BreastSegNet: Multi-label Segmentation of Breast MRI

2025-07-17

AI Summary

Existing breast MRI segmentation methods typically focus on a single class or a limited number of anatomical structures (e.g., tumors or fibroglandular tissue), hindering comprehensive quantitative analysis. To address this, we propose the first multi-label segmentation framework covering nine critical anatomical structures—including fibroglandular tissue, vasculature, muscle, bone, and lesions. Leveraging 1,123 expert-annotated MRI slices from radiologists, we establish a high-quality, publicly available dataset and release both model code and pre-trained weights. Methodologically, we systematically evaluate nine state-of-the-art architectures within the nnU-Net framework and identify a ResNet-based encoder variant (nnU-Net ResEncM) as optimal. Our method achieves a mean Dice score of 0.694 across all classes, with particularly strong performance on heart, liver, muscle, bone, and fibroglandular tissue (Dice > 0.85 for several). This advancement significantly enhances quantitative capabilities for breast cancer screening and preoperative staging.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Intelligent Robots: Multi-Robot Systems

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Breast MRI provides high-resolution imaging critical for breast cancer screening and preoperative staging. However, existing segmentation methods for breast MRI remain limited in scope, often focusing on only a few anatomical structures, such as fibroglandular tissue or tumors, and do not cover the full range of tissues seen in scans. This narrows their utility for quantitative analysis. In this study, we present BreastSegNet, a multi-label segmentation algorithm for breast MRI that covers nine anatomical labels: fibroglandular tissue (FGT), vessel, muscle, bone, lesion, lymph node, heart, liver, and implant. We manually annotated a large set of 1123 MRI slices capturing these structures with detailed review and correction from an expert radiologist. Additionally, we benchmark nine segmentation models, including U-Net, SwinUNet, UNet++, SAM, MedSAM, and nnU-Net with multiple ResNet-based encoders. Among them, nnU-Net ResEncM achieves the highest average Dice scores of 0.694 across all labels. It performs especially well on heart, liver, muscle, FGT, and bone, with Dice scores exceeding 0.73, and approaching 0.90 for heart and liver. All model code and weights are publicly available, and we plan to release the data at a later date.

Problem

• Multi-label segmentation of breast MRI for nine anatomical structures
• Limited scope of existing breast MRI segmentation methods
• Benchmarking performance of nine segmentation models on annotated data

Innovation

• Multi-label segmentation algorithm for breast MRI
• Benchmarked nine segmentation models including nnU-Net
• Publicly released model code and weights