Study on ultrasound nomogram model based on BI-RADS in predicting malignant tumor of breast
10.3969/j.issn.1672-8270.2025.08.015
- VernacularTitle:基于乳腺影像报告和数据系统研发超声列线图模型预测乳腺恶性肿瘤研究
- Author:
Xiaoting CHEN
1
;
Xiaochuan ZHANG
;
Jian ZHOU
Author Information
1. 滁州市第一人民医院超声科 滁州 239000
- Publication Type:Journal Article
- Keywords:
Radiomics;
Breast imaging reporting and data system(BI-RADS);
Malignant tumor of breast;
Nomogram
- From:
China Medical Equipment
2025;22(8):77-82
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop a nomogram model that combined radiomics and the Breast Imaging Reporting and Data System(BI-RADS),which was used to predict breast cancer in category 4 or 5 lesions of BI-RADS ultrasound(US).Methods:A total of 315 female patients with breast tumors who admitted to The First People's Hospital of Chuzhou and Xinjiang Uygur Autonomous Region Hospital of Traditional Chinese Medicine from November 2021 to September 2024 were selected.In them,211 patients from January 2022 to August 2023 were included in the training group from Xinjiang Uygur Autonomous Region Hospital of Traditional Chinese Medicine,and 104 patients from November 2021 to September 2024 were included in the validation group from The First People's Hospital of Chuzhou.The nomogram model was developed according to the results of multiple regression analysis of the training group.Then,the discrimination,calibration and clinical practicality of nomogram were assessed in predicting breast cancer in the validation group.Results:Radiomics score and BI-RADS category were independent influencing factors in predicting malignant tumors of breast(OR=4.66,4.87,P<0.05).In the ROC curve analysis,the area under curve(AUC)values of the ROC curves of the nomogram models of training group and the validation group were respectively 0.928 and 0.883 in the ability of distinguishing malignant and benign lesions,which all were better than AUC value(0.791,0.864)of the radiomics score,and that(0.825,0.857)of the BI-RADS category,and the differences were statistically significant(Ztraining group=4.026,3.716,and Zverification group=3142,2.847,P<0.05).Conclusion:The nomogram model based on radiomic score and BI-RADS category has potential application value in the prediction for malignant tumor of breast that is BI-RADS US category 4 or 5.
