Résumé : OBJECT: In this paper, the authors' goal was to evaluate the impact of PET data on the clinical management of incidental brain lesions in children. METHODS: Between 1995 and 2007, 442 children with a newly diagnosed brain lesion were referred to the authors' department. Of these, 55 presented with an incidental brain lesion and were selected for study because MR imaging sequences revealed limitations in assessing the tumor, its evolving nature, and/or the malignant potential of the lesion diagnosed. Thirteen children were studied using FDG-PET and 42 with L-(methyl-(11)C)-methionine (MET)-PET; 3 children underwent both FDG-PET and MET-PET but only the MET-PET results were used in the analysis. The PET and MR images were combined in image fusion navigation planning. Drawing on their experience with PET in adults, the authors proposed the following treatment plans: 1) surgery in children with imaging evidence of increased PET tracer uptake, which is highly specific of tumor and/or malignant tumor tissue; or 2) conservative treatment in children in whom there was little or no tracer uptake on PET. The authors compared the PET data with the MR imaging-based diagnosis and either 1) the results of histological examination in surgically treated cases, or 2) the long-term outcome in untreated cases. They studied PET and MR imaging sensitivity and specificity in detecting tumor and malignant tissues, and evaluated whether PET data altered their clinical management. RESULTS: Seventeen children had increased PET tracer uptake and underwent surgery. Tumor diagnosis was confirmed in all cases (that is, there were no false-positive findings). Cases in which there was little or no PET tracer uptake supported conservative treatment in 38 children. However, because PET was under evaluation, 16 of 38 lesions that were judged accessible for resection were surgically treated. Histological examination results demonstrated neither malignant nor evolving tumor tissue but yielded 9 indolent tumors (6 dysembryoplastic neuroectodermal tumors, 2 low-grade astrocytomas, and 1 low-grade astrocytoma and dysplasia) and 7 nontumoral lesions (3 cases of vasculitis, 3 of gliosis, and 1 of sarcoidosis). In 22 of the untreated 38 children, stable disease was noted during follow-up (range 18-136 months). Although an absence of PET tracer uptake might not exclude tumor tissue, PET did not reveal any false-negative findings in malignant or evolving tumor tissue detection in cases in which MR imaging showed false-positive and -negative cases in > 35 and 25% of the cases, respectively. CONCLUSIONS: These data confirmed the high sensitivity and specificity of PET to detect tumor as well as malignant tissue. Regarding the treatment of the incidental brain lesions, the PET findings enabled the authors to make more appropriate decisions regarding treatment than those made on MR imaging findings alone. Therefore, the risk of surgically treating a nontumoral lesion was reduced as well as that for conservatively managing a malignant tumor. Nowadays, it is estimated that these data justify conservative management in incidental lesions with low or absent PET tracer uptake.