Original title: Generalized super-resolution 4D Flow MRI $\unicode{x2013}$ using ensemble learning to extend across the cardiovascular system
Authors: Leon Ericsson, Adam Hjalmarsson, Muhammad Usman Akbar, Edward Ferdian, Mia Bonini, Brandon Hardy, Jonas Schollenberger, Maria Aristova, Patrick Winter, Nicholas Burris, Alexander Fyrdahl, Andreas Sigfridsson, Susanne Schnell, C. Alberto Figueroa, David Nordsletten, Alistair A. Young, David Marlevi
The article focuses on improving 4D Flow MRI, a method for studying blood flow in the heart and vessels. Existing techniques often suffer from low resolution and image noise. Researchers explore using trained super-resolution networks after scanning to enhance image quality. Prior efforts were limited to specific areas of the cardiovascular system, ignoring the varied blood flow conditions. This study aims to broaden this enhancement method, using diverse training data from different cardiovascular regions. They test various learning methods and architectures on both computer-generated and real patient data. Results show that ensemble learning significantly boosts image quality across these areas, accurately predicting high-resolution details from lower-quality scans. The optimized networks successfully recover fine details from patient scans and show promise in reducing noise in clinical images. This study offers a promising way to enhance 4D Flow MRI across diverse cardiovascular regions, potentially aiding clinical diagnoses.
Original article: https://arxiv.org/abs/2311.11819