基于DCE-MRI及临床病理特征的模型预测乳腺癌前哨淋巴结状态  

Prediction of sentinel lymph node status in breast cancer based on DCE-MRI and clinicopathological models

作　　者：张冬雪 李卓琳[1] 李振辉[1] 吴建萍[1] 谢瑜[1] ZHANG Dong-xue;LI Zhuo-lin;LI Zhen-hui(Department of Radiology,Third Hospital of Kunming Medical University,Yunnan Cancer Hospital,Kunming 650018,China)

机构地区：[1]昆明医科大学第三附属医院,云南省肿瘤医院放射科,昆明650018

出　　处：《放射学实践》2022年第9期1104-1108,共5页Radiologic Practice

基　　金：国家自然科学基金(No.82001986);云南省卫生健康委员会医学后备人才培养计划(H-2018005);云南省应用基础研究优秀青年基金(202101AW070001);云南省生物资源数字化开发应用项目(202002AA100007);云南省教育厅科学研究基金项目(2021J0260);云南省教育厅科学研究基金项目(2022J0249);昆明医科大学胃肠肿瘤影像学智能研究科技创新团队(CXTD202110);云南省科技厅昆明医科大学应用基础研究联合专项(202201AY070001-148)。

摘　　要：目的:探讨基于DCE-MRI及临床病理特征建立的预测模型对乳腺癌前哨淋巴结(SLN)转移的预测价值,为临床个性化治疗提供依据。方法:回顾性分析我院2016年1月至2018年12月经手术病理证实为乳腺癌并符合入组标准的260例患者的临床、病理及MRI资料。根据SLN病理结果,将全部患者分为SLN转移阳性组(n=127)和SLN转移阴性组(n=133),比较两组患者的临床病理资料、MRI特征的差异。利用logistic回归分析确定与SLN转移相关的预测因子并构建预测模型,采用受试者工作特征(ROC)曲线评价预测模型的效能。结果:单因素分析结果显示,ER、PR、HER2、分子分型、多灶、增强形态与乳腺癌SLN转移相关(P<0.05)。多因素logistic回归分析结果显示PR、HER2、增强形态、ADC值(P<0.05)是乳腺癌SLN转移的独立危险因素,预测模型的曲线下面积为0.710(P<0.05),其预测乳腺癌前哨淋巴结转移的敏感度为77.2%,特异度为58.6%。结论:ER、PR、HER2、分子分型、多灶、增强形态与乳腺癌SLN转移相关,基于PR、HER2、增强形态、ADC值等参数构建的预测模型对乳腺癌SLN转移有较高的预测价值。Objective:To explore the predictive value of the predictive model based on DCE-MRI and clinicopathological features for sentinel lymph node(SLN)metastasis of breast cancer and to provide the basis for personalized clinical treatment.Methods:The clinical,pathological,and MRI data of 260 patients with breast cancer confirmed by surgery and pathology in our hospital from January 2016 to December 2018 and who met the inclusion criteria were retrospectively collected.According to the SLN pathological results,all patients were divided into SLN metastasis-positive group(n=127)and SLN metastasis-negative group(n=133).The difference in MRI features and clinicopathological data of the two groups were compared.The predictive factors related to SLN metastasis were determined by logistic regression analysis,and the predictive model was constructed.The efficacy of the predictive model was evaluated by the receiver operating characteristic(ROC)curves.Results:Univariate analysis showed that ER,PR,HER2,molecular type,multifocal,and enhancement morphology were associated with SLN metastasis in breast cancer(P<0.05).Multivariate logistic regression analysis showed that PR,HER2,enhanced morphology,and ADC value were independent risk factors for SLN metastasis of breast cancer(P<0.05),and the ROC curve of the predictive model was 0.710(P<0.05).The sensitivity and specificity for predicting sentinel lymph node metastasis of breast cancer were 77.2%and 58.6%.Conclusion:ER,PR,HER2,molecular type,multifocal,and enhanced morphology are associated with SLN metastasis in breast cancer.The prediction model based on the PR,HER2,enhanced morphological,and ADC value has a high predictive value for SLN metastasis in breast cancer.

关 键 词：乳腺癌 前哨淋巴结 淋巴结转移 磁共振成像 预测模型 

分 类 号：R737.9[医药卫生—肿瘤] R445.2[医药卫生—临床医学]