Context: Changes in the practice of diagnosing brain tumors are formally reflected in the evolution of the World Health Organization classification. Beyond this classification, the practice of diagnostic pathology is also changing with the availability of new tests and the introduction of new treatment options.
Objective: Glioblastomas, oligodendrogliomas, glioneuronal tumors, and primitive pediatric tumors are discussed in an exemplary way to illustrate these changes.
Data sources: Review of relevant publications through Medline database searches.
Conclusions: The example of glioblastomas shows how new predictive markers may help identify subgroups of tumors that respond to certain therapy regimens. The development of new treatment strategies also leads to different questions in the assessment of brain tumors, as seen in the example of pseudoprogression or the changes in tumor growth pattern in patients taking bevacizumab. Oligodendrogliomas illustrate how the identification of 1p/19q loss as a cytogenetic aberration aids our understanding of these tumors and changes diagnostic practice but also introduces new challenges in classification. Glioneuronal tumors are an evolving group of lesions. Besides a growing list of usually low-grade entities with well-defined morphologic features, these also include more poorly defined cases in which a component of infiltrating glioma is often associated with focal neuronal elements. The latter is biologically interesting but of uncertain clinical significance. Oligodendrogliomas and glioneuronal tumors both illustrate the importance of effective communication between the pathologist and the treating oncologist in the discussion of these patients. Finally, the discussion of primitive pediatric tumors stresses the clinical importance of the distinction between different entities, like atypical teratoid rhabdoid tumor, "central" (supratentorial) primitive neuroectodermal tumor, "peripheral" primitive neuroectodermal tumor, and medulloblastoma. In medulloblastomas, the recognition of different variants is emerging as a prognostic factor that may in the future also predict therapy responsiveness.