Real‐time deep artifact suppression using recurrent U‐Nets for low‐latency cardiac MRI

Jaubert, Olivier; Montalt-Tordera, Javier; Knight, Delvin R.; Coghlan, Gerry; Arridge, Simon R.; Steeden, Jennifer A.; Muthurangu, Vivek

doi:10.1002/mrm.28834

Cited by 21 publications

(17 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It was able to remove artifacts and segment images without user intervention and with low enough latency to provide almost real‐time monitoring. Latency could be further reduced by including (1) initialization of the pipeline before the start of acquisition to reduce the initial latency, (2) parallelization of gridding and deep artifact suppression for higher frame rate, and (3) using a memory‐based network to reconstruct the latest frame rather than blocks, while still using temporal redundancies 30 …”

Section: Discussionmentioning

confidence: 99%

FReSCO: Flow Reconstruction and Segmentation for low‐latency Cardiac Output monitoring using deep artifact suppression and segmentation

Jaubert

Montalt-Tordera

Brown

et al. 2022

Magnetic Resonance in Med

Self Cite

View full text Add to dashboard Cite

Purpose Real‐time monitoring of cardiac output (CO) requires low‐latency reconstruction and segmentation of real‐time phase‐contrast MR, which has previously been difficult to perform. Here we propose a deep learning framework for “FReSCO” (Flow Reconstruction and Segmentation for low latency Cardiac Output monitoring). Methods Deep artifact suppression and segmentation U‐Nets were independently trained. Breath‐hold spiral phase‐contrast MR data (N = 516) were synthetically undersampled using a variable‐density spiral sampling pattern and gridded to create aliased data for training of the artifact suppression U‐net. A subset of the data (N = 96) was segmented and used to train the segmentation U‐net. Real‐time spiral phase‐contrast MR was prospectively acquired and then reconstructed and segmented using the trained models (FReSCO) at low latency at the scanner in 10 healthy subjects during rest, exercise, and recovery periods. Cardiac output obtained via FReSCO was compared with a reference rest CO and rest and exercise compressed‐sensing CO. Results The FReSCO framework was demonstrated prospectively at the scanner. Beat‐to‐beat heartrate, stroke volume, and CO could be visualized with a mean latency of 622 ms. No significant differences were noted when compared with reference at rest (bias = −0.21 ± 0.50 L/min, p = 0.246) or compressed sensing at peak exercise (bias = 0.12 ± 0.48 L/min, p = 0.458). Conclusions The FReSCO framework was successfully demonstrated for real‐time monitoring of CO during exercise and could provide a convenient tool for assessment of the hemodynamic response to a range of stressors.

show abstract

Section: Discussionmentioning

confidence: 99%

FReSCO: Flow Reconstruction and Segmentation for low‐latency Cardiac Output monitoring using deep artifact suppression and segmentation

Jaubert

Montalt-Tordera

Brown

et al. 2022

Magnetic Resonance in Med

Self Cite

View full text Add to dashboard Cite

show abstract

“…Küstner et al proposed the CINENet network for 3D+time cine CMR reconstruction and showed that it outperforms iterative reconstruction in terms of visual image quality and contrast [10]. These works clearly show the benefits of the proposed architectures and are now starting to be deployed into MR scanner software [15].…”

Section: A Acquisition and Reconstructionmentioning

confidence: 96%

“…To simulate the undersampled radial k-space acquisition, the images were organized into 3D matrices. The resulting matrix was Fourier transformed along the spatial domains and each (k x -k y ) space was masked by a radial pattern with a TR of 2.6 ms. For each sample, the number of projections per frame (t) was equal to P ∈ {7, 15,23,30,38,46,53,61,69,77,84,92,100,107,115,123,130,138,146,153,161,169,176,184,192,200,207,215,223, 230} corresponding to thirty different sampling rates. Accordingly, the corresponding acceleration factors R with respect to the radial fully-sampled data were {41.96, 19.58, 12.77, 9.79, 7.73, 6.39, 5.54, 4.82, 4.26, 3.81, 3.5, 3.19, 2.94, 2.74, 2.55, 2.39, 2.26, 2.13, 2.01, 1.92, 1.83, 1.74, 1.67, 1.60, 1.53 ,1.47, 1.42, 1.37, 1.32, 1.28}.…”

Section: Simulation Of Radial Acquisition Patternmentioning

confidence: 99%

A Deep Learning-based Integrated Framework for Quality-aware Undersampled Cine Cardiac MRI Reconstruction and Analysis

Machado¹,

Puyol‐Antón²,

Hammernik³

et al. 2022

Preprint

View full text Add to dashboard Cite

Cine cardiac magnetic resonance (CMR) imaging is considered the gold standard for cardiac function evaluation. However, cine CMR acquisition is inherently slow and in recent decades considerable effort has been put into accelerating scan times without compromising image quality or the accuracy of derived results. In this paper, we present a fully-automated, quality-controlled integrated framework for reconstruction, segmentation and downstream analysis of undersampled cine CMR data. The framework enables active acquisition of radial k-space data, in which acquisition can be stopped as soon as acquired data are sufficient to produce high quality reconstructions and segmentations. This results in reduced scan times and automated analysis, enabling robust and accurate estimation of functional biomarkers. To demonstrate the feasibility of the proposed approach, we perform realistic simulations of radial k-space acquisitions on a dataset of subjects from the UK Biobank and present results on in-vivo cine CMR k-space data collected from healthy subjects. The results demonstrate that our method can produce quality-controlled images in a mean scan time reduced

show abstract

“…It can be found from the leaderboard that most state-of-the-art methods (e.g., top-10 methods) employ U-Net 39 or its variants. 27,[40][41][42] Here, U-Net is the classical encoder-decoder form of a convolutional neural network. Despite great success, these methods are not optimal and have 4 problems: (1) classical U-Net is originally designed for data in the image domain, and directly applying the image-oriented U-Net in k-space data is not optimal for extracting features in the k-space domain 43,44 ; (2) classical U-Net is considered a heavyweight method (a large number of parameters) and hence is inefficient when cascaded many times for producing high-quality reconstruction 45 ;…”

Section: Introductionmentioning

confidence: 99%

“…It can be found from the leaderboard that most state‐of‐the‐art methods (e.g., top‐10 methods) employ U‐Net 39 or its variants 27,40–42 . Here, U‐Net is the classical encoder–decoder form of a convolutional neural network.…”

Section: Introductionmentioning

confidence: 99%

Dual‐domain reconstruction network with V‐Net and K‐Net for fast MRI

Liu

Pang

Jin

et al. 2022

Magnetic Resonance in Med

View full text Add to dashboard Cite

To introduce a dual-domain reconstruction network with V-Net and K-Net for accurate MR image reconstruction from undersampled k-space data.Methods: Most state-of-the-art reconstruction methods apply U-Net or cascaded U-Nets in the image domain and/or k-space domain. Nevertheless, these methods have the following problems: (1) directly applying U-Net in the k-space domain is not optimal for extracting features; (2) classical image-domain-oriented U-Net is heavyweighted and hence inefficient when cascaded many times to yield good reconstruction accuracy; (3) classical image-domain-oriented U-Net does not make full use of information of the encoder network for extracting features in the decoder network; and (4) existing methods are ineffective in simultaneously extracting and fusing features in the image domain and its dual k-space domain. To tackle these problems, we present 3 different methods: (1) V-Net, an image-domain encoder-decoder subnetwork that is more lightweight for cascading and effective in fully utilizing features in the encoder for decoding; (2) K-Net, a k-space domain subnetwork that is more suitable for extracting hierarchical features in the k-space domain, and (3) KV-Net, a dual-domain reconstruction network in which V-Nets and K-Nets are effectively combined and cascaded. Results: Extensive experimental results on the fastMRI dataset demonstrate that the proposed KV-Net can reconstruct high-quality images and outperform state-of-the-art approaches with fewer parameters. Conclusions: To reconstruct images effectively and efficiently from incomplete k-space data, we have presented a dual-domain KV-Net to combine K-Nets and V-Nets. The KV-Net achieves better results with 9% and 5% parameters than comparable methods (XPD-Net and i-RIM).

show abstract

Real‐time deep artifact suppression using recurrent U‐Nets for low‐latency cardiac MRI

Abstract: This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Cited by 21 publications

References 29 publications

FReSCO: Flow Reconstruction and Segmentation for low‐latency Cardiac Output monitoring using deep artifact suppression and segmentation

FReSCO: Flow Reconstruction and Segmentation for low‐latency Cardiac Output monitoring using deep artifact suppression and segmentation

A Deep Learning-based Integrated Framework for Quality-aware Undersampled Cine Cardiac MRI Reconstruction and Analysis

Dual‐domain reconstruction network with V‐Net and K‐Net for fast MRI

Contact Info

Product

Resources

About