背景：内质网（ER）应激在结直肠腺癌（COAD）中发挥促凋亡作用。本研究旨在开发一种新型的 COAD ER 应激相关预后风险模型，并为对免疫检查点抑制剂治疗具有不同风险评分反应的 COAD 队列提供支持。方法：TCGA-COAD 和 GSE39582 纳入本前瞻性研究。进行单变量和多变量 Cox 分析以确定预后 ER 应激相关基因 (ERSG)。因此，评估了不同风险组的免疫浸润情况和免疫治疗反应。最后，使用逆转录定量聚合酶链反应验证了 10 个正常组织样本和 10 个 COAD 组织样本中预后基因的表达。结果：选择八个预后基因在 TCGA-COAD 队列的训练集中建立基于 ERSG 的特征。使用 TCGA-COAD 和 GSE39582 队列的测试集证实了这一点的准确性。基因集变异分析表明，高低风险群体的差异功能与免疫相关途径有关。与此相对应，CD36、TIMP1和PTGIS与不同风险组之间比例不同的19种免疫细胞显着相关，例如中央记忆CD4T细胞和中央记忆CD8T细胞。此外，风险评分被认为可以有效预测免疫治疗的临床反应，并且免疫治疗反应与 COAD 个体的风险评分显着负相关。此外，免疫检查点抑制剂治疗在高危组中效果较差，高危组中PD-L1的表达水平以及肿瘤免疫功能障碍和排除评分显着升高。最后，实验结果表明临床样本中预后基因的表达趋势与公共数据库的结果一致。结论：我们的研究建立了一个新的风险特征来预测COAD患者的预后，为COAD的临床治疗提供理论支持。

Background: Endoplasmic reticulum (ER) stress plays a pro-apoptotic role in colorectal adenocarcinoma (COAD). This study aimed to develop a novel ER-stress-related prognostic risk model for COAD and provide support for COAD cohorts with different risk score responses to immune checkpoint inhibitor therapies. Methods: TCGA-COAD and GSE39582 were included in this prospective study. Univariate and multivariate Cox analyses were performed to identify prognostic ER stress-related genes (ERSGs). Accordingly, the immune infiltration landscape and immunotherapy response in different risk groups were assessed. Finally, the expression of prognostic genes in 10 normal and 10 COAD tissue samples was verified using reverse transcription-quantitative polymerase chain reaction. Results: Eight prognostic genes were selected to establish an ERSG-based signature in the training set of the TCGA-COAD cohort. The accuracy of this was confirmed using a testing set of TCGA-COAD and GSE39582 cohorts. Gene set variation analysis indicated that differential functionality in high-low-risk groups was related to immune-related pathways. Corresponding to this, CD36, TIMP1, and PTGIS were significantly associated with 19 immune cells with distinct proportions between the different risk groups, such as central memory CD4T cells and central memory CD8T cells. Moreover, the risk score was considered effective for predicting the clinical response to immunotherapy, and the immunotherapy response was significantly and negatively correlated with the risk score of individuals with COAD. Furthermore, the immune checkpoint inhibitor treatment was less effective in the high-risk group, where the expression levels of PD-L1 and tumor immune dysfunction and exclusion scores in the high-risk group were significantly increased. Finally, the experimental results demonstrated that the expression trends of prognostic genes in clinical samples were consistent with the results from public databases. Conclusion: Our study established a novel risk signature to predict the COAD prognosis of patients and provide theoretical support for the clinical treatment of COAD.