Effect of serum blood glucose on SUV measurements: implications in quantitative brain imaging and implementation of PERCIST criteria

Objectives Blood glucose is routinely measured prior to FDG administration in positron emission tomography (PET) imaging to identify hyperglycemia that may affect image quality . However, the effect of glucose levels upon quantitative standardized uptake value (SUV) measurements over a range otherwise considered acceptable for clinical PET imaging has not been extensively explored. Changes in target organ/tissue of interest FDG uptake based upon blood glucose may have significant implications in quantitative alogrithms used in the interpretation of PET imaging in dementia and in monitoring tumor treatment response with PET Response Evaluation Criteria in Solid Tumors (PERCIST) criteria.

Methods 436 FDG PET/CT studies from 230 patients obtained from the Ann Arbor VA were reviewed. All patients had blood glucose obtained prior to injection of FDG. SUV measurements were obtained of the liver (following PERCIST criteria), aortic arch blood pool, spleen, and the basal ganglia and overall brain. Additionally, patient demographic data was obtained to analyzed to explore confounding factors.

Results A negative inverse relationship of brain SUV with increasing serum glucose, both in terms of absolute and normalized (either to blood pool or liver) measurements was demonstrated. At higher blood glucose levels (> ~150 mg/dL), brain SUV approached serum blood pool. By contrast, there was a minimal significant dependence at higher blood glucose levels in both the liver and blood pool SUV. Spleen SUV was independent of blood glucose, but had the highest variability in measurements (deviation from linear regression). Normalized liver and spleen SUV to the blood pool SUV did not depend on blood glucose at time of injection.

Conclusions An inverse relationship between brain SUV and blood glucose levels otherwise considered acceptable for clinical PET imaging may affect PET scan interpretation, particularly when quantitative analyses are employed. Implementation of PERCIST criteria in assessing treatment responses may require modification to consider the effect of blood glucose upon both target organs and selection of reference tissues.

• Download figure
• Open in new tab
• Download powerpoint