Glial differentiation in medulloblastomas. Histogenetic insight, glial reaction, or invasion of brain?

The histogenesis of the medulloblastoma has been controversial since the original description of the tumor over 50 years ago. Recent debates have focused upon whether the tumor differentiates along glial or neuronal lines. In order to address this question, we reviewed the clinical histories, histopathologic features, and immunocytochemical staining for glial fibrillary acidic protein (GFAP) in 20 patients seen at our hospital from 1970 to 1980. Primary tumors were all included. Five of the 20 primary cerebellar tumors contained weakly GFAP-positive foci. Local recurrences occurred after treatment in seven patients and extracranial metastases in two, all of which were GFAP-negative. Interpretation of GFAP positivity was problematic. In most instances it occurred in areas where astrocytes might predictably be found: at the interface of tumor and adjacent brain and around blood vessels. Only rare single positive cells were identified in the midst of tumor nodules where cytologic distinction between neoplastic and reactive astrocytic cells was difficult. We conclude that GFAP postivity may not be intrinsic to medulloblastomas and must be interpreted with a degree of skepticism.