基于SEER数据库预测前列腺癌患者生存率列线图的建立和验证

Development and validation of a nomogram for predicting survival outcomes in prostate cancer patients based on SEER database

目的 基于SEER数据库建立并验证前列腺癌(PCa)患者总生存率(OS)的列线图模型。方法 从SEER数据库中提取2010-2015年诊断为PCa患者12 642例,将患者以7∶3的分割比随机分为建模组(8 850例)和验证组(3 792例),采用单因素Cox比例风险回归、lasso回归、多因素Cox比例风险回归分析出PCa患者OS的独立危险因素,最终构建列线图来预测PCa患者1、3、5年的OS。通过一致性指数(C-index)、绘制校准曲线及受试者工作特征(ROC)曲线来评估模型的预测能力。结果 多因素Cox回归分析显示年龄、T分期、N分期、M分期、骨转移、肝转移、是否行区域淋巴结切除为PCa患者OS的独立危险因素(P<0.05)。用这7个因素构建PCa患者OS列线图模型,最终建模组C指数为0.750,1、3、5年的ROC曲线下面积分别为0.77、0.77、0.76;验证组C指数为0.765,1、3、5年的ROC曲线下面积分别为0.83、0.79、0.76。建模组和验证组的校准曲线表明该模型生存预测率与实际的生存预测率有较好的一致性。基于此列线图模型将患者进行风险分层,可观察到高风险组患者的OS显著低于低风险组(P<0.001)。结论 本研究构建的列线图可以用于预测PCa患者的预后情况,对患者的个体化治疗具有重要意义。

Objective To establish and verify a nomogram model of overall survival(OS)of prostate cancer patients based on the SEER data.Methods A total of 12642 patients diagnosed with prostate cancer during 2010 and 2015 were extracted from the SEER database.Patients were randomly divided into the model group(n=8850)and validation group(n=3792).The independent risk factors for OS were analyzed with univariate Cox proportional risk regression,lasso regression and multivariate Cox proportional risk regression.A nomogram was constructed to predict the 1-year,3-year and 5-year OS.The prediction potential of the model was evaluated with the consistency index(C-index),calibration curve and receiver operating characteristic(ROC)curve.Results Multivariate Cox regression analysis showed that age,T stage,N stage,M stage,bone metastasis,liver metastasis and regional lymphadenectomy were independent risk factors for OS(P<0.05).The seven factors were used to construct an OS nomogram model.The C-index of the modeling set was 0.750,and the area under the ROC curve(AUC)at 1,3 and 5 years were 0.77,0.77 and 0.76,respectively;the C-index of the validation set was 0.765,and the AUC at 1,3 and 5 years were 0.83,0.79 and 0.76,respectively.The calibration curves of the modelling set and validation set showed a good agreement with the actual survival prediction rate.Risk stratification of patients based on the nomogram model showed that the OS of patients in the high-risk group was significantly lower than that in the low-risk group(P<0.001).Conclusion The nomogram can be used to predict the prognosis of prostate cancer patients,and is important for individualized treatment plans.