Deep Learning Applications in Magnetic Resonance Imaging: Has the Future Become Present?

Abstract

Deep learning technologies and applications demonstrate one of the most important upcoming developments in radiology. The impact and influence of these technologies on image acquisition and reporting might change daily clinical practice. The aim of this review was to present current deep learning technologies, with a focus on magnetic resonance image reconstruction. The first part of this manuscript concentrates on the basic technical principles that are necessary for deep learning image reconstruction. The second part highlights the translation of these techniques into clinical practice. The third part outlines the different aspects of image reconstruction techniques, and presents a review of the current literature regarding image reconstruction and image post-processing in MRI. The promising results of the most recent studies indicate that deep learning will be a major player in radiology in the upcoming years. Apart from decision and diagnosis support, the major advantages of deep learning magnetic resonance imaging reconstruction techniques are related to acquisition time reduction and the improvement of image quality. The implementation of these techniques may be the solution for the alleviation of limited scanner availability via workflow acceleration. It can be assumed that this disruptive technology will change daily routines and workflows permanently.

Reduction in Acquisition Time and Improvement in Image Quality in T2-Weighted MR Imaging of Musculoskeletal Tumors of the Extremities Using a Novel Deep Learning-Based Reconstruction Technique in a Turbo Spin Echo (TSE) Sequence

“…In recent years, new methods have been developed to reduce MRI acquisition time as well as the extent of artifacts and to achieve more precise imaging [ 4 , 5 , 6 ]. Recently, machine learning and artificial intelligence-based algorithms have found their way into clinical radiological imaging [ 7 , 8 ]. These deep learning algorithms (DL) are based on convolutional neural networks (CNN) that were developed on the basis of the function of animal neurobiology, resembling the human neural network [ 9 ].…”

Deep Learning-Enhanced Parallel Imaging and Simultaneous Multislice Acceleration Reconstruction in Knee MRI

“…The DL reconstruction provided improved image quality through the interplay of a PI model with a trained regularizer (in particular for higher accelerations), although a smoothing effect was observed in the images. 17 The conventional SNR measuring method based on 2 separate regions of interest from a single image, 25 one in the tissue of interest for signal intensity and the other in the background for noise, was inconsistent with the true SNR value obtained in accelerated imaging because the accelerating factor and phased-array coil systems could influence the noise distribution. 10,26,27 Methods to minimize this limitation by performing multiple acquisitions have been suggested in phantom studies, 10,11 although these methods are difficult to apply in practice.…”

“…The overall network typically comprises conventional data consistency steps (based on a signal model) interleaved with trainable regularization components. 16,17 A recently published study reported that DL with 3-fold PI maintained the image quality and diagnostic confidence of conventional 2-fold PI with remarkable acquisition time reduction in musculoskeletal TSE MRI. 16 In another study, DL-reconstructed 4-fold PI performed interchangeably with conventional PI in knee MRI 15 ; however, DL was applied only to PI with 3-or 4-fold acceleration, and the image quality was analyzed only subjectively.…”

“…Currently, DL-enhanced image reconstruction adopts supervised learning techniques within convolutional neural networks. The overall network typically comprises conventional data consistency steps (based on a signal model) interleaved with trainable regularization components 16,17 . A recently published study reported that DL with 3-fold PI maintained the image quality and diagnostic confidence of conventional 2-fold PI with remarkable acquisition time reduction in musculoskeletal TSE MRI 16 .…”

Biparametric prostate MRI: impact of a deep learning-based software and of quantitative ADC values on the inter-reader agreement of experienced and inexperienced readers

“…Due to lack of extensive real-world validation, however, there are no definitive data mandating their use in clinical practice to date [ 36 ]. Despite interesting results from earliest reports of AI implementations focusing on automated detection and characterization of PCa, currently there is increasing interest for using AI technology to improve quality control and workflow efficiency in radiology [ 37 ].…”