中国医学装备2026，Vol.23Issue(2)：52-57,6.DOI:10.3969/j.issn.1672-8270.2026.02.011
基于MRI影像组学模型对BI-RADS 4类乳腺肿块的诊断价值分析
Analysis for diagnostic value of MRI-based radiomics model for BI-RADS category 4 breast masses
摘要
Abstract
Objective:To explore the diagnostic value of magnetic resonance imaging(MRI)radiomics models for breast imaging reporting and data system(BI-RADS)category 4 breast masses.Methods:A retrospective selection was conducted on 157 patients with BI-RADS category 4 breast masses who admitted to the 83th Group Military Hospital of Chinese People's Liberation Group Force,and the First Affiliated Hospital of Henan Medical University from February 2023 to February 2025.According to the features of breast tumor,they were divided into benign group(33 cases)and malignant group(124 cases).A total of 853 features of MRI radiomics of breast masses were extracted.The Mann-Whitney U test was adopted to screen the features with significant differences between benign and malignant masses.Using stratified random sampling,157 patients with BI-RADS category 4 breast masses were divided into a training set(110 cases)and a test set(47 cases)as a ratio of 7 to 3,which was used to construct a multi-dimensional logistic regression model with MRI radiomics(abbreviation:radiomics model).Based on the difference of the malignant risk of BI-RADS category 4,the 157 patients were divided further into 3 subgroups:group 4A(89 cases,with a malignant risk from 2%to 10%),group 4B(45 cases,with a malignant risk from 10%to 50%),and group 4C(23 cases,with a malignant risk from 50%to 95%).The diagnostic efficacy of the model was verified in the test set.Compared with the diagnostic method of conventional BI-RADS classification radiomics,the area under curve(AUC)of the receiver operating characteristic(ROC)curve was used to analyze the diagnostic efficacy of the three groups.Results:A total of 56 features with significant difference were selected,and the finally constructed model can independently predict 10 independent radiomics features of benign and malignant features of breast masses.The AUC value of the radiomics model was 0.941 in the test set,which was significantly higher than 0.785 of the BI-RADS classification,and the difference was statistically significant(Z=3.856,P<0.001).The sensitivity,specificity,and accuracy of the radiomics model were respectively 92.7%,84.8%and 90.4%.The analysis of three-group showed that the AUC values of the radiomics model were respectively 0.938,0.945 and 0.951 in the group 4A,4B,and 4C of malignant risk,and all of them were>0.93,and the accuracy and positive predictive value of the group 4C were respectively 95.7%and 100%.Conclusion:The MRI radiomics model has a high diagnostic efficacy for BI-RADS category 4 of breast masses,which has especially exceptional performance in diagnosing high risk.It can provide references for clinically precise identification.关键词
乳腺肿块/乳腺影像报告和数据系统(BI-RADS)4类/磁共振成像(MRI)/影像组学/诊断Key words
Breast mass/Breast imaging reporting and date system(BI-RADS)category 4/Magnetic resonance imaging(MRI)/Radiomics/Diagnosis分类
医药卫生引用本文复制引用
王肖肖,梁长华,郭利茹,田方,孙运帮..基于MRI影像组学模型对BI-RADS 4类乳腺肿块的诊断价值分析[J].中国医学装备,2026,23(2):52-57,6.基金项目
河南省医学科技攻关计划(LHGJ20240494) Henan Province Medical Science and Technology Research Program(LHGJ20240494) （LHGJ20240494）
