基于超声特征构建肿块型肉芽肿性小叶性乳腺炎与浸润性小叶癌的分类模型  

Construction and Application of a Classification Model Based on Ultrasound Features for Mass-Type Granulomatous Lobular Mastitis and Invasive Lobular Carcinoma

作　　者：肖轶笙 何艳萍 刘洁琳 邱懿德[1] 蓝宁辉 甄朝炯 Xiao Yisheng;He Yanping;Liu Jielin;Qiu Yide;Lan Ninghui;Zhen Chaojiong(Ultrasonic Diagnosis and Treatment Center,The First People's Hospital of Foshan,Foshan,Guangdong 528000,China;Department of Ultrasound,The Sixth Affiliated Hospital of South China University of Technology(Nanhai District People's Hospital of Foshan),Foshan,Guangdong 528200,China)

机构地区：[1]佛山市第一人民医院超声诊疗中心,广东省佛山市528000 [2]华南理工大学附属第六医院(佛山市南海区人民医院)超声科,广东省佛山市528200

出　　处：《中国超声医学杂志》2025年第2期145-149,共5页Chinese Journal of Ultrasound in Medicine

摘　　要：目的探索基于超声特征的肿块型肉芽肿性小叶性乳腺炎(GLM)与肿块型浸润性小叶癌(ILC)分类模型的构建。方法回顾性收集进行超声检查并经病理证实为肿块型GLM或肿块型ILC的患者,分析其超声特征,通过多因素Logistic回归分析得出鉴别肿块型GLM与肿块型ILC的独立影响因素,并构建分类模型。绘制该模型的受试者工作特征曲线与临床决策分析曲线,并与医师经验诊断对比。结果共纳入164例患者,共172个乳腺肿块。多因素分析结果显示,鉴别肿块型GLM与肿块型ILC的独立影响因素包括:年龄、病灶位置和肿块的内部回声。根据上述独立影响因素构建分类模型1,该模型的曲线下面积(AUC)为0.978(95%CI:0.960~0.995),再结合美国放射学会乳腺影像报告和数据系统(ACR BI-RADS)分类构建分类模型2,其诊断性能最优(AUC=0.980,95%CI:0.965~0.996)。本研究构建的分类模型诊断效能均优于常规医师经验诊断与ACR BI-RADS分类(P<0.05)。结论基于超声特征构建的分类模型具有较优的鉴别肿块型GLM与肿块型ILC的诊断性能。Objective To explore the construction and application of a classification model based on ultrasound features for mass-type granulomatous lobular mastitis(GLM)and mass-type invasive lobular carcinoma(ILC).Methods A retrospective analysis was performed on patients who underwent ultrasound examination were pathologically confirmed as mass-type GLM or mass-type ILC.The ultrasound characteristics were analyzed.The independent influencing factors for distinguishing mass-type GLM from mass-type ILC were obtained by multiple Logistic regression analysis,and classification models were constructed.The receiver operating characteristic(ROC)curve and decision curve analysis(DCA)curve of the models were drawn,comparing with the physician experience diagnosis.Results A total of 164 patients with 172 breast masses were included.The results of multivariate analysis showed that the independent influencing factors for distinguishing mass-type GLM from mass-type ILC included age,lesion location and internal echo of the mass.According to the above independent influencing factors,the classification model 1 was constructed,and the area under the curve(AUC)value of the model was 0.978(95%CI:0.960-0.995).Combined with American College of Radiology breast imaging reporting and data system(ACR BI-RADS),the classification model 2 had the best diagnostic performance(AUC=0.980,95%CI:0.965-0.996),and its diagnostic efficacy was better than that of conventional physician experience diagnosis and ACR BI-RADS classification(P<0.05).Conclusions The classification model of mass-type GLM and mass-type ILC based on ultrasound features has good diagnostic performance.

关 键 词：超声 肿块型 肉芽肿性小叶性乳腺炎 浸润性小叶癌 

分 类 号：R445.1[医药卫生—影像医学与核医学] R737.9[医药卫生—诊断学] R655.8[医药卫生—临床医学]