基于临床及多模态超声特征风险评分模型鉴别乳腺BI-RADS 4A类良、恶性病变  

Risk score model based on clinical and multimodal ultrasound features for differentiating benign and malignant BI-RADS 4A breast lesions

作　　者：刘启凡 崔帅 郭文静 李伟[1] 王海龙[1] 李虎莎 姚俊东 张周龙[1] LIU Qifan;CUI Shuai;GUO Wenjing;LI Wei;WANG Hailong;LI Husha;YAO Jundong;ZHANG Zhoulong(Department of Ultrasound,the First Affiliated Hospital of Henan University of Science and Technology,Luoyang 471000,China)

机构地区：[1]河南科技大学第一附属医院超声科,河南洛阳471000

出　　处：《中国介入影像与治疗学》2025年第1期22-26,共5页Chinese Journal of Interventional Imaging and Therapy

摘　　要：目的 观察基于临床及多模态超声特征风险评分模型鉴别乳腺乳腺影像报告和数据系统(BI-RADS)4A类良、恶性病变的价值。方法 前瞻性纳入177例多模态乳腺超声提示BI-RADS 4A类病变患者,按7∶3比例分为训练集(n=123)与测试集(n=54)。以单因素及多因素logistic回归分析临床及灰阶超声、CDFI、弹性成像超声参数,筛选乳腺BI-RADS 4A类良、恶性病变的独立鉴别因素并以之构建风险评分模型。以术后病理或随访结果为标准,利用受试者工作特征(ROC)曲线获取模型鉴别乳腺BI-RADS 4A类良、恶性病变的最佳截断值,评估模型鉴别效能。结果177例中,良性39例、恶性138例。患者年龄>58岁,病变直径>15.1 mm、形状不规则、血流1或2级和病变周围2 mm环状区域(shell)声触诊弹性成像值标准差(shell-STESD)>16.33 kPa均为乳腺BI-RADS 4A类恶性病变的独立危险因素。以6.5为最佳截断值,风险评分模型鉴别训练集高危与低危病例的敏感度为84.69%(83/98)、特异度为88.00%(22/25),在测试集中分别为77.50%(31/40)及71.43%(10/14)。结论 基于临床及多模态超声特征风险评分模型可有效鉴别乳腺BI-RADS 4A类良、恶性病变。Objective To evaluate the value of risk score model based on clinical and multimodal ultrasound features for differentiating benign and malignant breast imaging reporting and data system(BI-RADS)4A lesions in breast.Methods Totally 177 patients with BI-RADS 4A lesions detected by multimodal breast ultrasound were prospectively enrolled,and the patients were divided into training set(n=123)and test set(n=54)at the ratio of 7∶3.Univariate and multivariate logistic regression analyses were used to analyze patients’clinical,gray-scale ultrasound,CDFI and elastography ultrasound parameters of lesions,and the independent risk factors for differentiating benign and malignant BI-RADS 4A lesions in breast were screened,so as a risk score model was constructed.Based on the results of sungical pathology or follow-up,the cut-off value of the model for differentiating benign and malignant BI-RADS 4A lesions in breast was obtained by receiver operating characteristic(ROC)curve,and the patients were divided into high-risk and low-risk subgroups according to the cut-off value,and the efficacy of the model was evaluated.Results Among 177 cases,39 were with benign lesions and 138 were with malignant lesions.Patients’age>58 years,diameter of lesion>15.1 mm,lesions with irregular shape,lesions with grade 1 or 2 of blood flow and standard deviation of sound touch elastography of a 2 mm circular area around the lesion(shell)(shell-STESD)>16.33 kPa were all independent risk factors of malignant BI-RADS 4A lesions in breast.The cut-off value of risk score model for differential diagnosis was 6.5 points,and its sensitivity was 84.69%(83/98)and specificity was 88.00%(22/25)in distinguishing high-risk and low-risk subgroups in training set,while its sensitivity was 77.50%(31/40)and specificity was 71.43%(10/14)in test set.Conclusion Risk score model based on clinical and multimodal ultrasound features could effectively differentiate benign and malignant BI-RADS 4A lesions in breast.

关 键 词：乳腺肿瘤 病理学 超声检查 乳腺影像报告和数据系统 前瞻性研究 

分 类 号：R737.9[医药卫生—肿瘤] R445.1[医药卫生—临床医学]