Related Experiment Video
Updated: May 15, 2026
15:48
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
Atlas-based probabilistic fibroglandular tissue segmentation in breast MRI
Shandong Wu1, Susan Weinstein, Despina Kontos
1Computational Breast Imaging Group, Department of Radiology, University of Pennsylvania, Philadelphia, PA, USA. shandong.wu@uphs.upenn.edu
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|January 5, 2013
Summary
This study introduces an automated method for segmenting fibroglandular tissue in breast MRI, enhancing breast cancer risk assessment. The novel atlas-aided approach accurately estimates tissue volume, aiding early detection and risk evaluation.
Area of Science:
- Medical Imaging
- Radiology
- Computational Anatomy
Background:
- Accurate segmentation of fibroglandular tissue in breast MRI is crucial for breast cancer risk assessment.
- Existing methods may lack precision in characterizing complex tissue patterns.
Purpose of the Study:
- To develop and validate an atlas-aided probabilistic model-based segmentation method for fibroglandular tissue in breast MRI.
- To improve the accuracy of fibroglandular tissue volume estimation for breast cancer risk assessment.
Main Methods:
- A novel fibroglandular tissue atlas was learned using deformable image warping and kernel density estimation.
- A mixture multivariate model characterized breast tissue based on MR image features.
- Segmentation was performed by incorporating the learned atlas as prior probability in a probabilistic model.
Main Results:
- The automated segmentation method achieved a Dice's Similarity Coefficient (DSC) of 0.85 when compared to manual segmentations.
- Experimental results on 10 cases demonstrated high agreement with expert radiologist segmentations.
- The method effectively estimates the volumetric amount of fibroglandular tissue.
Conclusions:
- The proposed atlas-aided segmentation method provides an accurate and automated approach for fibroglandular tissue estimation in breast MRI.
- This technique holds potential for enhancing breast cancer risk stratification and early detection strategies.
- The learned atlas significantly aids in improving segmentation accuracy and reliability.
