Editorial: Brain tumors

Abstract

For over a century physicians have tried to use what they see under the microscope to predict the behavior of brain tumors and to determine which treatment strategies would best benefit their patients. This approach has been particularly frustrating in pediatric neurooncology. What appears to be the same tumor in an infant may behave quite differently in an older child. That “same tumor” may have a very different prognosis in one anatomical location compared to another. A commonality in most pediatric brain tumor patients remains that what we accomplish as neurosurgeons in our attempts to remove a child’s tumor is presently the most important factor in that child’s prognosis. For most pediatric brain tumors, the difference between a grosstotal resection and a subtotal resection doubles the child’s chances of survival. As such, advances in neuroimaging and image-guided surgery are among the greatest adjuncts to surgery. The application of frameless stereotactic image guidance, functional neuroimaging, intraoperative MR imaging, early postoperative MR imaging, and second-look surgery has allowed the neurosurgeon to be more aggressive. This is evidenced by the fact that the number of children enrolled in national clinical trials with no or minimal residual tumor on their MR imaging has nearly doubled over the last 2 decades. Another major advance has been the routine banking of tumor tissue. This resource, coupled with more readily available molecular profiling techniques, has allowed us to understand that two ependymomas that look the same histologically are not the same tumor in the spinal cord as in the posterior fossa or the supratentorial space.8 The understanding of the molecular events giving rise to these tumors has allowed us to develop animal and in vitro models that are both histologically and molecularly the same as those in the human condition.4 The same holds true for small round blue-cell tumors. The recognition that molecular alterations in the INI-1 gene are associated with atypical teratoid rhabdoid tumors was a seminal occurrence.7 We now recognize that what we once called a medulloblastoma is one of several tumor subtypes.1 The Wnt-positive tumors, recognized by nuclear staining for b-catenin, are more likely to arise in younger children, are more likely to be located within the fourth ventricle, and in our last treatment trial were associated with 100% progression-free survival.2 Efforts are now underway to reduce the toxic side effects of treatment in this subgroup of children. Medulloblastomas associated with Sonic hedgehog (Shh) signaling abnormalities arise from a different cancer stem cell, are more likely to be located in the cerebellar hemisphere, and are more likely to be noted in infants and older children.3 With the development of GDC-0449, a specific Shh inhibitor, we now have targeted therapy for children with those medulloblastoma subtypes. For children with c-myc amplification or with large-cell anaplastic medulloblastomas, the prognosis remains dismal and we must develop novel targeted therapies for those children.5 In low-grade astrocytomas, similar progress has been made. The identification of BRAF abnormalities in pilocytic astrocytomas, the recognition that subependymal giant cell tumors respond quite nicely to mTOR inhibitors, and that antiangiogenesis agents such as bevacizumab offer specific molecular targets for treatment in these tumors, is further evidence that the understanding of the molecular pathways involved in tumor formation is our future. In this paper6 by Dr. Pollack, a pediatric neurosurgeon who has dedicated his life to better understanding childhood brain tumors, a very clear and readable overview of the progress made in the management of a variety of childhood brain tumors has been provided. A number of truths can be gleaned from the article. First, given the molecular studies that are available for these tumors, it behooves us at surgery not only to provide tissue for histology but also to bank fresh tumor tissue for future reference in each case. Second, children and their families should be offered enrollment in prospective clinical 133 134