On 10/25/2024 2:00 PM, Shen, Qian wrote:

External Email - Use Caution

Hi FreeSurfer Team,

We are looking to apply voxelwise PVC on our dynamic contrast-enhanced (DCE) MRI data using methods implemented in PETSurfer. Our DCE MRI data voxelsize =1x1x5mm, and we have used the Patlak model to estimate the transfer coefficient (Ktrans). We are interested in both GM and WM signals. I have a few specific questions regarding the applicability and execution of PVC in our case:

1.  Given that the M?ller-G?rtner (MG) method assumes constant WM signals and zero CSF signals (which is more suited to PET studies), it seems this may not be optimal for us. Would the RBV or Meltzer (MZ) methods be more appropriate for our DCE MRI data?

The FS version of MG (MGX) does not assume that CSF is 0. Don't use MZ. RBV may be ok; I have not worked with it that much.

2.  Should PVC be applied to the 4D dynamic DCE data before dynamic modeling, or directly to the 3D Ktrans map after modeling?

Probably to the raw data (that is how we do it in PET)

3.  Does the point spread function (PSF) of the MRI scanner cause more PVE than the TFE?

Probably not. For MRI, I would probably just set the psf to 0

Additionally, I have one more question about the outputs of the MG PVC method. Could you clarify the role of mgx.gm in the results? Does it the combination of mgx.ctxgm and mgx.subctxgm?

Yes, just there for convenience

Any help is appreciated. Many thanks! Shen

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer