The uptake of methionine in both conditions is different. As in primary gliomas, the uptake in recurrent tumor lesions is caused by active uptake of methionine in the tumor cells. On the other hand, radiation injury leads to passive diffusion across a broken blood-brain barrier. Theoretically, the uptake in tumor cells should exceed the uptake after radiation injury; however there is a possibility of overlap between this uptake.

In general, most studies agreed that MET-PET is useful in correctly identifying tumor recurrence after radiation. Li et al. recently published their findings of MET-PET in 36 postoperative patients with suspected residual/recurrent tumors, of which, after histopathologic examination and/or clinical follow-up for more than 6 months, 28 were verified having recurrent gliomas. The sensitivity, specificity, and accuracy of MET-PET were 96.4 %, 87.5 %, and 94.4 %, respectively (Li et al. 2012). Tripathi et al. scanned 24 patients with tumor recurrence and 11 patients without recurrence with MET. Using a cutoff T/N ratio of >1.9 to differentiate recurrence from no recurrence, sensitivity was 94.7 % and specificity 88.9 %. The interobserver agreement was good, with a kappa-value of 0.93 (Tripathi et al. 2012). Another study, from Grosu et al., reported a sensitivity of 91 % and a specificity of 100 % for differentiating tumor tissue and treatment-related changes in 29 patients with pretreated gliomas and 13 patients with pretreated brain metastases (Grosu et al. 2011). Okamoto et al. scanned 29 patients with 33 lesions suspected of recurrent brain tumors by MRI after radiation therapy. They found a significant higher T/N ratio in recurrence (1.98 ± 0.62) compared to necrosis (1.27 ± 0.28) and a high diagnostic value even for partial volume effect-affected small lesions (Okamoto et al. 2011). Terakawa et al. scanned in total 77 patients with MET, 51 patients who had been previously treated with radiotherapy after primary treatment for metastatic brain tumor and 26 after treatment for glioma. Best sensitivity and specificity results were found with a T/N ratio (by SUV_mean) of 1.58 for gliomas (sensitivity and specificity 75 %). For metastatic brain tumors, L/N greater than 1.41 provided best sensitivity and specificity (79 % and 75 %, respectively) (Terakawa et al. 2008). Yamane et al. asked the referring clinicians their treatment strategy before and after the MET-PET. A change in intended management by the scan results was found in 45 % of the patients (Yamane et al. 2010). Van Laere et al. scanned 30 patients with treated gliomas of whom 28 were positive, leading to a high sensitivity. Moreover, the interobserver agreement was 100 %, and MET-PET was the best prognostic predictor in a subgroup of astrocytomas (23 patients) (Van Laere et al. 2005). Earlier studies showed MET uptake in all recurrent tumors and one false-positive scan in seven patients with radionecrosis (Sonoda et al. 1998), and a case report mentioned high uptake of MET in a recurrent tumor with CT and MRI being normal (Viader et al. 1993). A study by Lilja et al., already performed in 1989, showed a larger tumor area than CT in two patients with residual tumor; however they also found nonspecific uptake in one clinically stable patient assumed to have no recurrence (Lilja et al. 1989).

Three other studies also published good sensitivity and specificity results; however they claimed that MET-PET was not better than conventional imaging. Pötzi et al. did find focally increased uptake in 24 out of 28 patients with treated glioblastoma multiforme, but showed no added prognostic information of the MET-PET, therefore in no way being superior than conventional imaging (Potzi et al. 2007). Kim et al. stated that perfusion MRI is superior in distinguishing a recurrence of high-grade glioma from radiation necrosis than MET-PET (Kim et al. 2010). Dandois found that perfusion MRI had an at least equal performance to MET-PET in differentiating tumor recurrence and necrosis in high-grade gliomas by using the values determined from regional cerebral blood flow (Dandois et al. 2010).