Deep learning for fast low-field MRI acquisitions

Abstract

Low-field (LF) MRI research currently gains momentum from its potential to offer reduced costs and reduced footprints translating into wider accessibility. However, the impeded signal-to-noise ratio inherent to lower magnetic fields can have a significant impact on acquisition times that challenges LF clinical relevance. Undersampling is an effective way to speed up acquisitions in MRI, and recent work has shown encouraging results when combined with deep learning (DL). Yet, training DL models generally requires large databases that are not yet available at LF regimes. Here, we demonstrate the capability of Residual U-net combined with data augmentation to reconstruct magnitude and phase information of undersampled LF MRI scans at 0.1 T with a limited training dataset (n = 10). The model performance was first evaluated in a retrospective study for different acceleration rates and sampling patterns. Ultimately, the DL approach was validated on prospectively acquired, fivefold undersampled LF data. With varying performances associated to the adopted sampling scheme, our results show that the approach investigated can preserve the global structure and the details sharpness in the reconstructed magnitude and phase images. Overall, promising results could be obtained on acquired LF MR images that may bring this research closer to clinical implementation.

Low‐field MRI: A report on the 2022 ISMRM workshop

“…He also showed examples of quantitative imaging and non-Cartesian imaging performed at 0.1T. 38,[58][59][60] Zheng Xu described work with a 2.1 MHz scanner (49 mT) for stroke imaging, which included a new permanent magnet design, active electromagnetic interference (EMI) cancelation, and excellent GRE and bSSFP image quality. 61 Session 13 focused on image reconstruction and processing methods for low-field MRI.…”

“…He emphasized that fast acquisitions must be paired with other methods to mitigate noise, and that it is important to be mindful of hardware to enable optimal, advanced acquisitions. He also showed examples of quantitative imaging and non‐Cartesian imaging performed at 0.1T 38,58–60 . Zheng Xu described work with a 2.1 MHz scanner (49 mT) for stroke imaging, which included a new permanent magnet design, active electromagnetic interference (EMI) cancelation, and excellent GRE and bSSFP image quality 61 …”

Deep learning enabled fast 3D brain MRI at 0.055 tesla

“…Alongside the increasing availability of large-scale, high-quality, and highfield human MRI data, e.g., from Human Connectome Project (HCP) Consortium and U.K. Biobank (31)(32)(33)(34), DL is expected to escalate the development of ULF MRI. Several earlier attempts have been made to reconstruct ULF images via DL (35)(36)(37). For example, one strategy is to exploit the omnipresent 3D structural features shared across humans for all organs including the brain.…”

Pushing the limits of low‐cost ultra‐low‐field MRI by dual‐acquisition deep learning 3D superresolution

“…Recently, several deep learning attempts have been made to improve ULF MRI image quality, but with limited success so far 24–27 . Image quality transfer (IQT) 28 is also proposed in a preliminary study to create artificially enhanced contrast and increase through‐plane resolution of multislice 2D T 1 ‐weighted MRI images at low field (0.36 T) with anisotropic U‐Net.…”

“…Recently, several deep learning attempts have been made to improve ULF MRI image quality, but with limited success so far. [24][25][26][27] Image quality transfer (IQT) 28 is also proposed in a preliminary study to create artificially enhanced contrast and increase through-plane resolution of multislice 2D T 1 -weighted MRI images at low field (0.36 T) with anisotropic U-Net. Some of these methods use synthetically degraded high-field images as training data for deep learning restoration of noisy and low-resolution MRI images.…”