I run a company in the medical imaging space. It's not as effective, yet. The article is wrong.

That's fine, and that happens today - in fact second opinions are a really important part of the process (both for you and the radiologists), but the AI is going to be of low value to you there, since it's accuracy is no better.

There is some value in having the AI sanity check the radiologist, but if they differ you are going to need two radiologists anyway.

I'm sorry to hear about your diagnosis, however, and I wish you a healthy future.

This is purely anecdotal on the doctors behalf. There are companies like Annalise.ai which are using hundreds of radiologists to train models for just one type of cancer, and the best they can do is sanity check, so far. This is likely one of those cases. There is a role for AI in this process, but it still requires a radiologist to interpret.

The bigger issue than finding cancers you missed is actually false positives, finding cancers that aren't really there. Thats a much harder problem. Imagine having to get a breast biopsy for a cancer that doesn't exist. That's the state of the art today.

Most of this tech is currently not as good as a real radiologist. Detecting cancers is a lot harder than it looks, and is often not just based on the imaging, but a raft of other factors.

Spoiler, it's not any more accurate than a rookie radiologist.