Noninvasive identification of HER2-zero, -low, or -overexpressing breast cancers: Multiparametric MRI-based quantitative characterization in predicting HER2-low status of breast cancer

• Published in: European journal of radiology Vol. 177; p. 111573
• Main Authors: Zhan, Ting, Dai, Jiankun, Li, Yan
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.08.2024
• Subjects: Breast cancer; Human epidermal growth factor receptor 2; Synthetic magnetic resonance imaging; ADC; PR; DCE-MRI; IHC; MRI; SPE; IDC; AUC; 2D; DWI; T1; T2; ACC; OR; CI; VEGF; ROC; Synthetic magnetic resonance imaging; ICC; CTRW; Breast cancer; FOV; ER; ROI; ADCs; TE; VOI; 3D; PD; Human epidermal growth factor receptor 2; FISH; DISCO; HER2; SEN; TR
• ISSN: 0720-048X; 1872-7727; 1872-7727
• DOI: 10.1016/j.ejrad.2024.111573

•SyMRI accurately determined the HER2 status of individuals with breast cancer, particularly those with low expression levels of HER2.•The PDe is superior to the conventional quantitative indicator in HER2 expression discrimination.•The quantitative indicator PDe may be an alternative marker suitable for patients undergoing invasive biopsy and can assist in identifying the appropriate biopsy site, and ultimately enhance the precision of the histological assessment of HER2 expression status. To evaluate the effectiveness of both synthetic magnetic resonance imaging (SyMRI) and conventional diffusion-weighted imaging (DWI) for identifying the human epidermal growth factor receptor 2 (HER2) status in breast cancer (BC) patients. In this retrospective study, 114 women with DWI and SyMRI were pathologically classified into three groups: HER2-overexpressing (n = 40), HER2-low-expressing (n = 53), and HER2-zero-expressing (n = 21). T1 and T2 relaxation times and proton density (PD) were assessed before and after enhancement, and the resulting quantitative parameters produced by SyMRI were recorded as T1, T2, and PD and T1e, T2e, and PDe. Logistic regression was used to identify the best indicators for classifying patients based on HER2 expression. The discriminative performance of the models was evaluated using receiver operating characteristic (ROC) curves. Our preliminary study revealed significant differences in progesterone receptor (PR) status, Ki-67 index, and axillary lymph node (ALN) count among the HER2-zero, -low, and -overexpressing groups (p < 0.001 to p = 0.03). SyMRI quantitative indices showed significant differences among BCs in the three HER2 subgroups, except for ΔT2 (p < 0.05). our results indicate that PDe achieved an area under the curve（AUC）of 0.849 (95 % CI: 0.760–0.915) for distinguishing HER2-low and -overexpressing BCs. Further investigation revealed that both the PDe and ADC were indicators for predicting differences among patients with HER2-zero and HER2-low-expressing BC, with AUCs of 0.765(95 % CI: 0.652–0.855) and 0.684(95 % CI: 0.565–0.787), respectively. The addition of the PDe to the ADC improved the AUC to 0.825(95 % CI: 0.719–0.903). SyMRI could noninvasively and robustly predict the HER2 expression status of patients with BC.