基于FFDM、超声特征及临床因素的列线图模型预测乳腺癌前哨淋巴结转移的价值  

The value of nomogram model for predicting sentinel lymph node metastasis in breast cancer based on FFDM,ultrasound features and clinical factors

作　　者：曹思薇 刘兴远 程晓英[1] 张五岳 郑洪彦[3] 金彦桐 赵明明[1] 阮野 高波[1] CAO Siwei;LIU Xingyuan;CHENG Xiaoying;ZHANG Wuyue;ZHENG Hongyan;JIN Yantong;ZHAO Mingming;RUAN Ye;GAO Bo(Department of Radiology,the Second Affiliated Hospital of Harbin Medical University,Heilongjiang Harbin 150001,China;Department of Ultrasound,the Second Affiliated Hospital of Harbin Medical University,Heilongjiang Harbin 150001,China;Department of Pathology,the Second Affiliated Hospital of Harbin Medical University,Heilongjiang Harbin 150001,China)

机构地区：[1]哈尔滨医科大学附属第二医院放射科,黑龙江哈尔滨150001 [2]哈尔滨医科大学附属第二医院超声科,黑龙江哈尔滨150001 [3]哈尔滨医科大学附属第二医院病理科,黑龙江哈尔滨150001

出　　处：《现代肿瘤医学》2024年第22期4304-4311,共8页Journal of Modern Oncology

基　　金：国家自然科学基金资助项目(编号:62172129)。

摘　　要：目的:探讨全视野数字化乳腺X线摄影(FFDM)、超声特征及临床病理因素对乳腺癌前哨淋巴结转移(SLNM)的预测价值并构建列线图预测模型。方法:回顾性分析416例乳腺癌患者的影像及临床资料,以7∶3的比例随机分为训练集(n=291)及测试集(n=125),对Lasso回归确定SLNM的预测因子行多因素Logistic回归分析并构建列线图模型,评价其预测SLNM的价值。结果:ROC曲线分析结果显示:当Youden=0.224时,NLR的最佳阈值为2.37。Lasso回归结合Logistic回归分析结果显示FFDM特征(肿块边缘呈星芒状、肿块内有可疑恶性钙化、腋窝异常淋巴结)、超声征象(肿块内血流特征、腋窝异常淋巴结)是乳腺癌SLNM的预测因子,训练集和测试集中模型的AUC分别为0.841和0.811,提示模型的区分度良好,基于训练集的ROC曲线确定列线图的最佳阈值为174.6分;校准曲线和临床决策曲线提示模型有良好的校准度和临床适用性。结论:基于FFDM、超声特征及临床因素的列线图模型为临床医生术前无创性预测乳腺癌SLNM提供了新方法。Objective:To explore the value of preoperative full-field digital mammography(FFDM),ultrasound features and clinicopathological factors for predicting sentinel lymph node metastasis(SLNM)in breast cancer and to construct a nomogram model.Methods:416 breast cancer patients were retrospectively analyzed,and they were randomly divided into training set(n=291)and testing set(n=125)at a ratio of 7∶3.Lasso regression was used to identify predictive factors for SLNM,multivariate Logistic regression analysis was performed on the selected variables and a nomogram model was constructed based on the above factors.Results:ROC curve analysis results showed that when Youden=0.224,the optimal threshold of NLR was 2.37.The results of Lasso regression and Logistic regression analysis showed that FFDM features(spiculated margin of tumor,suspicious malignant calcification in the tumor,FFDM reported abnormal lymph nodes),ultrasonic features(blood flow in the tumor,US reported abnormal lymph nodes)were predictors of SLNM in breast cancer,and the AUC of nomogram model in training set and testing set were 0.841 and 0.811 respectively,it indicated good discrimination of the model,the optimal threshold of the model based on the ROC curve of the training set was 174.6 points.The calibration curve and clinical decision curve indicated that the nomogram model had good calibration and clinical benefits.Conclusion:The nomogram model based on FFDM,ultrasound characteristics and clinical factors can provide valuable reference for clinicians to predict SLNM in breast cancer noninvasively and accurately before surgery.

关 键 词：乳腺癌 全视野数字化乳腺X线摄影 超声检查 前哨淋巴结转移 列线图 

分 类 号：R737.9[医药卫生—肿瘤]