Nomogram based on multimodal MRI radiomics for discriminating molecular subtypes of HER-2-negative breast cancer
10.19405/j.cnki.issn1000–1492.2026.04.016
- VernacularTitle:多模态MRI影像组学诺莫图区分HER-2阴性乳腺癌分子亚型
- Author:
Qun WANG
1
;
Hongli PAN
1
;
Xiaohu LI
1
;
Yongqiang YU
1
;
Yunwen YAN
2
;
Weishu HOU
1
Author Information
1. Dept of Medical Imaging, The First Affiliated Hospital of Anhui Medical University，Hefei 230022
2. Dept of Breast Surgery, The First Affiliated Hospital of Anhui Medical University，Hefei 230022
- Publication Type:Journal Article
- Keywords:
multimodal MRI;
radiomics;
breast cancer;
human epidermal growth factor receptor- 2;
nomogram
- From:
Acta Universitatis Medicinalis Anhui
2026;61(4):715-723
- CountryChina
- Language:Chinese
-
Abstract:
ObjectiveTo explore the value of a multimodal MRI-based radiomics nomogram for differentiating human epidermal growth factor receptor-2 （HER-2） negative breast cancer molecular subtypes.MethodsA retrospective analysis was conducted on 190 patients with HER-2 negative breast cancer who underwent multimodal MRI examination， and the patients were divided into two molecular subtype groups： a HER-2 low expression group （n=108） and a HER-2 zero expression group （n=82）. The cases were randomly stratified and sampled at a ratio of 7∶3 and divided into a training set of 133 cases and a testing set of 57 cases. The clinical and radiological features of the patients were collected， the radiomics features based on T2-weighted imaging （T2WI）， diffusion-weighted imaging （DWI）， and dynamic contrast-enhanced （DCE）-MRI were extracted， and the clinical-radiological model， unimodal radiomics model， multimodal radiomics model， and combined model were constructed respectively. Then the nomogram combined multimodal radiomics signature （radsocre） with clinical-radiological features was used to construct a visualized predictive model， and the area under the curve （AUC） was used to compare the effectiveness of different models in distinguishing HER-2 low expression and zero expression subtypes.ResultsA significant difference in radscore was demonstrated between the HER-2 low and HER-2 zero expression groups in both the training （P<0.000 1） and testing sets （P<0.01）. The AUC of the multimodal radiomics model in the training set and the testing set were 0.914 and 0.836， respectively， which was superior to any unimodal radiomics model. The nomogram demonstrated great diagnostic efficacy （AUC=0.930 in training set； AUC=0.865 in testing set）.ConclusionA multimodal MRI-based nomogram incorporating radsocre and clinical-radiological features can accurately distinguish the subtypes of HER-2 negative breast cancer.
