Harmonizing BPE Assessment in Contrast-Enhanced Mammography: A Critical Step Toward AI-Driven Breast Cancer Risk Stratification (Part 1)

Introduction Breast density (BD) influences breast cancer risk and back-ground parenchymal enhancement (BPE) in contrast-enhanced mammography (CEM), yet BPE assessment lacks standardization. This study establishes observational foun-dations for a unified classification sys-tem. Materials and Methods We analyzed 213 CEM cases (retrospective single-center co-hort, 2022–2023). BD was categorized via ACR BI-RADS (A–D), and BPE via a novel 4-tier scale (BPE-CEM Standard Scale, BCSS: Minimal/Light/Moderate/Marked). In-terobserver agreement (3 radiologists; 50 random cases) and BD-BPE correlations were assessed using Cohen’s kappa and regression analysis. ResultsHigher BD (BI-RADS C/D) significantly cor-related with elevated BPE (p < 0.05; R² = 0.144). BPE distribution: Minimal (57%), Light (31%), Moderate (10%), Marked (2%). The BCSS showed excellent reproducibility (κ = 0.85; 95% CI: 0.78–0.92). Conclusion BD is a key determinant of BPE in CEM. The BCSS provides a reproducible framework for standardized assessment, reducing interpretive variability. This obser-vational groundwork enables computational automation in Parte 2