Nomogram预测乳腺浸润性导管癌Ki-67表达水平  

Nomogram to Predict Ki-67 Expression Levels in Invasive Ductal Breast Cancer

作　　者：石红[1] 杨彦辉[1] 张亚萍[1] 钟跃[1] Shi Hong;Yang Yanhui;Zhang Yaping;Zhong Yue(Department of Ultrasound,The First People’s Hospital of Neijiang,Neijiang,Sichuan 641000,China)

机构地区：[1]内江市第一人民医院超声科,四川省内江市641000

出　　处：《中国超声医学杂志》2023年第1期17-21,共5页Chinese Journal of Ultrasound in Medicine

基　　金：内江市科技计划项目(No.Z201945)。

摘　　要：目的 基于患者临床及超声影像特征预测乳腺癌患者Ki-67表达水平。方法 回顾性分析甲乳外科收治的乳腺癌患者217例临床及超声影像资料,病理结果均为乳腺浸润性导管癌。病例一般资料包含年龄、肿瘤最大直径、二维超声表现、剪切波弹性成像资料。217例乳腺癌患者随机被分为训练组159例及验证组58例。先行单因素Logistic回归筛出可能影响乳腺癌Ki-67表达水平的变量,再行多因素Logistic回归构建乳腺癌Ki-67表达水平的临床预测模型。诺谟图可视化展示预测模型,采用受试者工作特征曲线、校准曲线及临床决策曲线对模型的校准度及实用性予以评价。结果 肿瘤最大直径、肿瘤微钙化及SWE最大弹性模量值。这3个临床组学特征与乳腺癌Ki-67表达水平呈显著相关性。训练集中ROC曲线下面积为0.892,提示该预测模型有能力区分Ki-67表达情况,同时在验证集中得以证实,AUC^(ROC)=0.717。诺谟图与临床决策曲线在后续分析中同样表现良好,证明了预测模型潜在临床价值。结论 本研究提供了以临床特征及超声影像资料相结合的Nomogram,可用于预测乳腺浸润性导管癌患者Ki-67表达水平。Objective To predict the Ki-67 expression level in breast cancer patients based on their clinical and ultrasound imaging characteristics.Methods The clinical and ultrasound imaging data of 217 breast cancer patients admitted to the Department of Breast Surgery of our hospital were retrospectively analyzed,and the pathological results were all invasive ductal carcinoma of the breast.The general data of the cases included age,tumor diameter,2 D ultrasound findings,and shear wave elastography data.A total of 217 breast cancer patients were randomly divided into the training group(n=159) and the validation group(n=58).First,univariate Logistic regression was used to screen out variables that might affect Ki-67 expression levels in breast cancer,and then multivariate Logistic regression was performed to construct a clinical predictive model of Ki-67 expression.The nomograph visualized the prediction model,and the receiver operating curve,calibration curve and clinical decision curve were used to evaluate the calibration degree and practicability of the model.Results The maximum tumor diameter,tumor microcalcification,and maximum elastic modulus value of SWE were significantly correlated with Ki-67 expression levels in breast cancer.The area under the ROC curve in the training set was 0.892,indicating that the prediction model had the ability to distinguish between Ki-67 expression conditions,and was simultaneously confirmed in the validation set with AUC^(ROC)=0.717.The nomograph and the clinical decision curve performed equally well in the subsequent analysis,proving the potential clinical value of the predictive model.Conclusions This study provides a nomogram that combines clinical features and ultrasound imaging data to predict the Ki-67 expression level in breast cancer patients.

关 键 词：诺谟图 乳腺癌 Ki-76 

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]