Mask-Guided Multi-Channel SwinUNETR Framework for Robust MRI Classification

2025-08-28

AI Summary

To address the limited sensitivity of mammography in women with dense breasts and high-risk populations, this paper proposes a robust classification method for multicenter breast MRI. The method builds upon the SwinUNETR architecture and introduces two key innovations: (i) a mask-guided mechanism incorporating breast region masks as auxiliary input channels, and (ii) a large-scale adaptive data augmentation strategy combined with ensemble learning. These enhancements collectively improve lesion feature sensitivity and cross-scanner, cross-institution generalizability. Evaluated on the ODELIA multicenter challenge, the method achieved second place, demonstrating superior clinical applicability—achieving a 6.2% absolute improvement in classification accuracy over baseline methods—alongside strong robustness and transferability across heterogeneous imaging protocols. The results validate its potential as a reliable AI-powered tool for early breast cancer diagnosis in real-world clinical settings.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Breast cancer is one of the leading causes of cancer-related mortality in women, and early detection is essential for improving outcomes. Magnetic resonance imaging (MRI) is a highly sensitive tool for breast cancer detection, particularly in women at high risk or with dense breast tissue, where mammography is less effective. The ODELIA consortium organized a multi-center challenge to foster AI-based solutions for breast cancer diagnosis and classification. The dataset included 511 studies from six European centers, acquired on scanners from multiple vendors at both 1.5 T and 3 T. Each study was labeled for the left and right breast as no lesion, benign lesion, or malignant lesion. We developed a SwinUNETR-based deep learning framework that incorporates breast region masking, extensive data augmentation, and ensemble learning to improve robustness and generalizability. Our method achieved second place on the challenge leaderboard, highlighting its potential to support clinical breast MRI interpretation. We publicly share our codebase at https://github.com/smriti-joshi/bcnaim-odelia-challenge.git.

Problem

• Classifying breast lesions as benign or malignant using MRI
• Improving robustness of AI models for multi-center MRI data
• Addressing limitations of mammography in dense breast tissue

Innovation

• SwinUNETR-based deep learning framework
• Incorporates breast region masking
• Extensive data augmentation and ensemble learning