Abstract: Background　Multiple inflammatory markers in lung cancer patients are associated with all-cause mortality risk, and these associations vary by pathological type and population characteristics.Objective　To explore the relationship between multiple inflammatory markers and all-cause mortality in different pathological types of lung cancer. Methods　Data of 2 113 lung cancer patients in the Kailuan study from 2006 to 2020 were collected, including 355 cases of lung adenocarcinoma, 215 cases of lung squamous cell carcinoma, 154 cases of small cell lung cancer, and 1 387 cases of other pathological types. Through questionnaire surveys (collecting age, gender, smoking history, education level, income level, family history of cancer, history of hypertension and diabetes, etc.), physical measurements (blood pressure, etc.) and laboratory tests (collecting fasting venous blood to detect high-sensitivity C-reactive protein, blood routine indicators and high-density lipoprotein cholesterol, etc.), And 15 inflammatory indicators such as the lymphocyte ratio (NLR), systemic inflammation index (SII), and comprehensive inflammation index (AISI) were calculated to collect data. Taking the date of lung cancer diagnosis as the starting point of follow-up and all-cause death (excluding accidental death) as the outcome event, the follow-up was conducted until December 31, 2023. The association between inflammation-related indicators and all-cause mortality from lung cancer was evaluated using a multivariate Cox proportional hazards regression model, and the error detection rate (FDR) of the primary results was corrected using the Benjamini-Hochberg method. For the pathological subtype analysis of competitive risk events, a Fine-Gray competitive risk model was constructed to evaluate the association between inflammation-related indicators and all-cause mortality of lung cancer, and FDR correction was performed simultaneously. Based on the main analysis results, further stratified analysis and sensitivity analysis were carried out. Results　Until the end of follow-up, a total of 1,309 all-cause deaths were observed, with a disease-free survival of (5.08±1.09) years and a cumulative incidence rate of 66.66%. Multivariable Cox proportional hazards regression analysis show that, in model I(adjusted for age, gender, smoking status, education level, income level, history of hypertension and diabetes mellitus), for each 1 Z-Score increase in the levels of white blood cells, neutrophils, NLR and SII, the all-cause mortality risk of overall lung cancer patients increases by 8.9%, 9.7%, 9.0% and 11.2%, respectively (FDR P<0.05); for each 1 Z-Score increase in NLR and SII, the all-cause mortality risk of patients with other pathological types of lung cancer increases by 9.1% and 10.9%, respectively (FDR P<0.05); no association was found between inflammatory indicators and all-cause mortality risk in patients with lung adenocarcinoma, lung squamous cell carcinoma or small cell lung cancer (FDR P>0.05). Subgroup analysis results indicate that among lung cancer patients with different pathological types, monocytes, LHR, NHR, PLR, SIRI, AISI, MLR, NLR, CLR, hs-CRP, lymphocytes, SII, white blood cells, neutrophils and MHR exhibit interaction effects with smoking status, sex, education level, income level, family cancer history, history of diabetes, history of hypertension and age, respectively (Pinteraction<0.10). Conclusion　Multiple inflammatory markers after lung cancer diagnosis are associated with the risk of all-cause death, of which SII, NLR, WBC and neutrophils are the most robust associations in overall lung cancer. Different pathological types and subgroups (female, smokers, and family history of cancer) had different association patterns of inflammatory markers.

Key words: Inflammation, Lung cancer, Different pathological types, Epidemiology

摘要： 背景　肺癌患者多种炎症指标与全因死亡风险相关，且关联因病理类型及人群特征而异。目的　探讨不同病理类型肺癌多种炎症指标与全因死亡的关系。方法　收集2006—2020年开滦研究队列中2 113例肺癌患者数据，其中肺腺癌355例、肺鳞癌215例、小细胞肺癌154例、其他病理类型1 387例。收集患者年龄、性别、吸烟史、教育水平、收入水平、癌症家族史、高血压病史及糖尿病病史等基本资料及实验室检查指标（包括高敏C反应蛋白、血常规指标及高密度脂蛋白胆固醇等），并计算淋巴细胞比值（NLR）、全身炎症指数（SII）、综合炎症指数（AISI）等15种炎症指标。以肺癌确诊日期为随访起点，全因死亡（除外意外死亡）为结局事件，随访截至2023-12-31。采用多因素Cox比例风险回归模型评估炎症相关指标与肺癌的全因死亡的关联，并采用Benjamini-Hochberg法对主要结果进行错误发现率（FDR）校正。构建Fine-Gray竞争风险模型评估不同病理类型肺癌相关炎症指标与全因死亡风险的关联，采用Benjamini-Hochberg错误发现率（FDR）方法进行P值校正。基于主要分析结果，进一步开展亚组分析。结果　截止随访终点，截止随访终点，共发生全因死亡1 309例（61.95%），其中肺腺癌151例（57.63%），肺鳞癌130例（58.56%），小细胞肺癌102例（64.15%），其他病理类型926例（62.99%），中位总生存期为（5.07±0.13）年，累积发病率为70.32%。多因素Cox比例风险回归分析结果显示，在模型I中（校正了年龄、性别、吸烟状态、教育水平、收入水平、高血压病史、糖尿病病史），白细胞、中性粒细胞、NLR和SII水平每升高1个Z-Score，总肺癌患者全因死亡风险分别增加8.9%、9.7%、9.0%、11.2%（P校正<0.05）；NLR、SII每增加1个Z-Score，其他病理类型患者的全因死亡风险分别增加9.1%、10.9%（P校正<0.05）；未发现炎症指标与肺腺癌、肺鳞癌、小细胞肺癌患者全因死亡风险的关联（P校正>0.05）。亚组分析结果显示，不同病理类型肺癌患者中单核细胞、LHR、NHR、PLR、SIRI、AISI、MLR、NLR、CLR、hs-CRP、淋巴细胞、SII、白细胞、中性粒细胞、MHR分别与吸烟状态、性别、教育水平、收入水平、癌症家族史、糖尿病病史、高血压病史、年龄均存在交互作用（P交互<0.10）。结论　肺癌患者多种炎症指标与全因死亡风险相关，不同病理类型患者的炎症指标预后效应存在显著差异，并受性别、教育水平、吸烟、糖尿病病史及癌症家族史等因素影响。临床应用时应结合病理类型与人口特征对肺癌进行预后评估，以实现更精准的个体化管理。

关键词: 炎症, 肺癌, 不同病理类型, 流行病学