Value of clinical and MRI features in predicting HER-2-positive and HER-2-low expression breast cancer

LI Sien; YANG Zhiqi; LIN Yulin; DENG Junliang; ZHANG Zhiqiang; LI Xiaoyuan; CHENG Fengyan; CHEN Xiaofeng

doi:10.19300/j.2025.L21656

Abstract：Objective To evaluate the value of models based on clinical and MRI features in predicting HER-2-positive and HER-2-low expression breast cancers.Methods A retrospective analysis was conducted on 213 female patients(mean age 50.8±10.6 years)with surgically and pathologically confirmed mass-forming breast cancer.Based on postoperative pathological results,patients were categorized into HER-2-zero(65 cases),HER-2-low(79 cases),and HER-2-positive(69 cases)groups.Clinical and MRI characteristics were compared among the three HER-2 expression groups using one-way ANOVA and chi-square tests,including estrogen receptor(ER)status,progesterone receptor(PR)status,T stage,clinical stage,maximum lesion diameter,and apparent diffusion coefficient(ADC)values.Multivariate logistic regression was used to identify independent predictive factors for the HER-2-positive and HER-2-low expression subtypes,followed by predictive model construction.Receiver operating characteristic(ROC)analysis was used to assess model performance.Results Significant differences were observed among the HER-2-zero,HER-2-low,and HER-2-positive groups in ER status,PR status,T stage,clinical stage,maximum lesion diameter,and ADC values(all P<0.05).Multivariate analysis demonstrated that clinical stage and ADC value were independent predictors for both HER-2-positive and HER-2-low breast cancer(both P<0.05).The model constructed using clinical stage and ADC demonstrated high predictive efficacy for both HER-2 positive and HER-2-low expressing breast cancers,with areas under the curve(AUC)of 0.899 and 0.861,respectively.Conclusion A model integrating clinical stage and ADC values shows high efficacy for the noninvasive prediction of HER-2-positive and HER-2-low expression breast cancers.