基于铂类耐药相关基因的卵巢癌预后风险模型构建与验证

Construction and Validation of a Prognostic Risk Model for Ovarian Cancer Based on Platinum-Resistance Related Genes

摘要: 目的：基于铂类耐药相关基因构建卵巢癌预后模型并探索其潜在的分子机制。方法：整合TCGA和GTEx数据库转录组数据及临床信息，联合HGSOC-Platinum数据库鉴定铂类耐药相关差异表达基因(DEG)。通过LASSO-Cox回归构建风险评分模型，利用Kaplan-Meier曲线、ROC曲线、列线图评估模型效能。采用GO和KEGG富集分析探讨生物学功能，并通过HPA和TISCH数据库进行外部验证。结果：共鉴定220个铂类耐药相关DEG，其中RAD51为关键节点基因。基于18个基因构建的风险评分模型可将患者分为高危、低危组，两组总生存期差异有统计学意义(P < 0.0001)。模型预测1、3、5年总生存期的AUC值分别为0.75、0.72、0.78，C指数为0.671。DEG主要富集于PI3K-Akt、p53信号通路及上皮间质转化相关过程。VIM主要表达于成纤维细胞亚群，且在卵巢癌组织中低表达。结论：本研究构建的铂类耐药相关基因模型可有效预测卵巢癌患者预后，RAD51和VIM可能通过调控DNA损伤修复及上皮间质转化参与铂类耐药，为后续机制研究和个体化治疗提供潜在靶点。

Abstract: Objective: To construct a prognostic risk model for ovarian cancer based on platinum-resistance related genes and to explore the underlying molecular mechanisms. Methods: Transcriptomic data and clinical information were integrated from TCGA and GTEx databases, and combined with the HGSOC-Platinum database to identify differentially expressed genes (DEGs) associated with platinum resistance. A risk score model was constructed using LASSO-Cox regression analysis. Model performance was evaluated by Kaplan-Meier curves, receiver operating characteristic (ROC) curves, and nomograms. GO/KEGG enrichment analyses were performed to explore biological functions, and external validation was conducted using the HPA and TISCH databases. Results: A total of 220 platinum resistance-related DEGs were identified, with RAD51 as a key node gene. A risk score model based on 18 genes stratified patients into high-risk and low-risk groups, with a statistically significant difference in overall survival between the two groups (P < 0.0001). The AUC values for predicting 1-, 3-, and 5-year overall survival were 0.75, 0.72, and 0.78, respectively, with a C-index of 0.671. Enrichment analyses indicated that DEGs were primarily enriched in the PI3K-Akt and p53 signaling pathways, as well as processes related to epithelial-mesenchymal transition (EMT). VIM was mainly expressed in fibroblast subpopulations and showed low expression in ovarian cancer tissues. Conclusion: The platinum resistance-related gene model constructed in this study effectively predicts the prognosis of ovarian cancer patients. RAD51 and VIM may participate in platinum resistance by regulating DNA damage repair and EMT, providing potential targets for further mechanistic research and personalized therapy.