Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Colonoscopies come with potential for complications.

There are actually a number of companies which have either released or plan to release non-invasive colorectal cancer screening based on blood dna sequencing.

GRAIL, Freenome, and Guardant all have tests out or coming down the pipeline.



I got a letter from the government last year saying they wanted some of my poo, so I sent them some. The whole thing was weird, but much better than visiting a doctor.

I'm in Australia.


They have your DNA now.


Colonoscopies also have a big advantage in that while the doctors are in there, if they see anything forming that is pre-cancerous they can take it out before it turns into a problem that would show up on a blood test.

* props to @nwellinghof whose response I stole and adapted


Which is why after another screening method sees something you are offered a colonoscopy to look further and if early enough do a simple removal which is much less of a risk and has less side effects than a full operation to remove parts of the bowel.


Sure but procedural complications are actually quite rare, as alluded to in this article.

Misses from operator error, suboptimal bowel prep, or inability to complete the examination are more common and where we have the most potential for benefit from novel screening tests.


> Sure but procedural complications are actually quite rare, as alluded to in this article.

article actually says that perforation (requires emergency surgery as per article) is between 1 to 100 and 1 to 20000 base on some studies, which sounds very high if it is not mistake.


The high end of that range is much higher than the baseline rates of colon cancer (also mentioned in the article), which are 90/100,000 in the USA.

That's one thing that I wish the author had spent more time on.


90/100k cancer rate is likely annual rate, while complication rate is likely per procedure, so you would need to multiply cancer rate by 3 (avg years between procedures) to have reasonable comparison.


You can't just multiply by 3 (it's smaller than that), but yes, you need to adjust both.

https://planetcalc.com/7044/

(Just to be clear: this is the wrong model. An annual rate of cancer (X%) at a population aggregate does not mean that you have X% independent probability of getting cancer per year. But even assuming that was true, you couldn't just multiply by 3.)


1 in 100 is old data, 1 to 20000 is not high. The study they cited had 0 in 12000.

The real number is closer to 20000, both from more recent literature and personal experience.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: