Related Experiment Video
Updated: May 21, 2026
15:48
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
Dynamic breast MRI with Flexible Temporal Resolution Aided by Deep Learning
Sungheon Gene Kim1, Jonghyun Bae1, Linda Moy2
1MRI Research Institute, Department of Radiology, Weill Cornell Medical College, Cornell University, 407 East 61st Street, New York, NY, USA.
Nature Communications
|May 19, 2026
Summary
Enhanced Locally low-rank Imaging for Tissue contrast Enhancement (ELITE) improves dynamic contrast-enhanced MRI (DCE-MRI) for breast cancer screening. This novel framework enhances image quality and allows for faster scan times, benefiting various MRI applications.
Area of Science:
- Radiology
- Medical Imaging
- Machine Learning
Background:
- Magnetic Resonance Imaging (MRI) is crucial for high-risk breast cancer screening.
- Current MRI methods primarily use morphological features, with limited use of contrast kinetic information.
- Balancing spatial resolution, temporal resolution, and scan time is a significant challenge in dynamic MRI.
Purpose of the Study:
- To introduce a novel radial MRI reconstruction framework, Enhanced Locally low-rank Imaging for Tissue contrast Enhancement (ELITE), for Dynamic Contrast Enhanced (DCE) imaging.
- To address the limitations of current dynamic MRI techniques in balancing resolution and scan time.
- To improve the effectiveness of MRI in screening and other applications.
Main Methods:
- Developed ELITE, a radial MRI reconstruction framework combining locally low-rank subspace modeling with deep learning.
- Utilized locally low-rank subspace modeling to capture spatially localized tissue dynamics.
- Evaluated ELITE using the fastMRI breast initiative dataset.
Main Results:
- Demonstrated substantial improvements in Contrast-to-Noise Ratio (CNR) and noise reduction.
- Enabled flexible temporal resolution down to 1 second.
- Showed benefits in neck and brain imaging, indicating broader applicability.
Conclusions:
- ELITE offers a viable alternative for DCE-MRI, significantly enhancing image quality and speed.
- The framework effectively addresses the challenges in dynamic MRI acquisition.
- ELITE shows promise for improving breast cancer screening and other DCE-MRI applications.
