Novel Considerations in the Approach to Glioblastoma

In this issueof Journal of Oncology Practice, NamanddeGroot discuss current practice in the management of glioblastoma, a near uniformly fatal disease with limited treatment options. Overall survival has not increased in recent years and is estimated at 15 months. In practice, we see a range of outcomes,withpatients often fallingoutside this median. Currently, we have few tools to predict which patients will suffer an aggressive course with rapid deterioration and a disease that progresses more indolently. Identification of important prognostic factors and targetable vulnerabilities is a major focus of current research efforts. In 2010, The Cancer Genome Atlas (TCGA) identified four categories of glioblastoma based on genomic markers. Unlike TCGA findings in other solid tumors, this subdivision of glioblastomas does not afford tailored therapy. However, it describes the proneural subgroup, which harbors the 5% to 10% of glioblastomas that carry an IDHmutation. The favorable prognostic importance of the IDHmutation is emphasized in the 2016 WHO diagnostic criteria, which classify glioblastomas first on their IDH status. In contrast, not only is the presence ofMGMT promoter methylation a prognostic factor, but multiple trials have alsodemonstrated that itpredicts response to alkylating chemotherapy. The benefit of chemotherapy in MGMT unmethylated glioblastoma is sufficiently minor that many current up-front clinical trials forgo a temozolomide cohort in these patients. Because unmethylated tumors constitute approximately 60% of all new glioblastomas, a large proportion of our patient population does not receive full benefit from the current standard of care. TCGA and other studies described recurrent patterns of alterations in glioblastoma, suggesting the existenceof targetable drivermutations.Althougha survival benefit fromasmall-molecule inhibitorhasnotbeen proven in a clinical trial, there are sufficient individual reports of response to warrant continuedinvestigation. Larger studies have been confounded by the inclusion of heavily pretreated patients whose genetic data may be out of date, as well as questions surrounding optimal drug delivery. Still, there is increasing excitement surrounding the common mutations and amplifications in genes such as EGFR, CDK4, PDGFRA, and FGFR, in addition to IDH1 and IDH2.Many trials exploring these options require sequencing data to select the optimal patients for benefit. In practice, our institution and others are now making routine use of modern sequencing techniques to identify the subset of patients who may be eligible for such studies. Genetic sequencing is typically done using a panel of cancerrelated genes; some information can be obtained using fluorescent in situ hybridization or immunohistochemistry, but inconsistencies between these techniques and sequencing data highlight their limitations. Whether done in house or through a commercial company, the process takes weeks, so it must be performed before the information is needed to guide treatment at recurrence. It has become our practice to sequence tumors at diagnosis, with the understanding that the informationwill not be used until failure of initial treatment.