原发性胸下段和腹部食管腺癌患者生存预后的列线图模型构建和验证

Construction and validation of a nomogram predictive model for survival and prognosis of patients with primary distal thoracic and abdominal esophageal adenocarcinoma

目的探讨原发性胸下段和腹部食管腺癌患者生存预后的独立影响因素,并构建列线图模型进行分析与验证。方法回顾性分析2008年1月至2018年12月在中山大学附属第八医院接受治疗的1537例原发性胸下段和腹部食管腺癌患者资料,按7∶3的比例分为训练集(n=1076)和验证集(n=461)。描述训练集的生存状况,并进行单因素和多因素分析。针对训练集数据,建立生存预测列线图模型,并进行内外部验证。结果训练集和验证集患者的年龄、性别等差异无统计学意义(P>0.05)。训练集中,5年生存率为17.19%(185/1076),平均生存期为38.00(22.00~56.00)个月;单因素分析结果显示,患者确诊时的年龄、组织分化程度、TNM分期、手术、放化疗情况以及淋巴结病理活检结果与患者的生存状况有关(P<0.05);多因素分析结果显示,患者确诊时的年龄、组织分化程度、TNM分期、手术、放化疗情况是生存状况的独立影响因素(P<0.05)。基于训练集构建的列线图风险预测模型C值为0.735±0.014,患者1年、3年、5年生存情况预测曲线下面积分别为0.720、0.740、0.765。结论基于原发性胸下段和腹部食管腺癌患者生存预后的独立影响因素构建的列线图模型对患者的生存预测具有良好的效能。

Objective:To investigate the independent influencing factorsof survival and prognosis of patients with distal thoracic and abdominal esophageal adenocarcinoma,and construct a nomogram model for analysis and verification.Methods:A retrospective analysis was conducted on 1537 patient records from the hospital between January 2008 and December 2018,divided into a training set(n=1076)and a validation set(n=461)at a 7∶3 ratio.Survival status was assessed,and univariate and multivariate analyses were performed on the training set.A survival prediction nomogram was then developed and validated internally and externally.Results:There were no significant differences in age or gender between the training and validation sets(P>0.05).The 5-year survival rate in the training set was 17.19%(185/1076),with a median survival time of 38.00 months(range:22.00-56.00 months).Univariate and multivariate analyses indicated that age at diagnosis,degree of tissue differentiation,TNM staging,surgical intervention and radiotherapy and chemotherapy,along with lymph node biopsy results,significantly influenced survival(P<0.05).The C value of the nomogram risk prediction model based on the training set was 0.735±0.014,and the areas under the curve for predicting the 1-year,3-year and 5-year survival conditions of thepatients were 0.720,0.740and 0.765,respectively.Conclusion:The nomogram model based on the independent influencing factors of survival prognosis in patients with primary lower thoracic segment and abdominal esophageal adenocarcinoma has good efficacy in predicting the survival of thepatients.