雌激素受体及孕激素受体双阴性且人表皮生长因子受体2阳性乳腺癌新辅助化疗病理完全缓解预测模型构建及验证  

Construction and validation of pathological complete response prediction model of neoadjuvant chemothera-py for ER-negative,PR-negative and HER2-positive breast cancer

作　　者：夏坤健 黄家良[1] 王琳[2] 欧阳永鹏 孙振华[1] 唐娜 XIA Kunjian;HUANG Jialiang;WANG Lin;OUYANG Yongpeng;SUN Zhenhua;TANG Na(Department of General Surgery,the Second Affiliated Hospital of Jiujiang College,Jiujiang 332005,Jiangxi,China;Department of Breast Surgery,the Second Affiliated Hospital of Nanchang University,Nanchang 330006,Jiangxi,China)

机构地区：[1]九江学院第二附属医院普通外科,九江332005 [2]南昌大学第二附属医院乳腺外科,南昌330006

出　　处：《医学研究与战创伤救治》2025年第3期290-296,共7页Journal of Medical Research & Combat Trauma Care

摘　　要：目的探究雌激素受体及孕激素受体双阴性且人表皮生长因子受体2(HER2)阳性乳腺癌新辅助化疗(NAC后病理完全缓解(pCR)的影响因素,构建HER2阳性乳腺癌NAC后p CR的预测模型并对其进行验证。方法回顾性分析2020年2月至2024年2月在南昌大学第二附属医院接受NAC的HER2阳性乳腺癌患者的临床资料,根据是否获得p CR分为p CR组(获得p CR)与对照组(未获得p CR)。收集并比较2组患者的一般临床病理资料包括年龄、手术方式、化疗方案、肿瘤临床TNM分级、病理、肿瘤Nottingham分级、Ki-67指数和月经状况,采用二元logistic回归分析NAC后p CR的影响因素并建立预测模型。采用Hosmer-Lemeshow检验及受试者工作特征(ROC)曲线评价回归模型的预测效能,最后对回归模型进行外部验证。结果共纳入了453例HER2阳性乳腺癌患者,其中验证队列151例、研究队列302例,研究队列302例患者分为p CR组143例与对照组159例。2组患者的年龄、化疗方案、cT分期、cN分期、肿瘤Nottingham分级及Ki-67指数等差异均有统计学意义(P>0.05)。二元logistic回归多因素分析显示,肿瘤c N2期(OR=0.409,95%CI:0.193~0.868)、肿瘤c N3期(OR=0.183,95%CI:0.057~0.583)、肿瘤NottinghamⅢ级(OR=4.202,95%CI:1.397~12.643)、Ki-67指数(OR=1.026,95%CI:1.011~1.041)为HER2阳性乳腺癌NAC后p CR的独立影响因素(P均<0.05)。Hosmer-Lemeshow检验的P值为0.431,ROC曲线下面积为0.723(95%CI:0.667~0.780)。外部验证显示,预测模型的灵敏度为74.7%,特异度为79.2%,准确度为76.8%,经Kappa检验具有较高的一致性(Kappa值为0.537,P<0.001)。结论构建的回归模型具有较高的预测价值,对临床医师预测HER2阳性乳腺癌NAC后是否出现p CR具有一定的参考价值。Objective To investigate the factors affecting pathological complete remission(pCR)after neoadjuvant chemo-therapy(NAC)for estrogen receptor and progesterone receptor double-negative and human epidermal growth factor receptor 2(HER2)positive breast cancers,to construct a prediction model of pCR after NAC for HER2-positive breast cancers and to validate it.Methods Clinical data of HER2-positive breast cancer patients who underwent NAC at the Second Affiliated Hospital of Nanchang University from February 2,2020 to February 2,2024 were retrospec-tively analyzed and divided into the pCR group(obtaining pCR)and the control group(not obtaining pCR)according to whether they ob-tained pCR or not.The general clinicopathological data of patients in the 2 groups were collected and compared including age,surgical modality,chemotherapy regimen,tumor clinical TNM grade,pathol-ogy,tumor Nottingham grade,Ki-67 index,and menstrual status,and binary logistic regression was used to analyze the influencing factors of pCR after NAC and to establish a prediction model.The predictive efficacy of the regression model was evaluated using the Hosmer Lemeshow test and subject work characteristics(ROC)curves,and finally the regression model was externally validated.Results A total of 453 patients with HER2-positive breast cancer were included,including 151 cases in the validation cohort and 302 cases in the study cohort.The 302 patients in the study cohort were divided into 143 cases in the pCR group and 159 cases in the con-trol group.There were significant differences in age,chemotherapy regimen,CT stage,CN stage,Nottingham grade and Ki-67 index between the two groups(P>0.05).Multivariate analysis of binary logistic regression showed that tumor cN2 stage(OR=0.409,95%CI:0.193-0.868),tumor cN3 stage(OR=0.183,95%CI:0.057-0.583),tumor Nottingham III grade(OR=4.202,95%CI:1.397-12.643),Ki-67 index(OR=1.026,95%CI:1.011-1.041)were the independent influencing factors of PCR after NAC in HER2 positive breast cancer(all P<0.05).The p value of Hosme

关 键 词：HER2阳性乳腺癌 新辅助化疗 病理完全缓解 预测模型 

分 类 号：R737.9[医药卫生—肿瘤]