Hello,

I am trying to get the relative size of a brain lesion in a template.

For this I created lesion masks that are normalized to the MNI brain.

I then opened a template added my lesion mask as an overlay to get me the size of my lesion in cc using batch descriptive.

I was hoping to get the total brain volume again from batch descriptive of the template but the size I am getting for different template is in the 1600cc area or 1700cc are which seems higher than the average, which makes me concerned that I am getting the right numbers.

Can I get the volume of a template this way and if not how can I get either the volume of a template or the relative volume of a lesion mask in the template compared to the whole brain.

Thank you!

Christian Schranz

You could use FSL:

fslstats input.nii.gz -V

I would derive these details using Python, which is very scriptable.  Note that you may want to consider partial volume effects when computing volumes for non-binary images, e.g. if you have a gray matter probability mask, you may only want to consider voxels that are over 50% GM. To be even more accurate, you could modulate volume by the probability, such that a voxel which is 50% GM only contributes half a voxels volume to the sum.

import nibabel as nb
import numpy as np
from nibabel.imagestats import mask_volume
fnm = 'input.nii.gz'
img = nb.load(fnm)
computed_volume = mask_volume(img)
print(f"{computed_volume}mm3")

Do remember that the MNI template is larger than the average brain, explaining the difference between SPM and MNI templates when normalizing data, e.g. Figure 1 here:

  https://www.ncbi.nlm.nih.gov/pmc/article...