Robot radiologist is coming

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

Dock1234

Full Member
10+ Year Member
Joined
Jun 1, 2011
Messages
52
Reaction score
4
http://www.wired.com/2015/10/robot-radiologists-are-going-to-start-analyzing-x-rays/

Are you worried yet? It seems that Enlitic has made some breakthroughs (at least according to them) with their medical pattern recognition and it's CEO has said in couple of speeches that in many cases their software is better than radiologist or pathologist.

As article states Enlitic isn't going to replace radiologists but make them more efficient. Isn't that alone gonna put many radiologists unemployed if one radiologist can analyze say 10x or maybe even 100x more scans? And not to mention that who knows how advanced this technology will be in let's say 10 years. Maybe they will replace all radiologists.

So what do you think?
 
Ah... Another episode in the quest for the robot radiologist. I'd definitely order one... After I buy my flying car.
 
Seems like computer engineers creating a product in a vacuum without input from people who actually do the job.

Half of the "advantages" already exist or are simply not how medicine works.

Routing the study to the correct sub specialty radiologist? Great it's called a PACS work list or better yet, it's all going to a general radiologist anyway.

Adjust the priority of the study? Ha! We all know it's the referring physician that drives priority not the acuity of pathology.

Probably end up like Mammo CAD, pretty much useless but upcharged for its use.
 
Seems like computer engineers creating a product in a vacuum without input from people who actually do the job.

Half of the "advantages" already exist or are simply not how medicine works.

Routing the study to the correct sub specialty radiologist? Great it's called a PACS work list or better yet, it's all going to a general radiologist anyway.

Adjust the priority of the study? Ha! We all know it's the referring physician that drives priority not the acuity of pathology.

Probably end up like Mammo CAD, pretty much useless but upcharged for its use.
Indeed, this sounds like an ex-vRad IT guy went and started this company to try and monetize their IT systems for other practices. I went to a vRad dinner and their systems sound amazing.
 
Their website is all marketing. No videos. No examples of their software in action.

http://www.enlitic.com/science.html

As the theranos drama unfolds, I continue to fail to be wowed by the Silicon Valley types who think they can disruptomate healthcare into oblivion. Let's see some results, shall we?

Yep. The tech guys don't realize how much of a PITA it is to work in a heavily regulated industry. Tech industry has little regulation and little in the way of infrastructure demands so growth is easier. They think that "inefficiency" in healthcare is merely because the people working in it are just too stupid to figure it all out.

Although I do chuckle at these types. They are typically left of center (usually far left) and enjoy the benefits of working in a lightly regulated industry all while voting/supporting politicians that regulated the s*** out of nearly every other industry.
 
Yep. The tech guys don't realize how much of a PITA it is to work in a heavily regulated industry. Tech industry has little regulation and little in the way of infrastructure demands so growth is easier. They think that "inefficiency" in healthcare is merely because the people working in it are just too stupid to figure it all out.

Although I do chuckle at these types. They are typically left of center (usually far left) and enjoy the benefits of working in a lightly regulated industry all while voting/supporting politicians that regulated the s*** out of nearly every other industry.


^ everything he/she just said, especially the last part.
 
Maybe we can reduce CT dose a lot and coupled with an actual useful body CAD continue to increase our volume! Long time before Watson finalizes reports without any review.
 
I'd have to agree...computers can't even read EKGs 100% accurately and cardiologists are still required to "over-read" them while getting paid for it. Long way away from a robot/computer automatically reading anything in rads.
 
Yep. The tech guys don't realize how much of a PITA it is to work in a heavily regulated industry. Tech industry has little regulation and little in the way of infrastructure demands so growth is easier. They think that "inefficiency" in healthcare is merely because the people working in it are just too stupid to figure it all out.

Although I do chuckle at these types. They are typically left of center (usually far left) and enjoy the benefits of working in a lightly regulated industry all while voting/supporting politicians that regulated the s*** out of nearly every other industry.



Is what I'm talking about.
 
Theranos is looking like it's going to be the biggest hype-fail of the millennium. Their CEO seems to really buy into other people's designation of her as the next Jobs. Haven't seen a picture of her yet when she's not wearing a black mock turtle. Substance is substantially lacking. The Journal expose was pretty damning to say the least.

I see Enlitic going much the same way. Buncha hype around the data set they have, but they have no idea what to do with it and their claims of Watson-style deep learning are vague, at best. Both of these strike me as the SV model of startups - healthcare style - except the minimum viable product idea doesn't really seem to work in healthcare. Unfortunately, the SV VCs backing them haven't quite figured that out yet.

Time will certainly reveal its share of contenders for the robot radiologist. However, people have been working on the problem of image analysis for decades and there hasn't been a lot of progress. I, for one, think Google's probably got the best shot at providing a useful image analysis software package in the near-term. Although, the addition of Kurzweil to their team may actually slow them down, given that he's now canonized HHMMs in his recent book outlining his model for brain function. But who knows? HHMMs certainly did wonders for voice recognition. PS360 doesn't get it right all of the time, but it's a damn good product.
 
INTRO SCENE:

Human radiologists in desperate combat with the machines for possession of the dead reading room. The human radiologists are a ragtag guerrilla force. PACSnet's weapons consist of drone sonographers, self-propelled computerized tomographinators, four-legged MRI-pods called LAVA-4-points, and humanoid mammographers in various forms.
 
Top