Abstract:

Objective　To investigate the relationship between immune-related long non-coding RNA (lncRNA) models in cervical cancer and patients' prognosis and response to immunotherapy. Methods　The transcriptome sequencing data and corresponding clinical information in 255 cervical cancer patients were obtained from the Cancer Genome Atlas (TCGA) database. The immune-related lncRNAs were extracted by bioinformatic method, and the prediction model was constructed using Kaplan-Meier analysis and multifactorial COX regression analysis. The receiver operating characteristic curve (ROC) was used to assess the efficacy of the prediction model. The patients were classified into high and low risk groups by the median value of the model score. The differences in immune infiltration between different subgroups were assessed using algorithms such as CIBERSORT and single sample gene set enrichment analysis (ssGSEA), and the association of the model with immune pathways was explored using gene ontology (GO) enrichment analysis, gene set enrichment analysis (GSEA), and gene set variation analysis (GSVA). The differences in immunotherapy response between different subgroups were also explored using the tumor immune dysfunction and rejection (TIDE) score and immune epistasis (IPS) score. Results　The immune-related lncRNA prediction models were constructed by Kaplan-Meier analysis and multifactorial COX regression analysis, and the survival curve showed a statistically significant difference in the prognosis between the two groups (P<0.05). The areas under the ROCs at 1, 2, and 3 years for the prediction models were 0.850, 0.796, and 0.702, respectively. The immune score in the low-risk group was significantly higher than that in the high-risk group (P<0.05). The results of CIBERSORT, an immune algorithm based on deconvolution, suggested that CD8⁺ T cells, helper T cells, B cells, and M0 and M1 macrophages differed between the two groups (all P<0.05). Immune pathways were significantly enriched in the low-risk group. TIDE and IPS scores were significantly different between different risk groups (both P<0.05). Conclusion　Immune-related lncRNA models for cervical cancer can be used as biomarkers to predict the response rate and prognosis in cervical cancer patients receiving immunotherapy.

Key words:

Cervical cancer, Immune microenvironment, Immunotherapy, Prognosis evaluation

摘要：

目的　探究宫颈癌中免疫相关长链非编码RNA（lncRNA）模型与患者预后及免疫治疗应答的关系。方法　在癌症基因组图谱（TCGA）数据库中获取255例宫颈癌患者的转录组测序数据及相应的临床信息，通过生物信息学方法提取免疫相关lncRNA，使用Kaplan-Meier分析和多因素COX回归分析构建预测模型。受试者工作特征曲线（ROC）用于评估预测模型的效能。以模型评分的中位值将患者分为高低风险两组。使用CIBERSORT、单样本基因富集分析（ssGSEA）等算法评估不同分组间免疫浸润的差异，使用基因本体（GO）富集分析、基因集合富集分析（GSEA）、基因集合变异分析（GSVA）分析探究模型与免疫通路的联系。同时使用肿瘤免疫功能障碍与排斥（TIDE）评分及免疫表观（IPS）评分探索不同分组间免疫治疗应答的差异。结果　通过Kaplan-Meier分析和多因素COX回归分析构建了免疫相关lncRNA预测模型，生存曲线显示出两组患者预后的差异有统计学意义（P<0.05），预测模型的1、2、3年ROC曲线下面积分别为0.850、0.796、0.702。低风险组的免疫得分显著高于高风险组（P<0.05）。基于反卷积的免疫算法CIBERSORT结果提示CD8⁺T细胞、辅助T细胞、B细胞和M0、M1型巨噬细胞在两组间差异均有统计学意义（均P<0.05）。低风险组中免疫通路被显著富集。TIDE和IPS评分在不同风险组间差异均有统计学意义（均P<0.05）。结论　宫颈癌免疫相关lncRNA模型可作为生物标志物用来预测宫颈癌患者接受免疫治疗的应答率及预后。

关键词:

宫颈癌, 免疫微环境, 免疫治疗, 预后评价