When
Where
Student: Brian Toner, Program in Applied Mathematics
Title: Advancing Quantitative Abdominal MRI: Methods for Highly Accelerated T2 Mapping with Radial Turbo Spin-Echo Sequences
Advisors: Maria Altbach, Radiology & Imaging Science
Ali Bilgin, Biomedical Engineering
Location: Math building, Room 402
Abstract: Magnetic resonance imaging (MRI) is an especially powerful medical imaging modality due to its ability to provide excellent contrast in soft tissue without the use of ionizing radiation. Major limitations to MRI include its sensitivity to motion and slow acquisition, the combination of which creates challenges in abdominal imaging, where respiratory motion is a concern. Although MRI is inherently a qualitative modality, meaning traditional images are unitless, quantitative MRI (qMRI) has emerged as a method for using MRI to measure biomarkers within the body. qMRI typically involves fitting multiple MR images of different contrasts of the same anatomy to a physical model, further exacerbating the limitation of sensitivity to motion and slow acquisitions.
T2 is one of the main parameters that controls contrast of traditional MRI, and the parameter values have been shown to hold significant clinical utility in diagnosing liver disease. The radial turbo spin-echo (RADTSE) sequence is ideally suited for T2 mapping of the abdomen due to its robustness to motion, ability to reconstruct a time series of co-registered images of different contrasts to fit to a T2 map, and its ability to be accelerated simply by collecting less data, which shifts the burden to the image reconstruction process to create high-quality images from incomplete datasets. We aim to improve highly accelerated abdominal T2 mapping using RADTSE via three main avenues. First, we will develop pulse sequences that will sample data more efficiently and strategically. Second, we will improve the image reconstruction process using deep learning and other advanced methods to obtain high-quality images from sparse datasets. Finally, we will make the parameter estimation process more robust and statistically interpretable to ensure T2 measurements are more reliable to be used for clinical decisions.