#JHSJC

March #JHSJC Transcript

Thank you to everyone who participated in the March #JHSJC! We looked at two great articles from Volume 45, Issue 3 of The Journal of Hand Surgery: Convolutional Neural Network for Second Metacarpal Radiographic Osteoporosis Screening and The Value of Intraoperative 3-Dimensional Fluoroscopy in the Treatment of Distal Radius Fractures: A Randomized Clinical Trial.

Feel free to continue the conversation below in the comments! And please check out previous #JHSJC transcripts (please scroll to the bottom of the page).


J Hand Surg Am- ASSH @JHandSurg:
March’s #JHSJC starts now! Let’s get the discussion going!

Brent Graham @BgTalkinman:
Welcome to March #jhsjc! #artificialintelligence #AI #osteoporosis #DRF Let’s start with the paper on convoluted neural networks #jhsjc
Be sure to use the hashtag #jhsjc

Here is the podcast interview I did with senior author @whammert https://www.jhandsurg.org/pb/assets/raw/Health%20Advance/journals/yjhsu/March_2020.mp3 #jhsjc
This was a very technical article so let’s focus on the clinical side. #jhsjc
As @whammert describes in out interview, the 1st task was to train their system to standardize hand radiographs #jhsjc

Warren @whammert to @BgTalkinman:
great podcast, but I am biased:) #jhsjc

Brent Graham @BgTalkinman to @whammert:
It was very informative! Recommended listening! #jhsjc

Brent Graham @BgTalkinman:
Once trained, the system had 82% sensitivity, 96% specificity. Is that adequate for a screening test? #jhsjc
I will answer my own question: yes! the sensitivity is high enough to screen and then move on to a more definitive test like DEXA #jhsjc
(Not sure if tchat.io is working. Be sure to check you Twitter feed for comments. #jhsjc)

Brent Graham @BgTalkinman:
#jhsjc The question might be whether this type of investigation should be the reference standard. In other words perhaps the standardized plain radiograph should replace the DEXA scan. Thoughts?
#jhsjc A DEXA scan doesn’t cost much but it does require effort while a plain radiograph is commonly obtained for a variety of indications. Will somebody study the feasibility of elevating the plain hand radiograph to reference standard status?
#jhsjc The main issue in #osteoporosis care is fragmentation – poor communication between orthodox surgeons and PCP. Is there any potential that this could help that?

Brent Graham @BgTalkinman:
Are there other areas where this type of #AI based standardization and screening could be useful in hand surgery? #jhsjc
#jhsjc Are there other tasks in clinical hand surgery that could be amenable to this kind of #AI based standardization/screening?

Warren @whammert to @BgTalkinman:
carpal alignment #jhsjc

Brent Graham @BgTalkinman:
#jhsjc Scanning of x-rays for subtle fractures, adequate reduction of #fractures, interpretation of #electrodiagnostic tests?

Warren @whammert to @BgTalkinman:
predicting need for imaging with computer algorithms based on probability for XX #jhsjc

Brent Graham @BgTalkinman:
Personally I think this is a groundbreaking article that has the potential to fundamentally change how we do things in the future! #jhsjc

Brent Graham @BgTalkinman:
#jhsjc Let’s move on two the other article – maybe that is more clearly on the radar screen of our participants!
#jhsjc This was a very well-planned an implemented RCT comparing standard intra-op fluoroscopy with 3D fluoro. There are some interesting images in the article showing what this look like.
#jhsjc The main outcome was a need for revision surgery. That was frequently required, but less for the 3D cases than the standard cases 24% vs 32%. This result was “not statistically significant”. Is it real, i.e. type I error?

Warren @whammert to @BgTalkinman:
are you concerned about it being underpowered? #jhsjc

Brent Graham @BgTalkinman to @whammert:
No – the sample was adequately powered. I misspoke in my last tweet – I meant type II error! #jhsjc
Even if the sample power is adequate, by the usual definition there is still a 20% chance that the conclusion is wrong. #jhsjc

Warren @whammert to @BgTalkinman:
they were still 46 patients short as they needed 250 and had 206. #jhsjc
I am not sure that would change anything, but worth noting #jhsjc

Brent Graham @BgTalkinman:
The 3D patients also had Intraoperative adjustments in 11% – 0 in the cases that used standard fluoro. #jhsjc

Brent Graham @BgTalkinman to @whammert:
You are correct – good pickup. I had forgotten that. #jhsjc

Brent Graham @BgTalkinman
#jhsjc Let’s assume that it is a type II error and that the difference between the groups is real. Does that mean that this is an innovation that should start to become the standard?
#jhsjc My guess is that this is a change that will take time but there seems to be a substantial advantage if the goal is to avoid revision surgery.

Warren @whammert to @BgTalkinman:
The most surprising point to me was the high revision rate in both groups – 4% and 10% in each group #jhsjc

Brent Graham @BgTalkinman
#jhsjc Are there other fluoroscopic view that would be better than the standard views used in this study?

RayRavenMD, MHCI, MBA @RayRavenMD:
Sorry for the late entry! Just finished with cases! #jhsjc

Warren @whammert to @BgTalkinman:
skyline and 20 degree lateral show articular surface #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD:
I like the standard views and skyline. #jhsjc
20 degree lateral is a standard view for me for DRFxs. #jhsjc
Anyone else use any other views? #jhsjc

Brent Graham @BgTalkinman to @whammert:
#jhsjc the rates were actually 31% for 2d and 24% for 3D

RayRavenMD, MHCI, MBA @RayRavenMD:
I have a question about the endpoint for this study: how objective do you believe it is? #jhsjc
I’m not sure I see the value in the primary outcome being revision surgery since this is so subjected to the surgeon. #jhsjc
Shouldn’t the primary outcome be the patient-reported outcome? #jhsjc

Brent Graham @BgTalkinman to @whammert:
#jhsjc Most of the revisions were important – intra-articular screw placement and articular surface malalignment.

Warren @whammert to @RayRavenMD:
yea, but I thought they were high – 4 % in the 3d and 10% in 2d are both high #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman and @whammert:
that’s right. #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD:
That is true but in the context of a RCT that effect should be neutralized. #jhsjc
The outcomes were equivalent but the patients in the 2D group actually required more care – about 50% more revisions #jhsjc
The results were equivalent but the 2D patients required 50% more revisions to get those results. That seems important. #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD:
I don’t do any volar plates. I treat all cases (except volar Barton) dorsally; more effective/easier. #jhsjc
No intra-articular screws. #jhsjc

Brent Graham @BgTalkinman:
I think there have been few problems with twitter and tchat.io tonight. I hope I haven’t ignored anyone with a comment. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman:
I still pin a lot of fractures but I also use volar plates. I am surprised to learn that you don’t use them at all, @BgTalkinman ! #jhsjc

Warren @whammert to @BgTalkinman and @RayRavenMD:
Not sure this is easier or there wouldn’t be so many companies making volar plates #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD:
They do not address the biomechanical need of most #DRF. Dorsal buttressing is simpler and more effective #jhsjc

Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
They make money! Buttressing the fracture is an AO fundamental #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman:
Makes sense biomechanically, but can see the utility in volar plating for some Fx types and less extensor tendon issues #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD:
I haven’t even done it once and I have treated literally hundreds of #DRF & a lot of osteotomies – hundreds of those too. #jhsjc
Warren @whammert to @BgTalkinman and @RayRavenMD:
yes, but extensor tendons, need for late removal, and anatomically alignment are dorsal issue #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman and @whammert:
They’ve created a huge market. I am guilty in this as well. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD:
I have three distal radius cases tomorrow. I am volar plating 2 and pinning 1. #jhsjc

Warren @whammert to @RayRavenMD and @BgTalkinman:
still a technique and the VLP have made it too easy, resulting in OIF rather than ORIF #jhsjc

Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
There is no credible evidence that volar plates are better. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD:
Did this study change your thinking in any way? #jhsjc

Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
Also very difficult to deal with intra-articular comminution. Treat those like a tibial plateau: graft/butress. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman and @whammert:
I agree. There is no credible evidence any basal joint reconstruction method is better than the other. #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD:
I don’t do those cases, so no. Intra-articular screws are a rare occurrence with dorsal plating. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman and @whammert:
I think surgeons often do what works best in their hands. #jhsjc

Warren @whammert to @BgTalkinman and @RayRavenMD:
tibial plateau different in there are not multiple glifding tendons right on the plate #jhsjc

Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
The cost of doing business there. No flexor tendon problems with dorsal plates. That’s more important IMHO #jhsjc

Brent Graham @BgTalkinman to @RayRavenMD and @whammert:
That’s fair but how many do you know who even know how to do a dorsal plating? #jhsjc

Warren @whammert to @BgTalkinman and @RayRavenMD:
comes back to technique – reduction and proper plate position avoids flexor problems #jhsjc

Brent Graham @BgTalkinman:
Okay! Thanks for tuning in all who were here! Back here next month. #jhsjc

RayRavenMD, MHCI, MBA @RayRavenMD to @BgTalkinman and @whammert:
Thanks, @BgTalkinman and @whammert! #jhsjc

Leave a Reply

Your email address will not be published. Required fields are marked *