A Novel Biomarker for Identifying HER2-low Breast Cancer Using Synthetic MRI

Abstract To evaluate the diagnostic utility of quantitative parameters derived from synthetic magnetic resonance imaging (SyMRI) in differentiating HER2-low-expressing breast cancer from non-low-expressing subtypes. This retrospective study included 70 patients with pathologically confirmed unilateral invasive breast cancer who underwent preoperative 3.0T MRI with SyMRI sequences. Based on IHC/FISH results, patients were classified into HER2-low (n = 48) and non-low (n = 22) groups. Two radiologists independently measured T1, T2, PD, and ADC values of the lesions. Logistic regression analysis was used to identify the most effective predictors of HER2 expression status. ROC curve analysis was conducted to assess the discriminative performance of these predictors. Univariate logistic regression revealed that the T2 value was a significant predictor for differentiating HER2-zero from HER2-low expression. T2 values demonstrated moderate diagnostic performance, with an AUC of 0.817. The optimal cutoff value was 96.167 milliseconds, yielding a sensitivity of 68.2% and a specificity of 85.4%. T2 quantification derived from SyMRI shows potential as a noninvasive biomarker for identifying HER2-low-expressing breast cancer, supporting its potential role in guiding individualized treatment strategies.