Résumé : OBJECT: In this paper, the authors' goal was to evaluate the impact of PET information on brain tumor surgery in children. METHODS: Between 1995 and 2007, 442 children were referred to the authors' institution for a newly diagnosed brain lesion. Of these, 85 were studied with FDG-PET and/or L-(methyl-(11)C)-methionine -PET in cases in which MR images were unable to assist in selecting accurate biopsy targets (35 patients) or to delineate tumors for maximal resection (50 patients). In surgical cases, PET and MR images were combined in image fusion planning for stereotactic biopsies or navigation-based resections. The preoperative planning images were compared postoperatively with MR imaging and PET findings and histological data for evaluating the clinical impact on the diagnostic yield and tumor resection. RESULTS: The PET data influenced surgical decisions or procedures in all cases. The use of PET helped to better differentiate indolent from active components in complex lesions (in 12 patients); improved target selection and diagnostic yield of stereotactic biopsies without increasing the sampling; provided additional prognostic information; reduced the amount of tissue needed for biopsy sampling in brainstem lesions (in 20 cases); better delineated lesions that were poorly delineated on MR imaging and that infiltrated functional cortex (in 50 cases); significantly increased the amount of tumor tissue removed in cases in which total resection influenced survival (in 20 cases); guided resection in hypermetabolic areas (in 15 cases); improved early postoperative detection of residual tumor (in 20 cases); avoided unnecessary reoperation (in 5 cases); and supported the decision to undertake early second-look resection (in 8 cases). CONCLUSIONS: The authors found that PET has a significant impact on the surgical decisions and procedures for managing pediatric brain tumors. Further studies may demonstrate whether PET improves outcomes in children.