鉴别活动性肺结核与肺恶性肿瘤的列线图预测模型构建  

Construction of a nomograph prediction model for differentiating active pulmonary tuberculosis from lung cancer

作　　者：魏帅娜 陈茜 张国俊[1] WEI Shuaina;CHEN Xi;ZHANG Guojun(Department of Respiration Diseases,the First Affiliated Hospital of Zhengzhou University,Zhengzhou Henan,450000,China;不详)

机构地区：[1]郑州大学第一附属医院呼吸与危重症医学科,河南郑州450000 [2]不详

出　　处：《国际检验医学杂志》2023年第7期769-774,780,共7页International Journal of Laboratory Medicine

基　　金：国家自然科学基金项目(81874042)。

摘　　要：目的探讨活动性肺结核与肺恶性肿瘤的差异因子,构建鉴别诊断的列线图预测模型并对其进行验证。方法回顾性分析郑州大学第一附属医院收治的463例活动性肺结核和肺恶性肿瘤患者的临床资料,根据入院时间将346例患者纳入训练组,117例患者纳入验证组。对训练组的临床资料进行单因素和多因素分析,筛选出差异因子并构建列线图预测模型。分别绘制训练组和验证组的受试者工作特征(ROC)曲线和校准曲线对模型进行评价。结果单因素分析和多因素分析结果表明,年龄、糖尿病、白细胞计数、血红蛋白、单核细胞与淋巴细胞比值、癌胚抗原、结核分枝杆菌感染T细胞斑点试验、肺内斑片或条絮影是鉴别活动性肺结核与肺恶性肿瘤的差异因子(P<0.05)。模型的灵敏度为81.0%,特异度为85.4%。列线图模型在训练组和验证组中ROC曲线下面积(AUC)分别为0.908和0.944,校准曲线与理想曲线较一致,模型区分度和校准度良好。结论鉴别活动性肺结核与肺恶性肿瘤的列线图预测模型,应用方便,准确度较好,具有一定的应用价值。Objective To explore the differential factors between active pulmonary tuberculosis and lung cancer and to construct and validate a nomograph prediction model.Methods The clinical data of totally 463 patients with active pulmonary tuberculosis and lung cancer admitted to the First Affiliated Hospital of Zhengzhou University were retrospectively analyzed.According to the admission time,346 patients were included in the training group and 117 patients in the validation group.The clinical data of the training group were analyzed by univariate analysis and multivariate analysis to screen out the difference factors and construct the nomogram prediction model.The receiver operating characteristic(ROC)curve and calibration curve of the training group and the verification group were drawn respectively to evaluate the model.Results The results of univariate analysis and multivariate analysis showed that age,diabetes,leukocyte count,hemoglobin,monocyte/lymphocyte ratio,carcinoembryonic antigen,T-SPOT.TB and patch and strip flocculent shadow in lung were the differential factors between active pulmonary tuberculosis and lung cancer(P<0.05).The sensitivity and specificity of the model were 81.0%and 85.4%,respectively.The area under ROC curve(AUC)of the prediction model in the training group and the validation group were 0.908 and 0.944.The calibration curve was consistent with the ideal curve,and the model had good differentiation and calibration.Conclusion The nomogram prediction model for differentiating active pulmonary tuberculosis from lung cancer is convenient and accurate,and has certain application value.

关 键 词：活动性肺结核 肺恶性肿瘤 列线图 

分 类 号：R734.2[医药卫生—肿瘤]