Dexamethasone (DEX) is frequently used in brain tumor management. This study investigated the effect of DEX treatment and plasma glucose levels on 18F-fluorodeoxyglucose (FDG) uptake in patients with malignant gliomas (16 glioblastoma, 3 anaplastic astrocytoma).
Methods: Fifteen DEX-treated patients (mean relative dose 0.23 +/- 0.15 mg(-1) x kg(-1) x day(-1), range 0.07-0.53), four patients not treated with DEX and nine healthy subjects were studied using PET and FDG. PET data obtained from tumors and the contralateral cortex were fitted to a standard two-tissue compartment model. The FDG transport and phosphorylation rates, distribution volume (DV), steady-state accumulation (Ki), deoxyglucose metabolism (R), plasma volume as well as standardized uptake values (SUVs) and tumor-to-brain ratios were determined. In addition, the tumor size was estimated from the maximal area of contrast-enhancing tumor on computed cranial tomography (CCT) scans or MRI.
Results: FDG uptake was depressed in the contralateral cortex of patients and was related to tumor size. With increasing relative DEX dose, a decrease in the DV of tumors (linear regression p = 0.021) and in the DV (p = 0.109) and plasma volume (p = 0.010) of contralateral cortex was found. R, Ki and SUVs in tumors and contralateral cortex were not related to the relative DEX dose. With increasing plasma glucose levels, differential decreases in Ki and SUVs in tumors (p = 0.057 and p = 0.733, respectively) and contralateral cortex (p = 0.001 and p = 0.029, respectively) were observed.
Conclusion: The data suggest that DEX affects FDG uptake in malignant gliomas through interaction with cerebral blood vessels and extracellular space, whereas FDG metabolism in tumors is not influenced substantially. This is of practical importance for patients having serial brain tumor imaging for treatment evaluation because patients may receive different DEX doses at different time points in the course of their disease. By contrast, the plasma glucose level must be considered a confounding variable when SUVs, tumor-to-brain ratios or Ki are used for treatment evaluation.