Medulloblastomas represent the most common malignant brain tumors of childhood. Recently, we have utilized DNA oligonucleotide microarray gene expression analysis to define the molecular profile of this tumor type. Medulloblastomas were found to be molecularly distinct from other CNS embryonal tumors, and their histological subtypes had unique and functionally significant gene expression patterns. Gene expression profiles were highly predictive of response to therapy, predicting survival in a retrospective analysis with much greater accuracy than current clinical staging criteria or single marker gene outcome predictors. These results must now be prospectively validated before molecular markers can be used for risk stratification in future clinical trials. The experiments of this proposal are correlative biological studies of Phase III medulloblastoma therapy trials of the Children's Oncology Group (COG), designed to optimize and validate prognostic molecular markers. Initially, a multi-molecular outcome predictor, based on gene expression profiles of tumors from 500 children treated in COG medulloblastoma trials over the next 5 years, will be optimized, validated and developed for "real time" analysis in the context of future clinical trials. This gene expression database will be used to develop a molecular taxonomy, identifying subclasses of medulloblastomas defined by the expression profiles of molecular signaling pathways that promote tumorigenesis. Gene expression will be linked to oncogenic genomic mutations by combining expression profiling with genome-wide mutation analysis obtained from DNA BAC clone microarray-based comparative genomic hybridization. In achieving these aims, we will create a comprehensive medulloblastoma gene expression and mutation database that is linked to clinical outcome of a large and well-characterized cohort of children. The data will be hosted on the website of the NCI/NINDS Glioma Molecular Diagnostic Initiative that is being developed by Howard Fine of the NCI/NINDS Neuro-Oncology Branch and the NIH Bioinformatics group so that investigators throughout the world can have free access to the molecular and clinical database generated by this project.
