Hi all!
We’re looking into an along-tract analysis between cervical levels C2 to C5 to compare healthy controls and persons with multiple sclerosis. We’re interested in investigating whether we can produce CSA values derived from the PAM50_continuous template at 0.1 intervals (i.e. CSA for level 3.1, 3.2, 3.3, 3.4, etc.). Is this an option with sct_process_segmentation and applying the PAM50_continuous level as the “vert” file, or would it be more effective to approach this differently? Open to discussing and looking forward to learning from you!
Sincerely,
Atlee
Hi Atlee,
Thank you for reaching out. Interesting idea. Unfortunately, you cannot use sct_process_segmentation and applying PAM50_levels_continuous.nii.gz as the “-vert” file, because the flag “-vert” is looking for a file with discrete, integer, values corresponding to vertebral levels.
One possibility is to use the new flag -normalize-PAM50 which was introduced in Valosek et al. 2024. It brings the CSA of the subject’s native space into the PAM50 space (only the superior-inferior axis is considered here). That way, you can compare a bunch of subjects at any coordinate along the superior-inferior axis, at a 0.5mm resolution.
For example, using these data from the SCT course: data_example.zip (10.6 KB)
sct_process_segmentation -i t2_seg.nii.gz -vertfile t2_seg_labeled.nii.gz -normalize-PAM50 1 -perslice 1 -o csa_in_pam50.csv
Here is the output CSV: csa_in_pam50.csv (141.0 KB)
1 Like