基于SEER数据库构建侵袭性导管内乳头状黏液性肿瘤的癌症特异性生存预测模型  

Construction of a cancer specific survival prediction model for invasive intraductal papillary mucinous neoplasm based on SEER database

作　　者：李微波 张正乐 汤志刚[1] LI Weibo;ZHANG Zhengle;TANG Zhigang(Dept.of General Surgery,Renmin Hospital of Wuhan University,Wuhan 430060,Hubei,China)

机构地区：[1]武汉大学人民医院普外科,湖北武汉430060

出　　处：《武汉大学学报（医学版）》2023年第9期1086-1092,1108,共8页Medical Journal of Wuhan University

基　　金：国家自然科学基金面上项目(编号:82172570)。

摘　　要：目的:确定影响导管内乳头状黏液性肿瘤(IPMN)患者生存的独立危险因素,并构建有效的生存预测模型指导临床实践。方法:从SEER数据库中回顾性提取2010—2015年间共727例侵袭性IPMN的相关资料。以单因素Cox回归分析联合多因素Cox回归分析确定与患者预后相关的独立影响因素,再以确定的独立影响因素构建诺模图;用C指数和曲线下面积(AUC)来评价模型的准确性和区分性;用校准曲线评价实际值和预测值之间的一致性;用以决策曲线分析(DCA)进一步评价模型的临床应用能力;最后采用Kaplan⁃Meier法绘制生存曲线,并以log⁃rank检验比较组间差异。结果:多因素Cox回归分析表明患者年龄、组织分化程度、AJCC分期以及淋巴结状态是影响侵袭性IPMN患者预后的独立危险因素;区域淋巴结清扫数量和放疗是影响侵袭性IPMN患者预后的保护性因素;构建预测患者1年、3年及5年生存的诺模图使得数据分析结果可视化;1年、3年及5年生存的受试者工作特征(ROC)曲线的AUC值分别为0.808、0.826和0.850;模型的C指数为0.741(95%CI:0.716~0.766);1年、3年及5年生存的校准曲线提示模型的实际值和预测值之间具有良好的一致性;1年、3年及5年生存的DCA表明模型具有的良好的应用潜能。结论:本研究建立的生存预测模型具有良好的预测能力和应用前景,可以为临床实践提供个性化的医疗决策。Objective:To identify independent risk factors for survival of patients with invasive intraductal papillary mucinous neoplasm(IPMN)and develop a valid predictive model to guide clinical practice.Methods:We screened 727 invasive IPMN patients from the Surveillance,Epidemiology,and End Results(SEER)database between 2010 and 2015.Univariate cox regression analysis combined with multivariate cox regression analysis were used to determine the independent influence factors for prog⁃nosis of invasive IPMN patients.A nomogram predicting the survival of the patients with invasive IPMN was constructed.Area under the curve(AUC)of receiver operating characteristic(ROC)curve and consistency⁃index(C⁃index)were used to determine the accuracy and discriminate ability of the no⁃mogram.Calibration curve was used to assess the agreement between the predicted probability of the model and the actual probability.In addition,decision curve analysis(DCA)were employed to further assess the clinical utility of the nomogram.Finally,survival curve was plotted by Kaplan⁃Meier meth⁃od,and the log⁃rank test was used to compare the differences between groups.Results:Multivariate cox regression analysis revealed that independent risk factors for prognosis of IPMN patients included age,histological grade,AJCC stage,and lymph node status.And the protective factors for prognosis of IPMN patients included the number of dissected lymph nodes and radiotherapy.A nomogram was constructed to visualize the results of the data analysis.The AUC of ROC curve for 1⁃year,3⁃year,and 5⁃year survival was 0.808,0.826 and 0.850 respectively;The C⁃index of the model was 0.741(95%CI:0.716⁃0.766);The calibration curve showed that the predicted probability of the model was in high agreement with the actual predicted probability;The DCA showed that the model had great po⁃tential clinical utility.Conclusion:The prognosis model established in this study has a good predictive performance and a promising potential application,which can provide pers

关 键 词：导管内乳头状黏液性肿瘤 生存分析 诺模图 SEER数据库 

分 类 号：R735.9[医药卫生—肿瘤]