开发和验证十四对固有免疫相关基因对预测头颈部鳞状细胞癌预后的标志。

BACKGROUND: Immune-related genes is closely related to the occurrence and prognosis of head and neck squamous cell carcinoma (HNSCC). At the same time, immune-related genes have great potential as prognostic markers in many types of cancer. The prognosis of HNSCC is still poor currently, and it may be effective to predict the clinical outcome of HNSCC by immunogenic analysis. METHODS: RNASeq and clinical follow-up information were downloaded from The Cancer Genome Atlas (TCGA), the MINiML format GSE65858 chip expression data was downloaded from NCBI, and immune-related genes was downloaded from the InnateDB database. Immune-related genes in 519 HNSC patients were integrated from TCGA dataset. By using multivariate COX analysis and Lasso regression, robust immune-related gene pairs (IRGPs) that predict clinical outcomes of HNSCC were identified. Finally, a risk prognostic model related to immune gene pair was established and verified by clinical features, test sets and GEO external validation set. RESULTS: A total of 699 IRGPs were significantly correlated with the prognosis of HNSCC patients. Fourteen robust IRGPs were finally obtained by Lasso regression and a prognostic risk prediction model was constructed. Risk score of each sample were calculated based on Risk models and divided into the high-risk group (Risk-H) and low Risk group (Risk-L). Risk models were able to stratify the risk in patients with TNM Stage, Age, gender, and smoking history, and the AUC > 0.65 in training set and test set, shows that 14-IRGPs signature in patients with HNSCC has excellent classification performance. In addition, 14-IRGPs had the highest average C index compared with the prognostic characteristics and T, N, and Age of the 3 previously reported HNSCC. CONCLUSION: This study constructed 14-IRGPs as a novel prognostic marker for predicting survival in HNSCC patients.

背景：免疫相关基因与头颈部鳞状细胞癌（HNSCC）的发生和预后密切相关。同时，免疫相关基因作为许多类型癌症的预后标志物具有很大的潜力。HNSCC 的预后仍然很差，通过免疫分析预测 HNSCC 的临床结局可能是有效的。

方法：从癌症基因组图谱（TCGA）下载 RNAseq 和临床随访信息，从 NCBI 下载 MINiML 格式 GSE65858 芯片表达数据，从 InnateDB 数据库下载免疫相关基因。从 TCGA 数据集整合 519 例 HNSC 患者的免疫相关基因。通过多变量 COX 分析和 Lasso 回归，确定预测 HNSCC 临床结局的稳健免疫相关基因对（IRGPs）。最后，根据临床特征、测试集和 GEO 外部验证集，建立并验证与免疫基因对相关的风险预后模型。

结果：共鉴定出 699 个与 HNSCC 患者预后显著相关的 IRGPs。通过 Lasso 回归最终获得 14 个稳健的 IRGPs，并构建了预后风险预测模型。根据风险模型计算每个样本的风险评分，并将其分为高风险组（Risk-H）和低风险组（Risk-L）。风险模型能够对 TNM 分期、年龄、性别和吸烟史的患者进行分层，并且在训练集和测试集中 AUC>0.65，表明 14-IRGPs 标志物在 HNSCC 患者中有很好的分类性能。此外，与之前报道的 3 种 HNSCC 的预后特征、T、N 和年龄相比，14-IRGPs 具有最高的平均 C 指数。

结论：本研究构建了 14-IRGPs 作为预测 HNSCC 患者生存的新的预后标志物。