AI method slashes breast MRI acquisition times
Researchers have developed an artificial intelligence-powered MRI technique that could significantly accelerate breast imaging while improving image quality and tumor detection, according to findings published in Nature Communications.
Experts from the Technion Israel Institute of Technology and the U.S. collaborated on creating the new approach, dubbed ELITE (Enhanced Locally low-rank Imaging for Tissue contrast Enhancement). The method combines deep learning with advanced mathematical modeling to generate dynamic breast MR images at a rate of just one image per second—a fraction of the time it takes to generate conventional dynamic MRI images, which can routinely take up to two minutes.
“MRI is the most effective method for screening high-risk breast cancer patients. While current exams rely on the qualitative evaluation of morphological features before and after contrast administration and less on contrast kinetic information, recent developments in fast acquisition methods aim to combine both,” lead author Eddy Solomon, PhD, MS, of the Technion’s Faculty of Biomedical Engineering, and colleagues explained in the study. “However, balancing spatial resolution, temporal resolution and scan time poses a considerable challenge in dynamic MRI."
To develop ELITE, researchers combined tissue-specific mathematical modeling with a ResNet deep neural network trained to remove image noise and distortions while reconstructing missing data from undersampled MRI acquisitions. The timing is what sets this method apart from other accelerated techniques, as the ability to track contrast as it moves through the body in real-time offers providers detailed visualization of anatomy.
When tested on 54 patients, the method proved to be superior at improving tumor visibility compared to current techniques; it also provided exceptionally high image quality with reduced noise and high diagnostic sensitivity.
The team is hopeful their technique can be used to improve access to breast MRI exams, which yield superior sensitivity for lesion detection compared to other modalities, by making them faster. However, the team also believes it could have utility for brain, head and neck MRI applications and could eventually be adapted to other imaging modalities.
Read more here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.