基于危险因素分析的难治性小细胞肺癌预测模型的构建与检验

Construction and validation of a predictive model for refractory small cell lung cancer based on risk factors analysis

目的:构建一项难治性小细胞肺癌(small cell lung cancer,SCLC)的预测模型并初步检验其预测效能。方法:收集2011年6月至2017年1月西安凤城医院、西安长安医院收治的333例小细胞肺癌患者为研究对象。根据一线化疗结束至疾病进展的间隔时间,将患者分为敏感组与难治组,分析可能影响难治性SCLC发生的相关因素。根据危险因素权重,构建预测模型,利用ROC曲线对其预测效能进行检验。结果:333例患者中,敏感性SCLC、难治性SCLC分别为241例(72.4%)、92例(27.6%)。多因素Logistic分析发现,吸烟指数、临床分期、化疗方案、是否合并SVCS、是否行胸部同步放疗是影响难治性SCLC发生的独立性因素,差异具有统计学意义(P<0.05)。以是否为敏感组为因变量,各独立性危险因素为自变量,建立回归模型。Logistic(P)=0.422-0.564×吸烟指数+0.921×SVCS-0.745×临床分期-0.540×化疗方案-0.807×胸部同步放疗。进行变换后,可得P=1/[1+e-Logistic(P)]。ROC曲线下面积为0.79,特异度为83.4%,灵敏度为79.4%。Hosmer-Lemeshow拟合优度检验显示,χ^2=1.52,P=0.93。模型拟合效度好,预测价值高。结论:多种危险因素参与了难治性SCLC的发生,临床应进行及时有效评估。构建的预测模型有较好的评估效能,有一定的临床应用价值。

Objective:To construct a predictive model of refractory small cell lung cancer(SCLC) and preliminarily test its predictive effectiveness.Methods:333 patients with SCLC were collected as the research subjects.According to the interval between the end of first-line chemotherapy and the progression of the disease,the patients were divided into sensitive and refractory groups,and the related factors that may affect the occurrence of refractory SCLC were analyzed.According to the weight of risk factors,a prediction model was constructed,and the ROC curve was used to test its predictive effectiveness.Results:Of the 333 patients,the sensitivity SCLC and refractory SCLC were 241 cases(72.4%) and 92 cases(27.6%) respectively.Multivariable analysis showed that smoking index,clinical stage,chemotherapy regimen,SVCS or not and chest synchronous radiotherapy or not were independent factors affecting the incidence of refractory SCLC(P<0.05).The area under the ROC curve of the model of P=1/[1+e-Logistic(P)].The area under the ROC curve was 0.79.The specificity was 83.4%,and the sensitivity was 79.4%.Goodness of fit test showed that χ^2=1.52,P=0.93.The model had good fitting validity and high prediction value.Conclusion:Multiple risk factors are involved in the occurrence of refractory SCLC,and clinical evaluation should be timely and effective.The prediction model constructed has good evaluation efficiency and has certain clinical application value.