Breast density in MRI: an AI-based quantification and relationship to assessment in mammography

2025-04-21

AI Summary

Conventional mammographic breast density assessment has limitations in sensitivity and specificity, particularly for dense breasts, necessitating complementary quantitative modalities. Method: We developed a novel AI-driven framework for fully automated 3D breast density quantification from multi-center T1- and T2-weighted MRI, incorporating a custom machine learning algorithm for precise breast tissue segmentation and volumetric density estimation; cross-dataset robustness was systematically validated. Contribution/Results: MRI-derived density exhibited high stability (0.104–0.114), age-dependent decline, and significant but systematically biased correlation with mammographic density. Crucially, MRI uniquely captured dense tissue components undetectable on mammography, demonstrating intrinsic modality complementarity. This work establishes a new paradigm for multimodal breast density integration and provides a validated computational tool to enhance breast cancer risk stratification beyond conventional mammographic assessment.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

Mammographic breast density is a well-established risk factor for breast cancer. Recently there has been interest in breast MRI as an adjunct to mammography, as this modality provides an orthogonal and highly quantitative assessment of breast tissue. However, its 3D nature poses analytic challenges related to delineating and aggregating complex structures across slices. Here, we applied an in-house machine-learning algorithm to assess breast density on normal breasts in three MRI datasets. Breast density was consistent across different datasets (0.104 - 0.114). Analysis across different age groups also demonstrated strong consistency across datasets and confirmed a trend of decreasing density with age as reported in previous studies. MR breast density was correlated with mammographic breast density, although some notable differences suggest that certain breast density components are captured only on MRI. Future work will determine how to integrate MR breast density with current tools to improve future breast cancer risk prediction.

Problem

• AI-based quantification of breast density in MRI
• Relationship between MRI and mammography density assessments
• Analytic challenges in 3D MRI breast density analysis

Innovation

• AI-based breast density quantification in MRI
• Machine-learning algorithm for 3D MRI analysis
• Correlating MRI and mammographic density metrics