Prognostic stratification of adult primary glioblastoma multiforme patients based on their tumor gene amplification profiles

// María González-Tablas 1 , Inês Crespo 2, 3 , Ana Luísa Vital 2, 3 , Álvaro Otero 4 , Ana Belén Nieto 5 , Pablo Sousa 4 , María Carmen Patino-Alonso 5 , Luis Antonio Corchete 6 , Hermínio Tão 7 , Olinda Rebelo 8 , Marcos Barbosa 7, 9 , Maria Rosário Almeida 2 , Ana Filipa Guedes 2 , María Celeste Lopes 2, 3 , Pim J. French 10 , Alberto Orfao 1, 11, * and María Dolores Tabernero 1, 11, * 1 Centre for Cancer Research (CIC IBMCC-CSIC/USAL), Department of Medicine, CIBERONC, University of Salamanca, Salamanca, Spain 2 Centre for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugal 3 Faculty of Pharmacy, University of Coimbra, Coimbra, Portugal 4 Servicio de Neurocirugía, Hospital Universitario e Instituto Biosanitario de Salamanca (IBSAL), Salamanca, Spain 5 Department of Statistics, University of Salamanca, Salamanca, Spain 6 Departamento de Hematología, Hospital Universitario, IBSAL, IBMCC (USAL-CSIC), Salamanca, Spain 7 Neurosurgery Service, University Hospital of Coimbra, Coimbra, Portugal 8 Neuropathology Laboratory, Neurology Service, University Hospital of Coimbra, Coimbra, Portugal 9 Faculty of Medicine, University of Coimbra, Coimbra, Portugal 10 Department of Neurology, Erasmus MC, Rotterdam, The Netherlands 11 Instituto Biosanitario de Salamanca (IBSAL), Salamanca, Spain * Both authors have equally contributed to this work and should be considered as last authors Correspondence to: María Dolores Tabernero, email: taberner@usal.es Keywords: glioblastoma; classification; subtypes; gene amplification; survival Received: February 28, 2018      Accepted: May 14, 2018      Published: June 15, 2018 ABSTRACT Several classification systems have been proposed to address genomic heterogeneity of glioblastoma multiforme, but they either showed limited prognostic value and/or are difficult to implement in routine diagnostics. Here we propose a prognostic stratification model for these primary tumors based on tumor gene amplification profiles, that might be easily implemented in routine diagnostics, and potentially improve the patients management. Gene amplification profiles were prospectively evaluated in 80 primary glioblastoma multiforme tumors using single-nucleotide polymorphism arrays and the results obtained validated in publicly available data from 267/347 cases. Gene amplification was detected in 45% of patients, and chromosome 7p11.2 including the EGFR gene, was the most frequently amplified chromosomal region – either alone (18%) or in combination with amplification of DNA sequences in other chromosomal regions (10% of cases). Other frequently amplified DNA sequences included regions in chromosomes 12q(10%), 4q12(7%) and 1q32.1(4%). Based on their gene amplification profiles, glioblastomas were subdivided into: i) tumors with no gene amplification (55%); ii) tumors with chromosome 7p/ EGFR gene amplification (with or without amplification of other chromosomal regions) (38%); and iii) glioblastoma multiforme with a single (11%) or multiple (6%) amplified DNA sequences in chromosomal regions other than chromosome 7p. From the prognostic point of view, these amplification profiles showed a significant impact on overall survival of glioblastoma multiforme patients (p>0.001). Based on these gene amplification profiles, a risk-stratification scoring system was built for prognostic stratification of glioblastoma which might be easily implemented in routine diagnostics, and potentially contribute to improved patient management.