Random stuff worth reading on the internet thread

[Northwestcoug]

There is a lady in Scotland that can smell Parkinsons before it presents medically:

https://www.npr.org/sections/health-...to-diagnose-di

Crazy.

[BigFatMeanie]

There is a lady in Scotland that can smell Parkinsons before it presents medically:

https://www.npr.org/sections/health-...to-diagnose-di

[Northwestcoug]

Listen to this on your way home.

https://www.pushkin.fm/podcasts/revi...ll-see-you-now

I haven't listened to Gladwell for awhile; I'd forgotten how enjoyable one of his good episodes can be.

Fittingly, I listened to this while walking my dogs. One of them would stop dead in her tracks to sniff my other dog's favorite pee rock, even a couple of weeks after he died. So you don't have to convince me of their olfactory excellence. And I really appreciated Gladwell's easy to understand discussion about sensitivity/specificity and negative/positive predictive values, without using those terms and getting bogged down in definitions.

I skimmed over the two most interesting articles he brought up, about prostate and colon cancers. Up front, he did oversell just how good dog sniffing was for these. For prostate cancer, the study used urine from men who had very high-grade cancer. It is safe to assume that these cancers are more extensive and of a higher stage than the more common prostate cancers, which would make sense if the tumor is developed enough to cause a detectable scent. But even for these cancers, the dog sniffing results were "71% sensitive and between 70–76% specific at detecting Gleason 9 prostate cancer".

The colon cancer paper was better designed. It included a bunch of non-cancer patients with colonic disease other than cancer, which conceivably could cause a detectable odor. And based on the paper, dogs were good at detecting the cancer patients via exhaled breath and poop smelling. The results were around 91-97% sensitivity and 99% specificity for both breath and stool samples, when compared to colonoscopy. That is pretty interesting, though I couldn't find the characteristics such as cancer stage of the cancer patients, which really does raise eyebrows. But still, it proves the point that cancer causes a detectable odor.

Having said this, colonoscopy serves a purpose other than just diagnosing cancer. It documents and removes pre-cancers, which according to that paper dogs couldn't detect. Also, the paper is from 2011. In the 15 years since fecal occult blood testing has progressed to DNA testing which has similar sensitivity and specificity to colonoscopy in many patients. It would be interesting to compare DNA-based tests with canine olfactory detection.

[Bo Diddley]

You know what else works really well at detecting cancer? Training dogs to smell it in humans. Research shows that they are remarkably accurate in determining if someone has cancer simply by giving them a sniff. Problem is, people don't think it is high tech or something so it isn't widely used.

I read about a while back. I thought there was also a push to objectively measure with technology as well.

[Jeff Lebowski]

While I'm a big proponent of the health benefits associated with dogs, I'm assuming that the point when dogs can detect cancer in humans is past the timeframe where other modalities catch it earlier. If you know of studies stating otherwise please let me know.

Listen to this on your way home.

https://www.pushkin.fm/podcasts/revi...ll-see-you-now

[Northwestcoug]

You know what else works really well at detecting cancer? Training dogs to smell it in humans. Research shows that they are remarkably accurate in determining if someone has cancer simply by giving them a sniff. Problem is, people don't think it is high tech or something so it isn't widely used.

While I'm a big proponent of the health benefits associated with dogs, I'm assuming that the point when dogs can detect cancer in humans is past the timeframe where other modalities catch it earlier. If you know of studies stating otherwise please let me know.

[Jeff Lebowski]

You know what else works really well at detecting cancer? Training dogs to smell it in humans. Research shows that they are remarkably accurate in determining if someone has cancer simply by giving them a sniff. Problem is, people don't think it is high tech or something so it isn't widely used.

[Bo Diddley]

Thats really cool. I love to hear about technological upgrades

[Northwestcoug]

This is really cool, and will likely be a defining paper for future AI in medicine.

https://www.thelancet.com/journals/l...267-X/fulltext

swedish women were randomized to AI-assisted versus standard radiologist interpretation of mammograms. AI-assisted treatment interpretations caught a significantly higher percentage of both pre-invasive and invasive breast cancers, and false positives were not significantly increased. This was a blinded prospective study, which is close to gold in medical research.

you should always pay attention to medical research in single payer nations. In general their level of standardization is difficult to achieve when studies deal with multiple independent centers in the US.

[Bo Diddley]

More political parties is a good thing. Would proportional representation work in the US? Some food for thought by this NYT article. No matter what you think about the merits of the article, the infographics are good. It starts of with this truism:



https://www.nytimes.com/interactive/...sentation.html

Agreed. This is perhaps the only thing I envy from the European governments.

[Northwestcoug]

More political parties is a good thing. Would proportional representation work in the US? Some food for thought by this NYT article. No matter what you think about the merits of the article, the infographics are good. It starts of with this truism:

In early 2020, Representative Alexandria Ocasio-Cortez, the progressive Democrat from New York, was asked to speculate about her role under a Joe Biden presidency. She groaned. “In any other country, Joe Biden and I would not be in the same party,” she said, “but in America, we are.”

https://www.nytimes.com/interactive/...sentation.html

[falafel]

The comments on this post are hilarious and spot on given the dorkiness of the product, essentially a onesie for men.

"I'm gay and I think this is too gay."

"Imagine going to the urinal as a man and seeing the guy next to you unbutton himself."

[falafel]

Communicable diseases are weird.

not sure this was worth reading on the internet

[PaloAltoCougar]

The comments on this post are hilarious and spot on given the dorkiness of the product, essentially a onesie for men.

[Pelado]

Communicable diseases are weird.