Brain partial volume correction with point spreading function reconstruction in high-resolution digital PET:...

Area of Science:

• Nuclear Medicine
• Radiology
• Neuroimaging

Background:

• Quantitative brain PET imaging is prone to errors from partial volume effects due to limited scanner resolution.
• Accurate measurement of tracer uptake, especially in small brain structures like the cerebral cortex, is crucial for diagnosis and research.

Purpose of the Study:

• To evaluate the effectiveness of point spread function (PSF) reconstruction for partial volume correction (PVC) in brain PET.
• To compare PSF reconstruction performance against MRI-based PVC using a digital PET scanner.

Main Methods:

• Ten healthy subjects underwent 18F-FDG PET scans on a digital PET/CT system and T1-weighted MRI.
• PSF reconstruction was applied with varying iterations (4-256).
• Cerebral cortical uptake was assessed using standardized uptake value ratio (SUVR) and compared between PSF reconstruction and MRI-based PVC.

Main Results:

• PSF reconstruction visualized cortical structures well, matching MR-derived segments.
• Higher iterations in PSF reconstruction increased cortical SUVRs, approaching MRI-based PVC values.
• Maximum iterations yielded SUVRs 16% lower than MRI-PVC, with an effective resolution of ~1.5-mm FWHM.

Conclusions:

• PSF reconstruction is a viable PVC technique for brain PET on modern digital scanners, despite suboptimal resolution recovery.
• PSF reconstruction offers an advantage in assessing small brain structures, including nuclei, which are challenging for MRI-based PVC.