With the availability of multiple positron emission tomography (PET) tracers for neurooncology, there is a need to define the appropriate tracer in a given clinical setting, and it is in this regard that we undertook this study to directly compare F-18 flurodeoxyglucose (FDG) PET and C-11 methionine (MET) PET for the evaluation of recurrence in primary brain tumors. Thirty-seven patients with a history of treated primary brain tumors referred for evaluation of recurrent disease were initially included in the study. Two patients had to be excluded because of insufficient follow-up. There were 23 males and 12 females, mean age: 33.7 ± 16.4 years; range: 5 to 65 years. All patients underwent the MET and FDG study on the same day. Visual image interpretation was performed independently by 2 PET physicians for each tracer using the plain PET and fused PET/CT images; the FDG images were evaluated first. Images were analyzed semiquantitatively using tumor to normal contralateral cortex ratios (T/N). Each patient was followed up for a minimum of 18 months. Imaging results were compared with histopathology on tumor excision or biopsy in 14 patients and with clinical follow-up and MRI/MRS at the end of 18 months in 21 patients. The final diagnosis was tumor recurrence in 24 patients and no recurrence/stable disease in 11 patients. On FDG, findings in 15/35 (42%) were suggestive of recurrent tumors. On MET, findings in 24/34 (70.5%) cases were suggestive of recurrent tumors. Spatially separated secondary lesions including intraventricular deposits were clearly delineated in 5 cases, 3 were glioblastoma multiforme (GBM) and 2 were anaplastic astrocytomas. One of the secondary lesions was missed on FDG PET. Using a cutoff for T/N ratio on FDG of >0.75 to differentiate recurrence from no recurrence, sensitivity of FDG was 81.2% (confidence interval [CI] = 54.4%-96%), whereas specificity was 88.9% (CI = 51.8%-99.7%). Area under the curve was 0.819 (CI = 0.615-0.943), P = 0.0003. Using a cutoff for T/N ratio of >1.9 to differentiate recurrence from no recurrence, sensitivity of MET was 94.7% (CI = 74.0%-99.9%), whereas specificity was 88.89% (CI = 51.8%-99.7%). Area under the curve was 0.942 (CI = 0.785-0.995), P < 0.0001. Interobserver agreement, κ coefficient, for MET was 0.93, suggesting good interobserver agreement, whereas for FDG, it was fair (0.23). MET should be the radiotracer of choice in the evaluation of recurrence of primary brain tumors because the sensitivity for detection and delineation of the possible recurrent tumor, as well as secondary deposits, is higher with MET. MET-PET is an easier technique to interpret, irrespective of the glioma grade, with less interobserver variability and straightforward localization of tumorous accumulation.
Read full abstract