基于SEER数据库肺腺鳞癌淋巴结转移预测模型的构建

Construction of Lymph Node Metastasis Prediction Model of Lung Adenosquamous Carcinoma Based on SEER Database

目的构建肺腺鳞癌淋巴结转移的预后模型,用于预测肺腺鳞癌淋巴结转移的危险程度。方法首先筛选SEER数据库中2669例肺腺鳞癌患者的临床资料,之后将总数据集按照7∶3的比例随机分为训练集(n=1869)和测试集(n=800)。对训练集的性别、年龄、种族、肿瘤位置、分化程度、T分期、M分期以及肿瘤最大径进行单因素和多因素logistic回归分析,筛选出P<0.05的变量用于模型的构建并绘制列线图,之后绘制受试者工作特征(ROC)曲线和校准曲线对模型的诊断价值进行评判。结果多因素logistic回归分析显示,年龄、肿瘤位置、分化程度、T分期、M分期以及肿瘤最大径均为肺腺鳞癌淋巴结转移的独立影响因素(P<0.05)。训练集和测试集的一致性指数分别为0.766(95%CI:0.744-0.788)、0.742(95%CI:0.707-0.777),两者的校准曲线均显示了模型具有良好的一致性。结论由年龄、肿瘤位置、分化程度、T分期、M分期以及肿瘤最大径构建的预测模型对肺腺鳞癌淋巴结转移的预测准确且敏感,具有一定的临床指导作用。

Objective To construct a prognostic model for lymph node metastasis of lung adenosquamous carcinoma to predict the risk degree of lymph node metastasis in lung adenosquamous carcinoma.Methods Firstly,the clinical data of 2669 patients with lung adenosquamous carcinoma in the SEER database were screened,and then the total data set was randomly divided into training set(n=1869)and test set(n=800)at a ratio of 7 to 3.The gender,age,race,tumor location,differentiation degree,T staging,M staging and maximum diameter of tumor of the training set were analyzed by univariate and multi variate logistic regression analysis,and the variables with P<0.05 were selected for model construction and plotted and listed line graph,and then draw receiver operating characteristic(ROC)curve and calibration curve to judge the diagnostic value of the model.Results Multivariate logistic regression analysis showed that age,tumor location,degree of differentiation,T stage,M stage and maximum diameter of tumor were independent influencing factors of lymph node metastasis of lung adenosquamous carcinoma(P<0.05).The consistency indexes of the training set and the test set were 0.766(95%CI:0.744 to 0.788)and 0.742(95%CI:0.707 to 0.777),respectively.The calibration curves of both of them showed that the model had good consistency.Condusion The prediction model constructed by age,tumor location,degree of differentiation,T stage,M stage and maximum diameter of tumor is accurate and sensitive in predicting lymph node metastasis of lung adenosquamous carcinoma,and has certain clinical guidance.