Related Experiment Video
Updated: May 17, 2026
10:31
Non-invasive Imaging and Analysis of Cerebral Ischemia in Living Rats Using Positron Emission Tomography with 18F-FDG
Published on: December 28, 2014
Anatomically guided voxel-based partial volume effect correction in brain PET: impact of MRI segmentation
Daniel Gutierrez1, Marie-Louise Montandon, Frédéric Assal
1Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital, CH-1211 Geneva, Switzerland.
Summary
Accurate brain PET imaging relies on MRI segmentation. Older SPM versions caused significant errors in partial volume correction (PVC), but SPM8 provides robust segmentation for reliable PET analysis.
Area of Science:
- Neuroimaging
- Medical Physics
- Radiology
Background:
- Partial volume effect (PVE) is a major limitation in brain PET imaging due to scanner resolution.
- Accurate PVE correction (PVC) in PET depends on precise MRI-based brain tissue segmentation (gray matter, white matter, CSF).
Purpose of the Study:
- To comparatively evaluate MRI segmentation algorithms from SPM releases (SPM99, SPM2, SPM5, SPM8).
- To assess the impact of these segmentation algorithms on PVE correction in (18)F-FDG brain PET imaging for neurodegenerative dementia patients.
Main Methods:
- Four SPM software releases' brain MRI segmentation algorithms were evaluated.
- Segmentation accuracy was assessed against clinical neurological examinations.
- PVE correction in (18)F-FDG PET was performed using MRI-derived gray matter volumes.
- SPM5 segmentation served as the reference standard.
Main Results:
- SPM99 and SPM2 significantly overestimated gray matter (~20%) and underestimated white matter (-17% to -6%).
- This led to substantial bias (30-90%) in PVE-corrected PET activity estimates.
- SPM8 demonstrated minimal differences (<1%) compared to SPM5, indicating improved accuracy.
Conclusions:
- The choice of MRI segmentation algorithm critically impacts the accuracy of MRI-guided PVC in brain PET.
- SPM8's segmentation algorithm provides robust and accurate results suitable for PVE correction in brain PET imaging.
- Accurate segmentation is essential for reliable interpretation of brain metabolism in PET studies, especially in cases of atrophy.
