The intent of non-invasive tests like this are that you give them to everyone, at scale, frequently.
And specifically tune them to minimize false negatives.
Worst case with a false positive? You're causing someone anxiety and giving them an extra scan.
Best case? You just saved someone's life by detecting an aggressive cancer early enough to do something about it.
At the end of the day aggressive, metastatic cancer is a time game. If the cancer is given time, it becomes progressively harder and harder to treat, and effective treatment options become tougher on the rest of the body, until finally there's nothing to be done.
> Worst case with a false positive? You're causing someone anxiety and giving them an extra scan.
It is arguable that this as minor an issue as you make it out to be. There has been work to try to assess this (google "cancer overdiagnosis").
The counterargument to what you state is that a false positive can not only lead to stress & unnecessary/more invasive screening, but a treatment plan that's a net negative. For instance: if a cancer were detected and it'd wind up being a cancer that someone dies with instead of from, and the treatment causes worse outcome than the cancer itself, that's not good. The hard issue here is that these things need to be determined at a population scale, and one can always cherry pick personal anecdotes in either direction to tug at heartstrings.
We seem to have found ourselves at a point where it's clear there's a balance that needs to be struck, but it's unclear what that balance is yet.
The existing downsides of a might-be-cancer hit on a test are real, for now - a statistically significant number of people with a breast cancer diagnosis end up killing themselves out of despair or doing radical surgery "to be sure", which detracts from the social benefit of mammograms. But there's no indication that they would persist if false positives were more common because proactive testing was more common. There would be a pipeline of followup testing and risk evaluation, which would be normalized by how common it was and how nonlethal diagnostic hits would become.
It would become "Oh, yeah, my cousin got a hit, but followup biopsy said it was a mutation that doesn't metastasize. Guess we'll see next week." rather than "My great-grandmother died of breast cancer, my mother died of breast cancer, my friend's aunt died of breast cancer, this feels like a death sentence", which is the information people who receive hits on their mammogram testing are acculturated with now.
Again, this is a well studied topic. Anyone claiming they've identified the one true answer isn't someone who should be believed. The topic is complex, at best. But these are not hypothetical situations. There are data to support the point to which you're countering.
You are correct that there exists a world where the problem is minimized and at that point obviously the math changes. But we're not there yet.
He did say that you tune them to minimize false negatives.
I do find it interesting that in the early days of HIV testing even most people who got a "positive" result were in fact negative. The tests have since greatly improved and the number of people taking a test due to hysteria rather than likelihood of infection declined.
But I can imagine a world in which we get very many forms of liquid biopsies like this every year, and false positives become a thing we understand and are used to.
Whether one should bias to allow more false positives or false negatives depends on the next steps after a false positive vs the risks after a false negative.
Well, it would seem hand-wavy to whichever scientists are tasked with improving it I'm sure! There's probably a lot of work involved.
But it's pretty standard for new tests. The first ones are never the most reliable. Obviously the rate of false positives and expense of follow-up testing determine whether it's overall a good idea, but it probably will start out with a positive ROI that becomes a highly positive ROI pretty fast.
I mean, yes, some people are going to freak out no matter how many times their doctor tells them about false negatives before they get the test. But overall this will save a lot of lives and as the test improves, a lot of money too.
Cancer testing is not benign. False-positive cancer diagnosis has a >0 mortality rate, because further testing and treatment is potentially fatal. Just a colonoscopy can be fatal, as can a biopsy that requires anesthesia.
The medical world weighs these things against each other and determines when the data shows that the risk of mortality from testing is smaller than the risk of cancer.
The emphasis there is on "further treatment". The test itself is broadly benign (except the general risks associated with phlebotomy and any risk of psychological harm).
The issue is that doctors often over-react to adverse results due to the risk of being sued if it did turn out to be a true positive.
I have two examples of this. One was during a routine blood test I had a liver enzyme flag up, which was then further investigated non-invasively with ultrasound and it was determined that I might either have moderately developed NAFLD (non-alcoholic fatty liver disease, I'm obese) or very early stage NASH (non-alcoholic steatohepatitis) associated fibrosis. The doctor wanted to perform a liver biopsy to confirm which is obviously an invasive procedure with a 1%-ish risk of complications.
My response was to ask how the treatment would differ between diagnoses, and he said in either case the treatment would be the same: lifestyle change. He agreed that from a risk perspective the biopsy was just inviting additional risk for no benefit, but that policy is to recommend the test and if I refuse it I'll need to sign an indemnification document saying I was refusing further diagnostics against medical advice. A few years go by, I've made efforts to improve my lifestyle, lost weight, and now my liver tests are all normal proving it was just NAFLD after all.
In another case, I had a suspicious finding in an eye test which (long story short) led to me getting two head CT scans which showed no problems. In hindsight, I think a double dose of brain radiation over a common minor finding with no other symptoms was a crazy over-reaction and I would have refused if I had all the facts, but it could have been a life threatening situation in some ridiculously tiny percentage of cases so it was all rush rush and I didn't have time to weigh it up.
Often the correct thing to do may be a combination of further non-invasive testing, repeating the test (possibly after a period of time), and "watchful waiting". Doctors often don't feel comfortable with the level of personal risk that could expose them to, and for good reason. That is the issue, not the test.
> The issue is that doctors often over-react to adverse results due to the risk of being sued if it did turn out to be a true positive.
Do you think a patient is going to receive a false positive and accept a response from a doctor of, "Oh, might be a false positive. Let's ignore it and not do any more risky testing."?
That scenario makes no sense from either perspective. If you get a positive, you do more testing (or skip to treatment).
Your examples are well-taken and I understand them, but they don't apply to cancer. When you detect cancer with any test, you immediately do something risky: either a further test that involves risk, or treatment that involves a lot of risk.
I didn’t say we do nothing, it’s just the next step doesn’t have to be invasive testing.
Obviously the particulars are going to come down to the specific test, but you could be looking at a set of MRIs, repeated a few weeks apart to determine, firstly if it is a false positive, and secondly, how aggressive it seems to be.
In other cases it might be just repeating the test periodically and watching out for other symptoms. We do this with cervical screenings for example, since it’s quite a slow moving disease, where a low grade result will just result in you being put on a more frequent testing pathway.
It has been suggested that we should move to a similar protocol with prostate testing too, where a high PSA shouldn’t result in any immediate action unless there are other symptoms, but rather the test should be repeated periodically to
monitor how the numbers are changing.
Now, that’s not to say you should do all tests. The test could still be pointless (i.e. there are no actions one can take in response to the result either way) or ,more commonly, simply not cost effective. However I fundamentally do not agree that more information itself is harmful, it’s just our response to it that’s lacking and that should be improved.
> Do you think a patient is going to receive a false positive and accept a response from a doctor of, "Oh, might be a false positive. Let's ignore it and not do any more risky testing."?
I mean, it depends on the test; a reasonable answer could be this is likely a false positive and confirmation tests have risks, here are some symptoms to look out for, and we'll test on a regular basis and see if anything changes. That's not appropriate for all positive results from screenings, but it is for some.
I feel like you're missing the point. Yes, false positives have risks. But if blood tests (or in the case of bowel cancer, fecal tests like cologuard) are effective, we can replace more invasive screening options with them. People have historically been encouraged to get colonoscopies once they reach a certain age because, for the general population, the risks of cancer are higher than the risks associated with a colonoscopy above the age of 50. Developing less invasive tests lets us lower that age, catching more cancers, while at the same time making screening safer for people already in the recommended screening window.
Also, pancreatic cancer, which is what the original article is about, has no alternate form of screening. Most people only find out they have it once it's already symptomatic, which is usually stage 3 or 4
> But if blood tests (or in the case of bowel cancer, fecal tests like cologuard) are effective, we can replace more invasive screening options with them.
That is not what I'm arguing against. People in this thread are talking about testing earlier and more often because we have these new tests, not "just" replacing existing tests.
The math is very unlikely to work out that we should do that.
Howso? Existing solutions vs more widely deployed and frequent blood+fecal tests followed up by existing solutions in the case of a potential positive?
These things are already taken into account via population scale statistics. In most cases, it's at best debatable whether more or less screening leads to an overall better outcome across the entire population aggregate. The argument against more screening is that it can (and the claim is it does) lead to overall worse outcomes in aggregate. For any individual case however, the story may be totally different.
What we need are better mechanisms to bin positive results to steer people towards a finer grained course of action. That'd change the math to be more of an overall net benefit.
> The medical world weighs these things against each other and determines when the data shows that the risk of mortality from testing is smaller than the risk of cancer.
Let the patient weigh the odds. Especially when they can afford retesting or may be otherwise in good health or whatever. Plus the test’s algorithms can be tuned to provide more or less false positives.
Patients are uninformed and emotional. Part of a doctor's job is to make the best decision for the patient using the information they have.
For example, if you tell people that daily aspirin reduces the risk of heart disease, you might get a 25yo with no heart issues starting to take it.
But if you look at actual data, all-cause mortality increases for people who have no risk of heart disease if they take daily aspirin, because aspirin can cause fatal internal bleeding.
That's the kind of thing doctors know and need to be firm about.
I disagree. Doctors are frequently incompetent, don’t spend enough time with patients to understand them deeply, and are not aware of the latest research or these nuances.
But leaving that aside, this is about patient control. Doctors should not be gatekeepers for diagnostics. I don’t even want them to be a gatekeeper for many relatively safe prescriptions, which is clearly a way to increase medical costs.
> Doctors should not be gatekeepers for diagnostics.
Doctors mostly aren't gatekeepers for risk-free diagnostics, like a blood test for a vitamin deficiency. Insurance companies are.
But for risky tests, doctors have a duty to "do no harm" and can't/shouldn't order something that they know causes an increase in all-cause mortality for their patients.
The concept is the same with antibiotics or anything else a patient might ask for without knowing the risks.
> Doctors mostly aren't gatekeepers for risk-free diagnostics
That's not entirely true. There are more than a few diagnostics in the US that the FDA explicitly discouraged companies from offering without physician referral. E.g. whole genome sequencing several years ago
the obvious answer is additional testing to reduce the likelihood of a false positive -- if additional tests are invasive those tests can be weighed on the balance of risk just the same. i think the real problem is cost, we cant afford/manage to do that correctly currently, which is exactly what improved non invasive tests could enable and disrupt this reactionary approach forever
> if additional tests are invasive those tests can be weighed on the balance of risk just the same
So let's say you have an extremely safe test. Let's use the "mail your poop to a lab" test for colon cancer as an example.
If that test (regardless of its accuracy) comes back with a positive, you're going to do one of two things: A) order a colonoscopy, perhaps with biopsy, to confirm the presence of a malignant tumor; B) start treatment immediately (if you trust the initial test enough).
So that brings you back to square one: you shouldn't do the test, regardless of the safety, if the math works out to make it riskier (due to false positives and unnecessary tests/treatment).
i think i agree about "if the math works out" -- but thats the hard question isnt it? isnt the accuracy of the poop test significant? if 2 or 3 poop tests increase our confidence in the result it changes the risk:benefit calculation for the colonoscopy or treatment. thats an easy win. but even if it did not, is the result from the (preventative) poop test less valid data than a patient complaining of pain? so generally speaking i agree we dont want to enable a path to risky procedures based on dubious evidence, but i think overall we are presently operating with a dearth of information (waiting for symptoms) and the math itself improves by having more proactive testing (information) in the first place
Worst case is the scan “sees something” which then puts them on a diagnostic anxiety roller coaster for the rest of their lives, “just to be safe”. When in the alternate universe they might not have gone another 60 years hardly ever seeing a doctor.
how is information making that worse? you either have a scan that shows a blip and you can use that information to inform your decision making, or you have no scan and go on without any kind of decision making. i know which option id prefer
Because you can wind up in a situation where the "blip" would have otherwise led to nothing problematic, but the followup for the "blip" actually does cause harm. And at an individual level it's impossible to tell which category you're in.
Every medical procedure after a scan has danger. Biopsies kill people. Colonospies kill people. The rate is extremely low, but test enough people who aren't at risk to have the disease and you will actively harm them.
They generally aren't going to give them to everyone. They will give them to everyone within a certain group, such as age 30+ since the under 30 group is very low risk (unless family history, etc). Similar to how they don't test most kids and younger people for cholesterol - it's just not a significant problem for that age group.
Traditionally, tests have been metered that way because of costs (expensive reagents, preparation, processing) or side effects (radiation from scans).
But the actual relevant equation is {cost of testing} vs {cost of delayed treatment}
If the cost of testing, in economic and health senses, decreases while the cost of delayed treatment holds constant, a different mass deployment optimal point is created.
Thankfully broader proactive testing is also in insurance companies' financial interests, given the high costs of late stage cancer treatment.
One way to test lots of patients where a) there's a low probability that an individual patient will have the disease and b) the test is expensive is to first mix some of the blood of each of N patients and do one test on the mix. If the batch tests negative then all patients are negative and you've only paid for one test. If the batch tests positive then you have to repeat the test on the remaining blood of each of the N patients to determine which were positive. Thus, with a high probability you only pay for 1 test, and with a low probability you have to pay for N+1 tests. The value of N is easily computed to minimize the overall cost, given the cost of each test and the percentage of patients that have the disease.
Which is good because "ASCVD is a disease that begins in childhood; hence, primordial prevention is an important target for improving cardiovascular morbidity and mortality later in life."
> Elevated LDL-C and triglyceride levels have a positive correlation with atherosclerotic lesion prevalence that persists from childhood through early adulthood... Follow-up data from the Young Finns cohort after 12 and 27 yr also demonstrated positive correlations between elevated childhood serum cholesterol and triglycerides to elevated levels in adulthood... Children from the i3C cohort with high and borderline-high total cholesterol have 1.5 to 2.13 times the risk of both fatal and nonfatal cardiovascular events in adulthood than children without. In addition, i3C children with high and borderline-high triglycerides have 1.69 to 2.47 times the risk than children with normal triglycerides.
Nearly everyone will reach 30 in their life, so it is safe to say we give them to everyone. It isn't a one and done test, cancer can form at any time in your life. To be useful we need to give this to everyone over 30 (40, 50....?) , on a regular schedule (yearly?). The article doesn't specify those details (or at least not before I hit the sign in wall)
I'm sorry but that is a very naive and honestly wrong take. An "extra scan" is not just giving anxiety. It raises cancer rates. It can discover underlying relatively bening conditions which affect insurance coverage, for example. It can cause anxiety. It takes away resources "just to make sure". At the scale you are proposing, false positives are a massive issue that you simply cannot ignore. It is all but trivial.
This is it, I am not sure how people can be so dismissive about the risks of over diagnosis.
However, usually there are studies done to carefully weigh the risks and benefits of testing likes this. I would expect tests like these to become the norm for screening at risk populations at some point (usually people beyond a certain age or people with family history).
And specifically tune them to minimize false negatives.
Worst case with a false positive? You're causing someone anxiety and giving them an extra scan.
Best case? You just saved someone's life by detecting an aggressive cancer early enough to do something about it.
At the end of the day aggressive, metastatic cancer is a time game. If the cancer is given time, it becomes progressively harder and harder to treat, and effective treatment options become tougher on the rest of the body, until finally there's nothing to be done.
So anything that gains time is critical.