Actually not. Flexion at the hips makes it much easier. Just fyi! 
2 Likes
But what doctor would want to put their mouth down th…? Oh. Nevermind.
2 Likes
Do we really need “AI” to analyze four numbers?
Seems superfluous which makes me worried this is publication bait
You hit on a really important point though: how do we tell which cancerous lesions are life-threatening and which are not a threat? Some cancers will grow so slowly or stay so localized that they won’t be a threat in the patient’s predicted lifespan while others are such aggressive bastards that they have to be dealt with immediately. Then you get into the problem of how to deal with the aggressive ones: if you blast them with huge doses of chemotherapy you can get rid of all the treatment-sensitive cells but that can leave behind a few resistant stragglers who then blow up and form a bigger, badder, harder-to-treat tumor.
For the issue of detection of non-threatening cancers, the story that gets told in radiation oncology classes is about how thyroid cancer rates spiked 200% in South Korea in the 90s and set off a bunch of alarm bells about environmental factors or undetected genetic issues driving that spike. A ton of Koreans were screened, diagnosed, and then treated aggressively. When they finally did retrospective analysis of the program, they found that they were grossly over-treating thyroid cancers: most 85yo+ patients had some dysplasia in their thyroids anyway but the vast majority of those dysplastic cells were never going to go on to become truly cancerous. It’s similar for prostate cancer: 90yo men without hyperplastic prostates are the exception, not the rule. This problem will get trickier over the next few years now that we have a bunch more biomarker tests coming online (e.g., Cologuard) for early cancer detection. Don’t get me wrong, better early detection is key to continuing to improve cancer survival rates, but just getting better early detection doesn’t actually solve the problem of knowing which cancers to treat and how aggressively to treat them.
If you have any other questions in this area, feel free to ask. This kind of stuff makes up the bulk of my research work.
6 Likes
Notably: their approach had a 100% detection rate on the (blinded) training set. From skimming the press release, it doesn’t appear to have been cross-validated in a different set of samples.
2 Likes
One question my doctors have never been able to answer is whether the amount of mutation in the bad prostate cells changes over time, that is, if you start with a Gleason score of 3+3 or 3+4, will they further mutate to higher values later? PSA can’t really tell you anything about the aggressiveness of the cancer, right now you only get that from biopsies. Not clear that post-surgical pathology reports even check for the presence of more advanced cancer cells.
If there are markers readable in this urine test that can characterize a reoccurrence as likely being less aggressive cancer than the original prostate tumor, that would dramatically change post-surgical treatments.
1 Like
Have not yet had one of those but my diverticulitis gave me the joy of 5 colonoscopies before my surgery
It’s a good question, it just comes down to different levels of biology. A Gleason score is based on histopathology: what do your cells look like under a microscope? There are some mutations that will change those cells’ appearance, but there are also a lot of mutations that won’t show up in appearance. And then, OK, cancer cells have mutations, yeah, but those mutations are alllllllllll over the place: some are pro-cancer (driver mutations), some are anti-cancer (tumor suppressor genes), and some are just neutral (passenger mutations).
I’m more optimistic about blood tests than I am about urinalysis, for prostate and many other types of cancers. We’re learning that there are a ton of different things that cancer cells plop out into the bloodstream and are beginning to find the signals within those things that give us a more detailed view of what’s going on in a cancer. There’s cell-free DNA, exosomes (little goo packets), and circulating tumor cells (which can be but aren’t necessarily pre-metastases). Another point of optimism is the arrival of multi-plexed biopsy imaging; traditional histopathology relies on chemical stains that don’t label specific proteins or genes that can be involved in cancer progression or the switch from benign to malignant. With multi-plexed biopsy imaging, we’re starting to get wayyyyyyy more info from each biopsy and I’m pretty excited to see where it goes.
3 Likes
Did I miss something? Where’s the link to the boingboing store?
2 Likes
Find it? Doc, I didn’t even realise it was lost!
This topic was automatically closed after 5 days. New replies are no longer allowed.
