In:
Journal of Clinical Oncology, American Society of Clinical Oncology (ASCO), Vol. 37, No. 15_suppl ( 2019-05-20), p. 2044-2044
Abstract:
2044 Background: Glioblastoma (GBM) is the most aggressive and common primary brain tumor. Nomograms are prediction models that help form individualized risk scores for cancer patients, which are valuable for treatment decision-making. The aim of this study is to create a refined nomogram by including novel molecular variables beyond MGMT promoter methylation. Methods: Clinical data and miRNA expression data were obtained from 226 newly diagnosed GBM patients. Clinical data included age at diagnosis, sex, Karnofsky performance status (KPS), extent of resection, O6-methylguanine-DNA methyltransferase ( MGMT) promoter methylation status, IDH mutation status and overall survival. Due to low representation of less than 13 cases each, IDH mutant glioblastomas and patients submitted to biopsy-only were excluded. Total RNA was isolated from formalin-fixed paraffin-embedded (FFPE) tissues; miRNA expression was subsequently measured using the NanoString human miRNA v3a assay. A Cox regression model was developed using glmnet R package with the elastic net penalty while adjusting for known prognostic factors. A dichotomized genomic score was created by finding the optimal cutpoint (maximum association with survival) of the linear combination of the selected. A nomogram was generated using known clinical prognostic factors, specifically age, sex, KPS, and MGMT status along with the dichotomized genomic score. Results: Four novel miRNAs were found to significantly correlate with overall survival and were used to create the dichotomized miRNA genomic score (GS). This score split the cohort into a poor performing group (GS_high) and a better performing group (GS_low) (p = 0.0031). A final nomogram was created using the Cox proportional hazards model (Figure 1). Factors that correlated with improved survival included younger age, KPS 〉 70, MGMT methylation and a low genomic score. Conclusions: This study is a proof of concept demonstrating that integration of molecular variables beyond MGMT methylation improve existing nomograms to provide individualized information about patient prognosis. Future directions include a more comprehensive analysis, including proteomic and methylation data, and subsequent validation in an external cohort. Finally, network analysis integrating molecular signatures of poor performers will help identify therapeutic targets.
Type of Medium:
Online Resource
ISSN:
0732-183X
,
1527-7755
DOI:
10.1200/JCO.2019.37.15_suppl.2044
Language:
English
Publisher:
American Society of Clinical Oncology (ASCO)
Publication Date:
2019
detail.hit.zdb_id:
2005181-5
Permalink
