Prediction of pathology and survival by FDG PET in gliomas

Objectives: Despite being in use for nearly two decades, the utility of [18F]2-fluoro-2deoxy-D-glucose positron emission tomography (FDG PET) in the evaluation and treatment of brain tumors remains controversial. We retrospectively analyzed all patients with histologically proven gliomas, between the years 1990 and 2000, who underwent FDG PET studies at various stages of their treatment and who were followed till either death or for a minimum period of 1 year in an attempt to bring resolution to this controversy. Methods: All PET scans prior to 1997 were acquired on an ECAT 951/31 scanner in 2D. Scans since 1997 were obtained on a Siemens HR+ scanner in 3D mode. The majority of FDG PET scans were co-registered with the magnetic resonance imaging (MRI) scans to aid in diagnosis and therapy. Based on independent visual inspection, two board certified nuclear medicine physicians graded the highest activity level of the tumor using the metabolic grading: 0=no uptake; 1=uptake less or equal to normal white matter; 2=uptake greater than normal white matter and less than gray matter; 3=uptake equal to or greater than gray mater. The measure of association of lambda (lambda) was used to measure the strength of predictive ability of FDG PET for pathological grading of the gliomas. The Cox proportional hazards regression model was used to assess the significance of grade of uptake on survival. Results: A total of 331 patients were analyzed of which 137 had a PET scan prior to histological diagnosis and therapeutic intervention (mean age=46.5 years; M:F=1.7:1). Eighty six percent (143/166) of the patients with low uptake (metabolic scores 0,1) had low-grade gliomas (grade I, II) and 14% (23/166) high-grade gliomas (grade III, IV) on histologic examination. Ninety four percent (154/165) of the patients with high uptake (metabolic scores 2,3) on PET had high-grade gliomas and 7% (11/165) had low-grade gliomas on histologic examination. The grade of uptake had increasing significance on survival as the level increased from 'low' to 'high' (P=0.0009). Ninety four percent (156/166) of the patients with low uptake survived for >1 year (median survival of 28 months) and 19% survived for >5 years. Only 29% (48/165) of patients with high uptake survived for >1 year, (median survival of 11 months) and none survived for >5 years. Irrespective of when the scan showed a high uptake of FDG, before or after intervention, the prognosis following that scan was poor. Conclusions: Our observations confirm the utility of FDG PET as a prognostic tool for the histological grading and survival in patients with gliomas and appears to more than complement pathological grading.