Relationships Between MRI Breast Imaging-Reporting and Data System (BI-RADS) Lexicon Descriptors and Breast Cancer...

The aim of this study was to determine the associations between breast MRI findings using the Breast Imaging-Reporting and Data System (BI-RADS) lexicon descriptors and breast cancer molecular subtypes. In this retrospective, IRB-approved, single institution study MRIs from 278 women with breast cancer were reviewed by one of six fellowship-trained breast imagers. Readers reported BI-RADS descriptors for breast masses (shape, margin, internal enhancement) and non-mass enhancement (distribution, internal enhancement). Pathology reports were reviewed for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2). Surrogates were used to categorize tumors by molecular subtype: ER/PR+, HER2- (luminal A); ER/PR+, HER2+ (luminal B); ER/PR-, HER2+ (HER2); ER/PR/HER2- (basal). A univariate logistic regression model was developed to identify associations between BI-RADS descriptors and molecular subtypes. Internal enhancement for mass and non-mass enhancement was combined for analysis. There was an association between mass shape and basal subtype (p = 0.039), which was more frequently round (17.1%) than other subtypes (range: 0-8.3%). In addition, there was an association between mass margin and HER2 subtype (p = 0.040), as HER2 cancers more frequently had a smooth margin (33.3%) than other subtypes (range: 4.2-17.1%). Finally, there was an association between internal enhancement and luminal B subtype (p = 0.003), with no cases of luminal B cancer demonstrating homogeneous internal enhancement versus a range of 10.9-23.5% for other subtypes. There are associations between breast cancer molecular subtypes and lesion appearance on MRI using the BI-RADS lexicon.