乳腺髓样癌临床预测模型的建立和验证:基于SEER数据库

Establishment and validation of a clinical prediction model for medullary carcinoma of breast based on the Surveillance,Epidemiology,and End Results database

目的 基于美国监测,流行病学和最终结果(Surveillance,Epidemiology,and End Results,SEER)数据库,建立并验证预测乳腺髓样癌(medullary carcinoma of breast,MCB)患者5年和10年生存率的列线图模型。方法 从SEER数据库中收集2000年至2015年诊断为MCB的患者临床资料,将患者按7∶3的比例随机分配至训练集和验证集。通过单因素和多因素Cox回归分析确定MCB患者预后的独立危险因素,并构建列线图预测MCB患者的5年和10年生存率。采用一致性指数(C指数)、受试者工作特征曲线下面积(area under the receiver operating characteristic curve,AUC)、校准图和决策曲线分析(decision curve analysis,DCA),比较列线图模型和美国癌症联合委员会(The American Joint Committee on Cancer,AJCC)分期系统之间的预测性能。结果共纳入2 086例MCB患者,其中训练集1 460例、验证集626例。单因素和多因素Cox回归分析显示诊断年龄、婚姻状况、AJCC分期和手术是影响MCB患者生存的危险因素,将它们纳入列线图并进行验证,训练集中,列线图显示5年总体生存率AUC为0.698,10年总体生存率AUC为0.707;验证集中,列线图显示5年总体生存率AUC为0.748,10年总体生存率AUC为0.729。校准图显示列线图模型的预测结果与实际观察结果有较好的一致性,DCA结果显示列线图在训练集和验证集中产生了较好的净收益。结论 通过Cox回归分析构建了MCB患者5年和10年生存率的列线图预测模型,经验证列线图有较好的预测性能。该模型不仅为患者生存预后提供更加个性化的评估,也能为临床医生选择个性化治疗方案提供参考。

Objective To establish and validate a nomogram model for predicting 5-year and 10-year survival rates of patients with medullary carcinoma of breast based on the Surveillance,Epidemiology,and End Results(SEER)database.Methods The clinical data of patients diagnosed with MCB between 2000 to 2015 were collected from the SEER database,and the patients were randomized to the training set and validation set in a 7/3 ratio.Independent risk factors for the prognosis of MCB patients were determined by univariate and multivariate Cox regression analysis,and the nomogram was constructed based on the independent risk factors to predict the patients’5-year and 10-year survival rates.Using the index of concordance,area under the receiver operating characteristic curve(AUC),the calibration plots and the decision curve analysis(DCA),the prediction performance between the nomogram model and the American Joint Committee on Cancer(AJCC)staging system were compared.Results A total of 2,086 MCB patients were included:1,460 in the training set and 626 in the validation set.Univariate and multivariate Cox regression analysis showed that age of diagnosis,marital status,AJCC stage and surgery were risk factors associated with survival in patients with MCB,these risk factors were included in the nomogram and validated.In the training set,the nomogram showed 5-year overall survival AUC=0.698,10-year overall survival AUC=0.707;in the validation set,the nomogram showed 5-year overall survival AUC=0.748,and 10-year overall survival AUC=0.729.The calibration plot showed that the prediction results of the nomogram model were in good agreement with the actual observations,and the clinical decision curve analysis yielded good net gains in the training and validation sets.Conclusion The nomogram prediction model of 5-and 10-year survival of MCB patients constructed using Cox regression analysis has good predictive performance.This model not only provides a more personalized assessment of patient survival,but also serves as a reference for clinicians to choose personalized treatment plans.