Medulloblastoma: a potpourri of distinct entities

Abstract

In the recent years it has become more and more evident that small round blue cell tumors of the cerebellum, which we all used to call or still call medulloblastoma, may in fact be as biologically distinct as different gastro-intestinal cancers, e.g., liver cancer, gastric cancer, esophageal cancer, and colon cancer, which are of course considered different entities and treated as such. This issue of Acta Neuropathologica hosts a series of seminal papers exclusively aiming to further characterize the distinctive biology of medulloblastoma subgroups. The first unambiguous entity that was separated from medulloblastoma about 10 years ago, based on its characteristic morphology and predominance in infants, but mainly by the quasi-defining loss of INI-1 protein expression (encoded by the SMARCB1 gene), was termed atypical teratoid/rhabdoid tumor (AT/RT) [1]. That this distinction was useful may best be illustrated by the fact that while these patients universally died within a few years after diagnosis when treated with medulloblastoma regimens, survival rates could be considerably improved by treating them with chemotherapy protocols more similar to sarcoma treatments. Interestingly, these treatment regimens contain agents such as anthracyclins that have previously been considered not to cross the blood–brain barrier, i.e., being ineffective in brain tumors. Leading medulloblastoma groups world-wide met in Boston, MA, USA at the end of 2010 to reach a consensus on the number and characteristics of biologically, clinically, and demographically distinct medulloblastoma subgroups. The consensus is that, as revealed by gene expression profiling, there are four core subgroups, and that some of the additional subgroups identified in certain studies essentially define subsets within these four core subgroups. The result of this consensus conference, which we propose as a basis for a clinically and biologically meaningful subclassification of medulloblastoma, e.g., in the upcoming WHO classification of CNS tumors [3], is presented in the manuscript by Taylor et al. in this issue of Acta Neuropathologica. The exact clinical, demographic, and biological characteristics of the four medulloblastoma subgroups (termed WNT, SHH, Group 3 and Group 4) are further detailed in this paper, and in the meta-analysis on all published medulloblastoma gene expression profiling datasets and complementary cytogenetic information conducted by Kool et al. Since these four core subgroups can most reliably be detected by gene expression profiling, which on the other hand is impractical and prone to logistical problems in a clinical setting, Northcott et al. developed a rapid and costeffective nanoString assay (based on hybridization of unique RNA molecules to multiplexed molecularly barcoded target-specific probes) by which robust sub-grouping can be achieved even from formalin-fixed paraffinembedded (FFPE) samples, as long as they have not been stored for excessive periods. In their study on alternative splicing in medulloblastoma, Dubuc et al. nicely demonstrate that alternative splicing events also occur in a subgroup-specific manner, and that primary tumor samples can correctly be assigned to their respective subgroups by their splicing patterns— further strengthening the dissimilarity of these subgroups. Furthermore, subgroup-specific genetic events, such as focal amplification of MYC family oncogenes, may serve as independent prognostic markers by themselves as demonstrated in the paper by Ryan et al. Further specifying the S. Pfister (&) German Cancer Research Center (DKFZ) and Heidelberg University Hospital, Heidelberg, Germany e-mail: s.pfister@dkfz-heidelberg.de