基于SEER数据库构建并验证腹膜后肿瘤的预后评价预测模型

Construction and validation of prediction model for retroperitoneal tumor based on SEER database

目的基于监测、流行病学和最终结果数据库(SEER)记录的腹膜后肿瘤数据, 构建并验证腹膜后肿瘤患者生存预后预测模型。方法收集2000至2019年腹膜后肿瘤临床数据, 使用X-title软件计算年龄的最佳截断值, 使用R语言将数据分为建模组和验证组。使用单因素和多因素COX回归分析构建腹膜后肿瘤患者预后预测模型。采用一致性指数(C指数)、受试者工作特征曲线(ROC曲线)和校准曲线对列线图进行评估。将本研究构建的模型与第7版美国癌症联合委员会(AJCC)TNM分期系统进行比较, 采用净重分类改善指数(NRI)和综合判别改善指数(IDI)对模型性能进行评价。结果纳入6个风险因素构建列线图用以预测患者3年和5年存活率。建模组C指数0.742, 验证组C指数0.737。建模组预测3年和5年生存率的ROC曲线下面积(AUC)分别为0.817和0.794, 验证组分别是0.779和0.789。校准曲线显示建模组和验证组预测3年和5年生存率和实际生存率拟合较好。NRI和IDI结果说明本研究预测模型与第7版AJCC的TNM分期比较具有正改善。以此模型构建的风险分层模型中低风险组预后好于高风险组(P=0), 差异有统计学意义。结论本研究成功构建腹膜后肿瘤预测模型并进行验证, 为患者3年和5年生存率预测提供一个简便可靠的工具。同时, 以本预测模型构建的风险分层模型可以区分不同风险的患者, 对腹膜后肿瘤患者的个体化治疗具有重大意义。

Objective Based on the data of retroperitoneal tumor recorded in Surveillance,Epidemiology,and End Results(SEER)database,to construct and verify the survival prediction model for retroperitoneal tumor.Methods The clinical data from patients diagnosed with retroperitoneal tumors between 2000 and 2019 were collecte.The X-title software was used to calculate the optimal cut-off value of age.The subjects were divided into a training group and a validation group using The R Programming Language.The independent prognostic factors in retroperitoneal tumor were determined and the prediction model for retroperitoneal tumor was constructed using univariate Cox regression and multivariate Cox regression.C index,receiver operating characteristic(ROC)curve and calibration curve were used to evaluate the accuracy.Net reclassification index(NRI)and integrated discrimination improvement(IDI)were used to compare the model constructed in this study with the 7th American Joint Committee on cancer(AJCC)TNM staging system.Results A total of 6 risk factors affecting the prognosis of retroperitoneal tumor patients were used according to COX regression analysis,and a nomogram was constructed to predict the 3-and 5-year survival.The C-index of training group was 0.742 and that of validation group was 0.737.The area under curve(AUC)for predicting 3-and 5-year survival was 0.817 and 0.794 in training group and 0.779 and 0.789 in validation group,respectively.The calibration curves showed a good fit between the predicted 3-year and 5-year survival and the actual survival.NRI and IDI results indicated that the prediction model in this study was better than 7th AJCC TNM staging system.The prognosis of the risk stratification model constructed by this model was significantly better in the low-risk group than in the high-risk group(P=0),which could distinguish patients with different risks.Conclusion In this study,the prediction model of retroperitoneal tumor was successfully constructed and verified,which provides a simple and reliable tool for predicting 3-and 5-year survival of patients with retroperitoneal tumor.At the same time,the risk stratification model constructed by this prediction model can screen patients with different risks,which is of great significance for the individualized treatment of patients with retroperitoneal tumor.