A simplified approach for molecular classification of glioblastomas (GBMs): experience from a tertiary care center in...

This study aims to establish a simplified molecular classification of glioblastomas (GBMs) based on molecular genetic alterations. GBM cases (n-114) were evaluated for IDH-1 and TP53 mutation by Sanger sequencing, PDGFRA and EGFR amplification by FISH, NF1 and YKL40 expression by qRT-PCR. Subsequently they were classified into four subgroups: classical like (CL), proneural like (PN), mesenchymal like (MES) and neural like (NEU). CL subtype was most frequent (39 %), followed by PN (32 %) and MES (20 %) subtypes. PN subtype had significantly younger age at presentation and longest survival (median PFS-82.5 weeks; 1 and 2 years OS-90.6 and 71.3 %). Other three subgroups had equally poor prognosis and hence, clubbed together as non-proneural (Non-PN) (median PFS-39 weeks; 1 and 2 years OS-66 and 0 %). Hence, we recommended this relatively easy method of subclassifying GBMs into PN and Non-PN which are statistically different in prognosis (both OS and PFS on uni and multivariate analysis). Although evaluation of six molecular alterations for identifying these two subgroups is still cumbersome, we propose segregation of PN subtype alone based on assessment of IDH1, TP53 and PDGFRA status, which is relatively easy and may be amenable to routine practice.