基于BI-RADS的多参数MRI列线图模型对乳腺黏液癌和黏液样纤维腺瘤的鉴别价值  

The Diagnositic Value of A Multi-parameter MRI Nomogram Model Based on BI-RADS in Discriminating Mucinous Breast Carcinoma from Mucinous Fibroadenoma

作　　者：蒋瑞生 高健 刘淑珍 董海霞 JIANG Ruisheng;GAO Jian;LIU Shuzhen;DONG Haixia(Department of Medical Imaging,Yidu Central Hospital,Shandong Second Medical University,Qingzhou 262500,China;Department of Medical maging,Qingzhou People's hospital)

机构地区：[1]山东省第二医科大学附属益都中心医院医学影像科,青州262500 [2]青州市人民医院医学影像科

出　　处：《中国医学计算机成像杂志》2025年第1期43-48,共6页Chinese Computed Medical Imaging

摘　　要：目的:探索基于乳腺影像报告和数据系统(BI-RADS)的多参数MRI构建的列线图模型对乳腺黏液癌(MBC)和黏液样纤维腺瘤(MFA)的鉴别价值。方法:回顾性分析经手术病理证实的29例MBC和38例MFA的平扫及动态增强MRI资料,比较2组的发病年龄、发病部位、经过优化的形态学特征、最高表观弥散系数(ADC)值和最低ADC值及其对应的早期强化率和延迟强化曲线,经多因素logistic回归分析筛选具有显著鉴别价值的独立预测变量,构建多参数列线图模型,采用受试者工作特征(ROC)曲线评估模型的鉴别效能。结果:年龄、病变长径及短径、形状、边缘特征、内部强化特征、间隔是否>1.2 mm、皂泡状外观、最高及最低ADC值及其对应的早期强化率在2组间存在统计学差异(P<0.05)。其中病变的边缘特征、皂泡状外观及最高ADC这3个变量是鉴别MBC和MFA的独立预测因素,对应的列线图模型ROC曲线下面积(AUC)为0.987,灵敏度93.10%,特异度94.74%,阳性及阴性预测值和准确率分别为93.10%、94.74%、94.03%。结论:基于BI-RADS的多参数MRI列线图模型对鉴别MBC和MFA具有很高的参考价值。Purpose:To explore the diagnostic efficacy of a multi-parameter MRI nomogram model based on Breast Imaging Reporting and Data System(BI-RADS)for distinguishing mucinous breast carcinoma(MBC)from myxoid fibroadenoma(MFA).Methods:The plain and dynamic contrast-enhanced MR images of 29 cases of MBC and 38 cases of MFA were retrospectively analyzed.The age,location,optimal morphological features,the highest and lowest apparent diffusion coefficient(ADC)values,as well as the corresponding early enhancement rates and delayed enhancement curves were measured and compared between the two groups.The significant independent risk factors were analyzed by multivariate logistic regression analysis.A nomogram model incorporating all independent risk factors was developed,and receiver operating characteristic(ROC)curves were used to evaluate the predictive performance of this model.Results:Significant differences were observed between the two groups in terms of age,long diameter and short diameter of the lesions,lesion shapes,edge characteristics,internal enhancement features,interval thickness,soap bubble-like appearance,highest and lowest ADC values as well as the corresponding early enhancement rates(P<0.05).Edge characteristics,soap bubble-like appearance and highest ADC values were identified as the independent predictive factors.The area under curve(AUC)for the nomogram model was 0.987 with a sensitivity of 93.10%and specificity of 94.74%.The positive predictive value,negative predictive value and overall accuracy for distinguishing MBC from MFA were 93.10%,94.74%and 94.03%,respectively.Conclusion:The multi-parameter MRI nomogram model based on BI-RADS demonstrates high diagnostic value in differentiating MBC from MFA.

关 键 词：乳腺黏液癌 乳腺纤维腺瘤 磁共振成像 列线图 

分 类 号：R445.2[医药卫生—影像医学与核医学]