Positron emission tomography (PET) with rubidium-82 (82Rb) and fluorine-18-fluorodeoxyglucose (18F-FDG) was used to diagnose active tumor recurrence and to differentiate this from radiation injury after interstitial irradiation of malignant gliomas. Patients were studied when they presented with radiological or clinical deterioration after an initial period of posttreatment stabilization. Forty studies were performed in 34 patients. The 82Rb was used as a blood-brain barrier tracer to localize the lesion. Uptake of 18F-FDG by the lesion was then compared to uptake by adjacent brain in order to make a diagnosis of active tumor recurrence (higher or equal lesion uptake) or no active tumor (lower uptake). Radiation injury was diagnosed by the exclusion of active tumor. A retrospective clinical diagnosis was established in 38 cases by following the patients' progress for 8 to 142 weeks after the PET study. In two cases, no follow-up diagnosis could be determined. The PET results agreed with the follow-up diagnosis in 15 of 17 cases of active tumor and 17 of 21 cases of radiation injury. Histological examination of surgically resected tissue obtained after the PET study was performed in 18 patients (nine with tumor regrowth and nine with radiation injury). This showed apparently viable tumor as well as necrosis in all cases, regardless of eventual clinical outcome. Some cells from the irradiated volume may appear morphologically intact, but have little or no metabolic or clinical activity. The functional nature of the PET-FDG technique allows diagnosis of tumor activity, which cannot be demonstrated by anatomic imaging studies or by histological examination. The addition of a blood-brain barrier tracer to the 18F-FDG study aids in differentiating normal brain uptake from tumor activity and improves the accuracy of the technique.