慢性阻塞性肺疾病合并肺癌的临床特征分析及列线图模型建立

Analysis of Clinical Characteristics and Construction of Nomogram Model in Patients of Lung Cancer Concomitant Chronic Obstructive Pulmonary Disease

摘要: 目的：分析COPD合并肺癌患者的高危因素，并建立列线图预测模型以辅助临床决策。方法：回顾性分析2018年1月至2024年12月收治于青岛市市立医院114例单纯COPD患者(COPD组)及102例COPD合并肺癌患者(COPD合并肺癌组)的临床资料，包括：性别、年龄、体重指数、吸烟指数、临床症状和体征、实验室指标及肺功能等，使用SPSS27.0软件和R软件(R4.3.3)对数据进行差异性分析及多因素logistic回归分析筛选出独立危险因素，并构建列线图预测模型。根据校准曲线及临床决策曲线对模型进行评估。结果：多因素logistic回归分析显示，持续性胸闷、咯血、胸腔积液、吸烟指数、白蛋白及血小板计数为独立危险因素(P < 0.05)，ROC曲线下面积为0.814 (95% CI: 0.756~0.871)，Bootstrap内部验证结果显示C-指数为0.806 (95% CI: 0.788~0.816)，校准曲线Hosmer-Lemeshow检验结果为χ² = 6.761，P = 0.579 (P > 0.05)，该模型拟合情况较好，临床决策曲线在阈值0.20~0.90内，能有较好的临床净获益。结论：根据上述的6项危险因素构建的预测模型，对COPD患者中可能罹患肺癌的高危人群具有一定的预测价值。

Abstract: Objective: To analyze high-risk factors in patients with concurrent COPD and lung cancer, and develop a nomogram-based predictive model to facilitate clinical decision-making. Methods: A retrospective analysis was conducted on the clinical data of 114 patients with COPD alone (COPD group) and 102 patients with COPD complicated with lung cancer (COPD with lung cancer group) admitted to Qingdao Municipal Hospital from January 2018 to December 2024. The clinical data included gender, age, body mass index, smoking index, clinical symptoms and signs, laboratory indicators, and pulmonary function. SPSS 27.0 software and R software (R4.3.3) were utilized for differential analysis and multivariate logistic regression analysis to screen independent risk factors, and construct a nomogram prediction model. The model was evaluated by using calibration curve and clinical decision curve analysis. Results: Multivariate logistic regression analysis revealed that persistent chest tightness, hemoptysis, pleural effusion, smoking index, albumin, and platelet count were independent risk factors (P < 0.05). The area under the ROC curve was 0.814 (95% CI: 0.756~0.871). Bootstrap internal validation showed a C-index of 0.806 (95% CI: 0.788~0.816). The Hosmer-Lemeshow test result of the calibration curve was χ² = 6.761, P = 0.579 (P > 0.05), The model fit was relatively good. The clinical decision curve demonstrated favorable clinical net benefit within the threshold range of 0.20~0.90. Conclusion: The prediction model constructed on the basis of these six risk factors has certain predictive value for identifying high-risk populations of lung cancer among COPD patients.