Keywords: Image Reconstruction, Image Reconstruction

Motivation: Using non-Cartesian trajectories allows for motion robustness, and a more efficient encoding. However, these non-Cartesian acquisitions necessitate the use of NUFFTs and field correction techniques, leading to costly reconstruction times.

Goal(s): We aim to remove the need for NUFFTs in non-Cartesian MRI, and drastically reduce the computational footprint of field correction.

Approach: Our approach is to correct the raw k-space data of phase due to field imperfections and off-grid sampling using an implicit representation of GRAPPA kernels.

Results: We show an order of magnitude increase in comparison to current standard techniques with near identical reconstructions quality.

Impact: This work aims to significantly reduce the computational requirement for reconstructing non-Cartesian data. This will help with the adoption of long readout non-Cartesian acquisitions, which naturally accelerate MRI exams.

Keywords: New Trajectories & Spatial Encoding Methods, Sparse & Low-Rank Models, Sparse sampling

Motivation: With the increasing demand for high-resolution and short MRI exams, especially at high and ultra-high field, there is a need for fast acquisition techniques.

Goal(s): To improve highly accelerated compressed-sensing by introducing a novel sampling named Alternating Basis Readout Imaging with COmpressed$$$\,$$$sensing with Three-dImensioNal Encoding (ABRICOTINE).

Approach: ABRICOTINE incorporates sparse phase-encoding in three orthogonal directions, achieving true three-dimensional undersampling of the Fourier domain. This differs from conventional compressed$$$\,$$$sensing, which only undersamples within two phase-encoding dimensions.

Results: We demonstrate significant enhancements in brain image quality through both simulations and true ABRICOTINE-accelerated acquisitions. It surpasses conventional compressed$$$\,$$$sensing methods and enables 0.5mm isotropic imaging in 4min.

Impact: ABRICOTINE allows for a substantial improvement in compressed-sensing acceleration compared to traditional sparse sampling techniques, especially when high acceleration factors are required. It thus shows great potential for further accelerating high resolution MRI acquisitions.

Keywords: Image Reconstruction, Image Reconstruction

Motivation: Estimating cardiac and respiratory inter-bin deformation fields from 5D motion-resolved free-running data is particularly challenging due to a high level of undersampling.

Goal(s): To address this challenge through an innovative multi-resolution approach to estimate the deformation fields and reconstruct 5D motion-resolved images.

Approach: The approach consists of a sequence of compressed-sensing image reconstructions that iteratively progresses from low to high spatial resolutions, where one lower-resolution iteration’s output is exploited as input for the next higher resolution until target resolution is reached.

Results: Using optimized regularization weights, the proposed approach achieved left-ventricular ejection fraction within a 4% error margin compared the 2D cine.

Impact: This study presents a multi-resolution framework for estimating cardiac and respiratory inter-bin deformation fields aimed at improved motion-resolved whole-heart 5D-imaging. This multi-resolution compressed sensing framework has the potential to accurately estimate deformation fields and reduce compression artefacts.

Keywords: Image Reconstruction, MR Fingerprinting

Motivation: Fast, regularized subspace reconstruction would enable the acquisition and synthesis of a standard clinical brain protocol in mere minutes.

Goal(s): Our goal is to make high-resolution, image reconstruction faster and more robust to $$$B_0$$$ inhomogeneities.

Approach: Our reconstruction alternates between data consistency (DC) and spatio-temporal low rank regularization. We leverage coil sensitivities to "snap" non-Cartesian trajectories to the kspace grid, speeding up DC steps and enabling $$$B_0$$$-robust reconstructions with fewer time segments. Polynomial preconditioning enables convergence in up to 2x fewer iterations, reducing expensive proximal updates.

Results: Our method reduces the reconstruction time by an order of magnitude while retaining quality.

Impact: The ability to reconstruct MRF quickly should make integration of MRF into clinical workflows not just possible, but convenient. Additionally, the efficiency gains from this framework can make even the most sophisticated and expensive regularizers computationally feasible.

Keywords: Image Reconstruction, Image Reconstruction

Motivation: Longitudinal MRI scans performed on the same patient offer valuable temporal redundancy that can be exploited for image reconstruction. However, this wealth of information is usually ignored in current clinical practice, with data from different sessions typically reconstructed separately.

Goal(s): This study introduces a longitudinal dynamic MRI framework that leverages temporal correlations across multiple imaging sessions to improve image reconstruction.

Approach: Our reconstruction approach aims to reconstruct multi-session data jointly as a dynamic image series employing a combination of low-rank subspace and spatiotemporal constraints. 

Results: The initial results demonstrate that joint longitudinal reconstruction outperforms standard separate reconstructions, which may allow for additional acceleration. 

Impact: By exploiting image correlations across multiple sessions, our longitudinal dynamic MRI framework can improve image reconstruction and enable higher acceleration compared to standard separate reconstruction. 

Keywords: Image Reconstruction, Contrast Mechanisms, Phase

Motivation: Unlike magnitude images, even simple averaging is difficult with phase images, because of the circular nature of phase, spanning 2pi radians range.

Goal(s): In our research that uses mesoscopic imaging (< 0.5 mm isotropic) at 7 T, we need to average multiple acquisitions to increase SNR. Being unable to average straightforwardly together with the lack of natural zero point is a critical constraint.

Approach: To address this problem, we propose to operate on the magnitude of the second spatial derivative of phase images - called “phase jolt”.

Results: Our results show phase jolt offers benefits for processing associated with SWI imaging.

Impact: Phase jolt is an easy to implement new contrast where vessels and non brain tissue are highlighted and background bias field is mitigated. Therefore, phase jolt images have potential to be impactful in any setting where phase images are used.

Keywords: Quantitative Imaging, Sparse & Low-Rank Models

Motivation: While MR Multitasking shows initial promise as a free-breathing, non-ECG approach for multiparametric CMR, its precision and repeatability still require further improvement to match the widely adopted clinical protocols.

Goal(s): To improve precision and repeatability of multiparametric mapping by cardiovascular MR Multitasking.

Approach: A novel low-rank tensor reconstruction strategy was developed to improve the reconstruction performance. Numerical simulations and in-vivo studies on healthy volunteers and cardiomyopathy patients were used to evaluate the proposed technique.

Results: Compared to conventional recontruction, the proposed approach showed lower RMSE in numerical simulations, and improved precision by ~20% and repeatability by ~30% in in-vivo studies.

Impact: The improved cardiovascular MR Multitasking has the potential to be an efficient and subject friendly (free-breathing, non-ECG) alternative for diagnosis of CMR patients whose T1 and T2 changes are greater than 100 ms and 2 ms, e.g., amyloidosis patients.

Keywords: Quantitative Imaging, Quantitative Imaging, Deep Learning

Motivation: Quantitative MRI typically involves a multi-step imaging pipeline from data acquisition to parameter estimation. Deep learning holds great promise to improve and streamline the entire workflow for quantitative MRI.

Goal(s): This work presents a general deep learning-based rapid quantitative MRI framework, called DeepGrasp-Quant, for efficient and accurate quantification of MRI parameters based on Golden-angle RAdial Sparse Parallel (GRASP) MRI.

Approach: DeepGrasp-Quant was designed with cascaded deep learning modules for reconstruction and parameter fitting, enabling direct estimation of MR parameters from undersampled images.

Results: Two examples of DeepGrasp-Quant (DeepGrasp-T1 and DeepGrasp-T1-Dixon) were demonstrated for rapid accurate T1 mapping of the brain and the liver.

Impact: DeepGrasp-Quant is expected to be a promising technique for efficient and accurate quantification of MRI parameters from highly-accelerated free-breathing data acquisition. In addition to T1 mapping, it can also be integrated with other quantitative MRI methods for different clinical applications.

Keywords: Image Reconstruction, Perfusion

Motivation: Enhancing myocardial perfusion MRI with self-supervised learning is key to achieving higher image quality and fidelity, especially in patients with varying image matrix sizes and asymmetric echo.

Goal(s): To enhance perfusion MRI by increasing resolution and slice coverage using self-supervised learning, self-regularization, and spatial attention, tailored for varied image sizes and asymmetric echo.

Approach: Implemented an accelerated perfusion MRI sequence with asymmetric echo; collected data from 20 patients; developed self-LR with SAM to enhance image quality.

Results: Self-LR with SAM yielded superior image quality and fewer artifacts in varied sizes and asymmetric echo, outperforming other methods, confirmed by expert evaluations.

Impact: The integration of Spatial Attention Module (SAM) with Self-Supervised Learning and Self-Regularization significantly enhances myocardial perfusion MRI, enriching spatial resolution and slice coverage. This development could potentially improve diagnostic accuracy, facilitating non-invasive whole-heart assessments with improved image quality.

Keywords: Quantitative Imaging, MR Fingerprinting, Quantitative Susceptibility Mapping

Motivation: Incorporating MR phase into the MRF scheme can provide further diagnostic information, such as QSM. However, the MRF dictionary exponentially grows with the number of parameters to estimate.

Goal(s): To validate QSM and B1 mapping using MRF and PS-DRONE reconstruction network against conventional reference maps.

Approach: Data were acquired at 3T with an EPI-MRF, a multi-echo GRE sequence for QSM and a Bloch Siegert sequence for B1 mapping.

Results: PS-DRONE enabled simultaneous quantification of T1, T2, PD, B1 and maps in 2 minutes. Tissue parameter maps were reconstructed in 1 second. Strong correlations were observed to reference B1 and QSM maps.

Impact: The ability of PS-DRONE to quantitatively image T1, T2, PD, B1 and QSM with similar accuracy to conventional techniques, but in a fraction of the time, would promote the use of multiparametric quantitative MRI in clinical practice.

Keywords: Software Tools, Software Tools, Pipeline, MR Fingerprinting

Motivation:  A number of promising multiparameter mapping approaches have been developed but are not yet in routine-use in clinical and neuroscientific settings. 

Goal(s): To create a rapid and robust 3D-MRF acquisition/reconstruction package, suitable for large-scale neuroscientific and clinical applications.

Approach: 3D-MRF acquisition is developed for 1-mm whole-brain-mapping in 2.5-minutes, with build-in robustness to motion, and quantification bias from B0&B1 inhomogeneities. Highly-efficient reconstruction package is created for online generation of quantitative maps and synthesized-contrasts within 4 minutes of scan-completion, benchmarked using a consumer-grade GPU. 

Results: The pipeline has been validated on over 100-scans performed across clinical and neuroscientific settings with plan for open-source distribution soon.

Impact: The distribution of such an acquisition/reconstruction package should help facilitate wide-spread deployment of multiparameter mapping.

Keywords: Machine Learning/Artificial Intelligence, Machine Learning/Artificial Intelligence

Motivation: Direct acquisition of high resolution data is time-consuming and degrades SNR. Super-resolution reconstruction (SRR) is widely used to address these challenges. However, existing reconstruction tools use algorithms that are sensitive to noise and motion.

Goal(s): Our study aims to develop a training-free deep learning-based SRR method that integrates multi-view thick-slice data to reconstruct images with enhanced spatial resolution and high SNR.

Approach: We used an implicit neural representation (INR) network, leveraging data from scans at various views, to achieve high isotropic SRR.

Results: Our technique exhibited 30% better SNR and significant motion-robustness compared to existing techniques.

Impact: Implicit neural representations allow continuous functional representation of MRI images thereby being a natural candidate for performing SRR in low SNR regimes. Our study validates the feasibility of employing INRs to reduce scan time, motion artifacts, and achieve high-quality SRR.

Keywords: Image Reconstruction, Diffusion/other diffusion imaging techniques, Intravoxel Incoherent Motion, Model-Based Reconstruction, Subspace Reconstruction

Motivation: Intravoxel incoherent motion (IVIM) is a measure in MRI to quantify tissue perfusion. However, clinical applications are limited by noisy parameter estimates for the perfusion fraction (f) and pseudodiffusion coefficient (D*).

Goal(s): We sought to improve IVIM parameter estimation using a model-based reconstruction

Approach: We combined locally low-rank (LLR) and temporal subspace constraints to reliably perform joint reconstruction of IVIM images before fitting while correcting shot-to-shot phase variations between each b-value.

Results: Our method resulted in smoother signal decay curves before fitting and improved the estimation of IVIM parameter maps with less noise and fewer outliers.

Impact: A model-based reconstruction with low rank and temporal constraints improved IVIM image reconstruction, reducing noise and outliers in parameter estimates. Spline interpolation further facilitated reliable estimation of IVIM maps from just 5 b-values, benefiting clinical situations like stroke.

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction, self-supervised learning, subspace, dynamic MRI

Motivation: Supervised deep learning (DL) can reduce reconstruction time for CMR Multitasking, but the lack of ground truth limits the quality of supervised DL to that of iteratively reconstructed labels. 

Goal(s): Our goal was to develop a self-supervised learning (SSL) reconstruction method, whose performance is not limited by the iterative reconstruction. 

Approach: We developed a dual-domain subspace SSL reconstruction method for non-Cartesian dynamic MRI, applying it to CMR Multitasking.

Results: The proposed method can perform image reconstruction without reference images and shows better interscan consistency than supervised DL.

Impact: With the proposed method, image quality of DL reconstruction for CMR Multitasking can potentially surpass iterative reconstruction. We applied subspace constraints to SSL reconstruction, showing an efficient way to relieve the computational burden of dynamic MRI SSL reconstruction. 

Keywords: Image Reconstruction, Brain, Reconstruction

Motivation: Transformers excel in medical image processing but require many parameters and training data. We mitigated this issue with the POCS-Transformer method.

Goal(s): POCS-Transformer goal was to enhances MR image reconstruction along with preserving image quality with various under-sampling masks.

Approach: The POCS-Transformer, built on Swin-T using FastMRI data, employed binary undersampling, POCS augmentation, data consistency penalties, and was compared to VN and POCS-CycleGAN on test data.

Results: POCS-Transformer outperformed POCS-CycleGAN with superior image quality and less blurring. POCS-Transformer achieved higher mean PSNR and SSIM compared to both VN and POCS-CycleGAN in knee and brain image datasets.

Impact: The POCS-Transformer improves MR image reconstruction in terms of reducing blurring even under diverse under-sampling conditions. Its impact extends to healthcare and research. New questions involve its applications in medical imaging, merging traditional and modern methods to inspire further innovations.

Keywords: MR Fingerprinting, Radiotherapy

Motivation: Four-dimensional MR fingerprinting (4D-MRF) has the potential to improve precision and efficacy in abdominal radiotherapy (RT). However, the long reconstruction time of the state-of-the-art reconstruction method, compress-sensed-based respiratory-correlated 4D-MRF (CS-RC-4DMRF), limits its clinical application.

Goal(s): The study aims to develop a novel method to reduce the reconstruction time of CS-RC-4DMRF.

Approach: We developed PGDN-RC-4DMRF by integrating a deep proximal gradient descent network (PGDN) into 4D-MRF. Tumor motion tracking accuracy, tissue quantification accuracy, and image quality were evaluated.

Results: The proposed PGDN-RC-4DMRF method reduce the reconstruction time by a factor of 120, decreasing it from 8 min/slice to 4 s/slice while maintaining other metrics.

Impact: The improvement in the reconstruction speed of 4D-MRF through PGDN-RC-4DMRF may enhance the practicality of 4D-MRF in clinical settings for clinicians and potentially benefit RT outcomes for patients. 

Keywords: Image Reconstruction, Image Reconstruction

Motivation: In functional and diffusion MRI, multi-shot EPI enhances spatial resolution and minimizes distortion compared to single-shot scans. However, its vulnerability to shot-to-shot phase variations presents a significant challenge, with various proposed methods having drawbacks.

Goal(s): We propose a robust, navigator-free, computational efficient multi-shot method without SNR penalty.

Approach: In readout-segmented multi-shot EPI, we exploit the k-space overlapped regions between adjacent segments to extract relative phase fluctuations. This method, inspired by ESPIRiT and nonlinear gradient calibration, efficiently extracts shot-dependent phase variations in subspace.

Results: Our ex-vivo and in-vivo scans, including diffusion-weighted imaging, successfully achieves a high in-plane resolution of about 0.6mm without ghost artifacts.

Impact: Our proposed multi-shot technique eliminates the needs for time-consuming navigators, provides robust high-resolution diffusion and potentially functional imaging, and could be easily adapted for interleaved Cartesian and spiral EPI allowing robust phase error estimation from merely small k-space regions. 

Keywords: Image Reconstruction, Spectroscopy

Motivation: To further accelerate high-resolution MRSI acquisitions leveraging parallel imaging.

Goal(s): While standard parallel imaging techniques such as (k,t)-GRAPPA can interpolate the sparsely sampled (k,t)-space in MRSI, learning-based nonlinear interpolation has demonstrated better performance in parallel MRI. But these methods have not effectively utilized the time/free induction decay (FID) dimension, which should be leveraged to improve interpolation accuracy.

Approach: We adapted and extended the RAKI method by incorporating the FID dimension, via a 3D, complex-valued convolutional network, for MRSI reconstruction.

Results: Our method effectively reconstructed data for different undersampling designs in in vivo MRSI, leading to improved subsequent spatiospectral processing results.

Impact: We presented a self-supervised learning-based (k,t)-space interpolation method, (k,t)-RAKI, that is useful for further accelerating MRSI acquisition, in combination with subspace methods.

Keywords: Image Reconstruction, Image Reconstruction, Atlas, speech imaging

Motivation: Individuals across a population typically exhibit similar articulatory movements when performing speech tasks with specific speech samples. From an imaging experiment, we are interested in representing how an individual’s speech behavior is different from the ‘standard’ motion, which assists the preoperative planning of velopharyngeal surgery.

Goal(s): We expected to visualize velopharyngeal variations between individual subjects and the average population.

Approach: We have integrated an atlas into a low-rank residual reconstruction framework to capture the distinctive motion variations unique to each subject.

Results: We demonstrated the ability of the method to visualize velopharyngeal variations as well as enhancing the quality of the reconstruction process.

Impact: By applying a spatio-temporal atlas-driven reconstruction method, we were able to visualize and analysis velopharyngeal variations between individuals and the average population which will specifically benefit the surgical planning of individual cleft palate patients.

Keywords: Image Reconstruction, Motion Correction

Motivation: Real-time MRI offers a continuous and dynamic view of the object being imaged. Researchers have applied real-time MRI to speech tracking, which allows for the visualization of the vocal tract during speech production. 

Goal(s): In this study, we propose applying self-navigated subspace reconstruction to real-time MRI for speech tracking.

Approach: During reconstruction, 1000 frames were compressed to a few principal components, and iterative low-rank approximation was performed on compressed k-space, greatly reducing computation costs.

Results: The proposed method allows for the joint reconstruction of all time frames and provides the dynamic motion pattern of the vocal tract at a high frame rate. 

Impact: Our study presented a subspace reconstruction technique that does not require a navigator echo, which can be used for real-time MRI, particularly in speech tracking applications.