结直肠癌合并肝骨转移病人预后评分系统的开发和验证:一项回顾性队列研究

Development and validation of a prognostic scoring system for colorectal cancer patients with Hepato-bone metastasis:a retrospective study

目的 建立高效预测结直肠癌伴肝骨转移(CRCHBM)病人总生存期(OS)和癌症特异性生存率(CSS)的列线图模型。方法 回顾性分析来自美国国立癌症研究所和武汉协和医院肿瘤中心(WUHCC)数据库2010年至2019年的2239例CRCHBM病人。SEER随机分配至训练及内部验证队列,武汉数据库作为外部验证。采用单因素和多因素Cox回归分析确定影响病人OS和CSS的独立预后因素,并构建预测病人OS和CSS的列线图。使用校准曲线、曲线下面积(AUC)和决策曲线分析(DCA)评估列线图的临床实用价值。结果 基于年龄、肿瘤位置、分化程度、肿瘤直径、TNM分期、化疗、原发灶手术、淋巴结采样数和血清癌胚抗原(CEA)水平9个独立预测因素构建OS列线图。列线图预测1、3、5年OS的敏感度在训练组中为0.764、0.790、0.805,在内部验证组中为0.754、0.760、0.801,在外部验证组中为0.822、0.874、0.906。基于TNM分期、放疗和化疗3个独立预测因素构建CSS列线图。训练组1、3、5年的AUROC值分别为0.791、0.757、0.782;内部验证组分别为0.682、0.709、0.625,外部验证组分别为0.759、0.702、0.755。受试者工作特征(ROC)曲线、校准曲线和DCA结果显示,使用我们的模型预测OS和CSS比其他单一临床病理特征更有效。结论 基于显著临床病理特征构建的列线图可以方便地用于CRCHBM病人的术后个体化预测OS和CSS。

Objective To establish a nomogram model for efficiently predicting overall survival(OS)and cancer-specific survival(CSS)in patients with CRCHBM.Method 2239 patients from 2010 to 2019 were retrospectively analyzed from the Surveillance,Epidemiology,and End Results Program(SEER)databases and Wuhan Union Hospital Cancer Center.SEER is randomly assigned to the training and internal validation cohorts,and the Wuhan database serves as the external validation.Cox regression analyses were used to determine the independent clinicopathological prognosis factors affecting OS and CSS,and a nomogram was constructed to predict OS and CSS.The clinical utility of columnar plots was assessed using calibration curves,area under the curve(AUC),and decision curve analysis(DCA).Result OS column line graphs were constructed based on nine independent predictors:age,tumor location,degree of differentiation,tumor size,TNM stage,chemotherapy,primary focus surgery,number of lymph nodes sampled,and serum carcinoembryonic antigen(CEA)level.The C-index of the nomogram to predict the 1-,3-,and 5-year OS were 0.764,0.790,and 0.805 in the training group,0.754,0.760,and 0.801 in the internal validation group,and 0.822,0.874,and 0.906 in the external validation group.CSS column line graphs were constructed based on 3 independent predictors of TNM staging,radiotherapy and chemotherapy.The 1-,3-,and 5-year CSS AUROC values of the training group were 0.791,0.757,and 0.782,respectively.0.682,0.709,0.625 in the internal validation group and 0.759,0.702,0.755 in the external validation group,respectively.The results of receiver operating characteristic curve(ROC),ROC and DCA showed that the use of our model was more effective in predicting OS and CSS than other single clinicopathological features.Conclusion In summary,the nomogram based on significant clinicopathological features can be conveniently used to predict OS and CSS individually in patients with CRCHBM.