Three-Dimensional Quantitative Validation of Breast MRI Background Parenchymal Enhancement Assessments

Introduction The MRI background parenchymal enhancement (BPE) and its clinical significance as a bio-marker of breast cancer risk has been proposed based on qualitative studies. Previous BPE quantification studies lack appropriate correlation with BPE qualitative assessments. The purpose of this study is to validate our three-dimensional (3D) BPE quantification method with standardized BPE qualitative cases. Methods An IRB approved study reviewed 500 consecutive MRI cases (from 1/2013-12/2014) using a strict inclusion criteria and 120 cases that best represented each of the BPE qualitative categories (minimal/mild/moderate/marked) were selected. Blinded to the qualitative data, fibroglandular tissue (FGT) contours of pre-and post-contrast images were delineated using an in-house, proprietary segmentation algorithm. Metrics of BPE were calculated including % BPE [(ratio of BPE volume to FGT volume)x100] at multiple threshold levels to determine the optimal cut-off point for BPE quantification that best correlated with the reference BPE qualitative cases. Results The highest positive correlation was present at 1.5x pre-SI threshold level (r=0.84, p