body each time. Instead, focus on anomalous places that you wanted to monitor from the first high-res scan, or places that look anomalous in low res in the latest scan. Then dial up the resolution in those areas.
2. If better imaging algorithms existed that could account for very slight movements of the body (ala the heartbeat one above), perhaps prep time could be reduced by changing the physical layout of the scanner itself. The whole process of lying down and getting your head or whatever mounted in their plastic frame, then lifting the gurney and slowly sliding it into the machine.. it's all very slow.
Instead of laying down, what if you could just walk in and be upright and get scanned relatively quickly - basically a slightly longer but similar experience to a chest X-ray. There's upright MRIs right now but they're not very high res (0.7tesla instead of 1-3), very few exist, and I'm sure they still take some time to complete scans and have lots of error correction extra scans to correct for patient movement.
However, even with less time in scanner, there's a lot of fixed time cost of scheduling and patient prep (remove all metal things, wear this gown and these ear plugs, please don't freak out its really claustrophobic patient messaging), as well as still needing technicians.
I'm just a patient who's gotten some MRIs but it definitely feels like there's ways to cut the time down significantly.
Your first idea is pretty common. There's usually a very coarse "localizer" scan at the beginning of a session, which is used to set the field of view for subsequent runs. The whole body scans are (at least in theory) meant to find tiny things that are asymptotic though, so I'm not sure that searching through (say) 10mm slabs will help much.
Open, upright scanners do exist, but they're lousy. The goal of the (big) magnet is to produce an incredibly strong, uniform magnetic field. Due to physics, this is much, much easier to do with a torus-shaped magnet than a 'U'shaped one. Even so, there's one point ('isocenter') where the magnetic field is maximally flat and the quality will be best. The gurney moves to point the region of interest (e.g., your head) right to the isocenter. That's why the tech often uses a little light or laser to find your position, rather than just asking you to scooch. Motion is also, as you alluded, a huge problem.
I hate to be discouraging, but I am excited to see people actually thinking about MRI on HN!.
MRI is all about the protons. Under normal conditions, the protons in your body are all spinning ('precessing') around their own axis, but they're disorganized: each proton's axis is pointing in a different direction. They're like little wobbly tops drifting through space.
In an MRI scanner, the strong static magnetic field (B_0) forces the protons into alignment, so that their rotational axes are now lined up with the field's north-south axis. The field needs to be very strong for this to work, which is why MRI systems usually have expensive superconducting magnets.
Now that we've created a nice organized system, we're going to destroy it. A quick radio frequency burst energizes the protons and 'knocks them over' so they're no longer aligned with the field. Once the pulse ends, they 'relax' and realign themselves with the magnetic field, releasing some of that energy as they do so.
Sensitive detectors around the subjects' head detect those emissions and use it to determine how long it took for protons to realign themselves with the different components of the magnetic field. T1 is the time (or formally, the time constant) needed for relaxation parallel to the static field; T2 is the time needed for protons to relax to the transverse component. The T1 relaxation time for fluids is on the order of seconds, while fatty issue is more like 50-150ms. In the brain, grey matter has a relaxation time of 1.3 sec, but the fat-coated white matter relaxes much faster (~0.8), which makes T1 images very useful for examining brain anatomy.
Hopefully, this little crash course has revealed one of the bottlenecks in MRI: the actual signal being measured is slow.
Of course, I've massively oversimplified things and didn't explain at all how we localize these responses. The proton's precession frequency depends on the magnetic field, so by changing the static field slightly (across space), we can measure T1 at different locations, and sometimes even overlap measurements. You can't switch the field too fast though, or you'll start to induce currents in the subjects' nerves, which hurts. This is actually the principle behind a brain stimulation technique called transcranial magnetic stimulation.
There is still tons of room for improvement. Stronger fields lower the relaxation time, so the scans are faster (and the relaxation time estimates are better). Improvements in the RF coils help a lot too: the signal being measured is very faint and there's a lot of self-cancellation. On the software side of things, a lot of effort has already gone into designing clever pulse sequences—and the sophisticated signal processing needed to interpret their results.
Things will obviously continue to get better; I was just looking at some data from ten years ago and it looks awful compared to more recent scans.
That said, the "just use machine learning!!!" tone in some of the comments is kinda frustrating. Most of the people in this field aren't dummies--if it were as easy as downloading PyTorch, someone would have done it already. It turns out that the biology and physics are both stupendously complicated (and fascinating too).
“Because it’s too expensive,” seems like a great opportunity for some clever startup to figure out a way to make it less expensive. Computers used to take up entire rooms. Flying across the country used to be insanely expensive. Cell phones used to cost a ton of money per minute. Reducing the cost of scanning, or developing entirely new scanning tech isn’t science fiction, it’s the future.
The doctor gets $98. The $550/hr price I quoted is for a research scanner where no one is even trying to turn a profit; they just want to pay off the machine and its operating costs.
The machines themselves are not magically cheaper in Japan; they're just being paid for through some other route. If the scanner were somehow free (gov't grant?), 98$ seems pretty reasonable for the labor.
Again, that’s what you pay (near as I can tell), but that’s not what it costs.
They’re also “free” (at the point of service) in the UK, but the scanner and helium are not a generous gift from the fey folk, nor does the Queen volunteer as a tech.
These things cost money. Any study you do is going to have to cover expenses—-including their share of a temperamental, multimillion dollar machine (or find a way to dip into the same accounts that cover its clinical use).
A medical grade MRI scanner costs around $1 million, and maintenance costs are circa $100,000/year. With good utilisation, say 3000 hrs/year, the machine costs might be around $70/hr.
The $550/hr you are paying may be because you are using a particularly fancy scanner, or because utilisation rate is low, or because the amount is loaded with overheads, or w/e.
That’s in the ballpark for a 1.5T, but those are fairly old. About $1M per Tesla used to be a decent rule of thumb, but it’s come down a little at the low end. Still, I would be amazed to see a 3T for anything below $2M.
As for the $550/hr, it’s probably true that research scanners have lower utilization and higher costs to support all the weird stuff researchers want to do. An outpatient clinic specializing in knees can run much leaner. That said, that rate seems to be pretty standard across universities and I maintain that it's a very reasonable estimate of the cost. For example:
Most of these do not include F&A. It’s already coming out of the grants and external users pay a "Dean's Tax" on top of that to cover the missing overhead (which can often double the price).
> For example, detecting prostate cancer a few months earlier is often not a huge win on its own. It is even less of a success if doing so requires that four people have allergic reactions to the anesthetic and two others acquire some kind of infection at the biopsy site.
Then the doctor should recommend not attempting that type of detection.
I find it quite hard to believe that, in an ideal world (!), less data would actually be better. The doctor should still have the ability (at least in theory) to decide what information is worth investigating further, and what should be ignored.
Sure, in the not-for-profit Vulcan medical system, this would might be grand. Even in our world I'd agree that in isolation, more information is always better.
The issue is that this sort of information isn't free. The full body scan isn't free. Following up on it, which is the whole point, 'costs' increased risks of adverse outcomes (from biopsies), it 'costs' emotional anxiety, and it costs even more actual money.
The argument is that right now, these whole body scans aren't worth the cost. You're more likely to find a few things you should ignore (but will probably worry over) and a few things you should follow up on (but will be benign). You're less likely to catch something deadly and just in the nick of time.
In the US system with malpractice lawsuits, there will be followup in most cases. US Doctors are still furious over the amount of unnecessary testing that occurred as a result of the "whole body scan at third party facility" fad from a decade or two ago.
I think you're correct in the current context. However, this strikes me as a sign of some kind of paradigm shift waiting to happen, enabled by the reduced cost of data gathering, data processing, and improvements in AI.
There was a time when many things which are mass produced now were once handcrafted luxury items. Perhaps, as technology improves, it will one day be cost effective to take a whole body scan into account. I can imagine a day when an AI would digest the information from a full body scan, then autonomously plan robot keyhole surgery to take many of the biopsies.
It's certainly possible that more data would eventually let us avoid most invasive followups, that's nowhere near certain.
I think it's likely to happen in decades, if our civilization doesn't fall soon.
Perhaps, as technology improves, it
will one day be cost effective to
take a whole body scan into account.
That's the really funny thing about this: A colossal library of full-body imaging studies over time is exactly the kind of technology that would make that library useable. Think about Google 411 and similar projects: Voice recognition was unusable until people dumped a ton of money into bootstrapping the necessary datasets and now I talk to my phone every day. Similarly, I bet that a bunch of that "anomalies are indistinguishable on images" is solely because we don't have enough images about non-malignant anomalies over time. The data itself would be the technology that makes it usable.
It's nice that you think those are costs worth remaining ignorant over. However, when the ignorance is about things that literally might kill me, my personal opinion is that it's worth it.
Both options have risks and both options can kill you. Namely:
A) Doing nothing may allow a problem to fester until it either kills you or becomes detectable with standard treatments. However, you have no exposure whatsoever to side effects, infections, or medical errors for the things that are missed.
B) Doing whole-body MRI 'fishing expeditions' decreases the risk of missing something, but increases the likelihood of possible false positives. Each thing flagged by the scan now needs to be ruled out (otherwise, why bother?) and this usually requires invasive procedures with their own risks.
With our current imaging and understanding of biology, Option A may be less likely to kill people. Opinion doesn't even have to enter into this--you can literally run the numbers.
It's worth noting that we limit testing FOR EVERYTHING. When you get a blood test, you could be screened for everything from ADH to Zika. You aren't though, because you're almost certainly a perfectly healthy outlier on a few tests; that's just how statistics work (someone has to be 3 sigma above the mean). Instead, tests are picked and interpreted based on your symptoms and circumstances.
Speaking of blood tests - where does the uncertainty come in there? Is the measuring equipment noisy, so they might detect some marker when it is actually not present? Or is it that you can actually have a high concentration of whatever they are looking for in your blood but it somehow just be normal for you because you are an outlier?
I tend to think it is the former but can anyone clarify?
The normative values aren't handed down from God, or some definite evidence of health. Instead, they were the typical range (mean ± 3SD, percentiles, etc) from a sample of healthy people, possibly matched for demographics (e.g., sex).
Suppose we set each range so that it contains 99/100 healthy people. It would only take about 70 independent tests before you have a 50:50 chance of being outside that range on one of them--and that's assuming you're an exact match for the reference population.
There's also error in the measurements themselves. Some of it is errors in the actual procedure or preparation: cross-contamination between samples, dust blows into the well, etc. However, many of the tests are stochastic too. Some measure binding between stuff in your blood (e.g., antibodies) and a "probe" that's designed to detect them. This usually works, but non-specific binding can cause false positives (the probe binds something that is similar to, but not quite the same as, the target). Other conditions can cause false negatives. The "front line" screenings are usually meant to be cheap, fast, and biased towards false positives.
You might be interested in a fairly common statistics question about screening. Suppose you had a test that is 99% reliable: 99% of sick people test positive; 99% of healthy people test negative. The disease itself is somewhat rare--only 1% of people have it. If you test positive, what are the odds you're actually sick? The answer[0] will explain why we don't test or scan people for tons of rare diseases, even with very reliable (and free) tests.
The idea that lots of full body MRIs will necessarily lead to better discrimination of malignant vs benign anomalies assumes that the imaging is able to detect differences at all, which often is not true. Some benign and malignant abnormalities can look _exactly_ the same on imaging, particularly at an early stage, when you could make a difference, for example lung nodules. Nevermind the fact that many benign entities can undergo malignant transformation at some point in the future. Or that MRI is only a good test for a subset of cancers, CT is better for others. All these issues help explain why doctors are averse to full-body scanning everybody. It is neither cost-effective nor medically sensible.
I think their statement is more general than what you're making it out to be. The goal of receiving better and more data puts pressure to have better imaging technology.
I think there are perverse incentives at work, too. If running the MRI machine cost $100, people would do it. If it can bill insurance $100k, and then any follow up might cost $500k, there is a rational aversion to doing it as it might make the _system_ less good for everyone if insurance cracked down on exploratory imaging as a billable thing (obviously, they already do this to some extent).
Better technology is unlikely to be cheaper, unless it is so new and novel that it can get its own unique billing codes, and if it becomes ubiquitous so it cheaper and more available than existing machines which are not yet depreciated.
>If running the MRI machine cost $100, people would do it.
There are different kinds of MRIs, but the actual direct costs for running an MRI without preparatory procedures is around $100 (preparatory dye injection may be another $50 cost). Marginal costs are lower still.
> Had we enforced a full-body MRI for every patient, we would quickly amass enough imaging data to know which anomalies are malignant and which are not to a very high degree
How, pray tell, are we actually determining which are malignant and which are not?? I’ll give you a hint: the charitable explanation is you are completely ignorant, the less charitable one is that you consider Mengele and Ishii modern medicine heroes. And that is even if the determination by imaging alone is even possible.
Once again I have to point out on HN... not everyone in the medical world is an idiot incapable of analytical thinking, it is just that a lot of us care about ethics. Our patients, even in research, are more than just numbers.
I don't even think it's about the ethics that necessarily make this statement wrong (though, you are correct in that HN needs to consider basic ethics on a much more frequent basis).
In the case of these detections, the ground truth can be very subjective and is certainly not a simple explanation. Is the ground truth read by a radiologist? A radiologist with fellowship training? Multiple radiologists? Biopsied? Followed for 5 years to see if the patient actually died of their tumor?
The complexity is a lot more than the underlying algorithms and computer science (though they are important!).
> How, pray tell, are we actually determining which are malignant and which are not?
When people on this forum speculate about things like this, I assume they are thinking of some future ML application. Of course you're right though, that is not doable today. The point of the OP is that if we collected more data perhaps it would be possible.
> the less charitable one is that you consider Mengele and Ishii modern medicine heroes.
The ad hominem attacks against the OP are completely unnecessary.
> I am actually an MD.
The start of your post feels like "appeal to authority."
What I mean is, just because the person is "actually an MD" doesn't mean the person has expertise in this specific area yet the person is using it to claim such. I.E. some might call me a computer scientist, but that doesn't mean I have enough expertise in machine learning, a sub-discipline, to say definitively that the OP's idea about using full body scan data is feasible.
Should I have started my post "I am actually a computer scientist." No, right? That would be an appeal to authority that doesn't prove I know anything about what I'm talking about. Now, if the person had started with "I am a radiologist who has looked at empirical data about full body scans versus more selective scans." that would be a more useful and valid statement.
It’s not a big deal. I only mentioned as a response because the op started off saying they were from a family of MDs.
Also the vast majority of practicing physicians are expected to have detailed expertise to understand the indications, application, and interpretation of imaging within their speciality. Most imaging orders are issued by medical or surgical staff, not radiologists.
My problem with your post is not that you're not a radiologist (that was just an example—and maybe I should've just written "person who has experience doing empirical studies of body scans"). It's that you use your status as "an actual MD" but that is clearly not enough and a real "appeal to authority" or "appeal to anecdotes." And then you combined that with some awful things that you compared OP to.
Looking through your post history you often jump to "appeal to authority" combined with name calling and it's a pattern that reeks of arrogance to me.
For example one of your posts starts out along the lines of, I'm not a 747 pilot, but I am a pilot... in a post about something specific to the 747 and you say the other guy was "talking out their ass." Again, not super relevant appeal to authority and some mean-spirited name calling.
> Not a 747 pilot (not sure why that matters here)... but a pilot. The flight positions are equal. The plane is designed to be flown the same from either flight position and this is routinely done. The guy saying otherwise is talking out their ass.
You just seem to love to tell everyone you're super special and smarter than them because of your accomplishments as a means of arguing and then put them down. It gives you instant validity to some of course, but I think it's just lazy. Your arguments should stand on their own merit. Not request to authority or anecdotes from your kind-of-related experience.
> How, pray tell, are we actually determining which are malignant and which are not??
A biopsy? Whichever method you would normally use to find this out?
You could even ignore all the information and just train your algorithm on whether the patient died in the next 5 years (after the scan/s).
Obviously you need data before you can determine anything automatically. But I don’t think everyone would be averse to this even if you do tell them it might be more likely to kill them due to excessive testing.
Had we enforced a full-body MRI for every patient, we would quickly amass enough imaging data to know which anomalies are malignant and which are not to a very high degree
These sorts of statements come across as naive optimism at best. And I like our long term odds of increasing clinical efficacy, and have designed and worked on several systems in this space that are in clinical use.
For those interested, I have posted a relevant screening-exam paper on the main thread that asks the question, "what do we find if we scan the brains of 2000 people?":
However, this is also one of the places where it's quite disappointing to see human squishiness in action. When we opt to not receive data because we know there'll likely be non-malignant anomalies, we are also depriving ourselves of the raw data that allows us to discern non-malignant from malignant, data which would help very much in deciding on the cases we do receive.
A stream of known false positives is worth its weight in gold in any detection system, it's just that this specific system happens to contain squishy feely human doctors that are susceptible to short-term human 'Please don't spend more money than necessary, we're already in the red for this quarter' kind of pressures from their superiors.
Had we enforced a full-body MRI for every patient, we would quickly amass enough imaging data to know which anomalies are malignant and which are not to a very high degree, which would counteract the temporary cost increase in chasing false positives. The fact that we don't want to eat the short term initial cost means some of those things that could be uncovered by the scan, things that could save a person from a malignant deviation, is left unseen, untreated, to its natural end.