Large-Scale Analysis Reveals Gene Signature for Survival Prediction in Primary Glioblastoma
- PDF / 1,470,479 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 14 Downloads / 154 Views
Large-Scale Analysis Reveals Gene Signature for Survival Prediction in Primary Glioblastoma Birbal Prasad 1 & Yongji Tian 2 & Xinzhong Li 1 Received: 30 March 2020 / Accepted: 24 August 2020 # The Author(s) 2020
Abstract Glioblastoma multiforme (GBM) is the most aggressive and common primary central nervous system tumour. Despite extensive therapy, GBM patients usually have poor prognosis with a median survival of 12–15 months. Novel molecular biomarkers that can improve survival prediction and help with treatment strategies are still urgently required. Here we aimed to robustly identify a gene signature panel for improved survival prediction in primary GBM patients. We identified 2166 differentially expressed genes (DEGs) using meta-analysis of microarray datasets comprising of 955 samples (biggest primary GBM cohort for such studies as per our knowledge) and 3368 DEGs from RNA-seq dataset with 165 samples. Based on the 1443 common DEGs, using univariate Cox and least absolute shrinkage and selection operator (LASSO) with multivariate Cox regression, we identified a survival associated 4-gene signature panel including IGFBP2, PTPRN, STEAP2 and SLC39A10 and thereafter established a risk score model that performed well in survival prediction. High-risk group patients had significantly poorer survival as compared with those in the low-risk group (AUC = 0.766 for 1-year prediction). Multivariate analysis demonstrated that predictive value of the 4-gene signature panel was independent of other clinical and pathological features and hence is a potential prognostic biomarker. More importantly, we validated this signature in three independent GBM cohorts to test its generality. In conclusion, our integrated analysis using meta-analysis approach maximizes the use of the available gene expression data and robustly identified a 4-gene panel for predicting survival in primary GBM. Keywords Glioblastoma . Prognosis . Biomarker . Survival analysis . Meta-analysis
Introduction Globally, there were about 330000 incident cases of central nervous system (CNS) cancers with a significant increase in age-standardized incidence rate (17.3%) between 1990 and 2016. However, there was no significant change in ageElectronic supplementary material The online version of this article (https://doi.org/10.1007/s12035-020-02088-w) contains supplementary material, which is available to authorized users. * Xinzhong Li [email protected] Birbal Prasad [email protected] Yongji Tian [email protected] 1
National Horizons Centre, School of Health and Life Sciences, Teesside University, Darlington DL1 1HG, UK
2
Beijing Tiantan Hospital, Capital Medical University, Beijing 100070, People’s Republic of China
standardized death rate (2.2%) globally between 1990 and 2016 when about 227,000 deaths were reported due to CNS cancers [1]. In particular, CNS cancer incidence was about 5053 in the UK in 2016 with a 21.6% change in agestandardized incidence rates between 1990 and 2016 [1]. Among these cancers, brain tumour incidence rates in the UK are expected
Data Loading...
