目的 探讨非小细胞肺癌（NSCLC）患者癌组织中热休克蛋白肽gp96（HSP gp96）以及免疫微环境指标表达特征并探讨指标与铂类药物化疗敏感性的关系。方法 选取2021年12月—2024年12月烟台毓璜顶医院收治的80例NSCLC患者为研究对象，根据患者治疗4个周期后的疗效，将患者分为化疗敏感组（n=48）和不敏感组（n=32）。分层回归模型分析不同临床特征与HSP gp96阳性细胞百分比的关系；COX回归分析HSP gp96阳性细胞百分比与化疗敏感性的关系；双重差分法（DID）结合广义线性混合效应回归（GLMER）模型，评估HSP gp96阳性细胞百分比与化疗敏感的反应关联，采用Stata软件中的“med4way”包进行中介分析；线性回归分析HSP gp96与免疫微环境的关系；受试者工作特征（ROC）曲线分析HSP gp96阳性细胞百分比、免疫微环境及联合预测对化疗敏感性的诊断效能。结果 鳞癌、淋巴结转移、TNM IV期、肿瘤大小≥5 cm患者的HSP gp96阳性细胞百分比高于腺癌、淋巴结无转移、TNM Ⅲ期、肿瘤大小<5 cm患者，差异具有统计学意义（P<0.05）。病理类型、淋巴结转移、TNM分期、肿瘤大小均会对HSP gp96阳性细胞百分比产生正向影响（β>0,P<0.05）。化疗敏感组鳞癌患者占比、淋巴结转移人数占比、TNM IV期患者占比、肿瘤大小≥5 cm患者占比、HSP gp96阳性细胞百分比低于不敏感组，差异具有统计学意义（P<0.05）。NK、CD3⁺、CD4⁺化疗敏感组水平高于不敏感组，CD8⁺水平低于不敏感组，差异具有统计学意义（P<0.05）。Logistic回归分析结果显示，病理类型、淋巴结转移、TNM分期、肿瘤大小、HSP gp96阳性细胞百分比、CD8⁺均为患者化疗敏感的独立危险因素（OR>1,P<0.05）,NK、CD3⁺、CD4⁺为患者化疗敏感的保护因素（OR<1,P<0.05）。随着HSP gp96阳性细胞百分比的上升，化疗敏感性降低，HSP gp96阳性细胞百分比与化疗敏感性呈负相关（P<0.05），存在非线性反应关系，HSP gp96阳性细胞百分比每增加1%，化疗敏感的风险的超额风险（ER）值为-0.078[95%置信区间（CI）：-0.150～-0.023]。HSP gp96阳性细胞百分比与NK、CD3⁺、CD4⁺存在负相关关系（β<0,P<0.05），与CD8⁺存在正相关关系（β>0,P<0.05）。NK、CD3⁺、CD4⁺、CD8⁺具有显著的中介效应，β（95%CI）分别为：0.221（0.078～0.364）、0.309（0.145～0.473）、0.278（0.121～0.435）、1.221（0.905～1.537），分别介导了14.41%、17.65%、16.56%、33.29%HSP gp96阳性细胞百分比对化疗敏感性的关联。HSP gp96阳性细胞百分比、NK、CD3⁺、CD4⁺、CD8⁺联合对化疗敏感的诊断效能大于各指标单独预测，灵敏度为79.40%，特异度为82.30%，曲线下面积（AUC）为0.88。结论 HSP gp96阳性细胞百分比与免疫微环境指标之间存在相关性，同时HSP gp96阳性细胞百分比与免疫微环境对化疗敏感性具有交互作用，为开发新的NSCLC治疗策略提供了理论基础。

Objective To investigate the expression characteristics of heat shock protein peptide gp96（HSP gp96） and immune microenvironment indicators in cancer tissues of patients with non-small cell lung cancer（NSCLC）, as well as their relationship with the sensitivity to platinum-based drug chemotherapy. Methods A total of 80 NSCLC patients admitted to Yantai Yuhuangding Hospital from December 2021 to December 2024 were selected as the research subjects. According to the therapeutic effect after four cycles of treatment, the patients were divided into the chemotherapy-sensitive group（n= 48） and the insensitive group（n= 32）. The hierarchical regression model was used to analyze the relationship between different clinical features and the percentage of HSP gp96 positive cells. COX regression analysis was used to analyze the relationship between the percentage of HSP gp96 positive cells and chemotherapy sensitivity; The difference-in-differences（DID） method combined with the generalized linear mixed-effects regression（GLMER） model was used to evaluate the association between the percentage of HSP gp96 positive cells and chemotherapy-sensitive responses, and mediating analysis was performed using the “med4 way” package in Stata software. Linear regression analysis of the relationship between HSP gp96 and the immune microenvironment; ROC curve was used to analyze the diagnostic efficacy of the percentage of HSP gp96 positive cells, the immune microenvironment and the combined prediction for chemotherapy sensitivity. Results The percentage of HSP gp96 positive cells in patients with squamous cell carcinoma, lymph node metastasis, TNM stage IV, and tumor ≥ 5 cm was higher than that in patients with adenocarcinoma, no lymph node metastasis, TNM stage III, and tumor < 5 cm, and the difference was statistically significant（P < 0.05）. Pathological type, lymph node metastasis, TNM stage and tumor size all have a positive impact on the percentage of HSP gp96 positive cells（β > 0, P < 0.05）. The proportions of squamous cell carcinoma patients, the proportion of lymph node metastases, the proportion of TNM stage IV patients, the proportion of patients with tumors ≥ 5 cm, and the percentage of HSP gp96 positive cells in the chemotherapy-sensitive group were lower than those in the insensitive group, and the differences were statistically significant（P < 0.05）. The levels of NK, CD3⁺ and CD4⁺ in the chemotherapy-sensitive group were higher than those in the insensitive group, while the level of CD8⁺ was lower than that in the insensitive group. The differences were statistically significant（P < 0.05）. The results of Logistic regression analysis showed that pathological type, lymph node metastasis, TNM stage, tumor size, percentage of HSP gp96 positive cells, and CD8⁺ were all independent risk factors for chemotherapy sensitivity in patients（OR > 1, P < 0.05）, while NK, CD3⁺, and CD4⁺ were protective factors for chemotherapy sensitivity in patients（OR < 1, P < 0.05）. With the increase of the percentage of HSP gp96 positive cells, the chemosensitivity decreased. The percentage of HSP gp96 positive cells is negatively correlated with chemotherapy sensitivity（P < 0.05）, and there is a nonlinear reaction relationship. For every 1% increase in the percentage of HSP gp96 positive cells, The ER value for chemotherapy-sensitive risk was-0.078（95%CI:-0.150—-0.023）. The percentage of HSP gp96 positive cells was negatively correlated with NK, CD3⁺, and CD4⁺（β < 0, P < 0.05）, and positively correlated with CD8⁺（β > 0, P < 0.05）. NK, CD3⁺, CD4⁺, and CD8⁺ have significant mediating effects, with β（95% CI） being: 0.221（0.078—0.364）, 0.309（0.145—0.473）, 0.278（0.121—0.435）, 1.221（0.905—1.537） It respectively mediated the associations of 14.41%, 17.65%, 16.56%, and 33.29% of the percentages of HSP gp96 positive cells with chemotherapy sensitivity. The combined diagnostic efficacy of the percentage of HSP gp96 positive cells, NK, CD3⁺, CD4⁺, and CD8⁺ for chemotherapy sensitivity is greater than that predicted by each index alone. The sensitivity was 79.40%, the specificity was 82.30%, and the AUC was 0.88. Conclusion There is a correlation between the percentage of HSP gp96 positive cells and immune microenvironment indicators. Meanwhile, the percentage of HSP gp96 and the immune microenvironment have an interaction effect on chemotherapy sensitivity, providing a theoretical basis for the development of new NSCLC treatment strategies.

R974

山东省医药卫生科技发展计划项目(2022WSB15003)