目的 利用监测、流行病学和最终结果(the Surveillance,Epidemiology,and End Results,SEER)数据库建立并验证直肠印戒细胞癌发生远处转移风险的列线图预测模型.方法 回顾性分析SEER数据库中2004-2015年期间诊断为结直肠印戒细胞癌患者的临床病理资料,并按7∶3的比例将纳入研究的患者随机分为训练集和验证集.在训练集中通过单因素和多因素logistic回归分析筛选出结直肠印戒细胞癌发生远处转移(distant metastasis,DM)的独立风险因素并构建列线图预测模型.应用受试者工作(receiver operating characteristic,ROC)曲线下面积(area under curve,AUC)、校准曲线和临床决策曲线评价预测模型的区分度、校准度和临床实用性.结果 总共纳入2 595例结直肠印戒细胞癌患者,其中1 022例(39.4％)发生了 DM.根据单因素及多因素logistic回归分析发现性别、年龄、T分期、N分期、手术治疗、放疗和化疗是结直肠印戒细胞癌DM的独立风险因素(P＜0.05).根据以上独立风险因素构建了结直肠印戒细胞癌发生DM的列线图预测模型,该列线图模型的ROC曲线的AUC在训练集和验证集中分别为0.78[95％CI为(0.76,0.80)]和0.77[95％CI为(0.74,0.81];校准曲线在训练集和验证集中显示了良好的拟合度,Hosmer-Lemeshow检验结果分别为:x2=9.43,P=0.31和x2=12.47,P=0.13;临床决策曲线分析(decision curve analysis,DCA)显示,当训练集和验证集的阈概率分别在10％～95％和11％～90％范围内时,该模型具有净获益.结论 本研究所构建的列线图预测模型具有较高的准确性和可靠性,可用于结直肠印戒细胞癌DM的早期干预和风险预测.
Establishment and validation of a nomogram prediction model for distant metastasis risk of colorectal signet-ring cell carcinoma based on SEER database
Objective To establish and validate a predictive nomogram for predicting the risk of distant metastasis in colorectal signet-ring cell carcinoma based on the Surveillance,Epidemiology,and End Results(SEER)database.Methods A retrospective analysis was conducted on clinical and pathological data of patients diagnosed with colorectal signet-ring cell carcinoma in the SEER database from 2004 to 2015,and they were randomly divided into training and validation sets at a ratio of 7∶3.Independent risk factors for distant metastasis(DM)in colorectal signet-ring cell carcinoma were screened out in the training set through univariate and multivariate logistic regression analysis,and a nomogram was constructed.The area under the curve(AUC)of the receiver operating characteristic(ROC)curve,calibration curve,and decision curve analysis(DCA)were used to evaluate the discrimination,calibration,and clinical utility of the nomogram model.Results A total of 2 595 patients with colorectal signet-ring cell carcinoma were included,of whom 1 022(39.4％)had DM.According to the univariate and multivariate logistic regression analysis,gender,age,T stage,N stage,surgical treatment,radiotherapy and chemotherapy were independent risk factors for DM of colorectal signet-ring cell carcinoma(P＜0.05).Based on the above independent risk factors,a nomogram for DM of colorectal signet-ring cell carcinoma was constructed.The nomogram AUC of the ROC was 0.78[95％CI(0.76,0.80)]and 0.77[95％CI(0.74,0.81)]in the training and validation sets,respectively.The calibration curves showed a good fit in the training and validation sets,with the Hosmer-Lemeshow test results being x2=9.43,P=0.31 and x2=12.47,P=0.13,respectively.The DCA curves showed that the model had a net benefit when the threshold probabilities of the training and validation sets were in the range of 10％-95％and 11％-990％,respectively.Conclusion The nomogram constructed in this study exhibits higher accuracy and reliability,and can be used for early intervention and risk prediction of DM in colorectal signet-ring cell carcinoma.
NETL
NSTL
万方数据
