LungCTAnalyzer extension for lung CT segmentation and analysis for COVID-19 assessment

New extension is added for lung CT segmentation and analysis for COVID-19 assessment. It is developed by Prof. Rudolf Bumm (@rbumm), a clinician in KSGR (Chur, Switzerland; with some help from Slicer core developers to polish things up), which is a nice demonstration of how Python scripting in Slicer allows users to quickly develop clinically deployable tools from their ideas.

The extension offers a quick (1-minute) lung segmentation wizard and quantitative volumetric analysis of the lungs with quick slice and volume rendering visualization and PDF report generation.

Extension webpage: GitHub - rbumm/SlicerLungCTAnalyzer: This is a 3D Slicer extension for segmentation and spatial reconstruction of infiltrated, collapsed, and emphysematous areas in lung CT.

Demo video:

4 Likes

I just tried this out-amazing! Thank you

1 Like

Thanks for the mention, @lassoan, and thanks for the great cooperation. I pushed several small improvements to Github.

1 Like

Hi, I am very new and naive to 3D slicer. I just installed the last stable version on Mac OS
and installed the Chest Imaging Platform but i could not find the Lung CT analyser extension for lung CT segmentation and analysis. I checked the extension webpage on GitHub but I don’t know how to install the extensions in 3D Slicer from the files after dowonloading the zip files. By the way I also hardly see the modules in Module finder as they look as ghost icons. Is it a matter of OS ? I am using a Mac Book Pro with 16 Go ram under Mac OS High sierra 10.13.6
Sorry for being at a starter point and thank you for your help
The demo video looks amazing and I am impatient to try it

Lung CT analyzer is available for latest Slicer Preview Release (extension name: LungCTAnalyzer, in Chest Imaging Platform category).

This is a known issue, a workaround is to change your system to dark mode.

@rbumm This extension may bring in users that are new to Slicer and therefore it would be useful if you could add a few sentences about how to install the extension to the extension documentation page. Until now, the extension was only submitted to the Slicer-4.13 extensions index, but I’ve added it not to the Slicer-4.11 extensions index now, so from tomorrow it should show up for the latest stable Slicer version (Slicer-4.11.20200930).

Hi Andras

Thank you very much for you so quick reply
Very clear !

I am a nuclear medicine physician and my aim is to use 3D slicer and CIP for lobar quantification of Lung ventilation/perfusion scans by combining lobar lung segmentation and quantification extensions. If you have suggestions I’ll be glad to follow them

Congratulations for the great job you did and are doing with 3D slicer
Khaldoun

1 Like

@lassoan I added a description on the extension documentation page. I will add a few screenshots later. Hope I did it right, would we have to comment on https://slicer.kitware.com/midas3/slicerappstore too ?

Oh @lassoan I just see that you added it to the extension index, so it will appear “normally” if you “Install slicer extensions” from 3D slicer “Welcome” page ?

Yes, the extension is already available in the extension manager for 4.13 and will be available for 4.11 from tomorrow. So, there is no need to describe the manual installation steps, only installation via the Extensions Manager in Slicer. Thank you!

Thanks for the introduction. Slicer and its extension should be able to help you with these. Post separate topics with specific questions if you need help.

What parameters should I look at to know if the patient has covid?

After running the Lung CT Analyzer, there is an option (button) to display the “COVID-19 Results Table”. In Slicer’s COVID-19 demo dataset you would get:

Results right (ml) right (%) left (ml) left (%) total (ml) total (%)
Lung volume 3857 51 3675 49 7532 100
Functional volume 2104 55 2874 78 4978 66
Affected volume 1846 48 854 23 2700 36
CovidQ (affected / functional) 0.88 -1 0.3 -1 0.54 -1

COVID-Q is higher in the right lung compared to the left lung. This corresponds well with the extent of infiltrations on the right side.

image

Please understand that this computation is for scientific use only, it has been implemented only recently and needs validation in clinical studies.

1 Like

hi, It is great work. I think the CT is segmented by threshold or other traditional tech,right? I found its still have some cavity in the segmentation.Is there any way to segment the lobes and PA,Aorta?

Hi,

The “Lung CT Segmenter” currently produces lung masks only by a grow from seeds function. The masks are used in “Lung CT Analyzer” for creating thresholded segments of lung parenchyma.
If the lung masks have dents, missing parts or leaks you can place additional fiducials (right lung, left lung, other = outside) before you press “Apply” to correct this. You can also change the lung intensity range slider to prevent leaks.

If you need to segment lung lobes you may try “Interactive Lobe Segmentation” from the “Chest Imaging Platform”. This is from different developers and also works quite well.

Best regards
Rudolf

For PA, Aorta, PV you would need a very good contrast CT and could segment the vessels by producing a vessel mask and then use “Grow from Seeds” like shown in this video:

Hi, I was trying to use this extension in order to predict covid, but the “COVID results table” button doesn’t appears in my slicer extension installation. Instead of this appears a button for “extended result table” (Screenshot - d0be934391961b2497048209faef2979 - Gyazo). It is an error? Or it has been removed?
Thanks.

Hi Fran,

the COVID results table was renamed because it could also be used in conditions like Pneumonia, measuring infiltrations or ARDS, and be useful in COPD. The data in the table have not changed much and can be used in COVID.19 disease. Systematic evaluations are under way. A high relative affection rate (infiltrated + collapsed, > 50 %) in % of total lung tissue indicates advanced cases. High collapse rates (10-20%) are often seen in fatal cases.

Best regards
rudolf

Ok, I get that it is better to look at the percentage of several injuries than in a single parameter of the covid. Thanks for the answer, the work is great!

Fran

1 Like

thans prof rbumm
please can you give your email to important disscution in this extention

Please keep the discussion open, to allow others to learn from and contribute to the discussion. Switch to private messages only if you need to talk about personal or confidential matters.

3 Likes