A nomogram model based on CGGA database to predict the prognosis of glioblastoma patients

GUO Shewei; ZHEN Yingwei; LIN Fa

Abstract：Objective To predict the prognosis of glioblastoma(GBM)patients based on the analysis of the Chinese Glioma Genome Atlas(CGGA)database data.Methods The whole gene transcriptome_693(mRNAseq_693)sequencing data and clinical data from CGGA database were acquired and downloaded.LASSO-COX regression was utilized to analyze the influencing factors and related functional genes.Patients were divided into the high-risk group and the low-risk group according to the risk score.Functional annotation of related genes as well as Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)analysis were carried out.A nomogram prediction model was built to calibrate the efficacy of curves in evaluating the nomogram.Results Amidst the 249 patients with GBM,140 were primary while 109 were recurrent.237 genes were identified as elevated(screening criteria:r/p>2 and P<0.005)and 119 genes as decreased(screening criteria:r/p<0.5 and P<0.005)from 23974 genes.9 candidate genes(CPLX1,DPT,FAM225B,GNAS_AS1,GRM5,HAR1B,NGFR,RP11_403I13.4 and RP11_498C9.13)were selected by LASSO-COX analysis of dimensionality reduction,and the RiskSocre of each patient was calculated according to their corresponding λ values with the median RiskSocre critical value being 0.013.With the increase of RiskSocre score,common characteristics of asymmetric distribution were displayed in PRS,IDH mutation state,1p19q co-deletion state and 9 candidate genes.The proportion of recurrence,IDH wild type and 1p19q non-deletion type in the high-risk group was upregulated compared with those in the low-risk group(P<0.05).Kaplan-Meier curve showed that the OS of high-risk patients was lower than that of low-risk patients(P<0.05).The RiskSocre of patients with recurrent GBM significantly increased while the RiskSocre of female patients slightly increased,and the RiskSocre was not correlated with age,postoperative radiotherapy,postoperative chemotherapy,IDH mutation status,1p19q co-deletion status,and the methylation status of MGMTp promoter.GO analysis indicated that the biological processes most relevant to the RiskSocre include immune response,immune response regulation,and T cell co-stimulation.KEGG analysis showed that the most correlative signal pathway with the RiskSocre was T cell receptor signal pathway,including cell adhesion molecule as well as antigen processing and presentation.The probability of tumor recurrence in GBM patients(1 year,2 years,3 years,5 years,and 9 years)could be estimated using individualized prediction models,and the nomogram in the calibration curve demonstrated satisfactory overlap with the actual observation results,indicating favorable prediction accuracy.Conclusions Establishing a nomogram model to predict the 1-year,2-year,3-year,5-year,and 9-year survival of GBM patients could serve as a potential tool for clinical guidance of postoperative treatment.