HER2阴性乳腺癌患者预后的危险因素及其预后预测模型的构建和验证  

Risk factors for prognosis in HER2-negative breast cancer patients and the construction and validation of a prognostic prediction model

作　　者：张欣华 张秀明 豆小文 ZHANG Xin-hua;ZHANG Xiu-ming;DOU Xiao-wen(School of Medicine,Anhui University of Science and Technology,Huainan 232000,Anhui,CHINA;Luohu Hospital Group Medical Laboratory Center,Shenzhen 518000,Guangdong,CHINA)

机构地区：[1]安徽理工大学医学院,安徽淮南232000 [2]罗湖医院集团医学检验中心,广东深圳518000

出　　处：《海南医学》2025年第8期1075-1081,共7页Hainan Medical Journal

基　　金：广东省深圳市医学重点学科项目(编号:SZXK054)。

摘　　要：目的探究影响HER2阴性乳腺癌患者预后的危险因素,构建其预后列线图预测模型并进行验证。方法选取SEER数据库中2010年1月至2015年12月期间诊断为HER2阴性乳腺癌患者11486例作为研究对象。将患者按7∶3比例随机分为训练组(n=8040)和验证组(n=3446),比较两组患者的临床基本特征。在训练组中通过单因素和多因素Cox回归分析筛选出影响预后的关键变量,建立HER2阴性乳腺癌预后预测模型。通过C指数(C-index)、受试者工作特征(ROC)曲线及曲线下面积(AUC)、校准曲线和临床决策曲线(DCA)评价模型的预测能力和临床适用性。结果两组患者的各项临床基本特征比较差异均无统计学意义(P>0.05)。经Cox回归分析筛选变量,其中年龄、种族、病理分级、雌激素受体(estrogenreceptor,ER)、孕激素受体(progesteronereceptor,PR)、局部淋巴结状态、临床分期、TNM分期、HER2状态、婚姻状态及治疗情况均与预后显著相关(P<0.05),基于上述变量建立HER2阴性乳腺癌预后的列线图预测模型。模型在训练组和验证组的C指数分别为0.740和0.720,ROC曲线分析显示训练组与验证组3年、5年生存率的曲线下面积(AUC)分别为0.795、0.777和0.785、0.745,均大于0.7,表明模型具有较好的区分度和准确性。校准曲线显示预测生存率与实际观测值高度吻合,表明模型预测效能良好;DCA分析提示模型在合适阈值范围内具有临床净获益。结论本研究建立的列线图模型能有效预测HER2阴性乳腺癌患者的个体化预后,为临床分层管理和治疗决策提供量化工具。Objective To explore the risk factors affecting the prognosis of HER2-negative breast cancer pa-tients,and to construct and validate a prognostic nomogram prediction model.Methods A total of 11486 HER2-nega-tive breast cancer patients diagnosed between January 2010 and December 2015 were selected from the SEER database as the study subjects.The patients were randomly divided into a training group(n=8040)and a validation group(n=3446)in a 7∶3 ratio,and the clinical characteristics of the two groups were compared.In the training group,univariate and multivariate Cox regression analyses were used to identify key variables affecting prognosis,and a prognostic prediction model for HER2-negative breast cancer was established.The predictive ability and clinical applicability of the model were evaluated using the C-index,receiver operating characteristic(ROC)curve,area under the curve(AUC),calibra-tion curve,and decision curve analysis(DCA).Results There were no statistically significant differences in the clinical characteristics between the two groups(P>0.05).Cox regression analysis identified age,race,pathological grade,estro-gen receptor(ER)status,progesterone receptor(PR)status,regional lymph node status,clinical stage,TNM stage,HER2 status,marital status,and treatment as significant prognostic factors(P<0.05).Based on these variables,a nomo-gram prediction model for HER2-negative breast cancer prognosis was constructed.The C-index values for the training and validation groups were 0.740 and 0.720,respectively.ROC curve analysis showed that the AUC values for 3-year and 5-year survival rates in the training and validation groups were 0.795,0.777 and 0.785,0.745,respectively,all ex-ceeding 0.7,indicating good discrimination and accuracy of the model.The calibration curve demonstrated a high consis-tency between predicted and observed survival rates,indicating excellent predictive performance.DCA analysis suggest-ed that the model provided clinical net benefit within an appropriate threshold range.Conclusion The nom

关 键 词：乳腺癌 HER2-low SEER数据库 预后模型 列线图 

分 类 号：R737.9[医药卫生—肿瘤]