Construction of a Prognostic Model for Liver Cancer Patients Based on Basal Membrane Genes Associated with Immune Checkpoint Genes,the Related Functions and Potential Drugs

TAN Tianhua; XU Jingyong; SONG Jinghai

doi:10.3969/j.issn.1002-3429.2024.11.006

Abstract：Objective To explore the feasibility of constructing a prognostic model of liver cancer patients based on the basal membrane genes associated with immune checkpoint genes and to analyze their correlation and differences in drug sensitivity.Methods Large-scale data of liver cancer patients were obtained from the Cancer Genome Atlas(TCGA)database,including transcriptome data,clinical characteristic data and mutation data.By integrating these data,a prognostic model was constructed and basement membrane genes associated with immune checkpoint genes were selected as independent prog-nostic proteins.The independence of the model was verified by independent prognostic analysis.Immunocorrelation analyses were also performed to assess differential gene and immune-related functions between high-risk and low-risk patients.Finally,drug sensitivity analysis was performed to evaluate the difference in sensitivity of patients in different risk groups to common-ly used drugs.Results According to multivariate Cox regression analysis,the prognostic model formula was obtained:Risk score=(1.25968690844101×ADAMTS5 expression)+(0.332985729856301×CTSA expression)+(-0.31683170952125×FBLN5 expression)+(0.346521502281786×P3H1 expression).Immunocorrelation analysis revealed differential genes and immune-related functions between high-risk and low-risk groups,and differential gens included Cytolytic_activity,Type_Ⅰ_IFN_Reponse,Type_Ⅱ_IFN_Reponse and MHC_class_Ⅰ.In univariate prognostic analysis,HR of risk score and 95％CI were 1.201 and 1.132,1.274,respectively(P<0.01).In multivariate prognostic analysis,HR of risk score and 95％CI were 1.174 and 1.102,1.251,respectively(P<0.01).The receiver operating characteristic(ROC)curve showed that the area un-der the ROC curve(AUC)of 1,3 and 5 years was greater than 0.5.The consistency index(C-index)curve showed that the con-structed model had the largest C-index value.Drug sensitivity analysis found that there were significant differences in the sensi-tivity of Cyclopa-mine,Epothilone B,Foretinib,Imatinib,Midostaurin and Obatoclax,Mesylate,Paclitaxel,Pyrimethamine,Sorafenib,Thapsi-gargin,Tipifarnib and Vinorelbine in high-risk and low-risk groups.Conclusion The prognostic model based on the basal membrane genes associated with immune checkpoint genes has potential predictive value for the prognosis of liver cancer.The immunocorrelation analysis and drug sensitivity analysis can provide information on the immune status and individualized treatment of liver cancer,which can provide new tools and guidance for prognostic assessment and treatment de-cision-making of liver cancer,and help to improve the survival rate and quality of life of patients.