Python Non-Uniform Fast Fourier Transform (PyNUFFT): An Accelerated Non-Cartesian MRI Package on a Heterogeneous Platform (CPU/GPU)

A vendor-neutral EPI sequence for hyperpolarized 13C MRI

PURPOSE

To develop a flexible, vendor-neutral EPI sequence for hyperpolarized 13C metabolic imaging.

METHODS

An open-source EPI sequence consisting of a metabolite-specific spectral-spatial RF excitation pulse and a customizable EPI readout was created using the Pulseq framework. To explore the flexibility of our sequence, we tested several versions of the sequence including a symmetric 3D readout with different spatial resolutions for each metabolite (1.0 cm3 and 1.5 cm3). A multichamber phantom constructed with a Shepp-Logan geometry, containing two chambers filled with either natural abundance 13C compounds or hyperpolarized (HP) [1-13C]pyruvate, was used to test each sequence. For experiments involving HP [1-13C]pyruvate, a single chamber was prefilled with nicotinamide adenine dinucleotide hydride and lactate dehydrogenase to facilitate the conversion of [1-13C]pyruvate to [1-13C]lactate. All experiments were performed on a Siemens Prisma 3T scanner.

RESULTS

All the sequence variations localized natural-abundance 13C ethylene glycol and methanol to the appropriate compartment of the multichamber phantom. [1-13C]pyruvate was detectable in both chambers following the injection of HP [1-13C]pyruvate, whereas [1-13C]lactate was only found in the chamber containing nicotinamide adenine dinucleotide hydride and lactate dehydrogenase. The conversion rate from [1-13C]pyruvate to [1-13C]lactate (kPL) was 0.01 s-1 (95% confidence interval [0.00, 0.02]).

CONCLUSION

We have developed and tested a vendor-neutral EPI sequence for imaging HP 13C agents. We have made all of our sequence creation and image reconstruction code freely available online for other investigators to use.

Sinogram upsampling using Primal-Dual UNet for undersampled CT and radial MRI reconstruction

Computed tomography (CT) and magnetic resonance imaging (MRI) are two widely used clinical imaging modalities for non-invasive diagnosis. However, both of these modalities come with certain problems. CT uses harmful ionising radiation, and MRI suffers from slow acquisition speed. Both problems can be tackled by undersampling, such as sparse sampling. However, such undersampled data leads to lower resolution and introduces artefacts. Several techniques, including deep learning based methods, have been proposed to reconstruct such data. However, the undersampled reconstruction problem for these two modalities was always considered as two different problems and tackled separately by different research works. This paper proposes a unified solution for both sparse CT and undersampled radial MRI reconstruction, achieved by applying Fourier transform-based pre-processing on the radial MRI and then finally reconstructing both modalities using sinogram upsampling combined with filtered back-projection. The Primal-Dual network is a deep learning based method for reconstructing sparsely-sampled CT data. This paper introduces Primal-Dual UNet, which improves the Primal-Dual network in terms of accuracy and reconstruction speed. The proposed method resulted in an average SSIM of 0.932±0.021 while performing sparse CT reconstruction for fan-beam geometry with a sparsity level of 16, achieving a statistically significant improvement over the previous model, which resulted in 0.919±0.016. Furthermore, the proposed model resulted in 0.903±0.019 and 0.957±0.023 average SSIM while reconstructing undersampled brain and abdominal MRI data with an acceleration factor of 16, respectively - statistically significant improvements over the original model, which resulted in 0.867±0.025 and 0.949±0.025. Finally, this paper shows that the proposed network not only improves the overall image quality, but also improves the image quality for the regions-of-interest: liver, kidneys, and spleen; as well as generalises better than the baselines in presence the of a needle.

Real-time radial reconstruction with domain transform manifold learning for MRI-guided radiotherapy

BACKGROUND

MRI-guidance techniques that dynamically adapt radiation beams to follow tumor motion in real time will lead to more accurate cancer treatments and reduced collateral healthy tissue damage. The gold-standard for reconstruction of undersampled MR data is compressed sensing (CS) which is computationally slow and limits the rate that images can be available for real-time adaptation.

PURPOSE

Once trained, neural networks can be used to accurately reconstruct raw MRI data with minimal latency. Here, we test the suitability of deep-learning-based image reconstruction for real-time tracking applications on MRI-Linacs.

METHODS

We use automated transform by manifold approximation (AUTOMAP), a generalized framework that maps raw MR signal to the target image domain, to rapidly reconstruct images from undersampled radial k-space data. The AUTOMAP neural network was trained to reconstruct images from a golden-angle radial acquisition, a benchmark for motion-sensitive imaging, on lung cancer patient data and generic images from ImageNet. Model training was subsequently augmented with motion-encoded k-space data derived from videos in the YouTube-8M dataset to encourage motion robust reconstruction.

RESULTS

AUTOMAP models fine-tuned on retrospectively acquired lung cancer patient data reconstructed radial k-space with equivalent accuracy to CS but with much shorter processing times. Validation of motion-trained models with a virtual dynamic lung tumor phantom showed that the generalized motion properties learned from YouTube lead to improved target tracking accuracy.

CONCLUSION

AUTOMAP can achieve real-time, accurate reconstruction of radial data. These findings imply that neural-network-based reconstruction is potentially superior to alternative approaches for real-time image guidance applications.

Bias field correction in hyperpolarized 129 Xe ventilation MRI using templates derived by RF-depolarization mapping

PURPOSE

To correct for RF inhomogeneity for in vivo 129 Xe ventilation MRI using flip-angle mapping enabled by randomized 3D radial acquisitions. To extend this RF-depolarization mapping approach to create a flip-angle map template applicable to arbitrary acquisition strategies, and to compare these approaches to conventional bias field correction.

METHODS

RF-depolarization mapping was evaluated first in digital simulations and then in 51 subjects who had undergone radial 129 Xe ventilation MRI in the supine position at 3T (views = 3600; samples/view = 128; TR/TE = 4.5/0.45 ms; flip angle = 1.5; FOV = 40 cm). The images were corrected using newly developed RF-depolarization and templated-based methods and the resulting quantitative ventilation metrics (mean, coefficient of variation, and gradient) were compared to those resulting from N4ITK correction.

RESULTS

RF-depolarization and template-based mapping methods yielded a pattern of RF-inhomogeneity consistent with the expected variation based on coil architecture. The resulting corrected images were visually similar, but meaningfully distinct from those generated using standard N4ITK correction. The N4ITK algorithm eliminated the physiologically expected anterior-posterior gradient (-0.04 ± 1.56%/cm, P < 0.001). These 2 newly introduced methods of RF-depolarization and template correction retained the physiologically expected anterior-posterior ventilation gradient in healthy subjects (2.77 ± 2.09%/cm and 2.01 ± 2.73%/cm, respectively).

CONCLUSIONS

Randomized 3D 129 Xe MRI ventilation acquisitions can inherently be corrected for bias field, and this technique can be extended to create flip angle templates capable of correcting images from a given coil regardless of acquisition strategy. These methods may be more favorable than the de facto standard N4ITK because they can remove undesirable heterogeneity caused by RF effects while retaining results from known physiology.

ReconResNet: Regularised residual learning for MR image reconstruction of Undersampled Cartesian and Radial data

MRI is an inherently slow process, which leads to long scan time for high-resolution imaging. The speed of acquisition can be increased by ignoring parts of the data (undersampling). Consequently, this leads to the degradation of image quality, such as loss of resolution or introduction of image artefacts. This work aims to reconstruct highly undersampled Cartesian or radial MR acquisitions, with better resolution and with less to no artefact compared to conventional techniques like compressed sensing. In recent times, deep learning has emerged as a very important area of research and has shown immense potential in solving inverse problems, e.g. MR image reconstruction. In this paper, a deep learning based MR image reconstruction framework is proposed, which includes a modified regularised version of ResNet as the network backbone to remove artefacts from the undersampled image, followed by data consistency steps that fusions the network output with the data already available from undersampled k-space in order to further improve reconstruction quality. The performance of this framework for various undersampling patterns has also been tested, and it has been observed that the framework is robust to deal with various sampling patterns, even when mixed together while training, and results in very high quality reconstruction, in terms of high SSIM (highest being 0.990 ± 0.006 for acceleration factor of 3.5), while being compared with the fully sampled reconstruction. It has been shown that the proposed framework can successfully reconstruct even for an acceleration factor of 20 for Cartesian (0.968 ± 0.005) and 17 for radially (0.962 ± 0.012) sampled data. Furthermore, it has been shown that the framework preserves brain pathology during reconstruction while being trained on healthy subjects.

Improved Image Quality for Static BLADE Magnetic Resonance Imaging Using the Total-Variation Regularized Least Absolute Deviation Solver

In order to improve the image quality of BLADE magnetic resonance imaging (MRI) using the index tensor solvers and to evaluate MRI image quality in a clinical setting, we implemented BLADE MRI reconstructions using two tensor solvers (the least-squares solver and the L1 total-variation regularized least absolute deviation (L1TV-LAD) solver) on a graphics processing unit (GPU). The BLADE raw data were prospectively acquired and presented in random order before being assessed by two independent radiologists. Evaluation scores were examined for consistency and then by repeated measures analysis of variance (ANOVA) to identify the superior algorithm. The simulation showed the structural similarity index (SSIM) of various tensor solvers ranged between 0.995 and 0.999. Inter-reader reliability was high (Intraclass correlation coefficient (ICC) = 0.845, 95% confidence interval: 0.817, 0.87). The image score of L1TV-LAD was significantly higher than that of vendor-provided image and the least-squares method. The image score of the least-squares method was significantly lower than that of the vendor-provided image. No significance was identified in L1TV-LAD with a regularization strength of λ= 0.4–1.0. The L1TV-LAD with a regularization strength of λ= 0.4–0.7 was found consistently better than least-squares and vendor-provided reconstruction in BLADE MRI with a SENSitivity Encoding (SENSE) factor of 2. This warrants further development of the integrated computing system with the scanner.

Short-Time Fourier Transform Based on Metaprogramming and the Stockham Optimization Method

The extension for high-performance STFT (Short-Time Fourier Transform) algorithm written entirely in Java language for non-parallel computations is presented in the current paper. This solution could compete with the best available and most common algorithms supplied by libraries such as FFTW or JTransform. The main idea was to move complex computations and expensive functions to the program generation phase. Thus, only core and essential operations were executed during the runtime phase. Furthermore, new approach allows to eliminate the necessity for a rearrangement operation that uses the bit-reversal permutation technique. This article presents a brief description of the STFT solution that was worked out as an extension for the original application, in order to increase its efficiency. The solution remains a Stockham algorithm adapted using metaprogramming techniques and entails an additional reduction its execution time. Performance tests and experiments were conducted using a Java Platform and JMH library, which allowed for accurate execution time measurements. Major aspects of the Java VM like warm-up effects were also taken into consideration. Solution was applied into Electrical Capacitance Tomography measurement system in order to measure the material changes during the silo discharging industrial process.

Gradient-based and wavelet-based compressed sensing approaches for highly undersampled tomographic datasets

Electron tomography is widely employed for the 3D morphological characterization at the nanoscale. In recent years, there has been a growing interest in analytical electron tomography (AET) as it is capable of providing 3D information about the elemental composition, chemical bonding and optical/electronic properties of nanomaterials. AET requires advanced reconstruction algorithms as the datasets often consist of a very limited number of projections. Total variation (TV)-based compressed sensing approaches were shown to provide high-quality reconstructions from undersampled datasets, but staircasing artefacts can appear when the assumption about piecewise constancy does not hold. In this paper, we compare higher-order TV and wavelet-based approaches for AET applications and provide an open-source Python toolbox, Pyetomo, containing 2D and 3D implementations of both methods. A highly sampled STEM-HAADF dataset of an Er-doped porous Si sample and a heavily undersampled STEM-EELS dataset of a Ge-rich GeSbTe (GST) thin film annealed at 450°C are used to evaluate the performance of the different approaches. We show that polynomial annihilation with order 3 (HOTV₃) and the Bior4.4 wavelet outperform the classical TV minimization and the related Haar wavelet.

Ground-truth-free deep learning for artefacts reduction in 2D radial cardiac cine MRI using a synthetically generated dataset

In this work, we consider the task of image reconstruction in 2D radial cardiac cine MRI using deep learning (DL)-based regularization. As the regularization is achieved by employing an image-prior predicted by a pre-trained convolutional neural network (CNN), the quality of the image-prior is of essential importance. The achievable performance of any DL-based method is limited by the amount and the quality of the available training data. For fast dynamic processes, obtaining good-quality MR data is challenging because of technical and physiological reasons. In this work, we try to overcome these problems by a transfer-learning approach which is motivated by a previously presented DL-method (XT,YT U-Net). There, instead of training the network on the whole 2D dynamic images, it is trained on 2D spatio-temporal profiles (xt,yt-slices) which show the temporal changes of the imaged object. Therefore, for the training and test data, it is more important that their spatio-temporal profiles share similar local features rather than being images of the same anatomy. This allows us to equip arbitrary data with simulated motion that resembles the cardiac motion and use it as training data. By doing so, it is possible to train a CNN which is applicable to cardiac cine MR data without using ground-truth cine MR images for training. We demonstrate that combining XT,YT U-Net with the proposed transfer-learning strategy delivers comparable performance to CNNs trained on cardiac cine MR images and in some cases even qualitatively surpasses these. Additionally, the transfer-learning strategy was investigated for a 2D and 3D U-Net. The images processed by the the CNNs were used as image-priors in the CNN-regularized iterative reconstruction. The XT,YT U-Net yielded visibly better results than the 2D U-Net and slightly better results than the 3D U-Net when used in combination with the presented transfer learning-strategy.

Adaptive sparsity level and dictionary size estimation for image reconstruction in accelerated 2D radial cine MRI

PURPOSE

In the past, dictionary learning (DL) and sparse coding (SC) have been proposed for the regularization of image reconstruction problems. The regularization is given by a sparse approximation of all image patches using a learned dictionary, that is, an overcomplete set of basis functions learned from data. Despite its competitiveness, DL and SC require the tuning of two essential hyperparameters: the sparsity level S - the number of basis functions of the dictionary, called atoms, which are used to approximate each patch, and K - the overall number of such atoms in the dictionary. These two hyperparameters usually have to be chosen a priori and are determined by repetitive and computationally expensive experiments. Furthermore, the final reported values vary depending on the specific situation. As a result, the clinical application of the method is limited, as standardized reconstruction protocols have to be used.

METHODS

In this work, we use adaptive DL and propose a novel adaptive sparse coding algorithm for two-dimensional (2D) radial cine MR image reconstruction. Using adaptive DL and adaptive SC, the optimal dictionary size K as well as the optimal sparsity level S are chosen dependent on the considered data.

RESULTS

Our three main results are the following: First, adaptive DL and adaptive SC deliver results which are comparable or better than the most widely used nonadaptive version of DL and SC. Second, the time needed for the regularization is accelerated due to the fact that the sparsity level S is never overestimated. Finally, the a priori choice of S and K is no longer needed but is optimally chosen dependent on the data under consideration.

CONCLUSIONS

Adaptive DL and adaptive SC can highly facilitate the application of DL- and SC-based regularization methods. While in this work we focused on 2D radial cine MR image reconstruction, we expect the method to be applicable to different imaging modalities as well.

CG-SENSE revisited: Results from the first ISMRM reproducibility challenge

PURPOSE

The aim of this work is to shed light on the issue of reproducibility in MR image reconstruction in the context of a challenge. Participants had to recreate the results of "Advances in sensitivity encoding with arbitrary k-space trajectories" by Pruessmann et al. METHODS: The task of the challenge was to reconstruct radially acquired multicoil k-space data (brain/heart) following the method in the original paper, reproducing its key figures. Results were compared to consolidated reference implementations created after the challenge, accounting for the two most common programming languages used in the submissions (Matlab/Python).

RESULTS

Visually, differences between submissions were small. Pixel-wise differences originated from image orientation, assumed field-of-view, or resolution. The reference implementations were in good agreement, both visually and in terms of image similarity metrics.

DISCUSSION AND CONCLUSION

While the description level of the published algorithm enabled participants to reproduce CG-SENSE in general, details of the implementation varied, for example, density compensation or Tikhonov regularization. Implicit assumptions about the data lead to further differences, emphasizing the importance of sufficient metadata accompanying open datasets. Defining reproducibility quantitatively turned out to be nontrivial for this image reconstruction challenge, in the absence of ground-truth results. Typical similarity measures like NMSE of SSIM were misled by image intensity scaling and outlier pixels. Thus, to facilitate reproducibility, researchers are encouraged to publish code and data alongside the original paper. Future methodological papers on MR image reconstruction might benefit from the consolidated reference implementations of CG-SENSE presented here, as a benchmark for methods comparison.

Deep learning-based image reconstruction and motion estimation from undersampled radial k-space for real-time MRI-guided radiotherapy

To enable magnetic resonance imaging (MRI)-guided radiotherapy with real-time adaptation, motion must be quickly estimated with low latency. The motion estimate is used to adapt the radiation beam to the current anatomy, yielding a more conformal dose distribution. As the MR acquisition is the largest component of latency, deep learning (DL) may reduce the total latency by enabling much higher undersampling factors compared to conventional reconstruction and motion estimation methods. The benefit of DL on image reconstruction and motion estimation was investigated for obtaining accurate deformation vector fields (DVFs) with high temporal resolution and minimal latency. 2D cine MRI acquired at 1.5 T from 135 abdominal cancer patients were retrospectively included in this study. Undersampled radial golden angle acquisitions were retrospectively simulated. DVFs were computed using different combinations of conventional- and DL-based methods for image reconstruction and motion estimation, allowing a comparison of four approaches to achieve real-time motion estimation. The four approaches were evaluated based on the end-point-error and root-mean-square error compared to a ground-truth optical flow estimate on fully-sampled images, the structural similarity (SSIM) after registration and time necessary to acquire k-space, reconstruct an image and estimate motion. The lowest DVF error and highest SSIM were obtained using conventional methods up to [Formula: see text]. For undersampling factors [Formula: see text], the lowest DVF error and highest SSIM were obtained using conventional image reconstruction and DL-based motion estimation. We have found that, with this combination, accurate DVFs can be obtained up to [Formula: see text] with an average root-mean-square error up to 1 millimeter and an SSIM greater than 0.8 after registration, taking 60 milliseconds. High-quality 2D DVFs from highly undersampled k-space can be obtained with a high temporal resolution with conventional image reconstruction and a deep learning-based motion estimation approach for real-time adaptive MRI-guided radiotherapy.

Bulk motion-compensated DCE-MRI for functional imaging of kidneys in newborns

BACKGROUND

Evaluation of kidney function in newborns with hydronephrosis is important for clinical decisions. Dynamic contrast-enhanced (DCE) MRI can provide the necessary anatomical and functional information. Golden angle dynamic radial acquisition and compressed sensing reconstruction provides sufficient spatiotemporal resolution to achieve accurate parameter estimation for functional imaging of kidneys. However, bulk motion during imaging (rigid or nonrigid movement of the subject resulting in signal dropout) remains an unresolved challenge.

PURPOSE

To evaluate a motion-compensated (MoCo) DCE-MRI technique for robust evaluation of kidney function in newborns. Our method includes: 1) motion detection, 2) motion-robust image reconstruction, 3) joint realignment of the volumes, and 4) tracer-kinetic (TK) model fitting to evaluate kidney function parameters.

STUDY TYPE

Retrospective.

SUBJECTS

Eleven newborn patients (ages <6 months, 6 female).

FIELD STRENGTH/SEQUENCE

3T; dynamic "stack-of-stars" 3D fast low-angle shot (FLASH) sequence using a multichannel body-matrix coil.

ASSESSMENT

We evaluated the proposed technique in terms of the signal-to-noise ratio (SNR) of the reconstructed images, the presence of discontinuities in the contrast agent concentration time curves due to motion with a total variation (TV) metric and the goodness of fit of the TK model, and the standard variation of its parameters.

STATISTICAL TESTS

We used a paired t-test to compare the MoCo and no-MoCo results.

RESULTS

The proposed MoCo method successfully detected motion and improved the SNR by 3.3 (P = 0.012) and decreased TV by 0.374 (P = 0.017) across all subjects. Moreover, it decreased nRMSE of the TK model fit for the subjects with less than five isolated bulk motion events in 6 minutes (mean 1.53, P = 0.043), but not for the subjects with more frequent events or no motion (P = 0.745 and P = 0.683).

DATA CONCLUSION

Our results indicate that the proposed MoCo technique improves the image quality and accuracy of the TK model fit for subjects who present isolated bulk motion events.

LEVEL OF EVIDENCE

3 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;52:207-216.