Radical surgical resection of newly diagnosed glioblastoma multiforme (GBM) and anaplastic astrocytoma (AA) in children is the most powerful favorable predictor of outcome when followed by radiation therapy and chemotherapy. In the largest study of childhood malignant gliomas (Children's Cancer Group Study, CCG-945), which was conducted between 1985 and 1992, a radical surgical resection was defined as greater than 90% resection of tumor as seen on imaging studies, predominantly using MRI. Of note is that the training of the neurosurgeon (i.e. in pediatric versus adult neurosurgery) had a significant impact on the extent of surgical resection in patients enrolled on this study. In children with recurrent malignant glial tumors, radical surgical resection has been shown to predict a more favorable survival for children, both with GBM and AA, undergoing high-dose (marrow-ablative) chemotherapy with hematopoietic stem cell rescue. In these studies, radical surgical resection was defined as resection leaving less than 3 cm maximal diameter of enhancing tumor mass in place.