Radiology + Pathology = Future ? (JAMA editorial comment)

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
What a worthless freaking article. You think every radiologist wouldn't kill for an automated lung nodule detector? I've been hearing this crap for 10 years and the best the world of AI has come up with is a breast cancer cad that misses every cancer but flags every skin calcification.

If only people spent as much time actually working on radiology AI rather than predicting it's inevitable takeover, the technology may actually get somewhere.
 
The only thing that I'm wondering is, did breast cancer CAD also have the support of, what looks like Silicon Valley? It seems that historically, Silicon Valley has not really been involved with the healthcare industry because frankly, it's just not as sexy as starting the next Snapchat or Facebook. My only concern is if Silicon Valley people are really behind this AI bit in terms of how it'll affect radiology, it could end up resulting in a more significant impact than what has been seen with breast cancer CAD in the past.
 
Watson's the only real player because they are the only ones who can overcome hipaa to access medical images given their Partnership with merge.

That said, I saw their exibit at Rsna last year.

I wouldn't lose any sleep over this. We are years away from even implenting a simple, simple automated task like nodule hunting--which should theoretically be very easy--find the white spot surrounded by black.

When it eventually does succeed at automating tasks, then our lives will get a whole, whole lot easier.
 
Last edited by a moderator:
When it eventually does succeed at automating tasks, then our lives will get a whole, whole lot easier.


And then we will be gradually phased out by midlevels. -_-
 
Top