摘要
目的基于TCGA数据库探讨坏死性凋亡相关基因(necroptosis-related genes,NRGs)对结肠癌预后的预测价值。方法从TCGA数据库获取358例患者的RNA高通量测序(RNA-sequencing,RNA-seq)数据及相应临床资料,单因素Cox回归分析初步确定与预后相关的NRGs,使用非负矩阵分解(non-negative matrix factorization,NMF)算法将样本划分为具有相似基因表达模式的亚组,分析亚组间的生存情况、预后相关NRGs与临床病理因素的相关性。所有样本1∶1随机分为训练集和验证集,通过LASSO-Cox回归分析筛选预后相关NRGs,计算风险评分,建立基于NRGs的预后风险模型。所有样本按风险评分的中位数分为低、高风险组,进行生存分析,绘制风险曲线与ROC曲线。单因素、多因素Cox回归分析结肠癌预后独立预测因素,构建列线图预测模型,计算预测模型的一致性指数(concordance index,C-index),评价模型预测能力。对预后相关NRGs进行GO富集分析,通过GSEA与ssGSEA评估高风险和低风险组中KEGG、Hallmarks基因集信号通路活性。分析风险评分与结肠癌临床病理因素的相关性。结果从KEGG数据库获得了159个NRGs,初步筛选出15个与预后相关的NRGs。结肠癌患者被分为两个亚组,两组间总生存期、预后相关NRGs表达水平、N分期、M分期、癌症分期存在差异(P<0.05)。LASSO-Cox回归分析最终确定了9个具有预后价值的NRGs,随着风险分数提高,患者的死亡风险上升,总生存期缩短(P<0.05)。训练集、验证集以及全体样本的低风险组与高风险组生存率差异有统计学意义(P<0.05)。训练集1、3、5年ROC曲线下面积分别为0.762、0.825、0.854,验证集分别为0.683、0.629、0.686,全体样本分别为0.720、0.727、0.778。预后相关NRGs与细胞周期调控和细胞增殖过程密切相关,低风险组与高风险组在某些肿瘤、药物代谢和能量代谢相关通路方面存在差异。不同癌症分期、T分期、N分期、M分期的患者风险评分差异有统计学意义(P<0.05)。结论基于NRGs构建的预后风险模型能够有效预测结肠癌患者的预后情况,具有潜在的临床应用价值。
Objective To investigate the prognostic value of necroptosis-related genes(NRGs)in colon cancer based on TCGA database.Methods The RNA-sequencing(RNA-seq)data and corresponding clinical data of 358 patients were obtained from TCGA database.Non-negative matrix factorization(NMF)algorithm was subsequently applied to classify the samples into subgroups with similar gene expression profiles.The survival outcomes,prognostic-related NRGs,and correlations with clinical and pathological factors were then analyzed across these subgroups.All samples were randomly divided into training and verification sets in a 1∶1.LASSO-Cox regression analysis was performed to screen prognosis-related NRGs,calculate risk scores,and develop a prognostic risk model based on NRGs.All samples were categorized into low-and high-risk groups based on the median risk score,survival analysis was conducted,and risk curves along with ROC curves were plotted.Independent prognostic factors for colon cancer were analyzed,a nomogram prediction model was constructed,and the predictive performance of the model was evaluated.GO enrichment analysis was performed on prognosis-related NRGs,and the activity of KEGG and Hallmarks gene set signaling pathways in high-and low-risk groups was assessed using GSEA and ssGSEA.The correlation between risk scores and clinicopathological factors in colon cancer was subsequently analyzed.Results Colon cancer patients were categorized into two subgroups,with significant differences observed in overall survival,prognosis-related NRGs expression levels,N/M stages,and cancer staging between the two groups(P<0.05).LASSO-Cox regression analysis ultimately identified 9 NRGs with prognostic value;as risk scores increased,patients exhibited elevated mortality risk and reduced overall survival(P<0.05).Survival rates differed significantly between low-and high-risk groups across the training set,validation set,and entire cohort(P<0.05).The AUC values for the training set at 1-,3-and 5-year were 0.762,0.825 and 0.854,respectively;for the validation set,they were 0.683,0.629 and 0.686,respectively,and for the entire cohort,they reached 0.720,0.727 and 0.778,respectively.Prognosis-related NRGs were closely associated with cell cycle regulation and cellular proliferation processes,and the low-and high-risk groups exhibited differences in certain tumor-related,drug metabolism,and energy metabolism-associated pathways.Patients with varying cancer stages,T stages,N stages and M stages exhibited differences in risk scores(P<0.05).Conclusion The prognostic risk model based on NRGs can effectively predict the prognosis of patients with colon cancer and has potential clinical value.
作者
高丽娜
张炎
刘刚
GAO Lina;ZHANG Yan;LIU Gang(Department of General Surgery,the Sixth Medical Center of PLA General Hospital,Beijing 100048,China)
出处
《胃肠病学和肝病学杂志》
2025年第7期954-962,共9页
Chinese Journal of Gastroenterology and Hepatology
