Molecular Classification of Gliomas

Abstract

Molecular markers have been intensively explored to overcome the limitations in the histopathological diagnosis of gliomas. Gene expression profiling, i.e., genome wide analysis of gene expression, has given rise to new molecular classification schemes. In particular, diagnostic systems for differential diagnosis of anaplastic oligodendroglioma and glioblastoma, or for prediction of prognosis for displastic astrocytoma, anaplastic astrocytoma, anaplastic oligodendroglioma and glioblastoma have been constructed through studies employing machine learning. Classification by gene expression profiling has also revealed molecular classes with distinct biological characteristics not detected by histopathology. The promoter methylation of the O6-methylguanine methyltransferase gene has been reported as prognostic as well as predictive for alkylating agents such as temozolomide in glioblastoma. Partly due to technical difficulties in detection of methylation with PCR, the results of studies are not necessarily consistent. However, a recent study with bisulfite sequencing revealed good prognostic ability, which promises future clinical application. Among other molecular markers, 1p-/19q- has been established as a prognostic factor in oligodendroglial tumors, and is being used as a diagnostic test in several institutes. IDH1 and EGFR are being explored for differentiation of primary and secondary glioblastoma, and as a possible predictive factor for molecular targeted drugs, respectively. Adequate prospective studies will help evaluate the ability of the above classification schemes as diagnostics tests to support histopathological diagnoses.