基于美国国家癌症研究所监测、流行病学、结果数据库食管癌病人预后影响因素分析

Factors affecting prognosis of esophageal cancer patients based on SEER database

目的剖析影响食管癌病人生存的临床参数,构建列线图以预测食管癌病人的1年、3年及5年生存率。方法利用美国国家癌症研究所监测、流行病学、结果数据库(National Cancer Institute,The Surveillance Epidemiology and End Results pro⁃gram,SEER),共获2010年1月至2015年12月8863例食管癌病人的年龄、病理等临床资料和相关随访数据。按随机数字表法分为验证组(2656例)和列线图建模组(6207例)(分配比例是3∶7)。建模组用Kaplan-Meier进行单因素生存分析,用log-rank检验法评估生存率的差异;将在单因素分析中差异有统计学意义的变量纳入多因素Cox比例风险模型,寻找建模组病人预后独立影响因素,将确定因素纳入并构建列线图,预测食管癌病人1年、3年及5年生存率。通过一致性指数(C-index)和校正曲线评估其预测的准确性和判别能力。结果食管癌病人的年龄、性别、种族、组织分级、美国癌症联合协会(AJCC)第7版T分期、N分期、M分期、手术、化疗均是其预后独立影响因素(P<0.05),将以上因素纳入并构建列线图。C-index:建模组内部验证为0.726(95%CI:0.718~0.734),验证组外部验证为0.723(95%CI:0.711~0.735);同时两组的校正曲线一致性良好。结论分析及构建食管癌病人生存风险的列线图预测精度良好,可帮助临床医师根据食管癌病人综合临床资料作出较准确的预后评估。

Objective To predict the 1-year,3-year and 5-year survival rates of patients with esophageal cancer and construct a No⁃mogram by analyzing the clinical parameters that affect the survival.Methods The age,clinicopathological and follow-up data of 8863 patients with esophageal cancer between January 2010 and December 2015 were collected from the National Cancer Institute's Surveillance Epidemiology and End Results(SEER)database.They were assigned into two groups according to the method of random number table:verification group(2656 patients)and Nomogram modeling group(6207 patients)(the allocation ratio was 3∶7).In No⁃mogram modeling group,the Kaplan-Meier method was used to perform the single factor survival analysis,and log-rank test method was employed to assess the differences in survival rate.Then variables with statistically significant difference were included in the multivariate Cox proportional hazards model to find out the independent prognostic influencing factors,which were used to construct Nomogram to predict the 1-year,3-year and 5-year survival rates of patients with esophageal cancer.Then the Harrell's C-index and calibration curve were used to estimate the accuracy and discrimination ability of the Nomogram system.Results The independent prognostic factors were age,gender,race,tissue grade,The American Joint Committee on Cancer(AJCC)7th T stage,N stage,M stage,operation and chemotherapy(P<0.05),which were all included and construct Nomogram.Concordance index(C-index):the internal validation of the modeling group was 0.726(95%CI:0.718-0.734),and the external validation of the verification group was 0.723(95%CI:0.711-0.735);meanwhile,the calibration curves of the two groups were in good agreement.Conclusion The Nomogram showed good accuracy in pre⁃dicting survival rate of patients with esophageal cancer,which can help clinicians make more accurate prognostic evaluation.