Targeted next-generation sequencing of 565 neuro-oncology patients at UCLA: A single institution experience. Issue 2 (29th January 2020)

• Record Type: Journal Article
• Authors: Ji, Matthew S; Eldred, Blaine S C; Liu, Regina; Pianka, Sean T; Molaie, Donna; Kevan, Bryan; Pan, Stephanie; Lai, Thomas J; Nguyen, Nhung T; Chow, Frances E; Yong, William H; Cox, Christopher D; Reeh, Devin N; Li, Tie; Liau, Linda M; Nghiemphu, Phioanh L; Cloughesy, Timothy F; Li, Gang; Lai, Albert
• Is Part Of: Neuro-oncology advances. Volume 2:Issue 2(2020)
• Publication Date: 2020-01-29
• Subjects: Targeted next-generation sequencing -- CNS tumors -- genomic profiling -- glioma -- glioblastoma; 616.99481
• DOI: 10.1093/noajnl/vdaa009
• Languages: English
• ISSNs: 2632-2498

Abstract

Abstract: Background: Targeted next-generation sequencing (NGS) is frequently obtained at the University of California, Los Angeles (UCLA) for clinical characterization of CNS tumors. In this study, we describe the diagnostic reliability of the Foundation Medicine (FM) targeted NGS platform and its ability to explore and identify tumor characteristics of prognostic significance in gliomas. Methods: Neuro-oncology patients seen at UCLA who have received FM testing between August 2012 and March 2019 were included in this study, and all mutations from FM test reports were recorded. Initial tumor diagnoses and diagnostic markers found via standard clinical methods were obtained from pathology reports. With overall and progression-free survival data, the elastic net regularized Cox regression and the Cox proportional hazards models were used to determine whether any mutations of unknown significance detected by FM could predict patient outcome in GBM. Results: 603 samples tested by FM from 565 distinct patients were identified. Concordance of diagnostic markers was high between standard clinical testing methods and FM. Oligodendroglial markers detected via FM were highly correlated with 1p19q co-deletion in IDH mutated gliomas. FM testing of multiple tumor samples from the same patient demonstrated temporal and spatial mutational heterogeneity. BCORL1, ERBB4, and PALB2, mutations of unknown significance in GBM, were shown to be statistically significant in predicting patient outcome. Conclusions: In our large cohort, we found that targeted NGS can both reliably and efficiently detect important diagnostic markers in CNS tumors.