High-quality MR Images Research Articles

A noise-robust post-processing pipeline for accelerated phase-cycled 23Na Multi-Quantum Coherences MRI.

To develop an improved post-processing pipeline for noise-robust accelerated phase-cycled Cartesian Single (SQ) and Triple Quantum (TQ) sodium (23Na) Magnetic Resonance Imaging (MRI) of in vivo human brain at 7 T. Our pipeline aims to tackle the challenges of 23Na Multi-Quantum Coherences (MQC) MRI including low Signal-to-Noise Ratio (SNR) and time-consuming Radiofrequency (RF) phase-cycling. Our method combines low-rank k-space denoising for SNR enhancement with Dynamic Mode Decomposition (DMD) to robustly separate SQ and TQ signal components. This separation is crucial for computing the TQ/SQ ratio, a key parameter of 23Na MQC MRI. We validated our pipeline in silico, in vitro and in vivo in healthy volunteers, comparing it with conventional denoising and Fourier transform (FT) methods. Additionally, we assessed its robustness through ablation experiments simulating a corrupted RF phase-cycle step. Our denoising algorithm doubled SNR compared to non-denoised images and enhanced SNR by up to 29% compared to Wavelet denoising. The low-rank approach produced high-quality images even at later echo times, allowing reduced signal averaging. DMD effectively separated the SQ and TQ signals, even with missing RF phase cycle steps, resulting in superior Structural Similarity (SSIM) of 0.89±0.024 and lower Root Mean Squared Error (RMSE) of 0.055±0.008 compared to conventional FT methods (SSIM=0.71±0.061, RMSE=0.144±0.036). This pipeline enabled high-quality 8x8x15mm3 in vivo 23Na MQC MRI, with a reduction in acquisition time from 48 to 10 min at 7 T. The proposed pipeline improves robustness in 23Na MQC MRI by exploiting low-rank properties to denoise signals and DMD to effectively separate SQ and TQ signals. This approach ensures high-quality MR images of both SQ and TQ components, even in accelerated and incomplete RF phase-cycling cases.

Using High-Pass Filter to Enhance Scan Specific Learning for MRI Reconstruction without Any Extra Training Data

In accelerated MRI, the robust artificial-neural-network for k-space interpolation (RAKI) method is an attractive learning-based reconstruction that does not require additional training data. This study was focused on obtaining high quality MR images from regular under-sampled multi-coil k-space data using a high-pass filtered RAKI (HP-RAKI) reconstruction without any extra training data. MRI scan from human subjects was under-sampled with a regular pattern using skipped phase encoding and a fully sampled k-space center. A high-pass (HP) filter was applied in k-space to reduce image support to facilitate linear prediction. The HP filtered k-space center was used to train the RAKI network without any extra training data. The unacquired k-space data can be predicted from a trained RAKI network with optimized parameters. Final reconstruction was obtained after performing an inverse HP filtering for the predicted k-space data. This HP-RAKI method can be extended to corresponding residual structure (HP-rRAKI). HP-RAKI was compared with GRAPPA, HP-GRAPPA, RAKI and MW-RAKI algorithms, and HP-rRAKI was compared with corresponding residual extensions, including rRAKI and MW-rRAKI, all qualitatively and quantitatively using visual inspection and such metrics as SSIM and PSNR. HP-RAKI and HP-rRAKI were found to be effective in reconstructing MR images even at high acceleration factors. HP-RAKI and HP-rRAKI compared favorably with other algorithms. Using high-pass filtered central k-space data for training, HP-RAKI offers higher reconstruction quality for regularly under-sampled multi-coil k-space data without any extra training data. It has shown promising capabilities for fast MRI applications, especially those lacking fully sampled training data.

FCSSL: fusion enhanced contrastive self-supervised learning method for parallel MRI reconstruction

Objective. The implementation of deep learning in magnetic resonance imaging (MRI) has significantly advanced the reduction of data acquisition times. However, these techniques face substantial limitations in scenarios where acquiring fully sampled datasets is unfeasible or costly.Approach. To tackle this problem, we propose a fusion enhanced contrastive self-supervised learning (FCSSL) method for parallel MRI reconstruction, eliminating the need for fully sampledk-space training dataset and coil sensitivity maps. First, we introduce a strategy based on two pairs of re-undersampling masks within a contrastive learning framework, aimed at enhancing the representational capacity to achieve higher quality reconstruction. Subsequently, a novel adaptive fusion network, trained in a self-supervised learning manner, is designed to integrate the reconstruction results of the framework.Results. Experimental results on knee datasets under different sampling masks demonstrate that the proposed FCSSL achieves superior reconstruction performance compared to other self-supervised learning methods. Moreover,the performance of FCSSL approaches that of the supervised methods, especially under the 2DRU and RADU masks, but no need for fully sample data. The proposed FCSSL, trained under the 3× 1DRU and 2DRU masks, can effectively generalize to unseen 1D and 2D undersampling masks, respectively. For target domain data that exhibit significant differences from source domain data, the proposed model, fine-tuned with just a few dozen instances of undersampled data in the target domain, achieves reconstruction performance comparable to that achieved by the model trained with the entire set of undersampled data.Significance. The novel FCSSL model offers a viable solution for reconstructing high-quality MR images without needing fully sampled datasets, thereby overcoming a major hurdle in scenarios where acquiring fully sampled MR data is difficult.

Motion-Compensated MR CINE Reconstruction With Reconstruction-Driven Motion Estimation.

In cardiac CINE, motion-compensated MR reconstruction (MCMR) is an effective approach to address highly undersampled acquisitions by incorporating motion information between frames. In this work, we propose a novel perspective for addressing the MCMR problem and a more integrated and efficient solution to the MCMR field. Contrary to state-of-the-art (SOTA) MCMR methods which break the original problem into two sub-optimization problems, i.e. motion estimation and reconstruction, we formulate this problem as a single entity with one single optimization. Our approach is unique in that the motion estimation is directly driven by the ultimate goal, reconstruction, but not by the canonical motion-warping loss (similarity measurement between motion-warped images and target images). We align the objectives of motion estimation and reconstruction, eliminating the drawbacks of artifacts-affected motion estimation and therefore error-propagated reconstruction. Further, we can deliver high-quality reconstruction and realistic motion without applying any regularization/smoothness loss terms, circumventing the non-trivial weighting factor tuning. We evaluate our method on two datasets: 1) an in-house acquired 2D CINE dataset for the retrospective study and 2) the public OCMR cardiac dataset for the prospective study. The conducted experiments indicate that the proposed MCMR framework can deliver artifact-free motion estimation and high-quality MR images even for imaging accelerations up to 20x, outperforming SOTA non-MCMR and MCMR methods in both qualitative and quantitative evaluation across all experiments. The code is available at https://github.com/JZPeterPan/MCMR-Recon-Driven-Motion.

High-resolution 3T to 7T ADC map synthesis with a hybrid CNN-transformer model.

7 Tesla (7T) apparent diffusion coefficient (ADC) maps derived from diffusion-weighted imaging (DWI) demonstrate improved image quality and spatial resolution over 3 Tesla (3T) ADC maps. However, 7T magnetic resonance imaging (MRI) currently suffers from limited clinical unavailability, higher cost, and increased susceptibility to artifacts. To address these issues, we propose a hybrid CNN-transformer model to synthesize high-resolution 7T ADC maps from multimodal 3T MRI. The Vision CNN-Transformer (VCT), composed of both Vision Transformer (ViT) blocks and convolutional layers, is proposed to produce high-resolution synthetic 7T ADC maps from 3T ADC maps and 3T T1-weighted (T1w) MRI. ViT blocks enabled global image context while convolutional layers efficiently captured fine detail. The VCT model was validated on the publicly available Human Connectome Project Young Adult dataset, comprising 3T T1w, 3T DWI, and 7T DWI brain scans. The Diffusion Imaging in Python library was used to compute ADC maps from the DWI scans. A total of 171 patient cases were randomly divided into 130 training cases, 20 validation cases, and 21 test cases. The synthetic ADC maps were evaluated by comparing their similarity to the ground truth volumes with the following metrics: peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and mean squared error (MSE). In addition, RESULTS: The results are as follows: PSNR: 27.0±0.9dB, SSIM: 0.945±0.010, and MSE: 2.0E-3±0.4E-3. Both qualitative and quantitative results demonstrate that VCT performs favorably against other state-of-the-art methods. We have introduced various efficiency improvements, including the implementation of flash attention and training on 176×208 resolution images. These enhancements have resulted in the reduction of parameters and training time per epoch by 50% in comparison to ResViT. Specifically, the training time per epoch has been shortened from 7.67 min to 3.86 min. We propose a novel method to predict high-resolution 7T ADC maps from low-resolution 3T ADC maps and T1w MRI. Our predicted images demonstrate better spatial resolution and contrast compared to 3T MRI and prediction results made by ResViT and pix2pix. These high-quality synthetic 7T MR images could be beneficial for disease diagnosis and intervention, producing higher resolution and conformal contours, and as an intermediate step in generating synthetic CT for radiation therapy, especially when 7T MRI scanners are unavailable.

MFTN: Multi-Level Feature Transfer Network Based on MRI-Transformer for MR Image Super-resolution

Due to the unique environment and inherent properties of magnetic resonance imaging (MRI) instruments, MR images typically have lower resolution. Therefore, improving the resolution of MR images is beneficial for assisting doctors in diagnosing the condition. Currently, the existing MR image super-resolution (SR) methods still have the problem of insufficient detail reconstruction. To overcome this issue, this paper proposes a multi-level feature transfer network (MFTN) based on MRI-Transformer to realize SR of low-resolution MRI data. MFTN consists of a multi-scale feature reconstruction network (MFRN) and a multi-level feature extraction branch (MFEB). MFRN is constructed as a pyramid structure to gradually reconstruct image features at different scales by integrating the features obtained from MFEB, and MFEB is constructed to provide detail information at different scales for low resolution MR image SR reconstruction by constructing multiple MRI-Transformer modules. Each MRI-Transformer module is designed to learn the transfer features from the reference image by establishing feature correlations between the reference image and low-resolution MR image. In addition, a contrast learning constraint item is added to the loss function to enhance the texture details of the SR image. A large number of experiments show that our network can effectively reconstruct high-quality MR Images and achieves better performance compared to some state-of-the-art methods. The source code of this work will be released on GitHub.

What Is the Influence of Femoral Version on Size, Tear Location, and Tear Pattern of the Acetabular Labrum in Patients With FAI?

Femoral version deformities have recently been identified as a major contributor to femoroacetabular impingement (FAI). An in-depth understanding of the specific labral damage patterns caused by femoral version deformities may help to understand the underlying pathomorphologies in symptomatic patients and select the appropriate surgical treatment. We asked: (1) Is there a correlation between femoral version and the mean cross-sectional area of the acetabular labrum? (2) Is there a difference in the location of lesions of the acetabular labrum between hips with increased femoral version and hips with decreased femoral version? (3) Is there a difference in the pattern of lesions of the acetabular labrum between hips with increased femoral version and hips with decreased femoral version? This was a retrospective, comparative study. Between November 2009 and September 2016, we evaluated 640 hips with FAI. We considered patients with complete diagnostic imaging including magnetic resonance arthrography (MRA) of the affected hip with radial slices of the proximal femur and axial imaging of the distal femoral condyles (allowing for calculation of femoral version) as eligible. Based on that, 97% (620 of 640 hips) were eligible; a further 77% (491 of 640 hips) were excluded because they had either normal femoral version (384 hips), incomplete imaging (20 hips), a lateral center-edge angle < 22° (43 hips) or > 39° (16 hips), age > 50 years (8 hips), or a history of pediatric hip disease (20 hips), leaving 20% (129 of 640 hips) of patients with a mean age of 27 ± 9 years for analysis, and 61% (79 of 129 hips) were female. Patients were assigned to either the increased (> 30°) or decreased (< 5°) femoral version group. The labral cross-sectional area was measured on radial MR images in all patients. The location-dependent labral cross-sectional area, presence of labral tears, and labral tear patterns were assessed using the acetabular clockface system and compared among groups. In hips with increased femoral version, the labrum was normal in size (21 ± 6 mm 2 [95% confidence interval 20 to 23 mm 2 ]), whereas hips with decreased femoral version showed labral hypotrophy (14 ± 4 mm 2 [95% CI 13 to 15 mm 2 ]; p < 0.01). In hips with increased femoral version, labral tears were located more anteriorly (median 1:30 versus 12:00; p < 0.01). Hips with increased femoral version exhibited damage of the anterior labrum with more intrasubstance tears anterosuperiorly (17% [222 of 1322] versus 9% [93 of 1084]; p < 0.01) and partial tears anteroinferiorly (22% [36 of 165] versus 6% [8 of 126]; p < 0.01). Hips with decreased femoral version showed superior labral damage consisting primarily of partial labral tears. In the evaluation of patients with FAI, the term "labral tear" is not accurate enough to describe labral pathology. Based on high-quality radial MR images, surgeons should always evaluate the combination of labral tear location and labral tear pattern, because these may provide insight into associated femoral version abnormalities, which can inform appropriate surgical treatment. Future studies should examine symptomatic patients with normal femoral version, as well as an asymptomatic control group, to describe the effect of femoral version on labral morphology across the entire spectrum of pathomorphologies. Level III, prognostic study.

DC-SiamNet: Deep contrastive Siamese network for self-supervised MRI reconstruction

Reconstruction methods based on deep learning have greatly shortened the data acquisition time of magnetic resonance imaging (MRI). However, these methods typically utilize massive fully sampled data for supervised training, restricting their application in certain clinical scenarios and posing challenges to the reconstruction effect when high-quality MR images are unavailable. Recently, self-supervised methods have been developed that only undersampled MRI images participate in the network training. Nevertheless, due to the lack of complete referable MR image data, self-supervised reconstruction is prone to produce incorrect structure contents, such as unnatural texture details and over-smoothed tissue sites. To solve this problem, we propose a self-supervised Deep Contrastive Siamese Network (DC-SiamNet) for fast MR imaging. First, DC-SiamNet performs the reconstruction with a Siamese unrolled structure and obtains visual representations in different iterative phases. Particularly, an attention-weighted average pooling module is employed at the bottleneck layer of the U-shape regularization unit, which can effectively aggregate valuable local information of the underlying feature map in the generated representation vector. Then, a novel hybrid loss function is designed to drive the self-supervised reconstruction and contrastive learning simultaneously by forcing the output consistency across different branches in the frequency domain, the image domain, and the latent space. The proposed method is extensively evaluated with different sampling patterns on the IXI brain dataset and the MRINet knee dataset. Experimental results show that DC-SiamNet can achieve 0.93 in structural similarity and 33.984 dB in peak signal-to-noise ratio on the IXI brain dataset under 8x acceleration. It has better reconstruction accuracy than other methods, and the performance is close to the corresponding model trained with full supervision, especially when the sampling rate is low. In addition, generalization experiments verify that our method has a strong cross-domain reconstruction ability for different contrast brain images.

Brain metastasis resection: the impact of fluorescence guidance (MetResect study).

Maximal resection of brain metastases (BMs) improves both progression-free survival and overall survival (OS). Fluorescein sodium (FL) in combination with the YELLOW 560-nm filter is a safe and feasible method for visualizing residual tumor tissue during BM resection. The authors of this study aimed to show that use of FL would positively influence the volumetric extent of resection (EOR) and thus the survival outcome in patients undergoing BM resection. Analyzing their institution's prospective brain tumor registry, the authors identified 539 consecutive patients with BMs (247 women, mean age 62.8 years) by using preoperative high-quality MR images for volumetric analysis. BMs were resected under white light (WL) in 293 patients (54.4%; WL group) and under FL guidance in 246 patients (45.6%; FL group). Sex, age, presurgical Karnofsky Performance Status (KPS), recursive partitioning analysis class, and adjuvant treatment modalities were well balanced between the two groups. Volumetric analysis was performed in a blinded fashion by quantifying pre- and postoperative tumor volume based on gadolinium-enhanced T1-weighted sequences. In the FL group, the postoperative tumor volume was significantly smaller (p = 0.01), and hence the quantitative EOR was significantly larger (p = 0.024) and OS was significantly longer (p = 0.0001) (log-rank testing). Multivariate Cox regression modeling showed that age, presurgical KPS, metastasis status, and FL-guided resection are independent prognostic factors for survival. Compared with WL resection, FL-guided BM resection increased resection quality, significantly improved EOR, and prolonged OS.

Improving accelerated 3D imaging in MRI-guided radiotherapy for prostate cancer using a deep learning method

PurposeThis study was to improve image quality for high-speed MR imaging using a deep learning method for online adaptive radiotherapy in prostate cancer. We then evaluated its benefits on image registration.MethodsSixty pairs of 1.5 T MR images acquired with an MR-linac were enrolled. The data included low-speed, high-quality (LSHQ), and high-speed low-quality (HSLQ) MR images. We proposed a CycleGAN, which is based on the data augmentation technique, to learn the mapping between the HSLQ and LSHQ images and then generate synthetic LSHQ (synLSHQ) images from the HSLQ images. Five-fold cross-validation was employed to test the CycleGAN model. The normalized mean absolute error (nMAE), peak signal-to-noise ratio (PSNR), structural similarity index measurement (SSIM), and edge keeping index (EKI) were calculated to determine image quality. The Jacobian determinant value (JDV), Dice similarity coefficient (DSC), and mean distance to agreement (MDA) were used to analyze deformable registration.ResultsCompared with the LSHQ, the proposed synLSHQ achieved comparable image quality and reduced imaging time by ~ 66%. Compared with the HSLQ, the synLSHQ had better image quality with improvement of 57%, 3.4%, 26.9%, and 3.6% for nMAE, SSIM, PSNR, and EKI, respectively. Furthermore, the synLSHQ enhanced registration accuracy with a superior mean JDV (6%) and preferable DSC and MDA values compared with HSLQ.ConclusionThe proposed method can generate high-quality images from high-speed scanning sequences. As a result, it shows potential to shorten the scan time while ensuring the accuracy of radiotherapy.

Federated Learning of Generative Image Priors for MRI Reconstruction.

Multi-institutional efforts can facilitate training of deep MRI reconstruction models, albeit privacy risks arise during cross-site sharing of imaging data. Federated learning (FL) has recently been introduced to address privacy concerns by enabling distributed training without transfer of imaging data. Existing FL methods employ conditional reconstruction models to map from undersampled to fully-sampled acquisitions via explicit knowledge of the accelerated imaging operator. Since conditional models generalize poorly across different acceleration rates or sampling densities, imaging operators must be fixed between training and testing, and they are typically matched across sites. To improve patient privacy, performance and flexibility in multi-site collaborations, here we introduce Federated learning of Generative IMage Priors (FedGIMP) for MRI reconstruction. FedGIMP leverages a two-stage approach: cross-site learning of a generative MRI prior, and prior adaptation following injection of the imaging operator. The global MRI prior is learned via an unconditional adversarial model that synthesizes high-quality MR images based on latent variables. A novel mapper subnetwork produces site-specific latents to maintain specificity in the prior. During inference, the prior is first combined with subject-specific imaging operators to enable reconstruction, and it is then adapted to individual cross-sections by minimizing a data-consistency loss. Comprehensive experiments on multi-institutional datasets clearly demonstrate enhanced performance of FedGIMP against both centralized and FL methods based on conditional models.

Accelerating GRAPPA reconstruction using SoC design for real-time cardiac MRI

Real-time cardiac MRI is a rapidly developing area of research that has the potential to improve the diagnosis and treatment of cardiovascular diseases. However, the acquisition of high-quality real-time cardiac MR (CMR) images is challenging as it requires a high frame rate and temporal resolution. To overcome this challenge, there have been recent efforts on several approaches including hardware-based improvements and image reconstruction techniques such as compressed sensing and parallel MRI. The use of parallel MRI techniques such as GRAPPA (Generalized Autocalibrating Partial Parallel Acquisition) is a promising approach for improving the temporal resolution of MRI and expanding its applications in clinical practice. However, the GRAPPA algorithm involves a significant amount of computation, particularly for high acceleration factors and large datasets. This can result in long reconstruction times, which can limit the ability to achieve real-time imaging or high frame rates. One solution to this challenge is to use specialized hardware i.e. field-programmable gate arrays (FPGAs). In this work, a novel 32-bit floating-point FPGA-based GRAPPA accelerator is proposed with an aim to reconstruct high-quality cardiac MR images at higher frame rates, making it well suited for real-time clinical applications. The proposed FPGA-based accelerator consists of custom-designed data processing units named as dedicated computational engines (DCEs) that allow for a continuous flow of data between the calibration and synthesis stages of GRAPPA reconstruction process. This greatly increases the throughput and reduces the latency of the overall proposed system. Moreover, a high-speed memory module (DDR4-SDRAM) is integrated with the proposed architecture to store the multi-coil MR data. An on-chip quad-core ARM Cortex-A53 processor is used to manage access control information required for data transfer between the DCEs and DDR4-SDRAM. The proposed accelerator is implemented on Xilinx Zynq UltraScale + MPSoC using high-level synthesis (HLS) and hardware descriptive language (HDL) with an aim to explore the trade-offs between the reconstruction time, resource utilization and design effort. Several experiments have been performed using in-vivo cardiac datasets i.e. 18-receiver coil and 30-receiver coil to evaluate the performance of the proposed accelerator. A comparison is performed with the contemporary CPU and GPU-based GRAPPA reconstruction methods in terms of reconstruction time, frames-per-second and reconstruction accuracy (RMSE and SNR). The results show that the proposed accelerator achieves speed-up factors up to 121× and 9× as compared to the contemporary CPU-based and GPU-based GRAPPA reconstruction methods, respectively. Moreover, it has been demonstrated that the proposed accelerator can achieve reconstruction rates of up to ∼27 frames-per-second while maintaining the visual quality of the reconstructed images.

Fast MRI reconstruction using StrainNet with dual-domain loss on spatial and frequency spaces

One of the main challenges to obtain a high throughput in the MRI process is a slow signal acquisition. This process could be improved using a parallel imaging technique, where fewer raw data with multiple radio frequency (RF) coils are acquired simultaneously to reconstruct a final MR image. Nowadays, all multi-coil MRI machines have a parallel imaging technique for the image reconstruction. However, the parallel imaging still cannot accelerate sufficiently to reduce the overall acquisition time. In another way, this paper proposes a solution relying on a deep convolution neural network (CNN) to generate high-quality reconstruction MR images with higher acceleration factors. The proposed method, called StrainNet, performs the reconstructions by encoding the under-sampled data (i.e., for the speeding up process) into high-level features. Then the important part of the network, called Strainer, is applied to discard irrelevance information, and decodes remaining features back to reconstruct MR images. The proposed network could be trained end-to-end with a newly presented loss function, Dual-Domain Loss (DDL), combining both spatial and frequency losses. The experimental results are based on the fastMRI dataset and show that StrainNet outperforms the competing methods for both 4- and 8-fold accelerations.

Project-Oriented RF Coil Comparison and Optimization for Preclinical, Single-Voxel MR Spectroscopy of the Rat Visual Cortex at 9.4 T

As one of the most important cortical structures in the brain, the visual cortex plays a primary role in processing visual data. Observing the metabolic changes in this region using single-voxel MR spectroscopy (MRS) can provide a plethora of information on the disease stage and progress in vivo. For single-voxel MRS experiments, a target voxel is always chosen and specifically investigated. The volume of the voxel must be sufficiently small to avoid any interference from adjacent areas but large enough to maximise its signal-to-noise ratio (SNR). Of the factors used to improve SNR, e.g. enlarging voxel size or increasing the number of signal averaging, the optimisation of the RF coil can contribute significantly to the SNR improvement. In this work, we compared and optimised RF coils which include a circularly-polarised (CP) high-pass birdcage coil and two CP surface coils with different dimensions. The coils were particularly designed for the entire rat head, brain and the target region (visual cortex), respectively. The performance of these coils was evaluated based on the project-specific purpose of investigating metabolite concentration changes in the visual cortex of the in vivo rat brain, assessed by means of single-voxel MRS at 9.4 T. In the SNR comparison, the values obtained using the optimised small loop coil were found to be approximately 5 and 1.35 times higher than those using the birdcage volume head coil and the medium-sized loop coil, respectively. In other words, using the optimised coil rather than, for example, the birdcage coil allows the acquisition to be completed more quickly by a factor of 25 times. In conclusion, by using the optimised setup, we can i) benefit either from reduced measurement time or in the acquisition of high-quality MR images or ii) obtain spectra with a minimum quantification error.

Fast MRI Reconstruction via Edge Attention

Fast and accurate MRI reconstruction is a key concern in modern clinical practice. Recently, numerous Deep-Learning methods have been proposed for MRI reconstruction, however, they usually fail to reconstruct sharp details from the subsampled k-space data. To solve this problem, we propose a lightweight and accurate Edge Attention MRI Reconstruction Network (EAMRI) to reconstruct images with edge guidance. Specifically, we design an efficient Edge Prediction Network to directly predict accurate edges from the blurred image. Meanwhile, we propose a novel Edge Attention Module (EAM) to guide the image reconstruction utilizing the extracted edge priors, as inspired by the popular self-attention mechanism. EAM first projects the input image and edges into Q_image, K_edge, and V_image, respectively. Then EAM pairs the Q_image with K_edge along the channel dimension, such that 1) it can search globally for the high-frequency image features that are activated by the edge priors; 2) the overall computation burdens are largely reduced compared with the traditional spatial-wise attention. With the help of EAM, the predicted edge priors can effectively guide the model to reconstruct high-quality MR images with accurate edges. Extensive experiments show that our proposed EAMRI outperforms other methods with fewer parameters and can recover more accurate edges.

Virtual coil augmentation for MR coil extrapoltion via deep learning

Magnetic resonance imaging (MRI) is a widely used medical imaging modality. However, due to the limitations in hardware, scan time, and throughput, it is often clinically challenging to obtain high-quality MR images. In this article, we propose a method of using artificial intelligence to expand the coils to achieve the goal of generating the virtual coils. The main characteristic of our work is utilizing dummy variable technology to expand/extrapolate the receive coils in both image and k-space domains. The high-dimensional information formed by coil expansion is used as the prior information to improve the reconstruction performance of parallel imaging. Two main components are incorporated into the network design, namely variable augmentation technology and sum of squares (SOS) objective function. Variable augmentation provides the network with more high-dimensional prior information, which is helpful for the network to extract the deep feature information of the data. The SOS objective function is employed to solve the deficiency of k-space data training while speeding up convergence. Experimental results demonstrated its great potentials in accelerating parallel imaging reconstruction.

Multi-scale deformable transformer for multi-contrast knee MRI super-resolution

Background and objective:Magnetic resonance imaging can present the precise anatomic structure in clinical applications. Nevertheless, due to the limited scanning equipment cost, scanning time and so on, high-resolution knee MR images are difficult to obtain. So the super-resolution technique is developed to improve the image quality. Unfortunately, conventional CNN-based methods cannot explicitly learn the long-range dependencies in images and simply integrate the auxiliary contrast without considering the characteristics of medical images. To tackle this issue, our approach aims to adaptively capture and fuse the significant auxiliary information of the multi-contrast images to improve the knee magnetic resonance image quality. Methods:We propose a multi-scale deformable transformer network (MSDT) for multi-contrast knee magnetic resonance imaging super-resolution. First, we aggregate multi-scale patch embedding from the multi-contrast knee MR images to effectively preserve the local contextual details and global structure information. Then, the deformable transformer architecture is designed to learn the data-dependent sparse attention of the knee MR image, which can adaptively obtain the high-frequency foreground details according to the image content. Results:The proposed method is evaluated on the fastMRI dataset under 2× and 4× enlargements. Our MSDT achieves higher PSNR of 31.98 and SSIM of 0.713 at 2× upsampling factor and PSNR of 30.38 and SSIM of 0.615 at 4× upsampling factor. Moreover, our method can generate clear tissue structures and fine details. Conclusions:The experimental results show superior performance in comparison to the state-of-the-art super-resolution methods. This indicates that the MSDT can effectively reconstruct the high-quality knee MR images.

Style transfer in conditional GANs for cross-modality synthesis of brain magnetic resonance images

Magnetic resonance imaging (MRI) has become one of the most standardized and widely used neuroimaging protocols in the detection and diagnosis of neurodegenerative diseases. In clinical scenarios, multi-modality MR images can provide more comprehensive information than single modality images. However, high-quality multi-modality MR images can be difficult to obtain in the actual diagnostic process due to various uncertainties. Efficient methods of modality complement and synthesis have aroused increasing attention in the research community. In this article, style transfer is introduced into conditional generative adversarial networks (cGAN) architecture. A cGAN model with hierarchical feature mapping and fusion (ST-cGAN) is proposed to address the cross-modality synthesis of MR images. In order to surmount the sole focus on the pixel-wise similarity as most cGAN-based methods do, the proposed ST-cGAN takes advantage of the style information and applies it to the synthetic image’s content structure. Taking images of two modalities as conditional input, ST-cGAN extracts different levels of style features and integrates them with the content features to form the style-enhanced synthetic image. Furthermore, the proposed model is made robust to random noise by adding noise input to the generator. A comprehensive analysis is performed by comparing the proposed ST-cGAN with other state-of-the-art baselines based on four representative evaluation metrics. The experimental results on the IXI (Information eXtraction from Images) dataset verify the validity of the ST-cGAN from different evaluation perspectives.

Swin transformer for fast MRI

Magnetic resonance imaging (MRI) is an important non-invasive clinical tool that can produce high-resolution and reproducible images. However, a long scanning time is required for high-quality MR images, which leads to exhaustion and discomfort of patients, inducing more artefacts due to voluntary movements of the patients and involuntary physiological movements. To accelerate the scanning process, methods by k-space undersampling and deep learning based reconstruction have been popularised. This work introduced SwinMR, a novel Swin transformer based method for fast MRI reconstruction. The whole network consisted of an input module (IM), a feature extraction module (FEM) and an output module (OM). The IM and OM were 2D convolutional layers and the FEM was composed of a cascaded of residual Swin transformer blocks (RSTBs) and 2D convolutional layers. The RSTB consisted of a series of Swin transformer layers (STLs). The shifted windows multi-head self-attention (W-MSA/SW-MSA) of STL was performed in shifted windows rather than the multi-head self-attention (MSA) of the original transformer in the whole image space. A novel multi-channel loss was proposed by using the sensitivity maps, which was proved to reserve more textures and details. We performed a series of comparative studies and ablation studies in the Calgary-Campinas public brain MR dataset and conducted a downstream segmentation experiment in the Multi-modal Brain Tumour Segmentation Challenge 2017 dataset. The results demonstrate our SwinMR achieved high-quality reconstruction compared with other benchmark methods, and it shows great robustness with different undersampling masks, under noise interruption and on different datasets. The code is publicly available athttps://github.com/ayanglab/SwinMR.

Watch and wait approach for rectal cancer: lessons learned in a rectal cancer center

Background: Watch and wait approach (W&W) is now considered an acceptable option in a selected group of patients with low rectal cancer after neoadjuvant chemoradiation (nCRT). Several studies reported similar oncologic outcomes for patients that achieve a clinical complete response (cCR) and those with a pathologic complete response (pCR) after radical surgery. However, several limitations of the available data contributed to the lack of a widespread adoption of a W&W approach, including the unavailability of high-quality MR imaging from baseline and assessment of tumor response, heterogeneity in treatment strategies and surveillance protocols.