Glucose and insulin variations in patients during the time course of a FDG-PET study and implications for the "glucose-corrected" SUV

Introduction: 2-Deoxy-2[¹⁸F]fluoro-d-glucose (FDG) positron emission tomography (PET) has an established role in the evaluation of cancer. Generally, tumor uptake and response to treatment are evaluated using the standardized uptake value (SUV). Some authors have proposed correcting SUV for glucose levels. Insulin is also thought to influence tumor uptake by changing uptake in other tissues. However, little attention has been paid to understanding the variability of glucose or insulin during a single PET study. Method: We studied the biological and instrumental variability of glucose and insulin measurements in 71 nondiabetic patients undergoing FDG-PET studies. Multiple glucose measurements were obtained in all 71 subjects, and in 69 of these 71 subjects, multiple serum insulin measurements were made. We determined the coefficient of observed variation (CV_ow) and the coefficient of variation attributable to biological variability (CV_bv) for both glucose and insulin. Results: The mean glucose concentration was 78.9±13.5 mg/dl. The mean insulin value was 6.49±5.92 μU/ml. The weighted mean CV_ow and CV_bv was 5.0% and 3.6%, respectively, for glucose and 14.2% and 8.3%, respectively, for insulin. Conclusions: Variations in the range of 3.6% are observed in glucose measurements during the time course of an FDG scan even after accounting for analytical error; larger variations of 8.3% are observed in insulin levels. Therefore, corrections of SUV for blood glucose, especially if obtained from single measurements, can introduce additional errors of at least this much.