Construction And Evaluation Of A Novel Prognostic Model For Glioblastoma

Objective:The purpose of this study is to construct a prognostic model based on recurrent GBM to predict the OS of GBM patients.Methods:A total of 426 GBM patients were included in this study,including 282 from the CGGA database(training group)and 144 from the TCGA database(validation group).At the same time,clinical pathological information and transcriptome sequencing data were downloaded.The LASSO-COX algorithm was used to reduce the number of predictive genes and establish a recurrent signature.K-M curves and time-dependent ROC curves were drawing to evaluate its predictive ability.The relationship between recurrent score and clinicopathological characteristics was explored.GO and KEGG analyses were performed for function annotations of the recurrent signature.A prognostic prediction model was established using independent prognostic factors,and the accuracy of the model was evaluated using a calibration curve and C-index.Results:A recurrent signature consisting of 9 key prognostic genes was established,which was significantly correlated with the OS of GBM patients;GO and KEGG enrichment analysis showed that recurrent signature is associated with Hedgehog signaling pathway and necroptosis.Recurrent signature is potential to predict immune infiltration and the efficacy of ICIs.The nomogram combines recurrent signature and clinical prognostic predictors,showing strong predictive power in both the training and validation groups.Conclusion:This study has constructed a prognostic prediction model that combines recurrent signature and clinical prognostic predictors to predict 1-,2-,3-year survival.It may serve as a potential tool to guide postoperative individualized care,and also provide a theoretical basis for possible treatment targets for GBM.