> it's likely this result in innovations that would drive down costs, improve accuracy, as well as producing a much larger corpus of data with which to guide diagnosis and reduce false positives.

Why is it likely? We already have a lot of MRI data. There are already a lot of incidental findings. It might also be an issue of the MRI not being able to produce enough information to discriminate.

> To use a software analogy, if your downtime detection system kept producing false negatives, would your solution to be just turn it off? You'd get some better night's sleep, but you'd pay for it when the system really went down and you had no idea.

The analogy is rather something like this: your downtime detector is not just a "ping" but a full web browser that tests everything and it sometimes flags things that are not actually issues. So you don't turn it off, but you only use it when you have another signal that indicates that something might be going wrong.

> Why is it likely? We already have a lot of MRI data. There are already a lot of incidental findings. It might also be an issue of the MRI not being able to produce enough information to discriminate.

This is the main reason. Well technically the opposite of the main reason but more or less it's the same. MRIs are extremely high fidelity nowadays and as a result it's really really hard to read an MRI. Every person is different and there's a lot of variations and weird quirks. You get all the data rather than clearly identified problem areas like you get with say a CT w/ contrast, etc.

That's actually exactly why it's important to have MRIs more frequently to be able to establish baselines and identify trends as they develop.

> That's actually exactly why it's important to have MRIs more frequently to be able to establish baselines and identify trends as they develop.

How? How do you establish baselines? How do you build a classification of incidental findings? It's very possible that you'll find a lot of types and not a lot of representatives of each type. And then you have to correlate that to actual clinical results, but the population will be so heterogeneous that it'll be really hard to find an actual result.

It's not just "let's throw more data at the problem".

When I say establish baselines what I mean is to establish baselines for the individual.

If you have records of the locations and sizes of various atypical structures and forms throughout the body going back for years and all of a sudden one of them starts changing in size at a rate disproportionate to its history, that's probably cause to dig a little deeper.

It's certainly not "throw more data at the problem". Instead it's about giving the data a time axis with some decent fidelity.

> and all of a sudden one of them starts changing in size at a rate disproportionate to its history, that's probably cause to dig a little deeper.

That sentence is doing a lot of heavy lifting.

- What's "disproportionate to its history"? Obviously something going from 1mm to 10cm is worth checking out, but what about something going from 1mm to 2mm? Might be a tumor, might be that the position is just slightly different.

- What about other less measurable factors? Example, border features. That's harder to measure and things like movement or different machines can change how the borders of a feature look. How do you know what's a baseline and what's not.

- How frequently do you run these scans? It's likely that if something "starts changing in size" suddenly it will start giving symptoms before you have your next scheduled scan.

> It's certainly not "throw more data at the problem". Instead it's about giving the data a time axis with some decent fidelity.

It's definitely throwing more data at the problem, and you're assuming that it's viable to give "a time axis with decent fidelity". MRIs are much more complicated to interpret than people think, and screening is a much harder problem too. There are a lot of studies testing MRI imaging as a screening technique (among other techniques) and they don't always show an increase in survival rates.

We do not have a lot of MRI data. The average person probably gets a couple MRIs in their lifetime, and this is biased because we wait until something is clearly wrong to get the MRI. If you want to find an MRI scan of an early stage asymptomatic cancer, the only data on that will be the exceedingly rare case that someone has something else unrelated wrong with them in the same general area and gets an MRI for that, and then just by chance also has the early stage cancer at the same time.

> we wait until something is clearly wrong to get the MRI. f you want to find an MRI scan of an early stage asymptomatic cancer, the only data on that will be the exceedingly rare case that someone has something else unrelated wrong

Not always. There are bunch of studies for MRI screening in high-risk populations for specific cancers. There are scoring systems for a lot of them based on imaging features and they do find asymptomatic cancers.

In fact, if you add low-risk populations to the studies used to design imaging scores, you might end up adding more noise and making the study more difficult and the scoring less accurate.

> We already have a lot of MRI data.

That's true but not in a useful way for improving MRI screening.

What we have is lots of days from people who were sent to get an MRI because they had a complaint.

That's a very different group than people doing screening.

And the fact that they have a complaint (or have known risks) makes it easier to classify, compare and understand the data.

There's a softer component to healthcare which is that people can overreact to medical results. If a doctor administers a scan, finds a handful of likely benign things but wants to administer another scan later on down the line, I'm probably much more likely to look for a second opinion that tells me to cut them out (even if it may not be medically necessary) than trust my doctor that "it's probably fine".

It's probably more accurate to use a software analogy about performance metrics. We measure random request spikes now and again that strain the system. It's probably fine, but later on down the line, something could change that results in an outage during one of these spikes. Do we proactively fix the problem even if no change is expected? Or do we wait till there is definitely a problem before taking action?

But surely this would decrease as we learned more from more frequent MRI scanning. Doctors and patients would be less likely to overreact, and we'd settle in on something better?

I'm not an expert though.

No, this is more like disabling logging because people are concerned the server is going down.

“Don’t worry about it, I don’t think it’s a real issue so we’re just going to ignore it”

But if you’ve never looked at the log before all these WARNINGs might be normal operation. It’s not turning off logging, it’s saying log at ERROR level.

That's a good point. How many people have a baseline MRI from when they were at their healthiest?

And if everyone had scans at their healthiest we would know better

Not to mention the malpractice risk and potential for extra income, which -- depending on where you live -- may strongly (if subconsciously) influence your physicians interpretations and recommendations.

Do you think a doctor is more likely to call something "possible cancer" and recommend that you either have a specialist do a biopsy (keeping in mind that many of these will be... hard to reach) or at least have a follow up MRI in 3, 6, 9 months?

Or tell you it's "pretty unlikely to be cancer, I don't think we need to worry about it" and then get sued for 20M when they are wrong about 1 in 100 cases (not to mention missing out on all the potential income from above).

At least in the US, the incentives here are grossly misaligned.

Incentives shape behavior in every industry, including medicine. Systems drive outcomes.

That is exactly why builders focus on owning aligned systems, not depending on broken ones. More on this at foundniche.com

> Arguments against proactive MRI scanning always seem to have a whiff of status quo bias to them

More and more European countries are currently adopting Lung Cancer screening programs. It's usually limited to people with a certain amount of cigarette-pack-years, but still gives the opportunity for driving more of the innovation you're talking about. I think the main challenge at the moment is that nothing in healthcare is prepared of looking at those scans effectively, a radiologist has full medical education + additional specialization - without effective procedures you'll never be able to provide full-body scans with any meaningful impact.

It's helpful in justifying screening that non-small cell lung cancer treatments have greatly improved in recent years.

Not just cost also time. When I had my scan it took 45 minutes and two trained nurses.

Ofcourse in America poor people don't have access to healthcare so it's a lot easier. But in a universal healthcare system everything has to be rationed.

False positives trigger more diagnostics some of which can be harmful, not just psychologically but physiologically as well.

If false positives are ok, why not build a down time detector that rolls a die every 5 mins and alert on hitting a 6.

Additional diagnostics can also be very expensive. Articles like this don’t seem to understand the overall costs to a health system with decisions like these. And that cost eventually does go down into the pockets of patients one way or another.

I think the point of the conversation is that if we take the predatory capitalism out of the way, using MRIs could potentially be a net benefit overall for everyone.

OECD data (most recent available, around 2020–2022): MRI units per 100K population: United States ~3.6, Canada ~1.0, United Kingdom ~0.7

I would argue that getting "predatory capitalism" out of the way has sharply curtailed MRI availability where that's been tried. Maybe we should loosen the leash on capitalism a bit to get better care...

I'd argue that malpractice risk has at least as much negative influence on a physicians judgment.

It's perceived as much less (medico-legally) risky to "do something" (or more often "refer the patient to someone else to do something") than not do something.

Harmful to shareholder value.

Personally I’d rather have cancer checked out rather than have a “wait and die” approach

Insurance-based medical systems mean the patient has transferred responsibility for saying no to those actually paying the bills. They have to draw the line somewhere.

The United States Preventative Services Task Force[1] reviews studies and meta-studies to make recommendations about screening procedures. Their ratings are used by Medicare and Medicaid to decide what's covered and for which patients. In turn, many private insurance companies cover the same procedures. The USPSTF explicitly doesn't consider cost in their recommendations. Most often, they look at whether a screening reduces mortality rates.

1: https://www.uspreventiveservicestaskforce.org/

But you're not having cancer checked out, you're having a "spot" or a "nodule" or something checked out.

And the person that's making the recommendation on whether or not to check it out may get sued for $10M if they tell you it's probably nothing and they're wrong, but have no harm come to them if they tell you it's worth having some other doctor do a biopsy.

And they might make an extra couple hundred bucks every time you have to come back and see them to follow up on this spot.

And the radiologist interpreting the MRIs... the same perverse incentives regarding how they interpret a "spot."

> some of which can be harmful,

Like what?

I've seen instances where this is used as an excuse for what is, ostensibly, a trick to dimiss people using something that sounds vaguely professional. Like when doctors say they don't want to do additional x-rays because of the risk of radiation exposure, nvm that if it comes out slightly blurry they'll ask to redo it, or if you're cautious about it initially they'll tell you how it's not big deal and there's more radiation in a cluster of bananas.

Biopsies, radiation from additional scans, surgery, treatment for cancers that would not have developed further.

A lot of potential harms are at a societal level as well—from a public health perspective, if everyone starts having regular MRIs that produce incidental findings which require followup, you’re suddenly tying up lots of resources that would otherwise be available to actually-sick people. A person with symptomatic problems whose treatment is delayed because they can’t get an appointment because the specialists are booked full with incidental findings, that person is indirectly harmed by this.

The radiation from CT scans is not especially concerning at an individual level when there is a compelling reason for it, but, if we’re suddenly doing tons more to investigate incidental MRI findings, there may well be a point where those scans are causing a significant amount of cancer overall—a recent study suggested, I believe, that CT scans may be responsible for 5% of cancers already.

> Like what?

Well, from the original article: "And if someone is over-diagnosed, gets a biopsy and develops an infection, that's a direct harm."

Like when someone has an anaphylactic reaction to the contrast dye for their CT?

Contrast-induced nephropathy?

Gadolinium accumulation in the brain doesn't sound good for you...

Although I think this argument is usually talking about the risks of the resulting procedures (eg an injury or complication related to a biopsy done for a finding on imaging).

I worry particularly about damage to the kidneys, something that you mention in your comments.

It takes a lot to clear the dye.

A colonoscopy has the risk of going very wrong due to a puncture.

But medicine isn't quite the same feedback loop as downtime monitoringg

> ... if regular scans were widespread, it's likely this result in innovations that would drive down costs, improve accuracy, as well as producing a much larger corpus of data with which to guide diagnosis and reduce false positives.

And if there's one thing where AI models really do already excel at it's classifying and noticing patterns.

Many dermatologists (not all of them yet, at least not in the EU), for example already have software classifiers using pictures of one's skin and helping guide diagnosis. I've lots of moles/nevi and freckles on my skin: I'm one of those Gen X kids raised by parents that had no idea that sun exposure and sunburns was a bad thing so I regularly get warning shots and my body, especially my back, if full of scars for for my entire life dermatologists have regularly removed concerning little buggers and sent them to the lab for further analysis.

Nowadays my dermatologist is helped in her classification by hardware/software.

I don't see why that wouldn't be the way forward for full scan MRI: they'll begin to be more and more hooked up to AI classifiers.

It always takes time: it's not as if the tech comes out and in 48 hours every hospital/physician is equipped with it.

It's literally the future is here (classifiers helping dermatologists find concerning nevi), just not evenly distributed (many dermatologists still don't have access to these latest machines).

I can't imagine this taking strong hold in the US unless it shields physicians from legal consequences of false negatives or produces enough false positives to ensure revenue doesn't fall.

I don't see any way that the hospital systems running healthcare in the US would embrace a technology that reduces false positives (income) without decreasing false negatives (risk and lost income) at least as much.

This isn’t just about false positives.

Most people do have things “wrong” with their body, but they are asymptomatic. The human body can and does cope with a certain amount of failure and/or anomaly as a part of normal operation that we otherwise consider healthy.

The problem is that this information is often not actionable. An MRI is great at identifying which ways your body doesn’t look like a textbook reference body, but it doesn’t necessarily tell you what those things are or whether they will ever cause you problems. The way the body naturally responds to problems doesn’t always look perfect on a scan but if it results in no symptoms it is the best result. And for most asymptomatic findings, taking an invasive next step has more risk than the finding itself. And these findings will always be in the back of the patients head, whether relevant or not, and might complicate how they seek care for other real issues later on.

> An MRI is great at identifying which ways your body doesn’t look like a textbook reference body, but it doesn’t necessarily tell you what those things are or whether they will ever cause you problems.

Its the doctors doing this, not the MRI.

There's this weird definition switch that always happens with the "overdiagnosis" defense where the information gets blamed for the overdiagnosis. An MRI doesn't provide any diagnosis in any sense of the definition. A doctor does.

Claiming an overdiagnosis defense is essentually implying the medical industry is worse for most than doing nothing.

Yeah, but in the scenarios this article is talking about (Prenuvo, et al), these aren’t scans ordered by a doctor.

The scans are unprompted and ordered by the patient themselves. And they cost a lot of money so the patient is going to be interested in whatever the findings are. Now, of course a doctor understands that most of these findings are probably not clinically significant. But, for a patient it will prompt a lot of questions, because a layperson by definition doesn’t know whether the findings are significant or not.

MRIs are very useful as a diagnosis tool, but they’re not a diagnosis machine. That’s why they are the standard of care for accurately diagnosing conditions for which there are some other indications of. But the signal to noise ratio of MRI findings alone, aren’t good enough to be the standard of care for diagnosing anything.

The major reason not to do this is that you often get worse outcomes for patients. Oncology provides a lot of examples where “more knowledge” does not lead to better outcomes. Routine ovarian cancer screening, prostate screens, childhood neuronlastoma screens, and breast cancer screens all have shown that overuse will identify more cancer, but do not lead to better outcomes like reduced mortality.

The reasons are complex, but the short answer is that cancer treatment is extremely hard on your body, and even if you don’t treat, stress can literally make you sick. I recommend reading The Emperor of all Maladies if you want to really get a sense for how delicate the problem of early screening is.

I appreciate that, but do wonder, if this is an issue with too much data or how we act on that data. In other words, could there be a future where we do have tons more data, but also use the data in such a way to achieve an overall better outcome for patients?

Ultimately its a balancing act between what we can know and what we can do about it. If you can’t treat a cancer (or your treatment is not effective, cf the radical mastectomy) then knowing who has it doesn’t help. As technology progresses and more cancers become readily treatable, it will make more sense to do early screening, and potentially full body MRIs.

But right now it is likely to cause a huge waste of time, resources, and yes, human lives to know about every little lump in your body.

Knowing every lump won’t save lives if you don’t act effectively on it. Focus on building tools and treatments that actually move the needle, not busywork that wastes time and resources.

Maybe, but there's a human element that can make things worse too. Take prostate cancer as an example. Most men die with prostate cancer. Most men don't die _from_ prostate cancer. It isn't usually aggressive enough to matter. Most people aren't zen enough to accept that though, so just knowing that you have cancer can add stress to your life with measurably bad health impacts from the resulting hormonal changes (reduced immune function, impaired sleep, increased clotting tendency, slower wound healing, etc).

You’re assuming a future with highly competent specialists who also don’t make medical decisions based on insurance requirements.

Unfortunately too many radiologists and specialists are more focused on upping cash flow than medical care.

I’m married to someone running various prostate cancer studies in the UK. I hear the arguments against screening a lot and the issue really blew up recently in the news here.

The thing is, when researchers talk about “worse outcomes” they’re often comparing survival (or rather lack of) against terrible side-effects.

What this fails entirely to capture is that doing something to increase your odds of survival, damn the consequences, is an individual choice. It shouldn’t be up to a health economist to make that judgement.

> What this fails entirely to capture is that doing something to increase your odds of survival, damn the consequences, is an individual choice.

What you're failing to capture is that this is a hard problem because it's both an individual choice and a collective one as well. Those "terrible side effects" might actually end up killing someone. You're choosing between a high-chance lottery on a small population or a low chance lottery on a far larger one. It's not that simple.

Every choice carries a ripple. Survival isn’t just personal, it’s strategic. Operatives plan knowing their move can save or cost more than just themselves.

Population level policy optimizes averages. Individuals optimize their own risk tolerance.

The real lesson is this. Never outsource life changing decisions to systems built for efficiency. Think independently. Act intentionally. Build that mindset at foundniche.

Bot?

But who will pay for the hundreds of thousands of screening MRIs, along with the large number of incidental results that will require some sort of follow-up? Many patients will seek second opinions if not recommended to "cut it out", with additional costs also for the complications resulting from unnecessary biopsies. US medical care is already tremendously expensive; adding all of these costs will break the bank and for no real benefit.

More data is not always better decisions. Precision beats paranoia.

Same rule in business. Chasing every signal kills performance. Focus on what actually improves outcomes. Build smarter with foundniche.

Then they should decide not to treat certain things and have better criteria around that than choosing to bury their heads in the sand and letting people die out of their ignorance.

Same logic as not treating the cancer of an 80 year-old as a heuristic rather than looking at the health, genetics, and epigenetics of the individual beyond the chronological age. Whatever happened to "personalized medicine?"

This article sees methodological failure, I see training data. Training data that could ultimately be used to refine low resolution scans into targeted high resolution scans as needed driving down costs and driving up accuracy. We've already demonstrated AI upscaling, what's the blocker to doing the same for MRI?

And finally, who are any of us to tell people what they can do with their money? China has these things down to $70. And they're leaning in hard on improving them. Cue obligatory China cuts corners blah blah blah. Sigh.

It’s expensive because superconducting magnets cryogenically cooled in liquid helium are expensive to build and run. No amount of demand signal will overcome that reality.

> if your downtime detection system kept producing false negatives, would your solution to be just turn it off

It's more like if your CI build fails in 95% of runs, and in only 2% of runs it indicates a real bug, do you have a bad CI which is next to useless for detecting bugs? I'd say yes, that's exactly what you have. No developer is going to pay any attention to this CI, and if you tell the developers they must ensure tests pass on such CI, they would rebel.