Brain Parcellation Implementation in PNEURO

Brain parcellation in PNEURO supports two situations:

• PET and MR: The patient had a brain PET study as well as a corresponding T₁-weighted anatomical MR study. In this case the MR-derived brain VOIs are applied for calculating statistics of brain PET uptake.
• MR only: The patient had an anatomical T₁-weighted MR study covering the entire brain. In this case the MR-derived brain VOIs are the main outcome of the analysis and might be applied for morphometric or other purposes.

The corresponding PNEURO workflows are outlined in the following.

Analysis of Study with PET and T1-weighted MRI

1. Loading of the PET image series which may be static or dynamic.
2. Dynamic case only: Averaging of the PET series in a specified range of the acquisition. The averaged PET image is used in the following for all steps except for the statistics calculation.
3. Loading of the T₁-weighted MR image series.
4. Denoising of the MR images.
5. Segmentation of gray matter (GM), white matter (WM) and the cerebrospinal fluid (CSF) and splitting of the hemispheres. This requires the interactive definition of 4 anatomical landmarks.
6. Interactive adjustment of the boundaries between the segments and definition of the number of hemispheres in the knowledge base to compare with. Parcellation of the brain structures based on the hemispheres which are most similar to the patient's hemispheres.
7. Rigid matching of the PET image to the MR image.
8. Presentation of the brain contours in a VOI editor together with the MR images, so that the user can adjust them interactively and save the final VOI set.
9. Application of the VOIs to the matched PET series for calculating statistics. This results in TACs for a dynamic PET series, and simple statistics otherwise. Optionally, a geometric partial-volume correction can be applied during the statistics calculation.
10. Dynamic case only: The resulting TACs can directly be transferred to the kinetic modeling tool or saved.

There is an alternative with respect to the recommended workflow above. Instead of defining the VOIs in the MR space, they may be transformed to the PET space, and the statistics calculated using the PET series in the original space.

Analysis of MR-only Study

If the PET information is lacking, the processing sequence reduces to the following workflow steps:

1. Loading of the T₁-weighted MR image series.
2. Denoising of the MR images.
3. Segmentation of gray matter (GM), white matter (WM) and the cerebrospinal fluid (CSF) and splitting of the hemispheres. This requires the interactive definition of 4 anatomical landmarks.
4. Interactive adjustment of the boundaries between the segments and definition of the number of hemispheres in the knowledge base to compare with. Parcellation of the brain structures based on the hemispheres which are most similar to the patient's hemispheres.
5. Presentation of the brain contours in a VOI editor together with the MR images, so that the user can adjust them interactively and save the final VOI set
6. Optional: application of the VOIs to the MR series, mainly for calculating the VOI volumes.

Documentation of the Workflows

The implementation of the two workflows described above in PNEURO is very similar. Therefore, only the first workflow with PET and a T₁-weighted MRI will be described in full detail, while the MR-only workflow is restricted to the essential parts.