Blood test that finds 50 types of cancer is accurate enough to be rolled out • Hacker News Board

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#381

Post by vidarh » Sat, Jun 26, 2021, 8:11 AM UTC

radu_floricica wrote: post #379

vidarh wrote: post #326

Earlier quoted context omitted.

> Testing doesn't kills by itself, not in numbers worth mentioning. Mammography involves radiation and pressure. The radiation alone is a significant risk [1], and kills by causing cancer . It can also cause tumors to rupture and spread malignant cells. It's significant enough to significantly increase the hurdle where mammography is justified. It does not mean it never is - absolutely not. But it means screening pro…

The study reads like a success story to me. They started from a hypotheses of testing reducing deaths overall, which was confirmed. Once they had more data they were able to find ways to optimize the process, like delaying testing for large breasted women who have a worse risk/benefit profile. What I'm advocating for btw, just to make clear, is regularly showing up to a doctor who will recommend all the cost effectiv…

The way to optimise the process was to reduce the amount of screening.

It points out exactly why broad, indiscriminate screening can not be assumed to be safe, and that "optimizing the process" means moving away from thinking that more screening is automatically better, and towards identifying which groups where the benefits outweigh the dangers.

But the main reason for posting that link was to point out that the idea that testing does not kill is false for some types of testing, and as such you need to understand the risk vs. reward, or you may end up doing harm.

> What I'm advocating for btw, just to make clear, is regularly showing up to a doctor who will recommend all the cost effective non-intrusive tests (like blood work) plus the intrusive testing depending in your particular risk profile. This doesn't seem like a controversial opinion to me - more like a common sense default.

I haven't seen anyone here arguing against that. That's not what's been discussed by those of us here pointing out the dangers of mass-screening.

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#382

Post by jacquesm » Sat, Jun 26, 2021, 11:59 AM UTC

msrenee wrote: post #380

jacquesm wrote: post #270

Earlier quoted context omitted.

We do all of those, but they're all voluntary (as they should be), there are regular drives for all of these.

Did I reply to the wrong comment? I thought I was replying to one about how the Netherlands doesn't cover cancer screening under its national healthcare.

Which it does. That comment you replied to is complete bullshit.

For instance:

Breast cancer:

https://www.bevolkingsonderzoeknederland.nl/borstkanker/

Cervical cancer:

https://www.bevolkingsonderzoeknederland.nl/baarmoederhalska...

Colon cancer:

https://www.bevolkingsonderzoeknederland.nl/darmkanker/

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#383

Post by dekhn » Sat, Jun 26, 2021, 3:14 PM UTC

epicureanideal wrote: post #308

rackjack wrote: post #190

Earlier quoted context omitted.

Same vibe as "I would just simply..." What do you think the entire medical, pharmaceutical, bioinformatic, etc. industries have been doing for decades? Are you really so arrogant to think that other people are too lazy and stupid to think of your brilliant idea and that you are simply more intelligent than everybody else? Do you really think people haven't been trying?

> What do you think the entire medical, pharmaceutical, bioinformatic, etc. industries have been doing for decades? Well, let's remember that the AlphaFold team at Google solved the protein folding problem with a relatively small team in a relatively small number of years after extremely large, well funded companies whose primary business was drug development failed to do so for decades. So yeah, it's been demonstrat…

alpha fold was developed at deepmind, not google, they didn't solve the protein folding problem, and it's not surprising that drug discovery companies didn't reach a similar level of accuracy first.

Note that DeepMind drafted off the work of a protein folding guy from academia who had been doing this for decades.

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#384

Post by bick_nyers » Sat, Jun 26, 2021, 7:47 PM UTC

TaupeRanger wrote: post #133

throwawaynotmd wrote: post #109

Earlier quoted context omitted.

The argument you’re making has always bothered me because it’s hiding the ball. If finding something that otherwise carries no symptoms is best left untreated, then the fact that you found it should make no difference to the decision. The doctor should say that the best course of action is to do nothing. More information can never be harmful. If you know information is best not acted on in the abstract, then you also…

Lots of problematic assumptions in your reply. >If finding something that otherwise carries no symptoms is best left untreated, then the fact that you found it should make no difference to the decision. The doctor should say that the best course of action is to do nothing. More information can never be harmful. Absolutely and demonstratively false. There is an entire field of health policy that destroys this harmful…

I wonder if the psychological stress you describe is a result of societal expectations, which itself is a result of what has been done historically. I would think that to most people you are either healthy, or you have a problem, because that is often how it is viewed when it comes to healthcare. In the same way your doctor could say "you are a little overweight, and let's keep an eye on it to make sure it doesn't cause problems down the road", a similar perspective could be had for what people commonly associate with more serious issues. You can look at what the correlations say for health policy, or you can consider the actual reasoning as to why people view healthcare as such in the first place. It's normalized for people to not go to a doctor unless they have a problem, it's normalized to not take regular preventative measures and screenings. In addition, there are financial barriers. While insurance companies may benefit in the long run from more regular, personalized, preventative care, they are likely very resistant to paying for it. With the costs of healthcare, only well off people would consider this approach, as most of their costs may be out of pocket because they don't have a problem to point to for these kinds of screenings. Like getting blood work done maybe every 6 months just to make sure everything is doing ok, hormones, etc. If it was normalized, and people less afraid of the doctor, needles, and of the costs attached to being unhealthy, their sources of psychological stress would likely be non-existent.

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#385

Post by bick_nyers » Sat, Jun 26, 2021, 8:06 PM UTC

throwawaytcp wrote: post #348

throwawaynotmd wrote: post #109

Earlier quoted context omitted.

The argument you’re making has always bothered me because it’s hiding the ball. If finding something that otherwise carries no symptoms is best left untreated, then the fact that you found it should make no difference to the decision. The doctor should say that the best course of action is to do nothing. More information can never be harmful. If you know information is best not acted on in the abstract, then you also…

If I can step in with a personal anecdote: I was diagnosed in my late 20s with papillary thyroid cancer. PTC is a very survivable cancer, with a near-100% survival rate (death usually only occurs in rare cases where the disease is diagnosed very late, the progession is atypical, or there are comorbidities at play). It is very easy to screen for and diagnose: a neck ultrasound identifies thyroid nodules, and if the no…

My thinking is that the tool was not the problem, it was the decision to do something. With what you say about how common and insignificant is often is, then the right choice in a lot of cases is to probably monitor, seeing how much it grows over time. This isn't a cost effective solution if the ultrasounds are expensive or not high resolution enough, or if insurance won't allow you to take this strategy. If there's no benchmarks or thresholds of size or growth of the tumor before cutting it out should be considered, then you are right, it would likely overwhelm the system with uneccessary treatments. The monitoring doesn't feel like the problem to me, it's the decisions/recommendations of the doctor, often being made as a result of professional wisdom, what research was financed, and the financial system that encompasses healthcare.

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#386

Post by bick_nyers » Sat, Jun 26, 2021, 8:07 PM UTC

nevinera wrote: post #159

0.5% false positive rate is actually quite high. I don't have the numbers, but I'd be surprised if more than 1% of the people that would receive such a screening actually already have cancer , which would mean that it produces about as many false positives as false negatives. Picture a population of a million people, all receiving the test. 1% of them have cancer (unknown to them) - half of those people get a negativ…

If it's cheap enough then get it performed 3 times.

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#387

Post by bick_nyers » Sat, Jun 26, 2021, 8:12 PM UTC

dmitryminkovsky wrote: post #65

One of my great fears in life is that employers/insurers will impose this en masse on people, causing immeasurable harm to the mental health of people who can’t handle the anxiety that this type of approach to medicine would cause. I’m definitely a big proponent of evidence-based medicine and using it to help living quality life as long as possible, but this sort of thing would just wreck my day to day existence and…

Maybe do an anxiety screening first

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#388

Post by inadequatespace » Mon, Jun 28, 2021, 12:50 AM UTC

Took me quite some time to find the paper that seemed to describe the ML algo[1]. Pretty disappointingly domain-specific and barebones ("logistic regression").

And indeed, model form was so unimportant that it was relegated to the supplement. [2]

Yet again evidence that access to data and domain knowledge trump fancy ML algos 10/10 times.

[1] https://www.sciencedirect.com/science/article/pii/S092375342...

[2] > Custom software was built to classify samples using source models that recognized methylation patterns per region as similar to those derived from a particular cancer type, fol- lowed by a pair of ensemble logistic regressions: one to determine cancer/non-cancer status and the other to resolve the TOO to one of the listed sites (supplementary information)

Re: Blood test that finds 50 types of cancer is accurate enough to be rolled out

#389

Post by nybble41 » Mon, Jun 28, 2021, 2:31 PM UTC

jacquesm wrote: post #268

nybble41 wrote: post #262

Earlier quoted context omitted.

> absent symptoms A positive blood test is a symptom. One which you'll miss if you don't do the blood test regularly. And looking for symptoms (other than a blood test) is a form of testing... one with much worse accuracy (both positive and negative) than the blood test, especially in the early stages.

Just so we are using the same definitions for the words we write, here is the definition of 'symptom' in a medical context: https://www.cancer.gov/publications/dictionaries/cancer-term... "A physical or mental problem that a person experiences that may indicate a disease or condition. Symptoms cannot be seen and do not show up on medical tests. Some examples of symptoms are headache, fatigue, nausea, and pain."

That still sounds like a subset of medical testing to me, where you're using the patient as your test equipment and ignoring anything the patient cannot sense directly (or fails to report). Perhaps you don't call the results of medical tests "symptoms", but to me this is an insignificant distinction. Either way we're talking about an observable result of some underlying condition; only the means of observation varies. Jargon aside, this approach still means waiting until after the condition has already affected the patient's quality of life, and perhaps beyond the point where it can be treated effectively, when it could have been treated and perhaps even cured before any symptoms manifested.