Generation of an Age-Dependent and Harmonized ¹⁸F-FDG Brain PET Atlas Using a High-Sensitivity Short-Axial FOV PET/CT...

Area of Science:

• Neuroimaging
• Radiochemistry
• Medical Physics

Background:

• ¹⁸F-FDG brain PET atlases are crucial for clinical and neuroscience research.
• Variations in PET scanner technology and effective image resolution (EIR) can lead to misinterpretations and quantification errors when using mismatched atlases.

Purpose of the Study:

• To develop high-resolution, age-stratified ¹⁸F-FDG brain atlases harmonized across different EIRs.
• To support the use of diverse PET/CT scanners by mitigating resolution-related discrepancies.

Main Methods:

• Retrospective enrollment of 140 adults (20-90 years) undergoing ¹⁸F-FDG PET/CT.
• Image smoothing to emulate 5.0-6.0 mm and 8.0-10.0 mm resolutions, normalization to MNI space, and calculation of Standardized Uptake Value Ratios (SUVRs).
• Creation of age-stratified and EIR-harmonized atlases, followed by analysis of gray/white matter contrast (GWC) and age correlations.

Main Results:

• Developed age-dependent, harmonized ¹⁸F-FDG PET atlases for both emulated resolutions (97 participants included after QC).
• Higher GWC observed at 5.0-6.0 mm resolution compared to 8.0-10.0 mm.
• SUVRs in cortical and subcortical regions showed negative correlation with age, with brainstem SUVRs correlating only at lower resolution; SUVRs were generally higher at higher resolution.

Conclusions:

• A novel set of age-dependent, harmonized ¹⁸F-FDG PET atlases was successfully developed.
• These atlases facilitate more accurate quantitative analysis across different PET scanner EIRs and age groups.
• The study characterized quantitative metrics across age and resolution, improving PET/CT data interpretation.