锥光束乳腺增强CT影像特征与乳腺癌激素受体及HER-2表达状态的相关性  

Correlation between cone-beam enhanced CT imaging features with hormone receptor and HER-2 expression in breast cancer

作　　者：代婷 杨鸿羽 赵青 苏桐 王锐 吕发金[1] 欧阳祖彬[1] DAI Ting;YANG Hong-yu;ZHAO Qing(Department of Radiology,The First Affiliated Hospital of Chongqing Medical University,Chongqing 400016,China)

机构地区：[1]重庆医科大学附属第一医院放射科,重庆400016 [2]重庆市长寿区人民医院放射科,重庆401220

出　　处：《放射学实践》2023年第12期1554-1561,共8页Radiologic Practice

基　　金：国家重点研发计划基金资助项目(2020YFA0714002)。

摘　　要：目的:探讨锥光束乳腺增强CT影像特征与乳腺癌激素受体及人表皮生长因子受体2(HER-2)表达状态的相关性。方法:回顾性搜集2019年1月-2022年6月在我院经病理证实的136例乳腺癌患者的临床(年龄及免疫组化结果)和术前锥光束乳腺增强CT(CBBCT)资料。其中,激素受体(PR、ER)阳性组82例,阴性组为54例;HER-2阳性组为47例,阴性组为89例。CBBCT影像特征主要包括肿块大小、形状、毛刺、可疑恶性钙化、肿块强化程度、内部强化特征、非肿块样强化、子灶、邻近皮肤回缩及乳头回缩。比较乳腺癌激素受体阳性组及阴性组、HER-2阳性组及阴性组之间CBBCT影像特征的差异。然后,采用二元logistic回归分析方法从影像特征中筛选出激素受体阳性及HER-2阳性乳腺癌的独立危险因素,构建相应的预测模型并绘制模型的列线图,使用ROC曲线、校正曲线及决策曲线分析(DCA)评估模型的预测效能。结果:与激素受体阴性组比较,阳性组的肿块较小,呈圆/卵圆形、毛刺征及皮肤回缩征更常见;二元logistic回归分析结果显示肿块大小(OR=0.964,95%CI:0.927~0.995)、形状(OR=0.327,95%CI:0.131~0.770)、毛刺征(OR=3.002,95%CI:1.313~7.105)和皮肤回缩征(OR=3.974,95%CI:1.537~11.361)是激素受体阳性乳腺癌的独立危险因素;联合上述4个征象构建激素受体阳性乳腺癌预测模型,其AUC为0.788(95%CI:0.711~0.865),敏感度为68.3%,特异度为85.2%。HER-2阳性组肿块强化程度较阴性组高,可疑恶性钙化、乳头回缩、非肿块样强化及子灶的出现率更高;将上述5个特征纳入二元logistic回归分析中,发现毛刺征(OR=0.410,95%CI:0.018~0.940)、可疑恶性钙化(OR=2.417,95%CI:1.067~5.560)、子灶(OR=2.848,95%CI:1.272~6.568)、乳头回缩征(OR=5.655,95%CI:1.421~28.597)以及肿块强化程度(OR=1.015,95%CI:1.004~1.027)是HER-2阳性乳腺癌的独立危险因素,所构建的HER-2阳性乳腺癌预测模型的AUC为0.794(95%CI:0.711~0.876),Objective:The purpose of this study was to explore the correlation between cone beam breast enhanced computed tomography(CBBCT)imaging features with hormone receptor and HER-2 expression in breast cancer.Methods:The clinical data(age and immunohistochemical results)and preoperative cone beam breast enhanced CT(CBBCT)features were retrospectively collected for 136 patients with pathologically confirmed breast cancer from January 2019 to June 2022 in our hospital.Among them,there were 82 cases in the hormone receptor positive group and 54 cases in the hormone receptor negative group,47 cases in the HER-2 positive group and 89 cases in the HER-2 negative group.The CBBCT imaging features mainly included lesion size,shape,spiculated margin,suspected malignant calcification,degree of mass enhancement,internal enhancement characteristics,non-mass enhancement,sub-focus,skin retraction and nipple retraction.Differences in CBBCT imaging features were compared between the hormone receptor positive breast cancer and negative group,HER-2 positive and negative group of breast cancer.Then,independent risk factors of hormone receptor positive and HER-2 positive breast cancer were screened out from the imaging features using binary logistic regression analysis,and the corresponding predictive models were constructed and plotted nomogram model.ROC curve,calibration curve and decision curve analysis(DCA)were used to evaluate the predictive efficiency of the models.Results:Compared with the hormone receptor negative group,the positive group showed smaller masses,round/oval shape,spiculated margin and skin retraction were more common.Binary logistic regression analysis showed that lesion size(OR=0.964,95%CI:0.927~0.995),shape(OR=0.327,95%CI:0.131~0.770),spiculated margin(OR=3.002,95%CI:1.313~7.105)and skin retraction(OR=3.974,95%CI:1.537~11.361)were the independent risk factors for hormone receptor positive breast cancer.The hormone receptor positive breast cancer predictive model was constructed by combining the above four features,

关 键 词：乳腺癌 激素受体 人类表皮生长因子受体2 锥光束CT 预测模型 诊断效能 

分 类 号：R814.42[医药卫生—影像医学与核医学] R737.9[医药卫生—放射医学]