Glioblastoma gene expression subtypes and correlation with clinical, molecular and immunohistochemical characteristics in a homogenously treated cohort: GLIOCAT project.

O-6-methylguanine-DNA methyltransferase
DOI: 10.1200/jco.2019.37.15_suppl.2029 Publication Date: 2019-05-27T15:57:46Z
ABSTRACT
2029 Background: Glioblastoma (GBM) gene expression subtypes have been described in last years, data homogeneously treated patients is lacking. Methods: Clinical, molecular and immunohistochemistry (IHC) analysis from with newly diagnosed GBM standard radiochemotherapy were studied. Samples classified based on the profiles into three different (classical, mesenchymal, proneural) using Support Vector Machine (SVM), K-nearest neighbor (K-NN) single sample Gene Set Enrichment Analysis (ssGSEA) classification algorithms provided by GlioVis web application. Results: GLIOCAT Project recruited 432 6 catalan institutions, all of whom received first-line treatment (2004 -2015). Best paraffin tissue samples selected for RNAseq reliable obtained 124. 82 cases (66%) same subtype algorithms. SVM ssGEA obtain more similar results (87%). No differences clinical variables found between 3 subtypes. Proneural was enriched IDH1 mutated G-CIMP positive tumors. Mesenchymal (SVM) unmethylated MGMT tumors (p = 0.008), classical methylated 0.008). Long survivors ( > 30 months) rarely as mesenchymal (0-7.5%) frequently (23.1-26.). Clinical (age, resection, KPS) IDH1, MGMT) known prognostic factors confirmed this serie. Overall, no prognosis observed subtypes, but a trend to worse survival K-NN (9.6 vs 15 ). presented less Olig2 < 0.001) SOX2 0.003) IHC, YLK-40 0.023, SVM). On other hand, expressed Nestin 0.004) compared (K-NN). Conclusions: In our study we not correlation glioblastoma outcome. This large serie provides reproducible regarding clinical-molecular-immunohistochemistry features genetic
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (0)
CITATIONS (4)