The aim of the study was to evaluate the effects of elevated plasma glucose levels on tumor detection.
Methods: One-hundred and seventy-one fasted patients (100 malignant pancreatic tumors, 46 chronic pancreatitis and 25 patients with other benign pancreatic lesions) were studied with 18F-fluorodeoxyglucose (FDG) PET before planned resective pancreatic surgery. Nineteen of 171 patients had elevated plasma glucose levels above 130 mg/dl, and 24 of 171 had diabetes mellitus. Standard uptake values (SUVs) with and without glucose correction, tumor-to-muscle ratios and tumor-to-liver ratios were measured of the pancreatic lesion respective of the area with the highest uptake within the pancreas. The original qualitative PET reports concerning the dignity of the pancreatic lesion were translated into a five-point malignancy scale. Tumor detection rates and SUVs were compared according to plasma glucose levels above and below 130 mg/dl, the presence of diabetes and by using receiver operating characteristic (ROC) analysis.
Results: The detection rates (and mean SUVs) for pancreatic malignancies were 86% and 42% (4.2 and 2.3) if fasted plasma glucose levels were below and above 130 mg/dl, respectively. The sensitivities (and mean SUVs of malignant tumors) were 83% and 69% (3.3 and 2.5) for patients without and with known diabetes. Areas under ROC curves were nearly equal for glucose corrected SUV and visual qualitative results (0.86 and 0.85), followed by uncorrected SUV (0.83), tumor-to-liver ratios (0.80) and tumor-to-muscle ratios (0.79). SUVs for chronic pancreatitis, muscle and liver had a tendency to increase with elevated plasma glucose levels.
Conclusion: Negative PET results of patients with elevated plasma glucose should be interpreted with caution.