Rapid MR relaxometry using deep learning: An overview of current techniques and emerging trends

Abstract

Quantitative mapping of MR tissue parameters such as the spin‐lattice relaxation time (T1), the spin‐spin relaxation time (T2), and the spin‐lattice relaxation in the rotating frame (T1ρ), referred to as MR relaxometry in general, has demonstrated improved assessment in a wide range of clinical applications. Compared with conventional contrast‐weighted (eg T1‐, T2‐, or T1ρ‐weighted) MRI, MR relaxometry provides increased sensitivity to pathologies and delivers important information that can be more specific to tissue composition and microenvironment. The rise of deep learning in the past several years has been revolutionizing many aspects of MRI research, including image reconstruction, image analysis, and disease diagnosis and prognosis. Although deep learning has also shown great potential for MR relaxometry and quantitative MRI in general, this research direction has been much less explored to date. The goal of this paper is to discuss the applications of deep learning for rapid MR relaxometry and to review emerging deep‐learning‐based techniques that can be applied to improve MR relaxometry in terms of imaging speed, image quality, and quantification robustness. The paper is comprised of an introduction and four more sections. Section 2 describes a summary of the imaging models of quantitative MR relaxometry. In Section 3, we review existing “classical” methods for accelerating MR relaxometry, including state‐of‐the‐art spatiotemporal acceleration techniques, model‐based reconstruction methods, and efficient parameter generation approaches. Section 4 then presents how deep learning can be used to improve MR relaxometry and how it is linked to conventional techniques. The final section concludes the review by discussing the promise and existing challenges of deep learning for rapid MR relaxometry and potential solutions to address these challenges.

Acceleration of Magnetic Resonance Fingerprinting Reconstruction Using Denoising and Self-Attention Pyramidal Convolutional Neural Network

“…MRF is a favorable approach to obtaining quantitative MR relaxation measurements. In addition, quantitative MR relaxometry can synthesize conventional contrast weightings [ 2 , 3 ], which can be useful for adherence to current clinical diagnostic standards. Furthermore, quantitative MRI relaxometry–based tissue segmentation was reported to have favorable repeatability [ 39 ] and can be beneficial in clinical settings for tracking the time course of a disease.…”

“…Quantitative magnetic resonance (MR) relaxometry can quantify the relaxation time (e.g., T1, T2, T2* relaxation time) to clarify the physical and pathological properties of human tissues [ 1 ]. Quantitative MR relaxometry was reported to increase accuracy and precision compared with conventional weighted magnetic resonance imaging (MRI) in detecting lesions, and it can even synthesize traditional weighted images [ 2 , 3 ]. However, clinical applications of quantitative MR relaxometry are limited by the length of the imaging procedure required to estimate the tissue relaxation time; moreover, motion artifacts can interfere with the results, and the procedure does not meet the needs for clinical scheduling efficiency.…”

“…In addition to deep learning studies of MRF reconstruction, one study used generative adversarial networks to speed up the generation of simulation data [ 14 ]. As the graphics hardware and deep learning algorithms mature, MRI imaging techniques can be optimized with deep learning to improve computational performance and thus increase the feasibility of clinical applications [ 2 ].…”

Cardiac MR: From Theory to Practice

“…This review provides an overview of five essential aspects of CMR which have been covered separately in-depth in other review papers (2)(3)(4)(5)(6)(7)(8)(9)(10)(11). We address: (1) data acquisition sequences and common artifacts, (2) clinical applications, (3) clinical examination protocols, (4) image acceleration, reconstruction, and motion handling, (5) artificial intelligenceassisted reconstruction and analysis.…”

Fast calculation software for modified Look-Locker inversion recovery (MOLLI) T1 mapping

“…The interface setup facilitates the comparison between Python-based and C++-based methods. Moreover, the Python language serves as a framework for deep learning libraries [14][15][16] and is commonly adopted for the development of deep learning algorithms, which may have potential for improving the performance in cardiac T1 mapping [17]. In an earlier study, we demonstrated fast T1 map calculation using the LM-based method implemented in C++, as a module for a comprehensive quantitative cardiac MRI analysis tool [18].…”