Diagnostic accuracy of 11C-methionine PET in detecting neuropathologically confirmed recurrent brain tumor after radiation therapy.

ABSTRACT

OBJECTIVE: This study aims to determine the diagnostic test accuracy (DTA) of 11C-methionine (MET) PET in the discrimination between recurrent tumor and radiation-induced injury in neuropathologically confirmed cases. METHODS: A retrospective cohort of 30 patients with previously irradiated intracranial tumors (23 gliomas, 6 metastases, and 1 meningioma) was included. All patients underwent a preoperative MET PET and postoperative neuropathological analysis. Maximum and mean standardized uptake values (SUV) were obtained in the lesion, in the contralateral mirror region, and in the contralateral frontal cortex. Lesion-to-background SUV ratios (SUR mirror and SUR cortex) were then calculated. The Mann-Whitney U test was used to evaluate differences in SUV ratios between confirmed recurrent tumor and radiation injury. DTA was determined through receiver operating characteristic (ROC) analysis. RESULTS: Twenty-one patients had recurrent tumor and nine had radiation injury. The area under the ROC curve (AUC) was 0.89 for SURmaxmirror and 0.88 for SURmaxcortex. The mean (SD) of SURmaxmirror was 2.37 (0.58) in tumor recurrence and 1.57 (0.40) in radiation necrosis (P ≤ 0.001). The corresponding values for SURmaxcortex were 2.13 (0.50) and 1.45 (0.37) (P = 0.001). Clinically relevant cutoffs were SURmaxmirror ≥ 1.99 giving a specificity of 100% for tumor recurrence with a sensitivity of 76% and SURmaxcortex ≥ 1.58 giving a sensitivity and specificity of 90 and 78%, respectively. CONCLUSIONS: Based on neuropathologically confirmed cases, the DTA of SURmaxmirror and SURmaxcortex from 11C-methionine PET was high when discriminating recurrent tumor from radiation injury.