肿瘤微环境（TME）由各种异质细胞组成，通过广泛的相互作用促进肿瘤增殖、侵袭和转移。影响患者生存时间和免疫治疗反应的关键因素在于肿瘤内免疫环境。我们通过基于TCGA和TISIDB数据库的生物信息学分析获得了112个与T细胞介导的LUAD肿瘤杀伤相关的差异基因。然后，通过进行单变量LASSO和多变量Cox回归分析，构建了由6个基因（CA9，OIP5，TIMP1，SEC11C，FURIN和TLR10）组成的预后风险评分模型。以中位数风险评分作为阈值，将样本分为两组进行分类。生存分析表明，低风险组具有更有利的预后。随后，将Cox回归分析与LUAD患者的临床信息（年龄、性别和病理分期）和风险评分相结合，证明了该模型作为独立预后因子的潜力。基于临床信息和风险评分构建的刻度曲线显示，该模型具有良好的预测能力，差异基因在高风险组和低风险组中富集于免疫相关的过程或通路，如GO和KEGG富集分析所示。单样本基因富集分析（ssGSEA）和免疫表型得分（IPS）的联合应用显示，低风险组表现出较高的免疫浸润和IPS分数，表明免疫治疗可能在该组患者中显示出良好的疗效。总之，基于T细胞介导的细胞杀伤相关基因构建的LUAD预后模型不仅能筛选LUAD患者的预后，还可用于筛选对免疫治疗具有高敏感性的LUAD患者。我们的研究为LUAD患者的临床治疗和预后预测提供了新的见解。

Constituted by various heterogeneous cells, the tumor microenvironment (TME) is capable of promoting tumor proliferation, invasion, and metastasis through extensive crosstalk. The pivotal factor influencing the survival time of patients and their response to immunotherapy lies in the intratumoral immune environment. We obtained 112 differential genes related to T cell-mediated tumor killing in LUAD by employing bioinformatics analysis on the basis of the TCGA and TISIDB databases. Then the 6-gene prognostic risk score model (CA9, OIP5, TIMP1, SEC11C, FURIN, and TLR10) was constructed by conducting univariate LASSO as well as multivariate Cox regression analyses. The median risk score was taken as the threshold to classify the samples into two groups. Survival analysis revealed that the low-risk group exhibited a more favorable prognosis. Subsequently, the Cox regression analysis combined with clinical information (age, gender, and pathological stage) and the risk score of LUAD patients demonstrated the potential of this model as an independent prognostic factor. The nomogram established based on clinical information and a risk score in combination with the calibration curve indicated that this model had good predictive ability. Notable enrichment of the differential genes from the high- and low-risk groups was discovered in immune-associated processes or pathways, as shown by the GO and KEGG enrichment analyses. The combined use of single-sample gene enrichment analysis (ssGSEA) and immunophenoscore (IPS) demonstrated heightened immune infiltration and IPS scores in the low-risk group, indicating that immunotherapy was likely to show good efficacy in patients from this group. To sum up, the prognostic model of LUAD constructed based on T-cell-mediated cell killing-related genes was not only capable of screening the prognosis of LUAD patients but was also used for screening those LUAD patients with high sensitivity to immunotherapy. Our study offered novel insights into the clinical treatment and prognostic prediction of LUAD patients.