结直肠癌肝转移患者预后列线图模型实证研究

Empirical study of the prognostic nomogram model in patients with liver metastasis of colorectal cancer

目的 通过建立列线图生存预测模型,探讨影响结直肠癌肝转移患者的相关因素,更好的为结直肠癌肝转移患者的个性化治疗及预后判断提供指导依据。方法 回顾性分析2010-01-01-2019-12-31就诊于新疆医科大学附属肿瘤医院的1 131例同时性结直肠癌肝转移患者。使用Cox比例风险回归确定患者的独立预后因素。通过R软件采用随机抽样法将患者按照7∶3的比例分为建模组(792例)和验证组(339例),构建列线图预测总生存期,并绘制受试者工作特征(ROC)曲线和校准曲线进行验证。结果 年龄>60岁(HR=0.65,95%CI:0.46~0.91)、临床分期Ⅳ期(HR=0.66,95%CI:0.45~0.96),N分期N2b期(HR=2.54,95%CI:1.16~5.55)、淋巴结转移率(LNR)>0.6(HR=1.63,95%CI:0.95~2.78)和微卫星高度不稳定状态(HR=0.39,95%CI:0.14~1.09)为结直肠癌肝转移患者预后的独立危险因素,均P<0.1。纳入独立危险因素构建患者1、3和5年生存率列线图。建模组1、3和5年生存率ROC值分别为0.621、0.612和0.587;验证组分别为0.602、0.657和0.596。总生存期的校准曲线相对接近理想45°参考线,表明模型的预测值与实际值吻合良好,模型具有良好的临床有效性。结论 本研究构建的列线图模型具有较好的预测效果,有利于临床实践中为患者制定针对性、个性化的治疗方案,提高患者有效生存率。

Objective By establishing a column chart survival prediction model,this study explores the relevant factors affecting patients with liver metastasis from colorectal cancer,and provided better guidance for personalized treatment and prognosis judgment of patients with liver metastasis from colorectal cancer.Methods Retrospective analysis was conducted on 1131patients with simultaneous liver metastasis and colorectal cancer who were admitted in Xinjiang Cancer Center from January 1,2010to December 31,2019.Cox proportional risk regression was used to determine independent prognostic factors for patients.R software and random sampling method were used to divide patients into a modeling group(792cases)and a validation group(339cases)in a 7∶3ratio.A nomographic chart was constructed to predict overall survival,and receiver operating characteristic(ROC)curves and calibration curves were plotted for validation.Results Age>60years old(HR=0.65,95%CI:0.46-0.91),clinical stageⅣ(HR=0.66,95%CI:0.45-0.96),N_(2b)stage(HR=2.54,95%CI:1.16-5.55),lymph node metastasis rate(LNR)>0.6(HR=1.63,95%CI:0.95-2.78),and highly unstable microsatellite status(HR=0.39,95%CI:0.14-1.09)were independent risk factors for the prognosis of colorectal cancer patients with liver metastasis,all P<0.1.Independent risk factors were included to construct patients'1-year,3-year,and 5-year survival nomograms.The ROC values of overall survival for modeling groups at1,3,and 5years were 0.621,0.612and 0.587,respectively;Those in the validation groups were 0.602,0.657and0.596,respectively.The calibration curve of the overall survival was relatively close to the ideal 45°reference line,indicating that the predicted values of the model were in good agreement with the actual values,and the model has good clinical effectiveness.Conclusion The column chart model constructed in this study has good predictive performance,which is conducive to developing targeted and personalized treatment plans for patients in clinical practice and improving their effective survival rate.