目的 綜合運(yùn)用生物信息學(xué)方法進(jìn)行膠質(zhì)母細(xì)胞瘤銅死亡相關(guān)免疫檢查點(diǎn)基因（ICGs）特征分析及潛在候選藥物預(yù)測(cè)。方法 從癌癥基因組圖譜（TCGA）數(shù)據(jù)庫下載膠質(zhì)母細(xì)胞瘤患者mRNA表達(dá)和臨床信息。采用Cox回歸、Lasso回歸及交叉驗(yàn)證篩選預(yù)后相關(guān)的ICGs，構(gòu)建風(fēng)險(xiǎn)模型并對(duì)膠質(zhì)母細(xì)胞瘤患者進(jìn)行風(fēng)險(xiǎn)評(píng)分。利用風(fēng)險(xiǎn)評(píng)分及臨床特征構(gòu)建列線圖預(yù)測(cè)模型，通過校準(zhǔn)曲線及受試者特征曲線（ROC）曲線進(jìn)行驗(yàn)證。利用Kaplan-Meier曲線對(duì)高、低風(fēng)險(xiǎn)組進(jìn)行預(yù)后分析。利用DSigDB數(shù)據(jù)庫篩選ICGs相關(guān)的潛在候選藥物，并對(duì)高、低風(fēng)險(xiǎn)組藥物敏感性進(jìn)行分析。結(jié)果 共得到36個(gè)銅死亡相關(guān)的ICGs。通過Cox回歸、Lasso回歸篩選出4個(gè)預(yù)后相關(guān)ICGs，分別是CD276、TNFSF14、CD40、TNFSF9。多因素Cox回歸結(jié)果示，風(fēng)險(xiǎn)評(píng)分是膠質(zhì)母細(xì)胞瘤患者的獨(dú)立預(yù)后因子。ROC曲線圖預(yù)測(cè)1、3年生存率的ROC曲線下面積分別為0.726（95% CI：0.637～0.816）、0.731（95% CI：0.593～0.870）。GSEA分析結(jié)果顯示，DNA復(fù)制、間隙連接、糖胺聚糖生物合成等癌癥通路主要富集在高風(fēng)險(xiǎn)組。高風(fēng)險(xiǎn)組患者中CD4⁺ T細(xì)胞、中性粒細(xì)胞、巨噬細(xì)胞、髓樣樹突狀細(xì)胞免疫浸潤(rùn)水平較低風(fēng)險(xiǎn)組升高，達(dá)沙替尼、二甲基乙二酰氨基乙酸的半數(shù)抑制濃度（IC₅₀）較低風(fēng)險(xiǎn)組減小。從DSigDB數(shù)據(jù)庫得到8種可能對(duì)膠質(zhì)母細(xì)胞瘤患者有益的小分子藥物。結(jié)論 篩選出4個(gè)預(yù)后有關(guān)的銅死亡相關(guān)ICGs，以此構(gòu)建的風(fēng)險(xiǎn)模型具有較好的預(yù)后價(jià)值，并確定了8種ICGs相關(guān)治療膠質(zhì)母細(xì)胞瘤患者的潛在候選藥物。

Objective To comprehensively use bioinformatics methods to analyze the characteristics of copper death-related immune checkpoint genes (ICGs) and predict potential drug candidates in glioblastoma. Methods The mRNA expression and clinical information of glioblastoma patients were obtained from The Cancer Genome Atlas (TCGA) database. The Cox regression, Lasso regression and cross-validation were used to screen out the ICGs related to prognosis, then construct a risk model and score the risk of glioblastoma patients. A nomogram prediction model was constructed using risk scores and clinical characteristics, and verified by calibration curve and ROC curve. Prognostic analysis of high and low risk groups was performed using Kaplan-Meier curves. The DSigDB database was used to screen potential drug candidates related to ICGs, and the drug sensitivity of high and low risk groups was analyzed. Results A total of 36 copper death-related ICGs were obtained. Four prognosis-related ICGs were screened by Cox regression and Lasso regression, namely CD276, TNFSF14, CD40, and TNFSF9. Multivariate Cox regression results showed that risk score was an independent prognostic factor for glioblastoma patients. The results of the ROC curve showed that the area under the ROC curve of the nomogram to predict the 1-year and 3-year survival rates was 0.726 (95% CI: 0.637-0.816) and 0.731 (95% CI: 0.593-0.870), respectively. The results of GSEA analysis showed that cancer pathways such as DNA replication, gap junctions, and glycosaminoglycan biosynthesis were mainly enriched in high-risk groups. The immune infiltration levels of CD4⁺ T cells, neutrophils, macrophages, and myeloid dendritic cells in the high-risk group were significantly increased, while the IC₅₀s of dasatinib and dimethyloxalylglycine were significantly decreased. Eight small-molecule drugs that may benefit glioblastoma patients were obtained from the DSigDB database. Conclusion In this study, 4 prognostic copper death-related ICGs were screened, and the risk model constructed by these ICGs had good prognostic value, and 8 potential drug candidates related to ICGs were identified for the treatment of glioblastoma patients.

R965