A Path Towards Clinical Adaptation of Accelerated MRI

TL;DR

This paper proposes methods to improve the clinical applicability of accelerated MRI using deep learning, including artifact detection, training with variable acceleration, and pre-training with simulated data, to enhance image quality and robustness.

Contribution

It introduces novel augmentation techniques, a loss function for multi-anatomy training, and a pre-training approach with simulated data to facilitate clinical adoption of accelerated MRI.

Findings

Artifact detection classifier with 79.1% F2 score

Performance improvement of up to 2% during clinical scans

Effective pre-training using simulated phantom data

Abstract

Accelerated MRI reconstructs images of clinical anatomies from sparsely sampled signal data to reduce patient scan times. While recent works have leveraged deep learning to accomplish this task, such approaches have often only been explored in simulated environments where there is no signal corruption or resource limitations. In this work, we explore augmentations to neural network MRI image reconstructors to enhance their clinical relevancy. Namely, we propose a ConvNet model for detecting sources of image artifacts that achieves a classifier $F_2$ score of 79.1%. We also demonstrate that training reconstructors on MR signal data with variable acceleration factors can improve their average performance during a clinical patient scan by up to 2%. We offer a loss function to overcome catastrophic forgetting when models learn to reconstruct MR images of multiple anatomies and orientations. Finally, we propose a method for using simulated phantom data to pre-train reconstructors in situations with limited clinically acquired datasets and compute capabilities. Our results provide a potential path forward for clinical adaptation of accelerated MRI.