Hi Dilek,

Yes, "--save-posteriors Lesions" will save a probability map with path *yourOutputDir*/posteriors/Lesions.mgz

1 and 2) Lesions are labeled as 99 in seg.mgz. A value of 0.04 in the stats file indicate that the lesion volume is 0.04 mm^3, so it makes sense that seg.mgz has no voxels labeled as 99.  However, it means that something went wrong with the segmentation. What was the command you used to segment these DIR images? I'd probably use --lesion-mask-pattern 1 for them (although I've never tried to run SAMSEG on DIR images). Do you have a T1w scan to add for better whole-brain segmentation and consequently better lesion segmentation?
 
3) You can use the lesion probability map obtained with the --save-posteriors Lesions flag and use mri_binarize and mri_segstats to obtain binary segmentation and corresponding volumes at different thresholds. Note that you don't need to use the same threshold of run_samseg.

Hope it helps,
Stefano 

------------------------------

Message: 2
Date: Thu, 18 Aug 2022 19:45:14 +0300
From: Dilek ?zkan <dozkan521@gmail.com>
Subject: [Freesurfer] samseg lesions labels and values
To: freesurfer@nmr.mgh.harvard.edu
Message-ID:
        <CAPQSyvKJBirH9oOdKXeg=k8BUfdohYvMMoaLnscy-5X0GOMZ7w@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

        External Email - Use Caution       

Hello dear all,

I want to use Freesurfer-Samseg for MS lesion
segmentation(cross-sectional). I have read that I need to add
"--save-posteriors Lesions" to run_samseg to obtain lesion probabilities.

1)In the stats file I can see the value of "Lesions" (0.04) But when I try
to see on freeview, it says label=99 (lesion label) has no value. So, I
can't see the lesions.

2)There are a few labels related to lesions. For example wm lesions;
(498: wmsa, 499: other_wmsa) I have no volume related to these labels
neither in my stats file nor on freeview. Should I get only the "Lesions
label"  for the MS lesion mask?

3)As input, I am using only DIR ?mages. I w?ill try different thresholds. I
have read that by playing mri_binarize and mri_segstats, we can try
different thresholds without re-running anything. So, the threshold value
required for mri_binarize command refers to the same threshold input in the
initial the command of *run_samseg*?

Best,
Wish
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/attachments/20220818/0f62023a/attachment-0001.html