Lesion segmentation protocol #9

plbenveniste · 2024-03-21T10:36:02Z

This PR adds the lesion segmentation protocol which was first discussed in issue 77 from the CanProCo repo.

The lesion segmentation protocol adds an .md file at the base of this repository.

Feedback and improvements are welcomed !

… from issue 77 on canproco repo

lesion_segmentation_protocol.md

jcohenadad

This is a great start @plbenveniste ! Thank you for initiating this. I would add a few visual examples of lesion segmentation, so we can all agree on how 'conservative' we want to be with our segmentations. Maybe that could be reserved to an 'example' section.

lesion_segmentation_protocol.md

maxradx · 2024-04-10T15:14:52Z

lesion_segmentation_protocol.md

+
+The following details the protocol for Multiple Sclerosis (MS) lesion segmentation in the spinal cord.
+Imaging the spinal cord is often essential to confirm the diagnosis of MS. That is because the lesions of the spinal cord are included in the McDonald diagnosis criteria, which studies dissemination in space and in time [(Thompson et al. 2018)](https://pubmed.ncbi.nlm.nih.gov/29275977/). While the MAGNIMS-CMSC-NAIMS working group recommends to use at least two sagittal images for MS diagnosis, still, axial imaging is mentioned as optional in international imaging guidelines [(Wattjes et al. 2021)](https://pubmed.ncbi.nlm.nih.gov/34139157/).
+For detecting MS lesions in the spinal cord, two main contrasts emerge: PSIR and STIR contrasts. New studies [(Peters et al. 2024)](https://pubmed.ncbi.nlm.nih.gov/38289376/)[(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that using PSIR contrasts improved MS lesion detection in the spinal cord. [(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that the PSIR contrast showed a higher signal-to-noise (SNR) ratio compared to the STIR contrast. 


Maybe my suggestion is not up-to-date, but I have doubts/questions abouts STIR sequences being frequently used for assessing MS lesions, at least in standard/basics protocols (even if mentioned in the picture showing example of of clinically used sequences for MS). From my understanding, the most useful sequence for detecting MS lesions in clinical practice is proton density (PD), but probably that changes from sites/radiologists perspective.

According to the articles, PSIR sequences seem great! However, here are some potential precisions I would address :
1-Peters and al : PSIR compared to STIR (not the best...) and T2;
2-Fechner : PSIR compared to T2 and T1C+ sequences (not compared to PD sequences...)

Suggested change

For detecting MS lesions in the spinal cord, two main contrasts emerge: PSIR and STIR contrasts. New studies [(Peters et al. 2024)](https://pubmed.ncbi.nlm.nih.gov/38289376/)[(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that using PSIR contrasts improved MS lesion detection in the spinal cord. [(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that the PSIR contrast showed a higher signal-to-noise (SNR) ratio compared to the STIR contrast.

For detecting MS lesions in the spinal cord, two main contrasts emerge: PSIR and STIR contrasts. New studies [(Peters et al. 2024)](https://pubmed.ncbi.nlm.nih.gov/38289376/)[(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that using PSIR contrasts improved MS lesion detection in the spinal cord. [(Fechner et al. 2019)](https://pubmed.ncbi.nlm.nih.gov/30679225/) showed that the PSIR contrast showed a higher signal-to-noise (SNR) ratio compared to the STIR contrast.

maxradx · 2024-04-10T15:19:26Z

lesion_segmentation_protocol.md

+
+## Criteria to segment MS lesions in the spinal cord:
+
+Do not segment lesions in images with too many artifacts (such as this one : https://github.com/ivadomed/canproco/issues/53#issue-1938136790). Preferably, add the image to the exclude file so that it isn’t used for model training…


Question : in words, what would be an image considered being too artefacted? Any criteria?

Suggested change

Do not segment lesions in images with too many artifacts (such as this one : https://github.com/ivadomed/canproco/issues/53#issue-1938136790). Preferably, add the image to the exclude file so that it isn’t used for model training…

Do not segment lesions in images with too many artifacts (such as this one : https://github.com/ivadomed/canproco/issues/53#issue-1938136790). Preferably, add the image to the exclude file so that it isn’t used for model training…

lesion_segmentation_protocol.md

Co-authored-by: Maxime B <[email protected]>

Co-authored-by: Julien Cohen-Adad <[email protected]>

jcohenadad

i would also reference related initiatives-- eg

there are other important initiatives

lesion_segmentation_protocol.md

Co-authored-by: Julien Cohen-Adad <[email protected]>

lesion_segmentation_protocol.md

Co-authored-by: Julien Cohen-Adad <[email protected]>

plbenveniste · 2024-04-11T15:50:55Z

Look at MSSEG challenge's public data to find cases of MS lesions in the spinal cord. The objective is to build a dataset for lesion segmentation training.
Also look at the MSSEG challenge article to get inspiration from the segmentation protocol

plbenveniste · 2024-04-11T15:56:26Z

lesion_segmentation_protocol.md

+Furthermore, it is recommended to get familiar with SCT for creating QCs and for manual correction ([SCT tutorial](https://spinalcordtoolbox.com/user_section/tutorials.html)). 
+
+## Step 2: Spinal cord anatomy and lesion segmentation
+Before, moving on to MS lesion segmentation, trainees are advised to study the neuroanatomical structures of healthy spinal cords. Trainees should look at healthy spinal cords in MRI images of different contrasts: T2w, T1w, PSIR, STIR, MP2RAGE... A public dataset will be built for this purpose.


SpineGeneric could be used for this purpose

plbenveniste · 2024-04-11T15:59:51Z

lesion_segmentation_protocol.md

+## Step 2: Spinal cord anatomy and lesion segmentation
+Before, moving on to MS lesion segmentation, trainees are advised to study the neuroanatomical structures of healthy spinal cords. Trainees should look at healthy spinal cords in MRI images of different contrasts: T2w, T1w, PSIR, STIR, MP2RAGE... A public dataset will be built for this purpose.
+
+To learn the specificity of MS lesions, trainees should work on differentiating MS lesions from other diseases. 


Which disease should be used ? NMO ?

We should also look into building this public dataset: which public data do we have ?

plbenveniste · 2024-04-11T16:01:24Z

lesion_segmentation_protocol.md

+
+To learn the specificity of MS lesions, trainees should work on differentiating MS lesions from other diseases. 
+
+One of the most challenging task of MS lesion segmentation is to distinguish the border of a lesion and the cerebrospinal fluid (CSF). To learn where to draw the lesion border, a set of tricky examples validated by a NeuroRadioligist will be created. 


Which data should be used? MSSEG data? Ask a collaborator to help us on this?

Co-authored-by: Maxime B <[email protected]>

plbenveniste · 2024-04-11T19:09:18Z

lesion_segmentation_protocol.md

+
+To learn the specificity of MS lesions, trainees should work on differentiating MS lesions from other diseases. 
+
+One of the most challenging tasks of MS lesion segmentation is to distinguish the border of a lesion and the cerebrospinal fluid (CSF). To learn where to draw the lesion border, a set of tricky examples validated by a NeuroRadioligist will be created. 


Reference the videos made by Julien during work with Michelle Chen regarding MS lesion segmentation

plbenveniste · 2024-08-28T19:08:55Z

Referencing this lesion atlas which could be very useful in creating this lesion seg protocol: https://link.springer.com/book/10.1007/978-3-540-71372-2

plbenveniste added 2 commits March 21, 2024 11:31

initialised the lesion_segmentation_protocol.md file based on comment…

f0b3b34

… from issue 77 on canproco repo

fixed typo based on Julien's comment

973c5b2

plbenveniste added the documentation Improvements or additions to documentation label Mar 21, 2024

plbenveniste requested review from jcohenadad and Nilser3 March 21, 2024 10:36

plbenveniste self-assigned this Mar 21, 2024

plbenveniste mentioned this pull request Mar 21, 2024

Lesion protocol section ivadomed/canproco#77

Open

jcohenadad requested a review from maxradx April 10, 2024 13:55

jcohenadad reviewed Apr 10, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 10, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 10, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 10, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 10, 2024

View reviewed changes

plbenveniste added 2 commits April 10, 2024 11:22

changed criterias to bullet points

6d01e30

replace link to example by hyperlink

fddd8e3

maxradx reviewed Apr 10, 2024

View reviewed changes

plbenveniste and others added 4 commits April 10, 2024 11:28

typo correction in introduction

2e7d7fd

Co-authored-by: Maxime B <[email protected]>

changed json link to json example in the readme

3bbcc6a

added examples of complicated lesion segmentations

ffa1b0a

QC explanation reformatting

61291c7

Co-authored-by: Julien Cohen-Adad <[email protected]>

jcohenadad reviewed Apr 10, 2024

View reviewed changes

This comment was marked as resolved.

Sign in to view

plbenveniste mentioned this pull request Apr 11, 2024

Update QC report command and JSON standard in lesion segmentation protocol #13

Open

plbenveniste added 2 commits April 11, 2024 10:14

add sct_qc commands

fe015f5

added software section

7e18554

This comment was marked as outdated.

Sign in to view

added ref at the bottom to other segmentation protocols

b98087b

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

added step 2 : sc anatomy and lesion segmentation

77425fb

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

julien typo fix : remove":"

6e4a30b

Co-authored-by: Julien Cohen-Adad <[email protected]>

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

jcohenadad reviewed Apr 11, 2024

View reviewed changes

lesion_segmentation_protocol.md Outdated Show resolved Hide resolved

plbenveniste and others added 8 commits April 11, 2024 11:39

typo fix : remove ":" 2

7e0eac3

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite partial volume explanation

359f051

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite sentence on not segmenting above first vert

8928382

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite two options for seg

f9a370b

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite manual correction explanation

11ac1d6

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite QC

b525c0c

Co-authored-by: Julien Cohen-Adad <[email protected]>

rewrite protocol when unsure

5f11e4f

Co-authored-by: Julien Cohen-Adad <[email protected]>

change last section name

2982982

Co-authored-by: Julien Cohen-Adad <[email protected]>

update dataset of multiple contrasts

e403321

plbenveniste commented Apr 11, 2024

View reviewed changes

remove dcm and sci

be19641

plbenveniste commented Apr 11, 2024

View reviewed changes

remove comment

a021996

plbenveniste commented Apr 11, 2024

View reviewed changes

plbenveniste and others added 4 commits April 11, 2024 12:11

add step 3

657b9bb

lesion definition suggestion

63f305c

Co-authored-by: Maxime B <[email protected]>

typo fix

13ca692

changed step 1 title

888513d

plbenveniste commented Apr 11, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Lesion segmentation protocol #9

Lesion segmentation protocol #9

plbenveniste commented Mar 21, 2024

jcohenadad left a comment

maxradx Apr 10, 2024

maxradx Apr 10, 2024

jcohenadad left a comment

This comment was marked as resolved.

This comment was marked as outdated.

plbenveniste commented Apr 11, 2024 •

edited

Loading

plbenveniste Apr 11, 2024

plbenveniste Apr 11, 2024

plbenveniste Apr 11, 2024

plbenveniste Apr 11, 2024

plbenveniste commented Aug 28, 2024


		## Criteria to segment MS lesions in the spinal cord:

		Do not segment lesions in images with too many artifacts (such as this one : https://github.com/ivadomed/canproco/issues/53#issue-1938136790). Preferably, add the image to the exclude file so that it isn’t used for model training…


		To learn the specificity of MS lesions, trainees should work on differentiating MS lesions from other diseases.

		One of the most challenging task of MS lesion segmentation is to distinguish the border of a lesion and the cerebrospinal fluid (CSF). To learn where to draw the lesion border, a set of tricky examples validated by a NeuroRadioligist will be created.


		To learn the specificity of MS lesions, trainees should work on differentiating MS lesions from other diseases.

		One of the most challenging tasks of MS lesion segmentation is to distinguish the border of a lesion and the cerebrospinal fluid (CSF). To learn where to draw the lesion border, a set of tricky examples validated by a NeuroRadioligist will be created.

Lesion segmentation protocol #9

Are you sure you want to change the base?

Lesion segmentation protocol #9

Conversation

plbenveniste commented Mar 21, 2024

jcohenadad left a comment

Choose a reason for hiding this comment

maxradx Apr 10, 2024

Choose a reason for hiding this comment

maxradx Apr 10, 2024

Choose a reason for hiding this comment

jcohenadad left a comment

Choose a reason for hiding this comment

This comment was marked as resolved.

This comment was marked as outdated.

plbenveniste commented Apr 11, 2024 • edited Loading

plbenveniste Apr 11, 2024

Choose a reason for hiding this comment

plbenveniste Apr 11, 2024

Choose a reason for hiding this comment

plbenveniste Apr 11, 2024

Choose a reason for hiding this comment

plbenveniste Apr 11, 2024

Choose a reason for hiding this comment

plbenveniste commented Aug 28, 2024

plbenveniste commented Apr 11, 2024 •

edited

Loading