Li Chen-Li, Jiang Yu-Qian, Pan Wei, Yang Yan-Li
Clin Nephrol. 2025 Mar;103(3):200-212. doi: 10.5414/CN111509.
Utilizing expression data of clear cell renal cell carcinoma (ccRCC) genes from the Cancer Genome Atlas (TCGA) database, this study employs weighted gene co-expression network analysis (WGCNA) and Cox regression analysis to identify genes associated with the occurrence and development of ccRCC, thereby providing a scientific basis for its treatment.
Differentially expressed genes between tumor and control groups were identified by preprocessing and batch correction of ccRCC transcriptome data in the TCGA database using the Wilcoxon test. Prognostic prediction models were established through a combination of WGCNA analysis, univariate Cox regression analysis, and multivariate Cox regression analysis. The reliability of these prognostic models was evaluated by plotting Kaplan-Meier survival analysis and receiver operating characteristic (ROC) curves and by further analyzing the relationship between model gene expression levels, tumor staging, and tumor grading.
Post-batch correction, M2-type macrophage infiltration was pronounced in tumor tissue, and 13 out of 290 screened relevant differential genes were included in the prognostic model. The Kaplan-Meier survival curves indicated that the 3- and 5-year overall survival rates were significantly higher in the low-risk group compared with the high-risk group (83.7 vs. 69.1%; 75.7 vs. 52.6%, p = 1.169e-08). The area under the ROC curve was 0.732, signifying strong predictive power for the survival curve. In this model, the expression levels of 11 genes were positively correlated with tumor stage and pathological grade, whereas the remaining 2 genes were negatively correlated.
This model can predict the overall survival of patients with ccRCC and has the potential to become an important therapeutic target.
本研究利用癌症基因组图谱（TCGA）数据库中透明细胞肾细胞癌（ccRCC）基因的表达数据，采用加权基因共表达网络分析（WGCNA）和Cox回归分析来识别与ccRCC发生发展相关的基因，从而为其治疗提供科学依据。
使用Wilcoxon检验对TCGA数据库中的ccRCC转录组数据进行预处理和批次校正，以识别肿瘤组和对照组之间的差异表达基因。通过WGCNA分析、单变量Cox回归分析和多变量Cox回归分析相结合的方式建立预后预测模型。通过绘制Kaplan-Meier生存分析和受试者工作特征（ROC）曲线，并进一步分析模型基因表达水平、肿瘤分期和肿瘤分级之间的关系，来评估这些预后模型的可靠性。
批次校正后，肿瘤组织中M2型巨噬细胞浸润明显，在筛选出的290个相关差异基因中，有13个被纳入预后模型。Kaplan-Meier生存曲线表明，低风险组的3年和5年总生存率显著高于高风险组（83.7%对69.1%；75.7%对52.6%，p = 1.169e-08）。ROC曲线下面积为0.732，表明对生存曲线具有较强的预测能力。在该模型中，11个基因的表达水平与肿瘤分期和病理分级呈正相关，而其余2个基因呈负相关。
该模型可预测ccRCC患者的总生存率，有潜力成为重要的治疗靶点。
