Purpose H3 K27 mutations, most commonly in H3F3A, are common in diffuse midline glioma. The exact frequency of these mutations in adults with gliomas in the midline location is unknown. This study was conducted to define the incidence of H3 K27M mutations in this location and to compare clinicopathological features with those of patients who do not harbor this mutation. Methods Consecutive glioma cases from 2007 to 2017 were screened for gliomas in the midline location. Immunohistochemistry was performed on all available tissue for mutations of H3 K27M, IDH1, and ARTX. Results Of 850 gliomas screened, 163 cases had midline glioma on MRI. Sufficient FFPE tissue was available for 123 cases (75%). H3 K27M mutation was identified in 18 of 123 cases (15%). All except one H3 K27M-mutant tumors were WHO grade III or IV on histology, while non-mutant tumors encompassed all four grades. The most common midline locations for H3 K27M-mutated tumors were midbrain (2/3; 67%), pons (4/11; 36%), and cerebellum (6/24; 25%). As compared to H3 K27M-wildtype tumors, there were no differences in age at diagnosis, sex, tumor grade, contrast enhancement on MRI, extent of resection, or treatment received. In this cohort, median survival was longer for patients with H3 K27M-mutated tumors (n = 18; 17.6 months) compared with high-grade wildtype tumors (n = 74; 7.7 months, p = 0.03). Conclusions H3 K27M mutations are common in midline gliomas in adults and can present in all midline locations. Survival comparison between H3 K27M-mutant and wildtype midline gliomas suggests that survival may be similar or possibly improved if the mutation is present.

Gliomatosis cerebri (GC) is a rare, extensively infiltrating glioma involving multiple contiguous lobes of the brain. This lethal disease affects all age groups, and the majority of patients have a poor outcome despite aggressive treatment. Despite its initial recognition in 1938, GC remains a controversial entity with little consensus in its definition, histology, or treatment. The majority of GC tumors are astrocytic, although mixed phenotypes have been identified. Treatment of GC is challenging as surgery is generally not an option due to the extensive areas of brain involved, the benefit of radiation therapy is unclear, and no chemotherapy has proven efficacy. Due to the rarity of the disease and its heterogeneity, both at histopathological and molecular levels, it is difficult to conduct clinical trials tailored for this diagnosis. This review summarizes our current knowledge, examines clinical studies focusing on the treatment of GC, highlights ongoing challenges, and discusses the recent molecular insights into adult and pediatric GC. We conclude that, although no longer recognized as a distinct pathological entity, GC represents a unique disease phenotype. Given the histologic and molecular overlap with other diffuse gliomas, the research emphasis should be on investigating its unique invasive biology.