TCGA Glioblastoma Multiforme

Overview

The TCGA GBM study is the canonical genomic and transcriptomic dataset for glioblastoma multiforme, assembled by The Cancer Genome Atlas Research Network. It comprises multi-platform molecular profiling of glioblastoma tumours and is hosted on cBioPortal as gbm_tcga. The companion TCIA imaging collection (tcia-tcga-gbm) provides matched multi-parametric MRI scans with expert-revised tumour segmentation labels and extracted radiomic features published by Bakas et al. 2017. PMID:28872634

Composition

  • Cancer type: glioblastoma multiforme (GB), n=592.
  • Assays: whole-exome sequencing, bulk RNA-seq, copy-number arrays, methylation arrays, and reverse-phase protein arrays.
  • Sample Selection: Original interim analysis (n=206) required a minimum of 80% tumor nuclei and maximum 50% necrosis. 143 cases had matched normal peripheral blood DNA. PMID:18772890
  • Molecular counterpart: paired with TCIA-hosted multi-parametric MRI scans (T1, T1-Gd, T2, T2-FLAIR) in tcia-tcga-gbm (135 scans from 8 institutions).

Assays / panels (linked)

  • Whole-exome sequencing; bulk RNA-seq; copy-number arrays.

Papers using this cohort

  • PMID:28872634 — Bakas et al. 2017, Scientific Data: Expert-revised MRI segmentation labels and >700 radiomic features released for 135 TCGA-GBM cases, complementing the genomic data in gbm_tcga to enable radiogenomic studies.
  • PMID:26824661 — Ceccarelli et al. 2016, Cell: TCGA pan-glioma integrated analysis using 606 GBM samples from this cohort combined with 516 LGG samples into lgggbm_tcga_pub.

Notable findings derived from this cohort

  • Identified three core signaling pathways altered in nearly all GBMs: RTK/RAS/PI3K (88%), p53 (78%), and RB (88%). PMID:18772890
  • Characterized the spectrum of somatic mutations, identifying significantly mutated genes including TP53, PTEN, NF1, EGFR, ERBB2, RB1, PIK3R1, and PIK3CA. PMID:18772890
  • Discovered frequent activating mutations in PIK3R1 (p85α regulatory subunit of PI3K) that disrupt interaction with p110α. PMID:18772890
  • Established that MGMT promoter methylation, when combined with mismatch repair (MMR) gene mutations (e.g., MSH6), is associated with a hypermutator phenotype in GBMs treated with alkylating agents. PMID:18772890
  • Bakas et al. 2017 released expert-revised segmentation labels for 135 GBM cases (from the originating set of 262), achieving median DICE of 0.92 for whole tumour, 0.88 for tumour core, and 0.88 for enhancing tumour versus GLISTRboost automated labels. These labels became the reference for BraTS’17. PMID:28872634
  • A panel of >700 radiomic features was extracted volumetrically from the manually revised labels, spanning intensity, volumetric, morphologic, texture (GLCM/GLRLM/GLSZM/NGTDM), wavelet, and glioma growth-model parameters — enabling downstream radiogenomic correlation with TCGA molecular profiles. PMID:28872634
  • Used as a component cohort (606 GBM samples, grade IV, n=592 per cBioPortal) in the TCGA pan-glioma integrated analysis identifying 75 SMGs, six DNA-methylation subtypes (LGm1–6), and PA-like IDH-wildtype glioma as a distinct favorable-prognosis entity PMID:26824661

Sources

  • cBioPortal study: gbm_tcga — https://www.cbioportal.org/study/summary?id=gbm_tcga
  • TCGA GBM marker paper: Cancer Genome Atlas Research Network 2008, Nature (PMID 18772890).
  • PMID:28872634 — Bakas et al. 2017, Scientific Data, DOI 10.1038/sdata.2017.117.
  • PMID:26824661 — Ceccarelli et al. 2016, Cell, DOI 10.1016/j.cell.2015.12.028.

This page was processed by entity-page-writer on 2026-05-14.