radRounds Radiology Network

Connecting Radiology | Enabling collaboration and professional development

A quick question: How and when do people use 3D reconstructions for diagnosis and treatment planning? What do these reconstructions add to the process?

I realize that a gestalt understanding of the coronary anatomy and the nature of pathology is important, but has anyone seen something on these reconstructions that they did not see on the axials and MPRs?

I also see that 4D scans may quickly become a standard for cardiac imaging, does anyone have any thoughts on this?

Views: 90

Replies to This Discussion

actually 3d reconstruction provides almost nothing about the diagnosis. You can see the merged natural gross anatomy, with 3D. Actually it adds nothing for the diagnosis. When reporting, i first use simple live multiplanar reconstruction. I see any plaque/stenosis in long axis and then cut it axial to the lumen. Beside, I see the live MIP reconstrution all the way from beginning to the end, in all 3 planes. Then, I check the curved MPR images, to find out if i missed anything.

good to know, thanks!
Depends on what you mean by 3D reconstructions- If you mean spinning hearts like you see on every TV at the RSNA they add very little. They are nice for congenital variants, CABG overviews and getting a general feel for the anatomy. If you mean true sohisticated advanced viz that uses all of the tools available it can be very helpful. I find that on some cases sliding MPR slabs are all that I need (ie, little or no calcium). Other times CPR's are essential to get true center line imaging. True cross-sectional views are often essential. Of course that entails having excellent techs to provide quality 3D protocols to your PACS as well as the ability to get on the software yourself when needed. I don't know about you, but I have no time to be a computer graphics artist and spend hours a day to process CTA's myself. I want to read them on the PACS with all of my other cases and only access the software on an as needed basis, quickly and efficiently.
4D scanning again requires excellent functional software and a skilled tech to check and edit contours. In my practice we do a lot more functional cardiac MRI than CT but it varies from place to place. Both are quite accurate in skilled hands. CT lags behind MRI in the breadth of 4D work that can be done but has better resolution and is easier to acquire.
well the tech vs. clinician approach to 3D/4D work is a different discussion altogether :-D. But, that said the 4D CT/MR work would be a little difficult for techs to do especially when dealing with various congenital abnormalities that you don't know are there. They would be able to do the standard 4 and 2 chamber views, but further evaluation would have to be done by the radiologist.
I agree that MPRs/CPRs/MIPs are really the standard for post-acquisition processing (and that 4D MR has more breadth than CT), and you are correct in that I was mainly referring to the pretty pictures at RSNA

I am not sure what your experience is, but I found that volume rendering with MR is a little more difficult than CT because of differences between pulse sequences and the increased artifact, but you could make template rendering preference for each. Then again, I believe one doesn't really use MR for that sort of thing.
Yes. Volume rendering for MRI on a sophisticated level is next to impossible for the reasons you state. Having said that, there are standardized things a tech can do to improve reading efficiency- flow analysis, CPR's with diameter measurements, LV and RV segmentation, etc. It's more efficient for me to check their work and make any corrections as needed myself or have them redo something than for me to do the whole thing myself. The tech doing the case MUST understand the anatomy and have a basic understanding of the disease process. That's why most CT or MRI techs aren't cut out for being 3D techs, esp on a part time basis. My biggest slowdown in reading and protocolling complex congenital heart cases is not having adequate history and having to dig up the info I need on the front end! :(
"not having adequate history and having to dig up the info I need on the front end"
I think every clinician has that problem! :-D I see what you mean though.

With respect to LV and RV segmentation, have you used any automated methods that work?
For CT, the best I've seen for the LV is Visage Imaging- very accurate and reproducible with minimal hand editing. Nothing is good at the RV.
For MRI, there's a guy in Sweden named Einar Heiberg who has an inexpensive program available that is better than most for LV segmentation. You can use it for free for academic purposes if you cite him in any publications and he sells a clinical version. He's a super nice guy and helpful with questions. If you Google 'Heiberg segment' you'll get right to him. I've found it to be better than Medis at segmentation although Medis is easier to navigate around the program. ReportCard by GE is about the same as Medis as far as accuracy.
I agree with out about Visage platform. It is by far the best I've worked with, and I've worked with them all.
Hi Bob!
How's Heart Saver's doing? I know you guys from my 3D lab- 3DR. We were in discussions with TJ et al about working together on 3D about three years ago. How many sites do you have up? As I recall a few years ago you had two out west and one east coast.
3DR is doing great- busier each month and our customer sites are very happy. We also have 3DR Academy up and running online for tech training with college credit and tech certification.
It sounds like you have put together what TJ couldn't. I left Heart Savers in May and it's almost out of buisness. You were lucky to not get involved with TJ.


Sponsor Ad

© 2020   Created by radRounds Radiology Network.   Powered by

Badges  |  Report an Issue  |  Terms of Service