MRISIMUL: a GPU-based parallel approach to MRI simulations

Particle-based MR modeling with diffusion, microstructure, and enzymatic reactions

PURPOSE

To develop a novel particle-based in silico MR model and demonstrate applications of this model to signal mechanisms which are affected by the spatial organization of particles, including metabolic reaction kinetics, microstructural effects on diffusion, and radiofrequency (RF) refocusing effects in gradient-echo sequences.

METHODS

The model was developed by integrating a forward solution of the Bloch equations with a Brownian dynamics simulator. Simulation configurations were then designed to model MR signal dynamics of interest, with a primary focus on hyperpolarized 13C MRI methods. Phantom scans and spectrophotometric assays were conducted to validate model results in vitro.

RESULTS

The model accurately reproduced the reaction kinetics of enzyme-mediated conversion of pyruvate to lactate. When varying proportions of restrictive structure were added to the reaction volume, nonlinear changes in the reaction rate measured in vitro were replicated in silico. Modeling of RF refocusing effects characterized the degree of diffusion-weighted contribution from preserved residual magnetization in nonspoiled gradient-echo sequences.

CONCLUSIONS

These results show accurate reproduction of a range of MR signal mechanisms, establishing the model's capability to investigate the multifactorial signal dynamics such as those underlying hyperpolarized 13C MRI data.

Unravelling the mechanisms of CE-SSFP in imaging myocardium at risk: The effect of relaxation times on myocardial contrast

PURPOSE

The aim of this study was to investigate the contrast mechanisms of Contrast-enhanced steady-state free-precession (CE-SSFP) through the utilization of Bloch simulations in an experimental porcine model and in patients with acute myocardial infarction.

METHODS

Six pigs and ten patients with myocardial infarction underwent CMR and tissue characterization at 1.5 T whereas a Bloch simulation framework was utilized to simulate the CE-SSFP signal formation and compare it against the actual CE-SSFP signal acquired from the experimental porcine model and the patient population. The relaxation times of remote, salvaged, and infarcted myocardium were calculated after the injection of gadolinium, at the time of CE-SSFP acquisition. Simulations were performed using the same CE-SSFP pulse sequence as used on the scanner on a set of spins with the calculated relaxation times from the CMR scans.

RESULTS

The normalized signal intensities of salvaged and infarcted myocardium obtained with simulations were lower than the corresponding normalized signal intensities obtained in vivo in pigs (p < 0.05, 134% vs 153%) and in patients (p < 0.05, 126% vs 145%). The results from simulations showed a linear relationship to the results obtained in the experimental porcine model (r2 = 0.61) and in patients (r2 = 0.69).

CONCLUSION

The T1 and T2 values of remote, salvaged, and infarcted myocardium only partly explain the signal intensities in CE-SSFP images. Bloch simulations suggest that there may be more elements that contribute to the CE-SSFP contrast. Integration of other aspects of the MR experiment into the simulation model could further help to fully unravel the mechanisms of CE-SSFP.

PhoenixMR: A GPU-based MRI simulation framework with runtime-dynamic code execution

BACKGROUND

Simulations of physical processes and behavior can provide unique insights and understanding of real-world problems. Magnetic Resonance Imaging (MRI) is an imaging technique with several components of complexity. Several of these components have been characterized and simulated in the past. However, several computational challenges prevent simulations from being simultaneously fast, flexible, and accurate.

PURPOSE

The simulation of MRI experiments is underutilized by medical physicists and researchers using currently available simulators due to reasons including speed, accuracy, and extensibility constraints. This paper introduces an innovative MRI simulation engine and framework that aims to overcome these issues making available realistic and fast MRI simulation.

METHODS

Using the CUDA C/C++ programing language, an MRI simulation engine (PhoenixMR), incorporating a Turing-complete virtual machine (VM) to simulate abstract spatiotemporal complexities, was developed. This engine solves a set of time-discrete Bloch equations using the symmetric operator splitting technique. An extensible front-end framework package (written in Python) aids the use of PhoenixMR to simplify simulation development.

RESULTS

The PhoenixMR library and front-end codes have been developed and tested. A set of example simulations were performed to demonstrate the ease of use and flexibility of simulation components such as geometrical setup, pulse sequence design, phantom design, and so forth. Initial validation of PhoenixMR is performed by comparing its accuracy and performance against a widely used MRI simulator using identical simulation parameters. Validation results show PhoenixMR simulations are three orders of magnitude faster. There is also strong agreement between models.

CONCLUSIONS

A novel MRI simulation platform called PhoenixMR has been introduced. This research tool is designed to be usable by physicists and engineers interested in performing MRI simulations. Examples are shown demonstrating the accuracy, flexibility, and usability of PhoenixMR in several key areas of MRI simulation.

CMRsim-A python package for cardiovascular MR simulations incorporating complex motion and flow

PURPOSE

To present an open-source MR simulation framework that facilitates the incorporation of complex motion and flow for studying cardiovascular MR (CMR) acquisition and reconstruction.

METHODS

CMRsim is a Python package that allows simulation of CMR images using dynamic digital phantoms with complex motion as input. Two simulation paradigms are available, namely, numerical and analytical solutions to the Bloch equations, using a common motion representation. Competitive simulation speeds are achieved using TensorFlow for GPU acceleration. To demonstrate the capability of the package, one introductory and two advanced CMR simulation experiments are presented. The latter showcase phase-contrast imaging of turbulent flow downstream of a stenotic section and cardiac diffusion tensor imaging on a contracting left ventricle. Additionally, extensive documentation and example resources are provided.

RESULTS

The Bloch simulation with turbulent flow using approximately 1.5 million particles and a sequence duration of 710 ms for each of the seven different velocity encodings took a total of 29 min on a NVIDIA Titan RTX GPU. The results show characteristic phase contrast and magnitude modulation present in real data. The analytical simulation of cardiac diffusion tensor imaging with bulk-motion phase sensitivity took approximately 10 s per diffusion-weighted image, including preparation and loading steps. The results exhibit the expected alteration of diffusion metrics due to strain.

CONCLUSION

CMRsim is the first simulation framework that allows one to feasibly incorporate complex motion, including turbulent flow, to systematically study advanced CMR acquisition and reconstruction approaches. The open-source package features modularity and transparency, facilitating maintainability and extensibility in support of reproducible research.

Brain MR image simulation for deep learning based medical image analysis networks

BACKGROUND AND OBJECTIVE

As large sets of annotated MRI data are needed for training and validating deep learning based medical image analysis algorithms, the lack of sufficient annotated data is a critical problem. A possible solution is the generation of artificial data by means of physics-based simulations. Existing brain simulation data is limited in terms of anatomical models, tissue classes, fixed tissue characteristics, MR sequences and overall realism.

METHODS

We propose a realistic simulation framework by incorporating patient-specific phantoms and Bloch equations-based analytical solutions for fast and accurate MRI simulations. A large number of labels are derived from open-source high-resolution T1w MRI data using a fully automated brain classification tool. The brain labels are taken as ground truth (GT) on which MR images are simulated using our framework. Moreover, we demonstrate that the T1w MR images generated from our framework along with GT annotations can be utilized directly to train a 3D brain segmentation network. To evaluate our model further on larger set of real multi-source MRI data without GT, we compared our model to existing brain segmentation tools, FSL-FAST and SynthSeg.

RESULTS

Our framework generates 3D brain MRI for variable anatomy, sequence, contrast, SNR and resolution. The brain segmentation network for WM/GM/CSF trained only on T1w simulated data shows promising results on real MRI data from MRBrainS18 challenge dataset with a Dice scores of 0.818/0.832/0.828. On OASIS data, our model exhibits a close performance to FSL, both qualitatively and quantitatively with a Dice scores of 0.901/0.939/0.937.

CONCLUSIONS

Our proposed simulation framework is the initial step towards achieving truly physics-based MRI image generation, providing flexibility to generate large sets of variable MRI data for desired anatomy, sequence, contrast, SNR, and resolution. Furthermore, the generated images can effectively train 3D brain segmentation networks, mitigating the reliance on real 3D annotated data.

KomaMRI.jl: An open-source framework for general MRI simulations with GPU acceleration

PURPOSE

To develop an open-source, high-performance, easy-to-use, extensible, cross-platform, and general MRI simulation framework (Koma).

METHODS

Koma was developed using the Julia programming language. Like other MRI simulators, it solves the Bloch equations with CPU and GPU parallelization. The inputs are the scanner parameters, the phantom, and the pulse sequence that is Pulseq-compatible. The raw data is stored in the ISMRMRD format. For the reconstruction, MRIReco.jl is used. A graphical user interface utilizing web technologies was also designed. Two types of experiments were performed: one to compare the quality of the results and the execution speed, and the second to compare its usability. Finally, the use of Koma in quantitative imaging was demonstrated by simulating Magnetic Resonance Fingerprinting (MRF) acquisitions.

RESULTS

Koma was compared to two well-known open-source MRI simulators, JEMRIS and MRiLab. Highly accurate results (with mean absolute differences below 0.1% compared to JEMRIS) and better GPU performance than MRiLab were demonstrated. In an experiment with students, Koma was proved to be easy to use, eight times faster on personal computers than JEMRIS, and 65% of test subjects recommended it. The potential for designing acquisition and reconstruction techniques was also shown through the simulation of MRF acquisitions, with conclusions that agree with the literature.

CONCLUSIONS

Koma's speed and flexibility have the potential to make simulations more accessible for education and research. Koma is expected to be used for designing and testing novel pulse sequences before implementing them in the scanner with Pulseq files, and for creating synthetic data to train machine learning models.

A fast and practical computation method for magnetic resonance simulators

PURPOSE

This work aims to develop a fast and practical computation method for MR simulations. The computational cost of MR simulations is often high because magnetizations of many isochromats are updated using a small step size on the order of microseconds. There are two types of subsequences to be processed for the simulations: subsequences with and without RF pulses. While straightforward implementations spend most of their time calculating subsequences with RF pulses, there is a method which efficiently reuses the computation for repetitive RF pulses.

THEORY AND METHODS

A new method for efficiently processing subsequences with RF pulses is proposed. Rather than using an iterative update approach, the proposed method computes the combined transition which combines all transitions applied iteratively for each subsequence with RF pulses. The combined transition is used again when the same subsequence is used later. The combined transitions are cached and managed using a least recently used algorithm.

RESULTS

The proposed method was found to accelerate the simulation by ˜20 times when 3.9 million isochromats were simulated using spin-echo sequences. Even on a laptop computer, the proposed method was able to simulate these sequences in ˜3.5 min.

CONCLUSION

An efficient method for simulating pulse sequences is proposed. The proposed method computes and manages combined transitions, making MR simulation practical on a wide range of computers, including laptops.

A Fetal Brain magnetic resonance Acquisition Numerical phantom (FaBiAN)

Accurate characterization of in utero human brain maturation is critical as it involves complex and interconnected structural and functional processes that may influence health later in life. Magnetic resonance imaging is a powerful tool to investigate equivocal neurological patterns during fetal development. However, the number of acquisitions of satisfactory quality available in this cohort of sensitive subjects remains scarce, thus hindering the validation of advanced image processing techniques. Numerical phantoms can mitigate these limitations by providing a controlled environment with a known ground truth. In this work, we present FaBiAN, an open-source Fetal Brain magnetic resonance Acquisition Numerical phantom that simulates clinical T2-weighted fast spin echo sequences of the fetal brain. This unique tool is based on a general, flexible and realistic setup that includes stochastic fetal movements, thus providing images of the fetal brain throughout maturation comparable to clinical acquisitions. We demonstrate its value to evaluate the robustness and optimize the accuracy of an algorithm for super-resolution fetal brain magnetic resonance imaging from simulated motion-corrupted 2D low-resolution series compared to a synthetic high-resolution reference volume. We also show that the images generated can complement clinical datasets to support data-intensive deep learning methods for fetal brain tissue segmentation.

Evaluation of Pseudo-Random Number Generation on GPU Cards

Monte Carlo methods rely on sequences of random numbers to obtain solutions to many problems in science and engineering. In this work, we evaluate the performance of different pseudo-random number generators (PRNGs) of the Curand library on a number of modern Nvidia GPU cards. As a numerical test, we generate pseudo-random number (PRN) sequences and obtain non-uniform distributions using the acceptance-rejection method. We consider GPU, CPU, and hybrid CPU/GPU implementations. For the GPU, we additionally consider two different implementations using the host and device application programming interfaces (API). We study how the performance depends on implementation parameters, including the number of threads per block and the number of blocks per streaming multiprocessor. To achieve the fastest performance, one has to minimize the time consumed by PRNG seed setup and state update. The duration of seed setup time increases with the number of threads, while PRNG state update decreases. Hence, the fastest performance is achieved by the optimal balance of these opposing effects.

Magnetic Resonance Simulation in Education: Quantitative Evaluation of an Actual Classroom Experience

Magnetic resonance is an imaging modality that implies a high complexity for radiographers. Despite some simulators having been developed for training purposes, we are not aware of any attempt to quantitatively measure their educational performance. The present study gives an answer to the question: Does an MRI simulator built on specific functional and non-functional requirements help radiographers learn MRI theoretical and practical concepts better than traditional educational method based on lectures? Our study was carried out in a single day by a total of 60 students of a main hospital in Madrid, Spain. The experiment followed a randomized pre-test post-test design with a control group that used a traditional educational method, and an experimental group that used our simulator. Knowledge level was assessed by means of an instrument with evidence of validity in its format and content, while its reliability was analyzed after the experiment. Statistical differences between both groups were measured. Significant statistical differences were found in favor of the participants who used the simulator for both the post-test score and the gain (difference between post-test and pre-test scores). The effect size turned out to be significant as well. In this work we evaluated a magnetic resonance simulation paradigm as a tool to help in the training of radiographers. The study shows that a simulator built on specific design requirements is a valuable complement to traditional education procedures, backed up with significant quantitative results.

Accuracy and performance analysis for Bloch and Bloch-McConnell simulation methods

PURPOSE

To introduce new solution methods for the Bloch and Bloch-McConnell equations and compare them quantitatively to different known approaches.

THEORY AND METHODS

A new exact solution per time step is derived by means of eigenvalues and generalized eigenvectors. Fast numerical solution methods based on asymmetric and symmetric operator splitting, which are already known for the Bloch equations, are extended to the Bloch-McConnell equations. Those methods are compared to other numerical methods including spin domain, one-step and multi-step methods, and matrix exponential. Error metrics are introduced based on the exact solution method, which allows to assess the accuracy of each solution method quantitatively for arbitrary example data.

RESULTS

Accuracy and performance properties for nine different solution methods are analyzed and compared in extensive numerical experiments including various examples for non-selective and slice-selective MR imaging applications. The accuracy of the methods heavily varies, in particular for short relaxation times and long pulse durations.

CONCLUSION

In absence of relaxation effects, the numerical results confirm the rotation matrices approach as accurate and computationally efficient Bloch solution method. Otherwise, as well as for the Bloch-McConnell equations, symmetric operator splitting methods are recommended due to their excellent numerical accuracy paired with efficient run time.

Bloch Simulation of a Three-point Dixon Experiment Using a Four-dimensional Numerical Phantom

A 4D numerical phantom, which is defined in the 3D spatial axes and the resonance frequency axis, is indispensable for Bloch simulations of biological tissues with complex distribution of materials. In this study, a 4D numerical phantom was created using MR image datasets of a biological sample containing water and fat, and the Bloch simulations were performed using the 4D numerical phantom. As a result, 3D images of the sample containing water and fat were successfully reproduced, which demonstrated the usefulness of the concept of the 4D numerical phantom.

Numerical simulation of time-resolved 3D phase-contrast magnetic resonance imaging

A numerical approach is presented to efficiently simulate time-resolved 3D phase-contrast Magnetic resonance Imaging (or 4D Flow MRI) acquisitions under realistic flow conditions. The Navier-Stokes and Bloch equations are simultaneously solved with an Eulerian-Lagrangian formalism. A semi-analytic solution for the Bloch equations as well as a periodic particle seeding strategy are developed to reduce the computational cost. The velocity reconstruction pipeline is first validated by considering a Poiseuille flow configuration. The 4D Flow MRI simulation procedure is then applied to the flow within an in vitro flow phantom typical of the cardiovascular system. The simulated MR velocity images compare favorably to both the flow computed by solving the Navier-Stokes equations and experimental 4D Flow MRI measurements. A practical application is finally presented in which the MRI simulation framework is used to identify the origins of the MRI measurement errors.

Magnetic resonance imaging simulation with spin-lock preparations to detect tiny oscillatory magnetic fields

Spin-lock preparation was studied to detect tiny oscillatory magnetic fields such as a neural magnetic field without the blood oxygen level-dependent effect. This approach is a direct measurement and independent of static magnetic field strength. Accordingly, it is anticipated as a feasible functional magnetic resonance imaging (fMRI) in low and ultra-low-field MRI. Several reports have been published on spin-lock preparation but reports on imaging simulation are rare. Research in this area can assist in investigating magnetic resonance signal changes and, accordingly, can help to develop new spin-lock methods. In this study, we propose an imaging simulation method with an analytical solution using the Bloch equation. To demonstrate the feasibility of our proposed method, we compared simulated images with experimental results in which the number of sub-voxels and the amplitude and phase of the target oscillatory magnetic fields varied. In addition, we also applied graphics processing unit parallel computing and investigated the feasibility of avoiding an impracticable calculation time by doing so.

Simulator-generated training datasets as an alternative to using patient data for machine learning: An example in myocardial segmentation with MRI

BACKGROUND AND OBJECTIVE

Supervised Machine Learning techniques have shown significant potential in medical image analysis. However, the training data that need to be collected for these techniques in the field of MRI 1) may not be available, 2) may be available but the size is small, 3) may be available but not representative and 4) may be available but with weak labels. The aim of this study was to overcome these limitations through advanced MR simulations on a realistic computer model of human anatomy without using a real MRI scanner, without scanning patients and without having personnel and the associated expenses.

METHODS

The 4D-XCAT model was used with the coreMRI simulation platform for generating artificial short-axis MR-images for training a neural-network to automatic delineate the LV endocardium and epicardium. Its performance was assessed on real MR-images acquired from eight healthy volunteers. The neural-network was also trained on real MR-images from a publicly available dataset and its performance was assessed on the same volunteers' data.

RESULTS

The proposed solution demonstrated a performance of 94% (endocardium) and 90% DICE (epicardium) in real mid-ventricular slices, whereas a 10% addition of real MR-images in the artificial training dataset increased the performance to 97% DICE. The use of artificial MR-images that cover the entire LV yielded 85% (endocardium) and 88% DICE (epicardium) when combined with real MR data with an 80%-20% mix respectively.

CONCLUSIONS

This study suggests a low-cost solution for constructing artificial training datasets for supervised learning techniques in the field of MR by using advanced MR simulations without the use of a real MRI scanner, without scanning patients and without having to use specialized personnel, such as technologists and radiologists.

Towards real-time 4D radiation dosimetry on an MRI-Linac

4D radiation dosimetry using a highly radiation-sensitive polymer gel dosimeter with real-time quantitative magnetic resonance imaging (MRI) readout is presented as a technique to acquire the accumulated radiation dose distribution during image-guided radiotherapy on an MRI-Linac. Optimized T ₂-weighted Turbo-Spin-Echo (TSE) scans are converted into quantitative ΔR ₂ maps and subsequently to radiation dose maps. The concept of temporal uncertainty is introduced as a metric of effective temporal resolution. A mathematical framework is presented to optimize the echo time of the TSE sequence in terms of dose resolution, and the trade-off between temporal resolution and dose resolution is discussed. The current temporal uncertainty achieved with the MAGAT gel dosimeter on a 1 T MRI-Linac is 3.8 s which is an order of magnitude better than what has been achieved until now. The potential of real-time 4D radiation dosimetry in a theragnostic MRI-Linac is demonstrated for two scenarios: an irradiation with three coplanar beams on a head phantom and a dynamic arc treatment on a cylindrical gel phantom using a rotating couch. The dose maps acquired on the MRI-Linac are compared with a treatment plan and with dose maps acquired on a clinical 3 T MRI scanner. 3D gamma map evaluations for the different modalities are provided. While the presented method demonstrates the potential of gel dosimetry for tracking the dose delivery during radiotherapy in 4D, a shortcoming of the MAGAT gel dosimeter is a retarded dose response. The effect of non-ideal radiofrequency pulses resulting from limitations in the specific absorption rate or B₁-field inhomogeneity on the TSE acquired ΔR ₂ values is analysed experimentally and by use of computational modelling with a Bloch simulator.

Development of a method for the Bloch image simulation of biological tissues

PURPOSE

The purpose of this study is to develop a method for the Bloch image simulation of biological tissues including various chemical components and T₂* distribution.

METHODS

The nuclear spins in the object material were modeled as a spectral intensity function Sr→ω defined by superposition of Lorentz functions with various central precession frequencies and the half width of 1/(πT₂'), where 1/T₂' is a relaxation rate attributable to microscopic field inhomogeneity in a voxel. Four-dimensional numerical phantoms were created to simulate Sr→ω and used for MRI simulations of the phantoms containing water and fat protons. Single slice multiple (16) gradient-echo sequences (ΔTE = 2.2 and 1.384 ms) were used for experiments at 1.5 T and 3 T and MRI simulations to evaluate the validity of the approach.

RESULTS

Experimentally measured image intensities of the multiple gradient-echo imaging sequences were well reproduced by the MRI simulations. The correlation coefficients between the experimentally measured image intensities and those numerically simulated were 0.9895 to 0.9992 for the 4-component phantom at 1.5 T and 0.9580 to 0.9996 for the 7-component phantom at 3 T.

CONCLUSION

T₂* and chemical shift effects were successfully implemented in the MRI simulator (BlochSolver). Because this approach can be applied to other MRI simulators, the method developed in this study is useful for MRI simulation of biological tissues containing water and fat protons.

Integration of an intelligent tutoring system in a magnetic resonance simulator for education: Technical feasibility and user experience

BACKGROUND AND OBJECTIVE

In this paper we propose to include an intelligent tutoring system (ITS) within a magnetic resonance (MR) simulator that has been developed in house. With this, we intend to measure the impact, in terms of user experience, of including an ITS in our simulator.

METHODS

We thoroughly describe the integration procedure and we have tested the benefits of this integration by means of two actual educational experiences, with one of them using the simulator as a standalone tool, and the other with the joint use of simulator+ITS. The experiences have consisted of two online courses with a number of students around 180 in both of them, where measurements of usability, perceived utility and likelihood to recommend were collected.

RESULTS

We have observed that the three measurements improved noticeably in the second course with respect to the first one; specifically, overall usability improved by 22.3%, perceived utility by an average of 55.1% and likelihood to recommend by 13.7%. In addition, quantitative measurements are complemented with comments in free text format directly provided by the students. Results show evidence on the benefits of integrating an ITS in terms of quantitative user experience, as well as qualitative comparative comments directly by students of both courses.

CONCLUSIONS

This is the first time that an ITS is used within the scope of MR simulation for training purposes. Benefits of integrating an ITS within an MR simulator have been evaluated in terms of user experience, with satisfactory comparative results.

An Accurate Dictionary Creation Method for MR Fingerprinting Using a Fast Bloch Simulator

This study proposes an accurate method for creating a dictionary for magnetic resonance fingerprinting (MRF) using a fast Bloch image simulator. An MRF sequence based on a fast imaging with steady precession sequence and a numerical phantom were used for dictionary generation. Cartesian and spiral readout gradients were used for the Bloch image simulation. The validity and usefulness of the method for accurate dictionary creation were demonstrated by MRF parameter maps obtained by pattern matching with the dictionaries generated by the proposed method.

Virtual clinical trials in medical imaging: a review

The accelerating complexity and variety of medical imaging devices and methods have outpaced the ability to evaluate and optimize their design and clinical use. This is a significant and increasing challenge for both scientific investigations and clinical applications. Evaluations would ideally be done using clinical imaging trials. These experiments, however, are often not practical due to ethical limitations, expense, time requirements, or lack of ground truth. Virtual clinical trials (VCTs) (also known as in silico imaging trials or virtual imaging trials) offer an alternative means to efficiently evaluate medical imaging technologies virtually. They do so by simulating the patients, imaging systems, and interpreters. The field of VCTs has been constantly advanced over the past decades in multiple areas. We summarize the major developments and current status of the field of VCTs in medical imaging. We review the core components of a VCT: computational phantoms, simulators of different imaging modalities, and interpretation models. We also highlight some of the applications of VCTs across various imaging modalities.

High-resolution in vivo MR-STAT using a matrix-free and parallelized reconstruction algorithm

MR-STAT is a recently proposed framework that allows the reconstruction of multiple quantitative parameter maps from a single short scan by performing spatial localisation and parameter estimation on the time-domain data simultaneously, without relying on the fast Fourier transform (FFT). To do this at high resolution, specialized algorithms are required to solve the underlying large-scale nonlinear optimisation problem. We propose a matrix-free and parallelized inexact Gauss-Newton based reconstruction algorithm for this purpose. The proposed algorithm is implemented on a high-performance computing cluster and is demonstrated to be able to generate high-resolution (1 mm  ×  1 mm in-plane resolution) quantitative parameter maps in simulation, phantom, and in vivo brain experiments. Reconstructed T1 and T2 values for the gel phantoms are in agreement with results from gold standard measurements and, for the in vivo experiments, the quantitative values show good agreement with literature values. In all experiments, short pulse sequences with robust Cartesian sampling are used, for which MR fingerprinting reconstructions are shown to fail.

A Web-Based Educational Magnetic Resonance Simulator: Design, Implementation and Testing

A new web-based education-oriented magnetic resonance (MR) simulator is presented. We have identified the main requirements that this simulator should comply with, so that trainees can face useful practical tasks such as setting the exact slice position and its properties, selecting the correct protocol or fitting the parameters to acquire an image. The tool follows the client-server model. The client contains the interface that mimics the console of a real machine and several of its features. The server stores anatomical models and executes the bulk of the simulation. This cross-platform simulator has been used in two real educational scenarios. The acceptance of the tool has been measured using two criteria, namely, the System Usability Scale and the Likelihood to Recommend, both with satisfactory results. Therefore, we conclude that given the potential of the tool, it may play a relevant role for the training of MRI operators and other involved personnel.

coreMRI: A high-performance, publicly available MR simulation platform on the cloud

Introduction

A Cloud-ORiented Engine for advanced MRI simulations (coreMRI) is presented in this study. The aim was to develop the first advanced MR simulation platform delivered as a web service through an on-demand, scalable cloud-based and GPU-based infrastructure. We hypothesized that such an online MR simulation platform could be utilized as a virtual MRI scanner but also as a cloud-based, high-performance engine for advanced MR simulations in simulation-based quantitative MR (qMR) methods.

Methods and results

The simulation framework of coreMRI was based on the solution of the Bloch equations and utilized a ground-up-approach design based on the principles already published in the literature. The development of a front-end environment allowed the connection of the end-users to the GPU-equipped instances on the cloud. The coreMRI simulation platform was based on a modular design where individual modules (such as the Gadgetron reconstruction framework and a newly developed Pulse Sequence Designer) could be inserted in the main simulation framework. Different types and sources of pulse sequences and anatomical models were utilized in this study revealing the flexibility that the coreMRI simulation platform offers to the users. The performance and scalability of coreMRI were also examined on multi-GPU configurations on the cloud, showing that a multi-GPU computer on the cloud equipped with a newer generation of GPU cards could significantly mitigate the prolonged execution times that accompany more realistic MRI and qMR simulations.

Conclusions

coreMRI is available to the entire MR community, whereas its high performance and scalability allow its users to configure advanced MRI experiments without the constraints imposed by experimentation in a true MRI scanner (such as time constraint and limited availability of MR scanners), without upfront investment for purchasing advanced computer systems and without any user expertise on computer programming or MR physics. coreMRI is available to the users through the webpage https://www.coreMRI.org.

A Fast GPU-optimized 3D MRI Simulator for Arbitrary k-space Sampling

Purpose:

To develop a fast 3D MRI simulator for arbitrary k-space sampling using a graphical processing unit (GPU) and demonstrate its performance by comparing simulation and experimental results in a real MRI system.

Materials and Methods:

A fast 3D MRI simulator using a GeForce GTX 1080 GPU (NVIDIA Corporation, Santa Clara, CA, USA) was developed using C++ and the CUDA 8.0 platform (NVIDIA Corporation). The unique advantage of this simulator was that it could use the same pulse sequence as used in the experiment. The performance of the MRI simulator was measured using two GTX 1080 GPUs and 3D Cones sequences. The MRI simulation results for 3D non-Cartesian sampling trajectories like 3D Cones sequences using a numerical 3D phantom were compared with the experimental results obtained with a real MRI system and a real 3D phantom.

Results:

The performance of the MRI simulator was about 3800–4900 gigaflops for 128- to 4-shot 3D Cones sequences with 256³ voxels, which was about 60% of the performance of the previous MRI simulator optimized for Cartesian sampling calculated for a Cartesian sampling gradient-echo sequence with 256³ voxels. The effects of the static magnetic field inhomogeneity, radio-frequency field inhomogeneity, gradient field nonlinearity, and fast repetition times on the MR images were reproduced in the simulated images as observed in the experimental images.

Conclusion:

The 3D MRI simulator developed for arbitrary k-space sampling optimized using GPUs is a powerful tool for the development and evaluation of advanced imaging sequences including both Cartesian and non-Cartesian k-space sampling.

BlochSolver: A GPU-optimized fast 3D MRI simulator for experimentally compatible pulse sequences

A magnetic resonance imaging (MRI) simulator, which reproduces MRI experiments using computers, has been developed using two graphic-processor-unit (GPU) boards (GTX 1080). The MRI simulator was developed to run according to pulse sequences used in experiments. Experiments and simulations were performed to demonstrate the usefulness of the MRI simulator for three types of pulse sequences, namely, three-dimensional (3D) gradient-echo, 3D radio-frequency spoiled gradient-echo, and gradient-echo multislice with practical matrix sizes. The results demonstrated that the calculation speed using two GPU boards was typically about 7 TFLOPS and about 14 times faster than the calculation speed using CPUs (two 18-core Xeons). We also found that MR images acquired by experiment could be reproduced using an appropriate number of subvoxels, and that 3D isotropic and two-dimensional multislice imaging experiments for practical matrix sizes could be simulated using the MRI simulator. Therefore, we concluded that such powerful MRI simulators are expected to become an indispensable tool for MRI research and development.

Fast Realistic MRI Simulations Based on Generalized Multi-Pool Exchange Tissue Model

We present MRiLab, a new comprehensive simulator for large-scale realistic MRI simulations on a regular PC equipped with a modern graphical processing unit (GPU). MRiLab combines realistic tissue modeling with numerical virtualization of an MRI system and scanning experiment to enable assessment of a broad range of MRI approaches including advanced quantitative MRI methods inferring microstructure on a sub-voxel level. A flexible representation of tissue microstructure is achieved in MRiLab by employing the generalized tissue model with multiple exchanging water and macromolecular proton pools rather than a system of independent proton isochromats typically used in previous simulators. The computational power needed for simulation of the biologically relevant tissue models in large 3D objects is gained using parallelized execution on GPU. Three simulated and one actual MRI experiments were performed to demonstrate the ability of the new simulator to accommodate a wide variety of voxel composition scenarios and demonstrate detrimental effects of simplified treatment of tissue micro-organization adapted in previous simulators. GPU execution allowed  ∼ 200× improvement in computational speed over standard CPU. As a cross-platform, open-source, extensible environment for customizing virtual MRI experiments, MRiLab streamlines the development of new MRI methods, especially those aiming to infer quantitatively tissue composition and microstructure.

Cloud GPU-based simulations for SQUAREMR

Quantitative Magnetic Resonance Imaging (MRI) is a research tool, used more and more in clinical practice, as it provides objective information with respect to the tissues being imaged. Pixel-wise T₁ quantification (T₁ mapping) of the myocardium is one such application with diagnostic significance. A number of mapping sequences have been developed for myocardial T₁ mapping with a wide range in terms of measurement accuracy and precision. Furthermore, measurement results obtained with these pulse sequences are affected by errors introduced by the particular acquisition parameters used. SQUAREMR is a new method which has the potential of improving the accuracy of these mapping sequences through the use of massively parallel simulations on Graphical Processing Units (GPUs) by taking into account different acquisition parameter sets. This method has been shown to be effective in myocardial T₁ mapping; however, execution times may exceed 30min which is prohibitively long for clinical applications. The purpose of this study was to accelerate the construction of SQUAREMR's multi-parametric database to more clinically acceptable levels. The aim of this study was to develop a cloud-based cluster in order to distribute the computational load to several GPU-enabled nodes and accelerate SQUAREMR. This would accommodate high demands for computational resources without the need for major upfront equipment investment. Moreover, the parameter space explored by the simulations was optimized in order to reduce the computational load without compromising the T₁ estimates compared to a non-optimized parameter space approach. A cloud-based cluster with 16 nodes resulted in a speedup of up to 13.5 times compared to a single-node execution. Finally, the optimized parameter set approach allowed for an execution time of 28s using the 16-node cluster, without compromising the T₁ estimates by more than 10ms. The developed cloud-based cluster and optimization of the parameter set reduced the execution time of the simulations involved in constructing the SQUAREMR multi-parametric database thus bringing SQUAREMR's applicability within time frames that would be likely acceptable in the clinic.

Fast Fourier-based simulation of off-resonance artifacts in steady-state gradient echo MRI applied to metal object localization

PURPOSE

To accelerate simulation of off-resonance artifacts in steady-state gradient echo MRI by using fast Fourier transforms and demonstrate its applicability to metal object localization.

THEORY AND METHODS

By exploiting the repetitive nature of steady-state pulse sequences it is possible to use fast Fourier transforms to calculate the MR signal. Based on this principle, a method for fast simulation of off-resonance artifacts was designed. The method was validated against Bloch simulations and MRI scans. Its clinical relevance was demonstrated by employing it for template matching-based metal object localization, as applied to a titanium cylinder, an oxidized zirconium knee implant, and gold fiducials.

RESULTS

The fast simulations were accurate compared with actual MRI scans of the objects. The differences between the fast simulations and Bloch simulations were minor, while the acceleration scaled linearly with the number of phase-encoding lines. The object localization method accurately localized the various metal objects.

CONCLUSION

The proposed simulation methodology provided accurate 3D simulations of off-resonance artifacts with a lower computational complexity than Bloch simulations. The speed of the simulations opens up possibilities in image reconstructions involving off-resonance phenomena that were previously infeasible due to computational limitations, as demonstrated for metal object localization. Magn Reson Med 78:2035-2041, 2017. © 2016 The Authors Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine. This is an open access article under the terms of the Creative Commons Attribution NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

Parallel simulations for QUAntifying RElaxation magnetic resonance constants (SQUAREMR): an example towards accurate MOLLI T1 measurements

BACKGROUND

T1 mapping is widely used today in CMR, however, it underestimates true T1 values and its measurement error is influenced by several acquisition parameters. The purpose of this study was the extraction of accurate T1 data through the utilization of comprehensive, parallel Simulations for QUAntifying RElaxation Magnetic Resonance constants (SQUAREMR) of the MOLLI pulse sequence on a large population of spins with physiologically relevant tissue relaxation constants.

METHODS

A CMR protocol consisting of different MOLLI schemes was performed on phantoms and healthy human volunteers. For every MOLLI experiment, the identical pulse sequence was simulated for a large range of physiological combinations of relaxation constants, resulting in a database of all possible outcomes. The unknown relaxation constants were then determined by finding the simulated signals in the database that produced the least squared difference to the measured signal intensities.

RESULTS

SQUAREMR demonstrated improvement of accuracy in phantom studies and consistent mean T1 values and consistent variance across the different MOLLI schemes in humans. This was true even for tissues with long T1s and MOLLI schemes with no pause between modified-Look-Locker experiments.

CONCLUSIONS

SQUAREMR enables quantification of T1 data obtained by existing clinical pulse sequences. SQUAREMR allows for correction of quantitative CMR data that have already been acquired whereas it is expected that SQUAREMR may improve data consistency and advance quantitative MR across imaging centers, vendors and experimental configurations. While this study is focused on a MOLLI-based T1-mapping technique, it could however be extended in other types of quantitative MRI throughout the body.

Validation of a new T2* algorithm and its uncertainty value for cardiac and liver iron load determination from MRI magnitude images

Purpose

To validate an automatic algorithm for offline T2* measurements, providing robust, vendor‐independent T2*, and uncertainty estimates for iron load quantification in the heart and liver using clinically available imaging sequences.

Methods

A T2* region of interest (ROI)‐based algorithm was developed for robustness in an offline setting. Phantom imaging was performed on a 1.5 Tesla system, with clinically available multiecho gradient‐recalled‐echo (GRE) sequences for cardiac and liver imaging. A T2* single‐echo GRE sequence was used as reference. Simulations were performed to assess accuracy and precision from 2000 measurements. Inter‐ and intraobserver variability was obtained in a patient study (n = 23).

Results

Simulations: Accuracy, in terms of the mean differences between the proposed method and true T2* ranged from 0–0.73 ms. Precision, in terms of confidence intervals of repeated measurements, was 0.06–4.74 ms showing agreement between the proposed uncertainty estimate and simulations. Phantom study: Bias and variability were 0.26 ± 4.23 ms (cardiac sequence) and −0.23 ± 1.69 ms (liver sequence). Patient study: Intraobserver variability was similar for experienced and inexperienced observers (0.03 ± 1.44 ms versus 0.16 ± 2.33 ms). Interobserver variability was 1.0 ± 3.77 ms for the heart and −0.52 ± 2.75 ms for the liver.

Conclusion

The proposed algorithm was shown to provide robust T2* measurements and uncertainty estimates over the range of clinically relevant T2* values. Magn Reson Med, 2015. © 2015 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine. 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. Magn Reson Med 75:1717–1729, 2016. © 2015 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance.

High performance MRI simulations of motion on multi-GPU systems

BACKGROUND

MRI physics simulators have been developed in the past for optimizing imaging protocols and for training purposes. However, these simulators have only addressed motion within a limited scope. The purpose of this study was the incorporation of realistic motion, such as cardiac motion, respiratory motion and flow, within MRI simulations in a high performance multi-GPU environment.

METHODS

Three different motion models were introduced in the Magnetic Resonance Imaging SIMULator (MRISIMUL) of this study: cardiac motion, respiratory motion and flow. Simulation of a simple Gradient Echo pulse sequence and a CINE pulse sequence on the corresponding anatomical model was performed. Myocardial tagging was also investigated. In pulse sequence design, software crushers were introduced to accommodate the long execution times in order to avoid spurious echoes formation. The displacement of the anatomical model isochromats was calculated within the Graphics Processing Unit (GPU) kernel for every timestep of the pulse sequence. Experiments that would allow simulation of custom anatomical and motion models were also performed. Last, simulations of motion with MRISIMUL on single-node and multi-node multi-GPU systems were examined.

RESULTS

Gradient Echo and CINE images of the three motion models were produced and motion-related artifacts were demonstrated. The temporal evolution of the contractility of the heart was presented through the application of myocardial tagging. Better simulation performance and image quality were presented through the introduction of software crushers without the need to further increase the computational load and GPU resources. Last, MRISIMUL demonstrated an almost linear scalable performance with the increasing number of available GPU cards, in both single-node and multi-node multi-GPU computer systems.

CONCLUSIONS

MRISIMUL is the first MR physics simulator to have implemented motion with a 3D large computational load on a single computer multi-GPU configuration. The incorporation of realistic motion models, such as cardiac motion, respiratory motion and flow may benefit the design and optimization of existing or new MR pulse sequences, protocols and algorithms, which examine motion related MR applications.