Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129
I do not think it is easy to solve the problem, because often it is unknown were the exact GM/WM border is in the lesion. Manual editing may introduce a bias to the CT, so probably it is better to leave it as a limitation of the method.
Best, Martijn
Van: freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] Namens Granberg, Erik Tobias Verzonden: vrijdag 15 januari 2016 21:15 Aan: freesurfer@nmr.mgh.harvard.edu Onderwerp: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129
Hi Martijn,
Thank you for the response! I believe you are right that it is a complex problem. It is true that the GM/WM border can sometimes by hard to identify in type I (leukocortical) lesions, but in in purely intracortical lesions, it is still readily identifiable. We also have corresponding 7T scans to help guide our edits. In this case we are specifically interested in looking at the relationship of cortical pathology and cortical thickness and are therefore still looking for a possible workaround.
Kind regards, Tobias
On 15 jan 2016, at 15:49, Martijn Steenwijk <martijnsteenwijk@gmail.commailto:martijnsteenwijk@gmail.com> wrote:
I do not think it is easy to solve the problem, because often it is unknown were the exact GM/WM border is in the lesion. Manual editing may introduce a bias to the CT, so probably it is better to leave it as a limitation of the method.
Best, Martijn
Van: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] Namens Granberg, Erik Tobias Verzonden: vrijdag 15 januari 2016 21:15 Aan: freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Onderwerp: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129 _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
In that case it may be better to apply some form of GM lesion filling. Studies have shown that (at least in FAST and SPM) WM lesions affect the GM segmentation throughout the brain (i.e. Chard et al, Gelinau-Morel et al, Battaglini et al) – and I have the feeling that the same is true for FreeSurfer. The reason is that the WM (and so the GM) distribution is slightly different as a result of the lesions, thereby the border of the GM is drawn slightly different compared to the normal situation. So I would try to fill the WM (and GM) lesions before running FreeSurfer. Note that estimating the proper local GM intensity may be challenging as there is a lot of intensity variation in the GM.
Best, Martijn
Van: Granberg, Erik Tobias [mailto:EGRANBERG@mgh.harvard.edu] Verzonden: vrijdag 15 januari 2016 21:59 Aan: martijnsteenwijk@gmail.com; Freesurfer support list Onderwerp: Re: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Hi Martijn,
Thank you for the response! I believe you are right that it is a complex problem. It is true that the GM/WM border can sometimes by hard to identify in type I (leukocortical) lesions, but in in purely intracortical lesions, it is still readily identifiable. We also have corresponding 7T scans to help guide our edits. In this case we are specifically interested in looking at the relationship of cortical pathology and cortical thickness and are therefore still looking for a possible workaround.
Kind regards, Tobias
On 15 jan 2016, at 15:49, Martijn Steenwijk martijnsteenwijk@gmail.com wrote:
I do not think it is easy to solve the problem, because often it is unknown were the exact GM/WM border is in the lesion. Manual editing may introduce a bias to the CT, so probably it is better to leave it as a limitation of the method.
Best, Martijn
Van: mailto:freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu [ mailto:freesurfer-bounces@nmr.mgh.harvard.edu mailto:freesurfer-bounces@nmr.mgh.harvard.edu] Namens Granberg, Erik Tobias Verzonden: vrijdag 15 januari 2016 21:15 Aan: mailto:freesurfer@nmr.mgh.harvard.edu freesurfer@nmr.mgh.harvard.edu Onderwerp: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129
_______________________________________________ Freesurfer mailing list mailto:Freesurfer@nmr.mgh.harvard.edu Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Hi again,
Thank you again for the response. We have been discussing such an approach as a backup plan if it is not possible to adjust the pial surface with manual interventions. It does not feel straightforward though as the lesion filling techniques have only been used in WM before (AFAIK). As you say, the intensities across the cortex can vary quite a lot as well so choosing the intensity to fill with is also problematic (although some sort of averaging of neighboring voxels to each lesion seems logical).
Kind regards, Tobias
_________________________________________ Tobias Granberg, MD, PhD
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129
Resident in Radiology Karolinska University Hospital Stockholm, Sweden
On 15 jan 2016, at 16:10, Martijn Steenwijk <martijnsteenwijk@gmail.commailto:martijnsteenwijk@gmail.com> wrote:
In that case it may be better to apply some form of GM lesion filling. Studies have shown that (at least in FAST and SPM) WM lesions affect the GM segmentation throughout the brain (i.e. Chard et al, Gelinau-Morel et al, Battaglini et al) – and I have the feeling that the same is true for FreeSurfer. The reason is that the WM (and so the GM) distribution is slightly different as a result of the lesions, thereby the border of the GM is drawn slightly different compared to the normal situation. So I would try to fill the WM (and GM) lesions before running FreeSurfer. Note that estimating the proper local GM intensity may be challenging as there is a lot of intensity variation in the GM.
Best, Martijn
Van: Granberg, Erik Tobias [mailto:EGRANBERG@mgh.harvard.edu] Verzonden: vrijdag 15 januari 2016 21:59 Aan: martijnsteenwijk@gmail.commailto:martijnsteenwijk@gmail.com; Freesurfer support list Onderwerp: Re: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Hi Martijn,
Thank you for the response! I believe you are right that it is a complex problem. It is true that the GM/WM border can sometimes by hard to identify in type I (leukocortical) lesions, but in in purely intracortical lesions, it is still readily identifiable. We also have corresponding 7T scans to help guide our edits. In this case we are specifically interested in looking at the relationship of cortical pathology and cortical thickness and are therefore still looking for a possible workaround.
Kind regards, Tobias
On 15 jan 2016, at 15:49, Martijn Steenwijk <martijnsteenwijk@gmail.commailto:martijnsteenwijk@gmail.com> wrote:
I do not think it is easy to solve the problem, because often it is unknown were the exact GM/WM border is in the lesion. Manual editing may introduce a bias to the CT, so probably it is better to leave it as a limitation of the method.
Best, Martijn
Van: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] Namens Granberg, Erik Tobias Verzonden: vrijdag 15 januari 2016 21:15 Aan: freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Onderwerp: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129 _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Hello Tobias,
The lesion segmentation toolbox, which has been validated in MS patients (http://www.applied-statistics.de/lst.html) has a useful tool to fill in T1 hypointensities (based on Chard et al., 2010; http://onlinelibrary.wiley.com/doi/10.1002/jmri.22214/abstract ).
The toolbox generates lesion probability maps, which can then be used to 'fill in' the lesions, with normally appearing white matter. You could then run recon-all on these filled in T1 images. I haven't used such an approach myself, so I don't have any data regarding the validity of or accuracy of cortical measures obtained using such a method.
Best,
Prad
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Granberg, Erik Tobias EGRANBERG@mgh.harvard.edu Sent: Friday, January 15, 2016 1:15 PM To: freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We're working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces "dips" in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or "anchor" the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129
Hi Prad,
Thank you for your response! Unfortunately the problematic issue here is cortical MS lesions, which is not as easily fixed as WM lesions. If you plan on using LST (which is a nice tool indeed) be sure to be attentive to juxtacortical lesions as these, from my experience, are not always picked up as nicely using LST:s LPA algorithm and may still cause some issues. They would, however, still be editable as usual in FreeSurfer.
Have a nice weekend!
Kind regards, Tobias
On 15 jan 2016, at 16:10, Bharadwaj, Pradyumna - (prad) <prad@email.arizona.edumailto:prad@email.arizona.edu> wrote:
Hello Tobias,
The lesion segmentation toolbox, which has been validated in MS patients (http://www.applied-statistics.de/lst.html) has a useful tool to fill in T1 hypointensities (based on Chard et al., 2010; http://onlinelibrary.wiley.com/doi/10.1002/jmri.22214/abstract ).
The toolbox generates lesion probability maps, which can then be used to 'fill in' the lesions, with normally appearing white matter. You could then run recon-all on these filled in T1 images. I haven't used such an approach myself, so I don't have any data regarding the validity of or accuracy of cortical measures obtained using such a method.
Best, Prad ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of Granberg, Erik Tobias <EGRANBERG@mgh.harvard.edumailto:EGRANBERG@mgh.harvard.edu> Sent: Friday, January 15, 2016 1:15 PM To: freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Cortical MS lesions - correcting pial surfaces
Dear FreeSurfer colleagues,
We’re working with MS data and wanted your input on how we could best handle patients with extensive cortical pathology causing inaccurate reconstructions of the surfaces. In short the pial surfaces “dips” in too deep towards the WM surface due to the T1-hypointensity.
A lot of the problems can be mediated by doing wm.mgz edits and actually including the lesioned tissues, which pushes out both surfaces to an accurate position. Sometimes though, the cortical lesion affects the whole depth of the surface and this trick does not help.
Is there any way to directly push out or “anchor” the pial surface?
Thank you for the help!
Kind regards, Tobias
Post-doctoral research fellow Multiple Sclerosis Imaging Group MGH/HMS/MIT Athinoula A. Martinos Center for Biomedical Imaging 149 13th St Charlestown, MA 02129 _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Bharadwaj, Pradyumna - (prad) -
Granberg, Erik Tobias -
Martijn Steenwijk