Deep MRI reconstruction with radial subsampling

Yiasemis, George; Zhang, Chaoping; Sánchez, Clara I.; Sonke, Jan‐Jakob; Teuwen, Jonas

doi:10.1117/12.2609876

Cited by 7 publications

(8 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…More precisely, we repeated the scheme-specific experiments on the CC dataset by replacing the RecurrentVarNet with a RIM. The choice of hyper-parameters for each RIM was identical as in [19].…”

Section: Results -Model Dependencementioning

confidence: 99%

“…For instance, non-Cartesian patterns such as radial or spiral have been shown to be applied in real-time MRI acquisitions due to the fact that they are less susceptible to motion compared to Cartesian ones [18]. The authors in [19] by employing a deep neural network architecture, namely the Recurrent Inference Machine (RIM) [13], explored the effects of training RIMs by applying either rectilinear or radial retrospective subsampling and concluded that the RIM trained using the latter can produce higher-fidelity reconstructions.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction

Yiasemis¹,

Sánchez²,

Sonke³

et al. 2023

Preprint

View full text Add to dashboard Cite

Purpose: The MRI k-space acquisition is time consuming. Traditional techniques aim to acquire accelerated data, which in conjunction with recent DL methods, aid in producing high-fidelity images in truncated times. Conventionally, subsampling the k-space is performed by utilizing Cartesianrectilinear trajectories, which even with the use of DL, provide imprecise reconstructions, though, a plethora of non-rectilinear or non-Cartesian trajectories can be implemented in modern MRI scanners. This work investigates the effect of the k-space subsampling scheme on the quality of reconstructed accelerated MRI measurements produced by trained DL models. Methods:The RecurrentVarNet was used as the DL-based MRI-reconstruction architecture. Cartesian fully-sampled multi-coil k-space measurements from three datasets with different accelerations were retrospectively subsampled using eight distinct subsampling schemes (four Cartesian-rectilinear, two Cartesian non-rectilinear, two non-Cartesian). Experiments were conducted in two frameworks: Scheme-specific, where a distinct model was trained and evaluated for each dataset-subsampling scheme pair, and multi-scheme, where for each dataset a single model was trained on data randomly subsampled by any of the eight schemes and evaluated on data subsampled by all schemes.Results: In the scheme-specific setting RecurrentVarNets trained and evaluated on non-rectilinearly subsampled data demonstrated superior performance especially for high accelerations, whilst in the multi-scheme setting, reconstruction performance on rectilinearly subsampled data improved when compared to the scheme-specific experiments. Conclusion:Training DL-based MRI reconstruction algorithms on non-rectilinearly subsampled measurements can produce more faithful reconstructions. Our findings demonstrate the potential for using DL-based methods trained on prospective acquisitions with non-rectilinearly subsampled measurements to optimize scan time and image quality.

show abstract

Section: Results -Model Dependencementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction

Yiasemis¹,

Sánchez²,

Sonke³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…In some papers, it is either the case that the relevance of the incoherence of a measurement matrix required for subsampling is watered down like in [6], or as observed in [5] the subsampling techniques largely proposed for signal reconstruction still have an implicit significant probability of coherence within the sensing matrix, with an insufficient theoretical framework to provide certainty of a successful CS method. In [4] and [12] The rest of the paper is organized as follows;…”

Section: Review Of Related Papersmentioning

confidence: 99%

An Enhanced Subsampling Technique in Compressive Sensing using Linear Interpolation and Random Measurement Matrix

Osei-Wusu,

Ahene,

Muntaka

2024

Preprint

View full text Add to dashboard Cite

In Compressive Sensing, the incoherence of a measurement matrix during subsampling is a crucial requirement for the accurate reconstruction of a signal. However, such incoherence is only probable and not assured when subsampling is done with the widely used random measurement matrix. The study proposes an enhanced subsampling technique that integrates linear interpolation with the conventional random measurement matrix to provide assured incoherence during subsampling in Compressive Sensing. The experiments show that the proposed technique is less costly computationally and does a faster subsampling of an audio digital signal than when the traditional random measurement matrix is used solely. Additionally, the results demonstrated that the proposed technique outperformed state-of-the-art techniques with respect to the accuracy and speed of the signal reconstruction along with the L1 optimization. This was proven through the use of performance evaluation metrics such as computational complexity, execution time and Mean Square Error.

show abstract

“…DL has also been applied to reconstruct undersampled non‐Cartesian acquisitions, particularly radial trajectories. For nondynamic anatomical imaging, AUTOmated transform by Manifold APproximation (AUTOMAP) 10 was employed to directly map undersampled radial k‐space to unaliased image space, and Recurrent Inference Machine (RIM) architecture was used as an unrolled‐loop network with data consistency to reconstruct undersampled radial data 13 . For dynamic radial MRI, DL was employed to reconstruct 4D images, where the fourth dimension is respiratory motion 14–16 .…”

Section: Introductionmentioning

confidence: 99%

“…For nondynamic anatomical imaging, AUTOmated transform by Manifold APproximation (AUTOMAP) 10 was employed to directly map undersampled radial k-space to unaliased image space, and Recurrent Inference Machine (RIM) architecture was used as an unrolled-loop network with data consistency to reconstruct undersampled radial data. 13 For dynamic radial MRI, DL was employed to reconstruct 4D images, where the fourth dimension is respiratory motion. [14][15][16] Imaging of respiratory motion also allows for the acquisition of a reference image during multiple respiratory cycles.…”

mentioning

confidence: 99%

GRASPNET: Fast spatiotemporal deep learning reconstruction of golden‐angle radial data for free‐breathing dynamic contrast‐enhanced magnetic resonance imaging

et al. 2022

View full text Add to dashboard Cite

The purpose of the current study was to develop a deep learning technique called Golden‐angle RAdial Sparse Parallel Network (GRASPnet) for fast reconstruction of dynamic contrast‐enhanced 4D MRI acquired with golden‐angle radial k‐space trajectories. GRASPnet operates in the image‐time space and does not use explicit data consistency to minimize the reconstruction time. Three different network architectures were developed: (1) GRASPnet‐2D: 2D convolutional kernels (x,y) and coil and contrast dimensions collapsed into a single combined dimension; (2) GRASPnet‐3D: 3D kernels (x,y,t); and (3) GRASPnet‐2D + time: two 3D kernels to first exploit spatial correlations (x,y,1) followed by temporal correlations (1,1,t). The networks were trained using iterative GRASP reconstruction as the reference. Free‐breathing 3D abdominal imaging with contrast injection was performed on 33 patients with liver lesions using a T1‐weighted golden‐angle stack‐of‐stars pulse sequence. Ten datasets were used for testing. The three GRASPnet architectures were compared with iterative GRASP results using quantitative and qualitative analysis, including impressions from two body radiologists. The three GRASPnet techniques reduced the reconstruction time to about 13 s with similar results with respect to iterative GRASP. Among the GRASPnet techniques, GRASPnet‐2D + time compared favorably in the quantitative analysis. Spatiotemporal deep learning enables reconstruction of dynamic 4D contrast‐enhanced images in a few seconds, which would facilitate translation to clinical practice of compressed sensing methods that are currently limited by long reconstruction times.

show abstract

Deep MRI reconstruction with radial subsampling

Cited by 7 publications

References 29 publications

On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction

On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction

An Enhanced Subsampling Technique in Compressive Sensing using Linear Interpolation and Random Measurement Matrix

GRASPNET: Fast spatiotemporal deep learning reconstruction of golden‐angle radial data for free‐breathing dynamic contrast‐enhanced magnetic resonance imaging

Contact Info

Product

Resources

About