this post was submitted on 23 Mar 2025
32 points (86.4% liked)

Good News Everyone

1473 readers
230 users here now

A place to post good news and prevent doom scrolling!

Rules for now:

  1. posts must link from a reliable news source
  2. no reposts
  3. paywalled articles must be made available
  4. avoid politics

founded 2 years ago
MODERATORS
can
 

In what's expected to soon be commonplace, artificial intelligence is being harnessed to pick up signs of cancer more accurately than the trained human eye. This latest AI model has a near 100% success rate and serves as a clear sign of things to come.

Study - https://www.sciencedirect.com/science/article/pii/S2666990025000059

you are viewing a single comment's thread
view the rest of the comments
[โ€“] [email protected] 3 points 4 days ago (1 children)

The article doesn't spell out the broader context for laymen like me. Can anyone clarify some points. Are these images taken from biopsies of tissues that are already suspected of being cancerous? Is this work translatable to preventative screening in a way that I'm unfamiliar with, or is it limited to processing biopsies?

For automated tools like this, what sort of protocols are established to prevent doctors from being biased by the tool output? Does the person running the test provide their findings before they see the output of software detection?

[โ€“] Jiggle_Physics 3 points 4 days ago* (last edited 4 days ago)

Seems for the tests they mixed a number of known cancerous images, from biopsied tissue, and mixed a bunch of imaged tissues, known to be cancer free. They then randomized the images and processed them with this AI, and it was able to find cancerous vs non-cancerous tissue, at a rate significantly higher than doctors.

They are also seeing a lot of success in using similarly purposed AI for finding pathology presented in a wide variety of images, including MRI, CT, X-ray, etc.