Seminar June 20 TH, 2023

Anthony Christodoulou

Title: Embracing beats and breaths: MR Multitasking for motion-resolved quantitative cardiac imaging

Abstract: Quantitative MRI provides many benefits over traditional qualitative imaging: reproducible tissue characterization, diagnosis of diffuse disease, the potential for earlier disease detection, and more. The standard approach to quantitative MRI of moving organs (e.g., the heart or abdominal organs) has been to “freeze” motion using a complicated mixture of ECG triggering and repeated breath holds. That approach is difficult, unreliable, and most importantly, unsuitable for patients with irregular heartbeats or trouble breath-holding, preventing the wide clinical adoption of quantitative MR in many areas. This seminar describes a new class of approaches to quantitative imaging, which redesign the MR imaging process around the concept of multiple time dimensions. Rather than trying to avoid motion, these approaches “multitask”, capturing motion alongside multiple simultaneous tissue processes for quantification—each of which is assigned its own time dimension. This allows fast, accurate, and repeatable motion-resolved quantitative imaging, and enables non-ECG, free-breathing quantification of multiple tissue parameters at once.

Bio: Anthony Christodoulou is an Associate Professor of Biomedical Sciences and the Director of Magnetic Resonance Technology Innovations for the Biomedical Imaging Research Institute at Cedars-Sinai Medical Center, as well as faculty in Medicine and Bioengineering at the University of California, Los Angeles. He received his Ph.D. in Electrical and Computer Engineering from the University of Illinois at Urbana-Champaign and his B.S. and M.S. degrees in Electrical Engineering from the University of Southern California. Prof. Christodoulou’s research laboratory develops and translates novel magnetic resonance imaging (MRI) techniques through innovations in MR physics, machine learning, and image reconstruction. His group’s primary focus is on multidimensional quantitative imaging methods for the diagnosis, risk prediction, and treatment monitoring of cardiovascular diseases and cancer. Dr. Christodoulou is an Executive Board member of the SMRA, an Annual Meeting Program Committee member of the ISMRM, a Fellow of the SCMR, and a member of the IEEE and AHA.

Aaron Ponce Sandoval

Title: Improving Temporal Segmentation in 4D Flow MRI Images Using a Semantic Segmentation Network Cascade and VoxelMorph Registration Neural Network

Abstract: The 4D flow MRI acquisition technique allows non-invasive assessment of blood flow velocities within a volume in all three orthogonal directions. However, to analyze these
acquisitions, segmentation is necessary. Obtaining segmentation remains a challenge, due
to different issues, such as random acquisition noise, low spatial and temporal resolution,
velocity aliasing, breath motion, and lags. For this reason, the use of a single segmentation
to represent the geometry throughout all cardiac phases has been standardized. However,
recent studies have proposed the use of image registration to solve this problem. In this
paper we propose a post-processing pipeline for 4D flow magnetic resonance images. The
pipeline consists of obtaining a unique segmentation of the branches of the pulmonary artery
and sections of the aorta to improve segmentation over time in cardiac phases of systole.
This is obtained by adapting a 4D flow MRI imaging processing algorithm using a
VoxelMorph medical imaging recording neural network and a semantic segmentation neural
network cascade.

Bio: Machine learning engineer and phd student in applied informatics from the university
of valparaíso.

Galería