Abstract PS1-06-02: Breast MRI Biomarkers of Tumor Biology: Integrating Imaging with Pathology to Guide Clinical Care

Abstract Background: Breast cancer exhibits substantial biological heterogeneity, and biopsy-based molecular profiling is limited by cost, sampling error, and inability to capture spatial variability. Breast magnetic resonance imaging (MRI) offers multiparametric assessment of tumor morphology, vascularity, and peritumoral tissue, but its role as a noninvasive biomarker of tumor biology remains underexplored. Methods: We retrospectively evaluated 340 women with invasive ductal carcinoma of no special type or invasive lobular carcinoma who underwent pretreatment 3T breast MRI. Imaging features were classified according to BI-RADS and correlated with histopathological and immunohistochemical markers, including tumor grade, estrogen receptor (ER), progesterone receptor (PR), HER2, Ki-67, and molecular subtype. Multivariable logistic regression was performed, with bootstrap validation and false discovery rate correction. Results: Larger tumors (>5 cm) were strongly associated with high grade (74.4%), HER2 positivity (26.8%), elevated Ki-67 (75.0%), and HER2-enriched subtype (14.6%). Paradoxically, round/oval shape and circumscribed margins—features often deemed benign—were predictive of ER/PR negativity, high Ki-67, and triple-negative subtype (OR 3.83, 95% CI 1.48-9.94, P=0.006). In contrast, irregular/spiculated margins correlated with Luminal A tumors and low-grade biology. T2 hyperintensity and peritumoral edema were among the strongest predictors of aggressiveness, independently associated with high grade, lymphovascular invasion, HER2 positivity, and Ki-67 ≥20% (P