目的:构建并验证子宫颈癌手术患者无病生存率(DFS)和总生存率(OS)的列线图预测模型,为评估子宫颈癌手术患者预后提供参考依据.方法:回顾性分析2013 年3 月至2018 年10 月在空军军医大学西京医院行子宫颈癌根治性手术患者的临床、病理及随访资料.基于Cox回归分析、贝叶斯信息准则的向后逐步选择法和R2 筛选变量,使用净重新分类指数和综合判别改进指数比较后,选取预测效能较好的列线图作为预测模型.使用一致性指数和受试者工作特征曲线(ROC)检验该预测模型的效能.结果:共纳入 950 例子宫颈癌患者.构建DFS列线图的危险因素为国际妇产科联盟(FIGO)分期(2018)、宫旁浸润、浸润深度和肿瘤最大径线,训练集和验证集的一致性指数(C-index)分别为0.754 和0.720,训练集 1、3、5 年的ROC曲线下面积(AUC)分别为 0.74(95%CI 0.65~0.82)、0.77(95%CI 0.71~0.83)、0.79(95%CI 0.74~0.85),验证集1、3、5 年的AUC分别为0.72(95%CI 0.58~0.87)、0.75(95%CI 0.64~0.86)、0.72(95%CI 0.61~0.84).构建OS列线图的危险因素为FIGO分期(2018)、组织学类型、淋巴脉管间隙浸润(LVSI)、宫旁浸润、手术切缘和浸润深度,训练集和验证集的一致性指数分别为 0.737 和 0.759,训练集 3、5 年的AUC 分别为0.76(95%CI 0.69~0.83)、0.78(95%CI 0.72~0.84),验证集 3、5 年的AUC分别为 0.76(95%CI 0.65~0.87)、0.79(95%CI 0.69~0.88).结论:本研究基于真实世界大数据构建的子宫颈癌1、3、5年DFS的列线图和3、5 年OS的列线图,具有理想的预测效果,有助于临床医师正确评估子宫颈癌手术患者的预后,对患者诊疗和预后评价提供有力的参考依据.
Objective:To develop and verify a nomogram to predict disease-free survival(DFS)and overall survival(OS)for patients undergoing cervical cancer surgery,which may provide reference for evaluating the prognosis of cervical cancer patients undergoing surgery.Methods:The clinical,pathological and follow-up data of patients who underwent radical operation for cervical cancer in Xijing Hospital,Air Force Medical University from March 2013 to October 2018 were analyzed retrospectively.Based on Cox regression analysis,Bayesian Informa-tion Criterion(BIC)backward stepwise selection method and R square screening variables,Net Reclassification Index(NRI)and Integrated Discrimination Improvement(IDI)were used to compare the predictive efficiency of the model,and a nomogram with better predictive efficiency was selected.The consistency index(C-index)and the receiver operating characteristic curve(ROC)were used to test the efficiency of the nomogram.Results:A total of 950 patients with cervical cancer were enrolled in this study.The risk factors for constructing the DFS nomogram were FIGO stage(2018),parametrium invasion,invasion depth,and maximum tumor diameter.The C-index for DFS in the training cohort and the verification cohort were 0.754 and 0.720,respectively.The area under ROC of the training cohort for 1-,3-and 5-years was 0.74(95%CI 0.65-0.82),0.77(95%CI 0.71-0.83)and 0.79(95%CI0.74-0.85),and the areas under ROC of verification cohort 1-,3-and 5-years were 0.72(95%CI 0.58-0.87),0.75(95%CI 0.64-0.86)and 0.72(95%CI 0.61-0.84),respectively.The risk factors for con-structing the OS nomogram were FIGO stage(2018),histological type,LVSI,parametrium invasion,surgical mar-gin,and invasion depth.The C-index for OS in the training cohort and the verification cohort were 0.737 and 0.759,respectively.The area under ROC of the 3-and 5-year training cohort were 0.76(95%CI 0.69-0.83)and 0.78(95%CI 0.72-0.84),and the areas under ROC of verification cohort 3-and 5-years were 0.76(95%CI 0.65-0.87)and 0.79(95%CI 0.69-0.88),respectively.Conclusions:This study is based on real-world big data to construct nomogram of DFS for 1,3,and 5 years and OS for 3,and 5 years for cervical cancer,which have ideal predictive effects and help clinical physicians correctly evaluate the prognosis of cervical cancer surgery patients.It provides strong reference basis for diagnosis,treatment,and prognosis evaluation.