摘要
目的 基于SEER数据库构建局部进展期肾癌患者列线图生存预测模型,为未来局部进展期肾癌患者术后预后研究提供参考.方法 从SEER数据库中下载相关病例数据(n=7893),采用简单随机分组法将数据按7∶3的比例分为实验组和验证组,利用统计学方法分析数据中所包含的临床病例信息,筛选影响患者预后的独立危险因素,并绘制局部进展期肾癌患者总体生存时间(OS)及肿瘤特异性生存时间(CSS)列线图模型.运用一致性指数、曲线下面积、内外部验证以及校准曲线评估模型效能.结果 患者年龄、肿瘤大小、疾病进展类型、TNM分期、阳性淋巴结数目、婚姻状态以及病理类型与患者OS及CSS显著相关(P＜0.001),利用上述预测因子,构建患者OS列线图模型并内部验证1、3、5年受试者工作特征曲线下面积(AUC)分别为0.809、0.721及0.715,构建CSS列线图模型并内部验证1、3、5年AUC分别为0.802、0.745和0.735;患者OS列线图模型外部验证1、3、5年AUC分别为0.792、0.628及0.620,CSS列线图外部验证1、3、5年AUC分别为0.943、0.803和0.737,模型区分度与准确度良好.结论 本列线图模型预测性能良好,对临床个体化治疗具有一定指导意义.
Abstract
Objective To construct a nomogram survival prediction model for patients with locally advanced renal cell carcinoma based on SEER database(n=7893),so as to provide reference for future prognosis study.Methods Case data were downloaded from the SEER database,and divided into the experimental group and validation group with a ratio of 7∶3 by simple randomization.The clinical information was analyzed,independent risk factors influencing prognosis were screened,and the overall survival(OS)and tumor-specific survival(CSS)were mapped.Model performance was evaluated using consistency index,area under the receiver operating characteristic curve(AUC),internal and external validation,and calibration curves.Results Patients'age,tumor size,disease progression tpye,TNM stage,number of positive lymph nodes,marital status and pathological type were significantly correlated with OS and CSS(P＜0.01).Based on the above predictors,the internal verification AUC of the 1-,3-and 5-year OS nomogram model was 0.809,0.721 and 0.715,respectively.The internal validation AUC of the nomogram model for 1-,3-and 5-year CSS was 0.802,0.745 and 0.735,respectively.The external validation AUC of the OS nomogram model was 0.792,0.628 and 0.620 at 1,3 and 5 years,respectively,and the external validation AUC of CSS was 0.943,0.803 and 0.737 at 1.3 and 5 years,respectively,showing good model differentiation and accuracy.Conclusion The prediction performance of the nomogram model is good,and it can provide reference for individualized treatment.
NETL
NSTL
万方数据