摘要
目的 基于监测、流行病学和最终结果数据库(SEER)开发和验证列线图,以预测食管癌的早期死亡风险。方法 基于SEER数据库收集2010年至2015年诊断的食管癌,通过X-tile分析获得最佳年龄阈值,通过多因素Logistic回归分析确定与早期死亡(食管癌诊断后存活≤3个月)相关的危险因素,危险因素被用来构建预测列线图,采用受试者工作特征(ROC)曲线进行模型的验证。结果 SEER数据库中符合入组标准的食管癌患者中,1 261例(15.5%)涉及全因早期死亡,1 133例(13.9%)涉及肿瘤特异性早期死亡。多因素Logistic回归分析显示,高龄(P均<0.001)、未婚(P均<0.001)、鳞癌(P均=0.001)、较高组织学分级(P均<0.001)、脑转移(P=0.009、P=0.012)、肝转移(P均<0.001)、骨转移(P均=0.001)、未放疗(P均=0.001)、未化疗(P均<0.001)是预测全因早期死亡和肿瘤特异性早期死亡的危险因素,肺转移(P=0.039)仅是预测全因早期死亡的危险因素。全因早期死亡和肿瘤特异性早期死亡列线图的曲线下面积分别为0.830(95%CI:0.826~0.851)和0.833(95%CI:0.823~0.848)。结论 本研究成功构建了基于SEER数据库预测食管癌患者早期死亡的列线图,这可能有助于临床医生筛查高危病例并实施适当治疗。
Objective To develop and validate nomograms based on the surveillance,epidemiology,and end results(SEER)database,to predict early death from esophageal cancer.Methods All the esophageal cancer patients diagnosed between 2010 and 2015 were collected from SEER database.The best age threshold was obtained through X-tile analysis.Risk factors associated with early death(≤3 months survival after esophageal cancer diagnosis)were identified via multivariate Logistic regression analysis.Those deemed significant were used to construct prediction nomograms,and the model was validated by using receiver operating characteristic curve.Results Among the esophageal cancer cases on the SEER database that fulfilled our eligibility criteria,1261(15.5%)involved all-cause early death and 1133(13.9%)involved cancer-specific early death.The factors that significantly predicted all-cause and cancer-specific early death in the multivariate Logistic analysis were older age(P<0.001),unmarried status(P<0.001),histological type(squamous cell carcinoma)(P=0.001),higher histological grade(P<0.001),brain metastases(P=0.009,P=0.012),liver metastases(P<0.001),bone metastases(P<0.001),no radiotherapy(P=0.001),no chemotherapy(P<0.001);lung metastases was the factor that significantly predicted all-cause early death(P=0.039).The area under the curve for the all-cause and cancer-specific early death nomograms were 0.830(95%CI:0.826-0.851)and 0.833(95%CI:0.823-0.848),respectively.Conclusion Nomograms that predict early death in esophageal cancer patients is constructed based on SEER database,which may assist clinicians in screening for high-risk patieints and implementing appropriate treatment strategies.
作者
张小杰
王子兵
ZHANG Xiaojie;WANG Zibing(Department of Immunotherapy,Henan Provincial Cancer Hospital,Zhengzhou 450003,China)
出处
《肿瘤基础与临床》
2023年第4期297-302,共6页
journal of basic and clinical oncology
关键词
食管癌
早期死亡
列线图
监测、流行病学和最终结果数据库
预后
esophageal cancer
early death
nomogram
surveillance,epidemiology,and end results database
prognosis
