Artificial intelligence in digital pathology

[@ShivaniVijayapandian]

its really helpful, thank you so much

[@aleksandrazurawpathology]

You're welcome! Thank you so much for watching!

[@jackt9535]

The real utility of AI is reliably filtering out negative /uninteresting slides and let pathologsts concentrate on useful slides and gain productivity

[@aleksandrazurawpathology]

That would require an "abnormality determination " approach, which so far did not demonstrate enough success in pathology, so currently, it's one entity or several entities at a time. I am waiting for a working abnormality determination solution for pathology. That will really increase efficiency.

[@jackt9535]

@@aleksandrazurawpathology of course! But if you choose the right body part with the right model, it would work

[@jackt9535]

Predicting molecular statute is very random through histology alone. We must check papers raw data rather than statistics.

[@aleksandrazurawpathology]

Or validate the results with a standard of care test, like in the cited paper. I agree, I would blindly trust this kind of predictions.

[@jackt9535]

@@aleksandrazurawpathology thus I doubt the real role of Pathologist in routine workflow...

Less work is good.

[@aleksandrazurawpathology]

Haha, I think so too :)