Sources of Variability in Normative Cerebral FDG-PET imaging

The variability in normative cerebral FDG-PET imaging is influenced by several factors, including scanner manufacturer, reference region, age, and sex. A study analyzed the impact of these variables on regional [18F]FDG uptake in cognitively normal adults from the Alzheimer's Disease Neuroimaging Initiative (ADNI).

The research utilized three reference regions - whole cerebellum, pons, and cortical gray matter - to calculate regional SUVr values, which were then converted to Z-scores. Linear regression models were employed to estimate standardized regression coefficients (beta), and 10-fold cross-validation was performed to determine the predictive contribution of each covariate.

Results showed that scanner manufacturer had a significant impact, with Siemens systems yielding higher Z-scores compared to GE and Philips scanners. The difference in Z-scores in specific regions, such as frontal and medial temporal areas, approached one standard deviation. Age also showed region-specific associations, with the strongest biological predictor in cross-validation (R2 approximately 0.11), particularly affecting subcortical nuclei, medial temporal structures, and the posterior cingulate cortex. Sex effects, on the other hand, were found to be negligible (R2 approximately 0).