BTC1.04 Genomic profiling of IDH-wildtype and IDH-mutant initial and matched recurrent glioblastomas reveals clinically actionable mutations (e.g. BRCA1/2) and resistance signatures. (19th September 2018)

• Record Type: Journal Article
• Authors: McInerney, C E; Ellis, H P; Schrimpf, D; Sahm, F; Stupnikov, A; Wadsley, M; Wragg, C; White, P; Prise, K M; McArt, D G; Kurian, K M
• Is Part Of: Neuro-oncology. Volume 20(2018)Supplement 3
• Page Start: iii215
• Page End: iii216
• Publication Date: 2018-09-19
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noy139.003
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Background: Glioblastomas account for 90% of all adult brain tumours. Patient survival remains low because treatments are ineffective against this lethal disease. Understanding the genes and pathways altered in subtypes is necessary for therapeutic intervention. Material and Methods: Tumours from 41 patients, 8 with matched initial and recurrent glioblastomas were genomically profiled. Some 130 clinically relevant neuro-oncology genes were assessed using an NGS-based diagnostic panel. Single nucleotide variants (SNVs), copy number variations (CNVs) and potentially clinically actionable mutations were described for initial and recurrent IDH-wildtype (n=38) and IDH-mutant glioblastomas (n=3). Results: The mutational landscape revealed discrete differences and similarities between subtypes. TSC2, MSH6, TP53, CREBBP and IDH1 were co-mutated and putatively pathogenic in both subtypes, suggesting they are driver mutations. Recurrent tumours were not hypermutated and matched analysis revealed inter-tumour heterogeneity. IDH-wildtype: SNVs (145) impacted 55 genes in the RTK/Ras/PI(3)K (82%), p53 (63%), WNT (58%), SHH (13%), NOTCH (11%), Rb (5%) and G-protein (8%) pathways. SNV burden was a predictor of overall survival ( P =0.003) but no particular pathway was individually responsible. SNVs (40) in BRAF, DAXX, EGFR, FGFR2, JAK2, MYB, PIK3CA, PIK3R1, PTEN, ATM, BRCA1, CHEK2, PPM1D, PTCH1 and SMO were also putatively pathogenic. Many initial tumours had BRCA1 (21%) and BRCA2 (18%) variants, including previously confirmed somatic mutations in haemangioblastoma, and also WNT variants (58%). Recurrent tumours had fewer pathways (RTK/Ras/PI(3)K, p53, WNT, G-protein) impacted by genetic alterations. Possible tumour resistance signatures included a private SNV in PIK3C2G and CNV gains ( BRCA2, GNAS, EGFR ) and losses ( TERT, SMARCA4 ). Survival analysis suggested GNAS variation was prognostic ( P <0.001). Recurrent tumours (57%;4/7) harboured potentially actionable variation in PTEN, BRCA1 / 2, ATR and EGFR . Combination therapies with erlotinib, everolimus or dasatinib, olaparib, ATR inhibitors and EGFR -targeting antibodies, vaccines or TK inhibitors could provide therapeutic intervention. IDH-mutant: SNVs (15) impacted 11 genes in the RTK/Ras/PI(3)K (66%), p53 (100%) and WNT pathways (33%). SNVs were also putatively pathogenic in KLK1 exclusively in this subtype. The recurrent tumour had fewer pathways (p53, WNT, G-protein) impacted by genetic alterations and a private SNV in TCF4 . Potentially actionable variation in ATR could be targeted using ATR inhibitors. Conclusion: TCGA-GBM and GDC datasets corroborated results confirming the clinical significance of findings. Combination therapies targeting the potentially clinically actionable mutations described for IDH-wildtype and IDH-mutant subtypes may hold the best promise for patient oncological management.