Golden-angle radial sparse parallel MRI: combination of compressed sensing, parallel imaging, and golden-angle radial sampling for fast and flexible dynamic volumetric MRI

Technical Improvements in Head and Neck MR Imaging: At the Cutting Edge

Head and neck MR imaging is technically challenging because of magnetic field inhomogeneity, respiratory and swallowing motion, and necessity of high-resolution imaging to trace key anatomic structures. These challenges have been answered by advances in MR imaging technology, including isovolumetric three-dimensional imaging, robust fat-water separation techniques, and novel deep learning-based reconstruction algorithms. New applications of MR imaging have been advanced and functional imaging has been improved. Improvements in acquisition and reconstruction technique facilitate novel applications of morphologic and functional imaging. This results in opportunities to improve diagnosis, staging, and treatment selection through application of advanced MR imaging techniques.

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.

Whole-heart, ungated, free-breathing, cardiac-phase-resolved myocardial perfusion MRI by using Continuous Radial Interleaved simultaneous Multi-slice acquisitions at sPoiled steady-state (CRIMP)

PURPOSE

To develop a whole-heart, free-breathing, non-electrocardiograph (ECG)-gated, cardiac-phase-resolved myocardial perfusion MRI framework (CRIMP; Continuous Radial Interleaved simultaneous Multi-slice acquisitions at sPoiled steady-state) and test its quantification feasibility.

METHODS

CRIMP used interleaved radial simultaneous multi-slice (SMS) slice groups to cover the whole heart in 9 or 12 short-axis slices. The sequence continuously acquired data without magnetization preparation, ECG gating or breath-holding, and captured multiple cardiac phases. Images were reconstructed by a motion-compensated patch-based locally low-rank reconstruction. Bloch simulations were performed to study the signal-to-noise ratio/contrast-to-noise ratio (SNR/CNR) for CRIMP and to study the steady-state signal under motion. Seven patients were scanned with CRIMP at stress and rest to develop the sequence. One human and two dogs were scanned at rest with a dual-bolus method to test the quantification feasibility of CRIMP. The dual-bolus scans were performed using both CRIMP and an ungated radial SMS saturation recovery (SMS-SR) sequence with injection dose = 0.075 mmol/kg to compare the sequences in terms of SNR, cardiac phase resolution and quantitative myocardial blood flow (MBF).

RESULTS

Perfusion images with multiple cardiac phases in all image slices with a temporal resolution of 72 ms/frame were obtained. Simulations and in-vivo acquisitions showed CRIMP kept the inner slices in steady-state regardless of motion. CRIMP outperformed SMS-SR in slice coverage (9 over 6), SNR (mean 20% improvement), and provided cardiac phase resolution. CRIMP and SMS-SR sequences provided comparable MBF values (rest systolic CRIMP = 0.58 ± 0.07, SMS-SR = 0.61 ± 0.16).

CONCLUSION

CRIMP allows for whole-heart, cardiac-phase-resolved myocardial perfusion images without ECG-gating or breath-holding. The sequence can provide MBF if an accurate arterial input function is obtained separately.

Linear Time Invariant Model based Motion Correction (LiMo-MoCo) of Dynamic Radial Contrast Enhanced MRI

Early identification of kidney function deterioration is essential to determine which newborn patients with dilation of the renal pelvis (hydronephrosis) should undergo surgery. Kidney function can be measured by fitting a tracer kinetic (TK) model onto a series of Dynamic Contrast Enhanced (DCE) MR images and deriving the glomerular filtration rate (GFR) from the TK model. Unfortunately, heavy breathing and large bulk motion events create outliers and misalignments that introduce large errors in the TK estimates. Moreover, aligning the series of DCE images is not trivial due to the contrast differences between them and the undersampling artifacts due to fast imaging. We present a bulk motion detection and a linear time invariant (LTI) model-based motion correction approach for DCE-MRI alignment that leverages the temporal dynamics of the DCE data at each voxel. We evaluate our approach on 10 newborn patients that underwent DCE imaging without sedation. For each patient, we reconstructed the sequence of DCE images, detected and removed the volumes corrupted by motion using a self navigation approach, aligned the sequence using our approach and fitted the TK model to compute GFR. The results show that our approach correctly aligned all volumes and improved the TK model fit and, on average, reducing the normalized root-mean-squared error by 0.17.

Dynamic MR image reconstruction based on total generalized variation and low-rank decomposition

PURPOSE

Propose a novel decomposition-based model employing the total generalized variation (TGV) and the nuclear norm, which can be used in compressed sensing-based dynamic MR reconstructions.

THEORY AND METHODS

We employ the nuclear norm to represent the time-coherent background and the spatiotemporal TGV functional for the sparse dynamic component above. We first design an algorithm using the classical first-order primal-dual method for solving the proposed model and then give the norm estimation for the convergence condition. The proposed model is compared with the state-of-the-art methods on different data sets under different sampling schemes and acceleration factors.

RESULTS

The proposed model achieves higher SERs and SSIMs than kt-SLR, kt-RPCA, L+S, and ICTGV on cardiac perfusion and breast DCE-MRI data sets under both the pseudoradial and the Cartesian sampling schemes. In addition, the proposed model better suppresses the spatial artifacts and preserves the edges.

CONCLUSIONS

The proposed model outperforms the state-of-the-art methods and generates high-quality reconstructions under different sampling schemes and different acceleration factors.

CArtesian sampling with Variable density and Adjustable temporal resolution (CAVA)

PURPOSE

To develop a variable density Cartesian sampling method that allows retrospective adjustment of temporal resolution for dynamic MRI applications and to validate it in real-time phase contrast MRI (PC-MRI).

THEORY AND METHODS

The proposed method, called CArtesian sampling with Variable density and Adjustable temporal resolution (CAVA), begins by producing a sequence of phase encoding indices based on the golden ratio increment. Then, variable density is introduced by nonlinear stretching of the indices. Finally, the elements of the resulting sequence are rounded up to the nearest integer. The performance of CAVA is evaluated using PC-MRI data from a pulsatile flow phantom and real-time, free-breathing data from ten healthy volunteers.

RESULTS

CAVA enabled image recovery at various temporal resolutions that were selected retrospectively. For the pulsatile flow phantom, image quality and flow quantification accuracy from CAVA were comparable to that from another pseudo-random sampling pattern with fixed temporal resolution. In addition, flow quantification results based on CAVA were in good agreement with a breath-held segmented acquisition.

CONCLUSIONS

By allowing retrospective binning of the MRI data, CAVA provides an avenue to retrospectively adjust the temporal resolution of PC-MRI.

Evaluating feasibility of high resolution T1-perfusion MRI with whole brain coverage using compressed SENSE: Application to glioma grading

PURPOSE

To evaluate the efficacy of optimized T1-Perfusion MRI protocol (protocol-2) with whole brain coverage and improved spatial resolution using Compressed-SENSE (CSENSE) to differentiate high-grade-glioma (HGG) and low-grade-glioma (LGG) and to compare it with the conventional protocol (protocol-1) with partial brain coverage used in our center.

METHODS

This study included MRI data from 5 healthy volunteers, a phantom and 126 brain tumor patients. Current study had two parts: To analyze the effect of CSENSE on 3D-T1-weighted (W) fast-field-echo (FFE) images, T1-W, dual-PDT2-W turbo-spin-echo images and T1 maps, and to evaluate the performance of high resolution T1-Perfusion MRI protocol with whole brain coverage optimized using CSENSE. Coefficient-of-Variation (COV), Relative-Percentage-Error (RPE), Normalized-Mean-Squared-Error (NMSE) and qualitative scoring were used for the former study. Tracer-kinetic (K^trans,v_e,v_p) and hemodynamic (rCBV,rCBF) parameters computed from both protocols were used to differentiate LGG and HGG.

RESULTS

The image quality of all structural images was found to be of diagnostic quality till R = 4. NMSE in healthy T1-W-FFE images and COV in phantom images increased with-respect-to R and images provided optimum quality till R = 4. Structural images and maps exhibited artefacts from R = 6. All parameters in tumor tissue and hemodynamic parameters in healthy gray matter tissue computed from both protocols were not significantly different. Parameters computed from protocol-2 performed better in terms of glioma grading. For both protocols, rCBF performed least (AUC = 0.759 and 0.851) and combination of all parameters performed best (AUC = 0.890 and 0.964).

CONCLUSION

CSENSE (R = 4) can be used to improve the resolution and brain coverage for T1-Perfusion analysis used to differentiate gliomas.

Prospective pediatric study comparing glomerular filtration rate estimates based on motion-robust dynamic contrast-enhanced magnetic resonance imaging and serum creatinine (eGFR) to 99mTc DTPA

BACKGROUND

Current methods to estimate glomerular filtration rate (GFR) have shortcomings. Estimates based on serum creatinine are known to be inaccurate in the chronically ill and during acute changes in renal function. Gold standard methods such as inulin and ^99mTc diethylenetriamine pentaacetic acid (DTPA) require blood or urine sampling and thus can be difficult to perform in children. Motion-robust radial volumetric interpolated breath-hold examination (VIBE) dynamic contrast-enhanced MRI represents a novel tool for estimating GFR that has not been validated in children.

OBJECTIVE

The purpose of our study was to determine the feasibility and accuracy of GFR measured by motion-robust radial VIBE dynamic contrast-enhanced MRI compared to estimates by serum creatinine (eGFR) and ^99mTc DTPA in children.

MATERIALS AND METHODS

We enrolled children, 0-18 years of age, who were undergoing both a contrast-enhanced MRI and nuclear medicine ^99mTc DTPA glomerular filtration rate (NM-GFR) within 2 weeks of each other. Enrolled children consented to an additional 6-min dynamic contrast-enhanced MRI scan using the motion-robust high spatiotemporal resolution prototype dynamic radial VIBE sequence (Siemens, Erlangen, Germany) at 3 tesla (T). The images were reconstructed offline with high temporal resolution (~3 s/volume) using compressed sensing image reconstruction including regularization in temporal dimension to improve image quality and reduce streaking artifacts. Images were then automatically post-processed using in-house-developed software. Post-processing steps included automatic segmentation of kidney parenchyma and aorta using convolutional neural network techniques and tracer kinetic model fitting using the Sourbron two-compartment model to calculate the MR-based GFR (MR-GFR). The NM-GFR was compared to MR-GFR and estimated GFR based on serum creatinine (eGFR) using Pearson correlation coefficient and Bland-Altman analysis.

RESULTS

Twenty-one children (7 female, 14 male) were enrolled between February 2017 and May 2018. Data from six of these children were not further analyzed because of deviations from the MRI protocol. Fifteen patients were analyzed (5 female, 10 male; average age 5.9 years); the method was technically feasible in all children. The results showed that the MR-GFR correlated with NM-GFR with a Pearson correlation coefficient (r-value) of 0.98. Bland-Altman analysis (i.e. difference of MR-GFR and NM-GFR versus mean of NM-GFR and MR-GFR) showed a mean difference of -0.32 and reproducibility coefficient of 18 with 95% confidence interval, and the coefficient of variation of 6.7% with values between -19 (-1.96 standard deviation) and 18 (+1.96 standard deviation). In contrast, serum creatinine compared with NM-GFR yielded an r-value of 0.73. Bland-Altman analysis (i.e. difference of eGFR and NM-GFR versus mean of NM-GFR and eGFR) showed a mean difference of 2.9 and reproducibility coefficient of 70 with 95% confidence interval, and the coefficient of variation of 25% with values between -67 (-1.96 standard deviation) and 73 (+1.96 standard deviation).

CONCLUSION

MR-GFR is a technically feasible and reliable method of measuring GFR when compared to the reference standard, NM-GFR by serum ^99mTc DTPA, and MR-GFR is more reliable than estimates based on serum creatinine.

Respiratory motion compensation using data binning in dynamic contrast enhanced golden-angle radial MRI

GRASP (Golden-Angle Radial Sparse Parallel MRI) is a data acquisition and reconstruction technique that combines parallel imaging and golden-angle radial sampling. The continuously acquired free breathing Dynamic Contrast Enhanced (DCE) golden-angle radial MRI data of liver and abdomen has artifacts due to respiratory motion, resulting in low vessel-tissue contrast that makes GRASP reconstructed images less suitable for diagnosis. In this paper, DCE golden-angle radial MRI data of abdomen and liver perfusion is sorted into different motion states using the self-gating property of radial acquisition and then reconstructed using GRASP. Three methods of amplitude-based data binning namely uniform binning, adaptive binning and optimal binning are applied on the DCE golden-angle radial data to extract different motion states and a comparison is performed with the conventional GRASP reconstruction. Also, a comparison among the amplitude-based data binning techniques is performed and benefits of each of these binning techniques are discussed from a clinical perspective. The image quality assessment in terms of hepatic vessel clarity, liver edge sharpness, contrast enhancement clarity and streaking artifacts is performed by a certified radiologist. The results show that DCE golden-angle radial trajectories benefit from all the three types of amplitude-based data binning methods providing improved reconstruction results. The choice of binning technique depends upon the clinical application e.g. uniform and adaptive binning are helpful for a detailed analysis of lesion characteristic and contrast enhancement in different motion states while optimal binning can be used when clinical analysis requires a single image per contrast enhancement phase with no motion blurring artifacts.

Assessment of Hepatic Perfusion Using GRASP MRI: Bringing Liver MRI on a New Level

PURPOSE

The aim of this study was to demonstrate the feasibility of hepatic perfusion imaging using dynamic contrast-enhanced (DCE) golden-angle radial sparse parallel (GRASP) magnetic resonance imaging (MRI) for characterizing liver parenchyma and hepatocellular carcinoma (HCC) before and after transarterial chemoembolization (TACE) as a potential alternative to volume perfusion computed tomography (VPCT).

METHODS AND MATERIALS

Between November 2017 and September 2018, 10 patients (male = 8; mean age, 66.5 ± 8.6 years) with HCC were included in this prospective, institutional review board-approved study. All patients underwent DCE GRASP MRI with high spatiotemporal resolution after injection of liver-specific MR contrast agent before and after TACE. In addition, VPCT was acquired before TACE serving as standard of reference. From the dynamic imaging data of DCE MRI and VPCT, perfusion maps (arterial liver perfusion [mL/100 mL/min], portal liver perfusion [mL/100 mL/min], hepatic perfusion index [%]) were calculated using a dual-input maximum slope model and compared with assess perfusion measures, lesion characteristics, and treatment response using Wilcoxon signed-rank test. To evaluate interreader agreement for measurement repeatability, the interclass correlation coefficient (ICC) was calculated.

RESULTS

Perfusion maps could be successfully generated from all DCE MRI and VPCT data. The ICC was excellent for all perfusion maps (ICC ≥ 0.88; P ≤ 0.001). Image analyses revealed perfusion parameters for DCE MRI and VPCT within the same absolute range for tumor and liver tissue. Dynamic contrast-enhanced MRI further enabled quantitative assessment of treatment response showing a significant decrease (P ≤ 0.01) of arterial liver perfusion and hepatic perfusion index in the target lesion after TACE.

CONCLUSIONS

Dynamic contrast-enhanced GRASP MRI allows for a reliable and robust assessment of hepatic perfusion parameters providing quantitative results comparable to VPCT and enables characterization of HCC before and after TACE, thus posing the potential to serve as an alternative to VPCT.

Impact of temporal resolution and motion correction for dynamic contrast-enhanced MRI of the liver using an accelerated golden-angle radial sequence

This paper presents a prospective study evaluating the impact on image quality and quantitative dynamic contrast-enhanced (DCE)-MRI perfusion parameters when varying the number of respiratory motion states when using an eXtra-Dimensional Golden-Angle Radial Sparse Parallel (XD-GRASP) MRI sequence. DCE acquisition was performed using a 3D stack-of-stars gradient-echo golden-angle radial acquisition in free-breathing with 100 spokes per motion state and temporal resolution of 6 s/volume, and using a non-rigid motion compensation to align different motion states. Parametric analysis was conducted using a dual-input single-compartment model. Nonparametric analysis was performed on the time-intensity curves. A total of 22 hepatocellular carcinomas (size: 11-52 mm) were evaluated. XD-GRASP reconstructed with increasing number of spokes for each motion state increased the signal-to-noise ratio (SNR) (p < 0.05) but decreased temporal resolution (0.04 volume/s vs 0.17 volume/s for one motion state) (p < 0.05). A visual scoring by an experienced radiologist show no change between increasing number of motion states with same number of spokes using the Likert score. The normalized maximum intensity time ratio, peak enhancement ratio and tumor arterial fraction increased with decreasing number of motion states (p < 0.05) while the transfer constant from the portal venous plasma to the surrounding tissue significantly decreased (p < 0.05). These same perfusion parameters show a significant difference in case of tumor displacement more than 1 cm (p < 0.05) whereas in the opposite case there was no significant variation. While a higher number of motion states and higher number of spokes improves SNR, the resulting lower temporal resolution can influence quantitative parameters that capture rapid signal changes. Finally, fewer displacement compensation is advantageous with lower number of motion state due to the higher temporal resolution. XD-GRASP can be used to perform quantitative perfusion measures in the liver, but the number of motion states may significantly alter some quantitative parameters.

Free-breathing fat and R2 * quantification in the liver using a stack-of-stars multi-echo acquisition with respiratory-resolved model-based reconstruction

Purpose

To develop a free‐breathing hepatic fat and R2∗ quantification method by extending a previously described stack‐of‐stars model‐based fat‐water separation technique with additional modeling of the transverse relaxation rate R2∗.

Methods

The proposed technique combines motion‐robust radial sampling using a stack‐of‐stars bipolar multi‐echo 3D GRE acquisition with iterative model‐based fat‐water separation. Parallel‐Imaging and Compressed‐Sensing principles are incorporated through modeling of the coil‐sensitivity profiles and enforcement of total‐variation (TV) sparsity on estimated water, fat, and R2∗ parameter maps. Water and fat signals are used to estimate the confounder‐corrected proton‐density fat fraction (PDFF). Two strategies for handling respiratory motion are described: motion‐averaged and motion‐resolved reconstruction. Both techniques were evaluated in patients (n = 14) undergoing a hepatobiliary research protocol at 3T. PDFF and R2∗ parameter maps were compared to a breath‐holding Cartesian reference approach.

Results

Linear regression analyses demonstrated strong (r > 0.96) and significant (P ≪ .01) correlations between radial and Cartesian PDFF measurements for both the motion‐averaged reconstruction (slope: 0.90; intercept: 0.07%) and the motion‐resolved reconstruction (slope: 0.90; intercept: 0.11%). The motion‐averaged technique overestimated hepatic R2∗ values (slope: 0.35; intercept: 30.2 1/s) compared to the Cartesian reference. However, performing a respiratory‐resolved reconstruction led to better R2∗ value consistency (slope: 0.77; intercept: 7.5 1/s).

Conclusions

The proposed techniques are promising alternatives to conventional Cartesian imaging for fat and R2∗ quantification in patients with limited breath‐holding capabilities. For accurate R2∗ estimation, respiratory‐resolved reconstruction should be used.

Comparison of Objective Image Quality Metrics to Expert Radiologists' Scoring of Diagnostic Quality of MR Images

Image quality metrics (IQMs) such as root mean square error (RMSE) and structural similarity index (SSIM) are commonly used in the evaluation and optimization of accelerated magnetic resonance imaging (MRI) acquisition and reconstruction strategies. However, it is unknown how well these indices relate to a radiologist's perception of diagnostic image quality. In this study, we compare the image quality scores of five radiologists with the RMSE, SSIM, and other potentially useful IQMs: peak signal to noise ratio (PSNR) multi-scale SSIM (MSSSIM), information-weighted SSIM (IWSSIM), gradient magnitude similarity deviation (GMSD), feature similarity index (FSIM), high dynamic range visible difference predictor (HDRVDP), noise quality metric (NQM), and visual information fidelity (VIF). The comparison uses a database of MR images of the brain and abdomen that have been retrospectively degraded by noise, blurring, undersampling, motion, and wavelet compression for a total of 414 degraded images. A total of 1017 subjective scores were assigned by five radiologists. IQM performance was measured via the Spearman rank order correlation coefficient (SROCC) and statistically significant differences in the residuals of the IQM scores and radiologists' scores were tested. When considering SROCC calculated from combining scores from all radiologists across all image types, RMSE and SSIM had lower SROCC than six of the other IQMs included in the study (VIF, FSIM, NQM, GMSD, IWSSIM, and HDRVDP). In no case did SSIM have a higher SROCC or significantly smaller residuals than RMSE. These results should be considered when choosing an IQM in future imaging studies.

Predicting clinically significant prostate cancer from quantitative image features including compressed sensing radial MRI of prostate perfusion using machine learning: comparison with PI-RADS v2 assessment scores

Background

To investigate if supervised machine learning (ML) classifiers would be able to predict clinically significant cancer (sPC) from a set of quantitative image-features and to compare these results with established PI-RADS v2 assessment scores.

Methods

We retrospectively included 201, histopathologically-proven, peripheral zone (PZ) prostate cancer lesions. Gleason scores ≤3+3 were considered as clinically insignificant (inPC) and Gleason scores ≥3+4 as sPC and were encoded in a binary fashion, serving as ground-truth. MRI was performed at 3T with high spatiotemporal resolution DCE using Golden-angle RAdial SParse (GRASP) MRI. Perfusion maps (Ktrans, Kep, Ve), apparent diffusion coefficient (ADC), and absolute T2-signal intensities (SI) were determined in all lesions and served as input parameters for four supervised ML models: Gradient Boosting Machines (GBM), Neural Networks (NNet), Random Forest (RF) and Support Vector Machines (SVM). ML results and PI-RADS scores were compared with the ground-truth. Next ROC-curves and AUC values were calculated.

Results

All ML models outperformed PI-RADS v2 assessment scores in the prediction of sPC (RF, GBM, NNet and SVM vs. PI-RADS: AUC 0.899, 0.864, 0.884 and 0.874 vs. 0.595, all P<0.001).

Conclusions

Using quantitative imaging parameters as input, supervised ML models outperformed PI-RADS v2 assessment scores in the prediction of sPC. These results indicate that quantitative imagining parameters contain relevant information for the prediction of sPC from image features.

Bi-Linear Modeling of Data Manifolds for Dynamic-MRI Recovery

This paper puts forth a novel bi-linear modeling framework for data recovery via manifold-learning and sparse-approximation arguments and considers its application to dynamic magnetic-resonance imaging (dMRI). Each temporal-domain MR image is viewed as a point that lies onto or close to a smooth manifold, and landmark points are identified to describe the point cloud concisely. To facilitate computations, a dimensionality reduction module generates low-dimensional/compressed renditions of the landmark points. Recovery of high-fidelity MRI data is realized by solving a non-convex minimization task for the linear decompression operator and affine combinations of landmark points which locally approximate the latent manifold geometry. An algorithm with guaranteed convergence to stationary solutions of the non-convex minimization task is also provided. The aforementioned framework exploits the underlying spatio-temporal patterns and geometry of the acquired data without any prior training on external data or information. Extensive numerical results on simulated as well as real cardiac-cine MRI data illustrate noteworthy improvements of the advocated machine-learning framework over state-of-the-art reconstruction techniques.

From Compressed-Sensing to Artificial Intelligence-Based Cardiac MRI Reconstruction

Cardiac magnetic resonance (CMR) imaging is an important tool for the non-invasive assessment of cardiovascular disease. However, CMR suffers from long acquisition times due to the need of obtaining images with high temporal and spatial resolution, different contrasts, and/or whole-heart coverage. In addition, both cardiac and respiratory-induced motion of the heart during the acquisition need to be accounted for, further increasing the scan time. Several undersampling reconstruction techniques have been proposed during the last decades to speed up CMR acquisition. These techniques rely on acquiring less data than needed and estimating the non-acquired data exploiting some sort of prior information. Parallel imaging and compressed sensing undersampling reconstruction techniques have revolutionized the field, enabling 2- to 3-fold scan time accelerations to become standard in clinical practice. Recent scientific advances in CMR reconstruction hinge on the thriving field of artificial intelligence. Machine learning reconstruction approaches have been recently proposed to learn the non-linear optimization process employed in CMR reconstruction. Unlike analytical methods for which the reconstruction problem is explicitly defined into the optimization process, machine learning techniques make use of large data sets to learn the key reconstruction parameters and priors. In particular, deep learning techniques promise to use deep neural networks (DNN) to learn the reconstruction process from existing datasets in advance, providing a fast and efficient reconstruction that can be applied to all newly acquired data. However, before machine learning and DNN can realize their full potentials and enter widespread clinical routine for CMR image reconstruction, there are several technical hurdles that need to be addressed. In this article, we provide an overview of the recent developments in the area of artificial intelligence for CMR image reconstruction. The underlying assumptions of established techniques such as compressed sensing and low-rank reconstruction are briefly summarized, while a greater focus is given to recent advances in dictionary learning and deep learning based CMR reconstruction. In particular, approaches that exploit neural networks as implicit or explicit priors are discussed for 2D dynamic cardiac imaging and 3D whole-heart CMR imaging. Current limitations, challenges, and potential future directions of these techniques are also discussed.

MRSIGMA: Magnetic Resonance SIGnature MAtching for real-time volumetric imaging

PURPOSE

To propose a real-time 3D MRI technique called MR SIGnature MAtching (MRSIGMA) for high-resolution volumetric imaging and motion tracking with very low imaging latency.

METHODS

MRSIGMA consists of two steps: (1) offline learning of a database of possible 3D motion states and corresponding motion signature ranges and (2) online matching of new motion signatures acquired in real time with prelearned motion states. Specifically, the offline learning step (non-real-time) reconstructs motion-resolved 4D images representing different motion states and assigns a unique motion range to each state. The online matching step (real-time) acquires motion signatures only and selects one of the prelearned 3D motion states for each newly acquired signature, which generates 3D images efficiently in real time. The MRSIGMA technique was evaluated on 15 golden-angle stack-of-stars liver data sets, and the performance of respiratory motion tracking with the online-generated real-time 3D MRI was compared with the corresponding 2D projections acquired in real time.

RESULTS

The total latency of generating each 3D image during online matching was about 300 ms, including acquisition of the motion signature data (~138 ms) and corresponding matching process (~150 ms). Linear correlation assessment suggested excellent correlation (R²  = 0.948) between motion displacement measured from the online-generated real-time 3D images and the 2D real-time projections.

CONCLUSION

This proof-of-concept study demonstrates the feasibility of MRSIGMA for high-resolution real-time volumetric imaging, which shifts the acquisition and reconstruction burden to an offline learning step and leaves fast online matching for online imaging with very low imaging latency. The MRSIGMA technique can potentially be used for real-time motion tracking in MRI-guided radiation therapy.

Improved gradient-echo 3D magnetic resonance imaging using compressed sensing and Toeplitz encoding with phase-scrambled RF excitation

PURPOSE

To develop a novel three-dimensional (3D) hybrid-encoding framework using compressed sensing (CS) and Toeplitz encoding with variable phase-scrambled radio-frequency (RF) excitation, which has the following advantages: low power deposition of RF pulses, reduction of the signal dynamic range, no additional hardware requirement, and signal-to-noise ratio (SNR) improvement.

METHODS

In light of the actual imaging framework of magnetic resonance imaging (MRI) scanners, we applied specially tailored RF pulses with phase-scrambled RF excitation to implement a 3D hybrid Fourier-Toeplitz encoding method based on 3D gradient-recalled echo pulse (GRASS) sequence. This method exploits Toeplitz encoding along the phase encoding direction, while keeping Fourier encoding along the readout and slice encoding directions. Phantom experiments were conducted to optimize the amplitude of specially tailored RF pulses in the 3D GRASS sequence. In vivo experiments were conducted to validate the feasibility of the proposed method, and simulations were conducted to compare the 3D hybrid-encoding method with Fourier encoding and other non-Fourier encoding methods.

RESULTS

An optimized low RF amplitude was obtained in the phantom experiments. Using the optimized specially tailored RF pulses, both the watermelon and knee experiments demonstrated that the proposed method was able to preserve more image details than the conventional 3D Fourier-encoded methods at acceleration factors of 3.1 and 2.0. Additionally, SNR was improved because of no additional gradients and 3D volume encoding, when compared with single-slice scanning without 3D encoding. Simulation results demonstrated that the proposed scheme was superior to the conventional Fourier encoding method, and obtained comparative performance with other non-Fourier encoding methods in preserving details.

CONCLUSIONS

We developed a practical hybrid-encoding method for 3D MRI with specially tailored RF pulses of phase-scrambled RF excitation. The proposed method improves image SNR and detail preservation compared with the conventional Fourier encoding methods. Furthermore, our proposed method exhibits superior performance in terms of detail preservation, compared with the conventional Fourier encoding method.

MRI Techniques to Decrease Imaging Times in Children

Long acquisition times can limit the use of MRI in pediatric patients, and the use of sedation or general anesthesia is frequently necessary to facilitate diagnostic examinations. The use of sedation or anesthesia has disadvantages including increased cost and imaging time and potential risks to the patient. Reductions in imaging time may decrease or eliminate the need for sedation or general anesthesia. Over the past decade, a number of imaging techniques that can decrease imaging time have become commercially available. These products have been used increasingly in clinical practice and include parallel imaging, simultaneous multisection imaging, radial k-space acquisition, compressed sensing MRI reconstruction, and automated protocol selection software. The underlying concepts, supporting data, current clinical applications, and available products for each of these strategies are reviewed in this article. In addition, emerging techniques that are still under investigation may provide further reductions in imaging time, including artificial intelligence-based reconstruction, gradient-controlled aliasing sampling and reconstruction, three-dimensional MR spectroscopy, and prospective motion correction. The preliminary results for these techniques are also discussed. ^©RSNA, 2020 See discussion on this article by Greer and Vasanawala.

Incorporating prior knowledge via volumetric deep residual network to optimize the reconstruction of sparsely sampled MRI

For sparse sampling that accelerates magnetic resonance (MR) image acquisition, non-linear reconstruction algorithms have been developed, which incorporated patient specific a prior information. More generic a prior information could be acquired via deep learning and utilized for image reconstruction. In this study, we developed a volumetric hierarchical deep residual convolutional neural network, referred to as T-Net, to provide a data-driven end-to-end mapping from sparsely sampled MR images to fully sampled MR images, where cartilage MR images were acquired using an Ultra-short TE sequence and retrospectively undersampled using pseudo-random Cartesian and radial acquisition schemes. The network had a hierarchical architecture that promoted the sparsity of feature maps and increased the receptive field, which were valuable for signal synthesis and artifact suppression. Relatively dense local connections and global shortcuts were established to facilitate residual learning and compensate for details lost in hierarchical processing. Additionally, volumetric processing was adopted to fully exploit spatial continuity in three-dimensional space. Data consistency was further enforced. The network was trained with 336 three-dimensional images (each consisting of 32 slices) and tested by 24 images. The incorporation of a priori information acquired via deep learning facilitated high acceleration factors (as high as 8) while maintaining high image fidelity (quantitatively evaluated using the structural similarity index measurement). The proposed T-Net had an improved performance as compared to several state-of-the-art networks.

Feed and wrap magnetic resonance urography provides anatomic and functional imaging in infants without anesthesia

OBJECTIVE

To describe a technique for performing magnetic resonance urogram (MRU) in infants without sedation or anesthesia.

METHODS

Eighteen infants underwent MRU in the absence of sedating medications using a 'feed and wrap' technique (FW-MRU). Dynamic contrast enhanced images were obtained. Dynamic radial VIBE and compressed sensing image reconstruction were used to correct for motion artifact.

RESULTS

Seventeen of the 18 patients had successful FW-MRU. Feed and wrap' magnetic resonance urogram provided high-quality anatomic and functional renal data.

CONCLUSION

Initial experience with FW-MRU demonstrates it to be a promising anesthesia-free modality for obtaining anatomic and functional imaging of the urinary tract in infants.

Improving dynamic contrast-enhanced MRI of the lung using motion-weighted sparse reconstruction: Initial experiences in patients

PURPOSE

The purpose of this study was to evaluate the performance of motion-weighted Golden-angle RAdial Sparse Parallel MRI (motion-weighted GRASP) for free-breathing dynamic contrast-enhanced MRI (DCE-MRI) of the lung.

METHODS

Motion-weighted GRASP incorporates a soft-gating motion compensation algorithm into standard GRASP reconstruction, so that motion-corrupted motion k-space (e.g., k-space acquired in inspiratory phases) contributes less to the final reconstructed images. Lung MR data from 20 patients (mean age = 57.9 ± 13.5) with known pulmonary lesions were retrospectively collected for this study. Each subject underwent a free-breathing DCE-MR scan using a fat-statured T₁-weighted stack-of-stars golden-angle radial sequence and a post-contrast breath-hold MR scan using a Cartesian volumetric-interpolated imaging sequence (BH-VIBE). Each radial dataset was reconstructed using GRASP without motion compensation and motion-weighted GRASP. All MR images were visually evaluated by two experienced radiologists blinded to reconstruction and acquisition schemes independently. In addition, the influence of motion-weighted reconstruction on dynamic contrast-enhancement patterns was also investigated.

RESULTS

For image quality assessment, motion-weighted GRASP received significantly higher visual scores than GRASP (P < 0.05) for overall image quality (3.68 vs. 3.39), lesion conspicuity (3.54 vs. 3.18) and overall artifact level (3.53 vs. 3.15). There was no significant difference (P > 0.05) between the breath-hold BH-VIBE and motion-weighted GRASP images. For assessment of temporal fidelity, motion-weighted GRASP maintained a good agreement with respect to GRASP.

CONCLUSION

Motion-weighted GRASP achieved better reconstruction performance in free-breathing DCE-MRI of the lung compared to standard GRASP, and it may enable improved assessment of pulmonary lesions.

A 12-channel flexible receiver coil for accelerated tongue imaging

Objective

MRI of the tongue requires acceleration to minimise motion artefacts and to facilitate real-time imaging of swallowing. To accelerate tongue MRI, we designed a dedicated flexible receiver coil.

Materials and methods

We designed a flexible 12-channel receiver coil for tongue MRI at 3T and compared it to a conventional head-and-neck coil regarding SNR and g-factor. Furthermore, two accelerated imaging techniques were evaluated using both coils: multiband (MB) diffusion-tensor imaging (DTI) and real-time MRI of swallowing.

Results

The flexible coil had significantly higher SNR in the anterior (2.1 times higher, P = 0.002) and posterior (2.0 times higher, P < 0.001) parts of the tongue, while the g-factor was lower at higher acceleration. Unlike for the flexible coil, the apparent diffusion coefficient (P = 0.001) and fractional anisotropy (P = 0.008) deteriorated significantly while using the conventional coil after accelerating DTI with MB. The image quality of real-time MRI of swallowing was significantly better for hyoid elevation (P = 0.029) using the flexible coil.

Conclusion

Facilitated by higher SNR and lower g-factor values, our flexible tongue coil allows faster imaging, which was successfully demonstrated in MB DTI and real-time MRI of swallowing.

Electronic supplementary material

The online version of this article (10.1007/s10334-019-00824-5) contains supplementary material, which is available to authorized users.

Improvement of Fast Model-Based Acceleration of Parameter Look-Locker T1 Mapping

Quantitative mapping is desirable in many scientific and clinical magneric resonance imaging (MRI) applications. Recent inverse recovery-look locker sequence enables single-shot T₁ mapping with a time of a few seconds but the main computational load is directed into offline reconstruction, which can take from several minutes up to few hours. In this study we proposed improvement of model-based approach for T₁-mapping by introduction of two steps fitting procedure. We provided analysis of further reduction of k-space data, which lead us to decrease of computational time and perform simulation of multi-slice development. The region of interest (ROI) analysis of human brain measurements with two different initial models shows that the differences between mean values with respect to a reference approach are in white matter-0.3% and 1.1%, grey matter-0.4% and 1.78% and cerebrospinal fluid-2.8% and 11.1% respectively. With further improvements we were able to decrease the time of computational of single slice to 6.5 min and 23.5 min for different initial models, which has been already not achieved by any other algorithm. In result we obtained an accelerated novel method of model-based image reconstruction in which single iteration can be performed within few seconds on home computer.

Advances in Diffusion and Perfusion MRI for Quantitative Cancer Imaging

PURPOSE OF REVIEW

This article is to review recent technical developments and their clinical applications in cancer imaging quantitative measurement of cellular and vascular properties of the tumors.

RECENT FINDINGS

Rapid development of fast Magnetic Resonance Imaging (MRI) technologies over last decade brought new opportunities in quantitative MRI methods to measure both cellular and vascular properties of tumors simultaneously.

SUMMARY

Diffusion MRI (dMRI) and dynamic contrast enhanced (DCE)-MRI have become widely used to assess the tissue structural and vascular properties, respectively. However, the ultimate potential of these advanced imaging modalities has not been fully exploited. The dependency of dMRI on the diffusion weighting gradient strength and diffusion time can be utilized to measure tumor perfusion, cellular structure, and cellular membrane permeability. Similarly, DCE-MRI can be used to measure vascular and cellular membrane permeability along with cellular compartment volume fractions. To facilitate the understanding of these potentially important methods for quantitative cancer imaging, we discuss the basic concepts and recent developments, as well as future directions for further development.

Tracer kinetic models as temporal constraints during brain tumor DCE-MRI reconstruction

Purpose

To apply tracer kinetic models as temporal constraints during reconstruction of under‐sampled brain tumor dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI).

Methods

A library of concentration vs time profiles is simulated for a range of physiological kinetic parameters. The library is reduced to a dictionary of temporal bases, where each profile is approximated by a sparse linear combination of the bases. Image reconstruction is formulated as estimation of concentration profiles and sparse model coefficients with a fixed sparsity level. Simulations are performed to evaluate modeling error, and error statistics in kinetic parameter estimation in presence of noise. Retrospective under‐sampling experiments are performed on a brain tumor DCE digital reference object (DRO), and 12 brain tumor in‐vivo 3T datasets. The performances of the proposed under‐sampled reconstruction scheme and an existing compressed sensing‐based temporal finite‐difference (tFD) under‐sampled reconstruction were compared against the fully sampled inverse Fourier Transform‐based reconstruction.

Results

Simulations demonstrate that sparsity levels of 2 and 3 model the library profiles from the Patlak and extended Tofts‐Kety (ETK) models, respectively. Noise sensitivity analysis showed equivalent kinetic parameter estimation error statistics from noisy concentration profiles, and model approximated profiles. DRO‐based experiments showed good fidelity in recovery of kinetic maps from 20‐fold under‐sampled data. In‐vivo experiments demonstrated reduced bias and uncertainty in kinetic mapping with the proposed approach compared to tFD at under‐sampled reduction factors >= 20.

Conclusions

Tracer kinetic models can be applied as temporal constraints during brain tumor DCE‐MRI reconstruction. The proposed under‐sampled scheme resulted in model parameter estimates less biased with respect to conventional fully sampled DCE MRI reconstructions and parameter estimation. The approach is flexible, can use nonlinear kinetic models, and does not require tuning of regularization parameters.

Volumetric multicomponent T1ρ relaxation mapping of the human liver under free breathing at 3T

PURPOSE

To develop a 3D sequence for T_1ρ relaxation mapping using radial volumetric encoding (3D-T_1ρ -RAVE) and to evaluate the multi relaxation components in the liver of healthy controls and chronic liver disease (CLD) patients.

METHODS

Fat saturation and T_1ρ preparation modules were followed by a train of gradient-echo acquisitions and T₁ restoration delay. The series of T_1ρ -weighted images were fitted using mono-exponential, bi-exponential, and stretched-exponential models. The repeatability and reproducibility of the proposed technique were evaluated on National Institute of Standards and Technology phantom by calculating the coefficient of variation between test-retest scans on the same scanner and between two different 3T scanners, respectively. Mann-Whitney U-test was performed to assess differences in T_1ρ components among patients (n = 3) and a control group (n = 10).

RESULTS

The phantom study showed an error of 8.9% and 11.5% in mono T₂ relaxation time measurement relative to the reference on 2 different scanners. The coefficient of variation for test-retest scans performed on the same scanner was 5.7% and 2.4% for scans performed on 2 scanners. The comparison between healthy controls and CLD patients showed a significant difference (P < .05) in mono relaxation time (P = .002), stretched-exponential relaxation parameter (P = .04). The Akaike information criteria C criterion showed 2.53 ± 0.9% (2.3 ± 0.3% for CLD) of the voxels are bi-exponential while in 65.3 ± 5.8% (81.2 ± 0.06% for CLD) of the liver voxels, the stretched-exponential model was preferred.

CONCLUSION

The 3D-T_1ρ -RAVE sequence allows volumetric, multicomponent T_1ρ assessment of the liver during free breathing and can distinguish between healthy volunteers and CLD patients.

Perturbed spiral real-time phase-contrast MR with compressive sensing reconstruction for assessment of flow in children

PURPOSE

we implemented a golden-angle spiral phase contrast sequence. A commonly used uniform density spiral and a new 'perturbed' spiral that produces more incoherent aliases were assessed. The aim was to ascertain whether greater incoherence enabled more accurate Compressive Sensing reconstruction and superior measurement of flow and velocity.

METHODS

A range of 'perturbed' spiral trajectories based on a uniform spiral trajectory were formulated. The trajectory that produced the most noise-like aliases was selected for further testing. For in-silico and in-vivo experiments, data was reconstructed using total Variation L1 regularisation in the spatial and temporal domains. In-silico, the reconstruction accuracy of the 'perturbed' golden spiral was compared to uniform density golden-angle spiral. For the in-vivo experiment, stroke volume and peak mean velocity were measured in 20 children using 'perturbed' and uniform density golden-angle spiral sequences. These were compared to a reference standard gated Cartesian sequence.

RESULTS

In-silico, the perturbed spiral acquisition produced more accurate reconstructions with less temporal blurring (NRMSE ranging from 0.03 to 0.05) than the uniform density acquisition (NRMSE ranging from 0.06 to 0.12). This translated in more accurate results in-vivo with no significant bias in the peak mean velocity (bias: -0.1, limits: -4.4 to 4.1 cm/s; P = 0.98) or stroke volume (bias: -1.8, limits: -9.4 to 5.8 ml, P = 0.19).

CONCLUSION

We showed that a 'perturbed' golden-angle spiral approach is better suited to Compressive Sensing reconstruction due to more incoherent aliases. This enabled accurate real-time measurement of flow and peak velocity in children.

Topics on quantitative liver magnetic resonance imaging

Liver magnetic resonance imaging (MRI) is subject to continuous technical innovations through advances in hardware, sequence and novel contrast agent development. In order to utilize the abilities of liver MR to its full extent and perform high-quality efficient exams, it is mandatory to use the best imaging protocol, to minimize artifacts and to select the most adequate type of contrast agent. In this article, we review the routine clinical MR techniques applied currently and some latest developments of liver imaging techniques to help radiologists and technologists to better understand how to choose and optimize liver MRI protocols that can be used in clinical practice. This article covers topics on (I) fat signal suppression; (II) diffusion weighted imaging (DWI) and intravoxel incoherent motion (IVIM) analysis; (III) dynamic contrast-enhanced (DCE) MR imaging; (IV) liver fat quantification; (V) liver iron quantification; and (VI) scan speed acceleration.

Motion Compensated Dynamic MRI Reconstruction With Local Affine Optical Flow Estimation

This paper proposes a novel framework to reconstruct dynamic magnetic resonance imaging (DMRI) with motion compensation (MC). Specifically, by combining the intensity-based optical flow constraint with the traditional compressed sensing scheme, we are able to jointly reconstruct the DMRI sequences and estimate the interframe motion vectors. Then, the DMRI reconstruction can be refined through MC with the estimated motion field. By employing the coarse-to-fine multi-scale resolution strategy, we are able to update the motion field in different spatial scales. The estimated motion vectors need to be interpolated to the finest resolution scale to compensate the DMRI reconstruction. Moreover, the proposed framework is capable of handling a wide class of prior information (regularizations) for DMRI reconstruction, such as sparsity, low rank, and total variation. The formulated optimization problem is solved by a primal-dual algorithm with linesearch due to its efficiency when dealing with non-differentiable problems. Experiments on various DMRI datasets validate the reconstruction quality improvement using the proposed scheme in comparison to several state-of-the-art algorithms.

Opportunities in Interventional and Diagnostic Imaging by Using High-Performance Low-Field-Strength MRI

Background Commercial low-field-strength MRI systems are generally not equipped with state-of-the-art MRI hardware, and are not suitable for demanding imaging techniques. An MRI system was developed that combines low field strength (0.55 T) with high-performance imaging technology. Purpose To evaluate applications of a high-performance low-field-strength MRI system, specifically MRI-guided cardiovascular catheterizations with metallic devices, diagnostic imaging in high-susceptibility regions, and efficient image acquisition strategies. Materials and Methods A commercial 1.5-T MRI system was modified to operate at 0.55 T while maintaining high-performance hardware, shielded gradients (45 mT/m; 200 T/m/sec), and advanced imaging methods. MRI was performed between January 2018 and April 2019. T1, T2, and T2* were measured at 0.55 T; relaxivity of exogenous contrast agents was measured; and clinical applications advantageous at low field were evaluated. Results There were 83 0.55-T MRI examinations performed in study participants (45 women; mean age, 34 years ± 13). On average, T1 was 32% shorter, T2 was 26% longer, and T2* was 40% longer at 0.55 T compared with 1.5 T. Nine metallic interventional devices were found to be intrinsically safe at 0.55 T (<1°C heating) and MRI-guided right heart catheterization was performed in seven study participants with commercial metallic guidewires. Compared with 1.5 T, reduced image distortion was shown in lungs, upper airway, cranial sinuses, and intestines because of improved field homogeneity. Oxygen inhalation generated lung signal enhancement of 19% ± 11 (standard deviation) at 0.55 T compared with 7.6% ± 6.3 at 1.5 T (P = .02; five participants) because of the increased T1 relaxivity of oxygen (4.7e-4 mmHg-1sec-1). Efficient spiral image acquisitions were amenable to low field strength and generated increased signal-to-noise ratio compared with Cartesian acquisitions (P < .02). Representative imaging of the brain, spine, abdomen, and heart generated good image quality with this system. Conclusion This initial study suggests that high-performance low-field-strength MRI offers advantages for MRI-guided catheterizations with metal devices, MRI in high-susceptibility regions, and efficient imaging. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Grist in this issue.

SANTIS: Sampling-Augmented Neural neTwork with Incoherent Structure for MR image reconstruction

PURPOSE

To develop and evaluate a novel deep learning-based reconstruction framework called SANTIS (Sampling-Augmented Neural neTwork with Incoherent Structure) for efficient MR image reconstruction with improved robustness against sampling pattern discrepancy.

METHODS

With a combination of data cycle-consistent adversarial network, end-to-end convolutional neural network mapping, and data fidelity enforcement for reconstructing undersampled MR data, SANTIS additionally utilizes a sampling-augmented training strategy by extensively varying undersampling patterns during training, so that the network is capable of learning various aliasing structures and thereby removing undersampling artifacts more effectively and robustly. The performance of SANTIS was demonstrated for accelerated knee imaging and liver imaging using a Cartesian trajectory and a golden-angle radial trajectory, respectively. Quantitative metrics were used to assess its performance against different references. The feasibility of SANTIS in reconstructing dynamic contrast-enhanced images was also demonstrated using transfer learning.

RESULTS

Compared to conventional reconstruction that exploits image sparsity, SANTIS achieved consistently improved reconstruction performance (lower errors and greater image sharpness). Compared to standard learning-based methods without sampling augmentation (e.g., training with a fixed undersampling pattern), SANTIS provides comparable reconstruction performance, but significantly improved robustness, against sampling pattern discrepancy. SANTIS also achieved encouraging results for reconstructing liver images acquired at different contrast phases.

CONCLUSION

By extensively varying undersampling patterns, the sampling-augmented training strategy in SANTIS can remove undersampling artifacts more robustly. The novel concept behind SANTIS can particularly be useful for improving the robustness of deep learning-based image reconstruction against discrepancy between training and inference, an important, but currently less explored, topic.

Estimation of cellular-interstitial water exchange in dynamic contrast enhanced MRI using two flip angles

PURPOSE

To investigate the feasibility of using multiple flip angles in dynamic contrast enhanced (DCE) MRI to reduce the uncertainty in estimation of intracellular water lifetime (τ_i ).

METHODS

Numerical simulation studies were conducted to assess the uncertainty in estimation of τ_i using dynamic contrast enhanced MRI with one or two flip angles. In vivo experiments with a murine brain tumor model were conducted at 7T using two flip angles. The in vivo data were used to compare τ_i estimation using the single-flip-angle (SFA) protocol with that using the double-flip-angle (DFA) protocol. Data analysis was conducted using the two-compartment exchange model combined with the three-site-two-exchange model for water exchange.

RESULTS

In the numerical simulation studies with a range of contrast kinetic parameters and signal-to-noise ratio = 20, the median bias of τ_i estimation decreased from 72 ms with SFA to 65 ms with DFA, and the corresponding median inter-quartile range reduced from 523 ms to 156 ms. In the in vivo studies, τ_i estimation with SFA was not successful in most voxels in the tumors, as the estimated τ_i values reached the upper limit of the parameter range (2 s). In contrast, the estimated τ_i values with DFA were mostly between 0.2 and 1.5 s and homogeneously distributed spatially across the tumor. The τ_i estimation with DFA was less sensitive to arterial input function scaling but more sensitive to pre-contrast T₁ than the other contrast kinetic parameters.

CONCLUSION

This study results demonstrate the feasibility of using multiple flip angles to encode the post-contrast time-intensity curve with different weighting of water exchange effect to reduce the uncertainty in τ_i estimation.

Quantitative renal function assessment of atheroembolic renal disease using view-shared compressed sensing based dynamic-contrast enhanced MR imaging: An in vivo study

Atheroembolic renal disease (AERD) is the major cause of renal insufficiency in the elderly, and particularly, the diagnose of AERD is often delayed and even missed due to its nonspecific presentation and the sudden occurrence of an embolic event. To investigate the feasibility of the view-shared compressed sensing (VCS) based dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) in the assessment of AERD in animal models. The reproducibility of VCS DCE-MRI based glomerular filtration rate (GFR) estimation was first evaluated using the three healthy rabbits. Animal models of unilateral AERD were then conducted. All the rabbits underwent VCS DCE-MRI and the GFR maps were estimated by a commonly used cortical-compartment model. The whole kidney and suspicious lesion region GFR values of embolized kidneys were then compared with the corresponding values of normal kidneys. Finally, the suspicious lesion regions were confirmed by the corresponding renal specimens and histological findings. The reproducibility of GFR measurements was analyzed using the coefficient of variation and Bland-Altman analysis. The GFR values of normal and embolized kidneys were compared using the Student t-test. Contrast-enhanced images with sufficient diagnostic quality and reduced motion artifacts are obtained at a temporal resolution of 2.5 s. The Bland-Altman plot indicated close agreement between the GFR values estimated from between-day scans in healthy rabbits. Besides, there existed significant differences between the pixel-wise GFR values of normal and AERD kidneys in region-based comparison(P < 0.0001). The suspicious lesions are consistent well with the renal specimen and histological findings. The preliminary animal study verified the feasibility of VCS DCE-MRI for renal function evaluation, and the strategy could potentially provide a valuable tool to identify AERD.

Model-Based High-Definition Dynamic Contrast Enhanced MRI for Concurrent Estimation of Perfusion and Microvascular Permeability

This work introduces a model-based, high-definition dynamic contrast enhanced (DCE) MRI for concurrent estimation of perfusion and microvascular permeability over the whole brain. A time series of reference-subtracted signals is decomposed into one component that reflects main contrast dynamics and the other one that includes residual contrast agents (CA) and background signals. The former is described by linear superposition of a finite number of basic vectors trained from an augmented set of data that consists of tracer-kinetic model driven signal vectors and patient-specific measured ones. Contrast dynamics is estimated by solving a constrained optimization problem that incorporates the linearized signal decomposition into the measurement model of DCE MRI and then combining the main component with the background-suppressed, residual CA signals. To the best of our knowledge, this is the first work that prospectively enables rapid temporal sampling with 1.5 s (3 ∼ 4 times higher than clinical routines) while simultaneously achieving high isotropic spatial resolution with 1.0 mm³ (4 ∼ 6 times higher than routines), enhancing estimation of both patient-specific inputs and outputs for quantification of microvascular functions. Simulations and experiments are performed to demonstrate the effectiveness of the proposed method in patients with brain cancer.

Accelerated mono- and biexponential 3D-T1ρ relaxation mapping of knee cartilage using golden angle radial acquisitions and compressed sensing

PURPOSE

To use golden-angle radial sampling and compressed sensing (CS) for accelerating mono- and biexponential 3D spin-lattice relaxation time in the rotating frame (T_1ρ ) mapping of knee cartilage.

METHODS

Golden-angle radial stack-of-stars and Cartesian 3D-T_1ρ -weighted knee cartilage datasets (n = 12) were retrospectively undersampled by acceleration factors (AFs) 2-10. CS-based reconstruction using 8 different sparsifying transforms were compared for mono- and biexponential T_1ρ -mapping of knee cartilage, including spatio-temporal finite differences, wavelets, dictionary from principal component analysis, and exponential decay models, and also low rank and low rank plus sparse models (L+S). Complex-valued fitting was used and Marchenko-Pastur principal component analysis filtering also tested.

RESULTS

Most CS methods performed well for an AF of 2, with relative median normalized absolute deviation below 10% for monoexponential and biexponential mapping. For monoexponential mapping, radial sampling obtained a median normalized absolute deviation below 10% up to AF of 10, while Cartesian obtained this level of error only up to AF of 4. Radial sampling was also better with biexponential T_1ρ mapping, with median normalized absolute deviation below 10% up to AF of 6.

CONCLUSION

Golden-angle radial acquisitions combined with CS outperformed Cartesian acquisitions for 3D-T_1ρ mapping of knee cartilage, being it is a good alternative to Cartesian sampling for reducing scan time and/or improving image and mapping quality. The methods exponential decay models, spatio-temporal finite differences, and low rank obtained the best results for radial sampling patterns.

A Convex Optimization Algorithm for Compressed Sensing in a Complex Domain: The Complex-Valued Split Bregman Method

The Split Bregman method (SBM), a popular and universal CS reconstruction algorithm for inverse problems with both l₁-norm and TV-norm regularization, has been extensively applied in complex domains through the complex-to-real transforming technique, e.g., MRI imaging and radar. However, SBM still has great potential in complex applications due to the following two points; Bregman Iteration (BI), employed in SBM, may not make good use of the phase information for complex variables. In addition, the converting technique may consume more time. To address that, this paper presents the complex-valued Split Bregman method (CV-SBM), which theoretically generalizes the original SBM into the complex domain. The complex-valued Bregman distance (CV-BD) is first defined by replacing the corresponding regularization in the inverse problem. Then, we propose the complex-valued Bregman Iteration (CV-BI) to solve this new problem. How well-defined and the convergence of CV-BI are analyzed in detail according to the complex-valued calculation rules and optimization theory. These properties prove that CV-BI is able to solve inverse problems if the regularization is convex. Nevertheless, CV-BI needs the help of other algorithms for various kinds of regularization. To avoid the dependence on extra algorithms and simplify the iteration process simultaneously, we adopt the variable separation technique and propose CV-SBM for resolving convex inverse problems. Simulation results on complex-valued l₁-norm problems illustrate the effectiveness of the proposed CV-SBM. CV-SBM exhibits remarkable superiority compared with SBM in the complex-to-real transforming technique. Specifically, in the case of large signal scale n = 512, CV-SBM yields 18.2%, 17.6%, and 26.7% lower mean square error (MSE) as well as takes 28.8%, 25.6%, and 23.6% less time cost than the original SBM in 10 dB, 15 dB, and 20 dB SNR situations, respectively.

Non-Cartesian slice-GRAPPA and slice-SPIRiT reconstruction methods for multiband spiral cardiac MRI

PURPOSE

Spiral MRI has advantages for cardiac imaging, and multiband (MB) spiral MRI of the heart shows promise. However, current reconstruction methods for MB spiral imaging have limitations. We sought to develop improved reconstruction methods for MB spiral cardiac MRI.

METHODS

Two reconstruction methods were developed. The first is non-Cartesian slice-GRAPPA (NCSG), which uses phase demodulation and gridding operations before application of a Cartesian slice-separating kernel. The second method, slice-SPIRiT, formulates the reconstruction as a minimization problem that enforces in-plane coil consistency and consistency with the acquired MB data, and uses through-plane coil sensitivity information in the iterative solution. These methods were compared with conjugate-gradient SENSE in phantoms and volunteers. Temporal alternation of CAIPIRINHA (controlled aliasing in parallel imaging results in higher acceleration) phase and the use of a temporal filter were also investigated.

RESULTS

Phantom experiments with 3 simultaneous slices (MB = 3) showed that mean artifact power was highest for conjugate-gradient SENSE, lower for NCSG, and lowest for slice-SPIRiT. For volunteer cine imaging (MB = 3, N = 5), the artifact power was 0.182 ± 0.037, 0.148 ± 0.036, and 0.139 ± 0.034 for conjugate-gradient SENSE, NCSG, and slice-SPIRiT, respectively (P < .05, analysis of variance). Temporal alternation of CAIPIRINHA reduced artifacts for both NCSG and slice-SPIRiT.

CONCLUSION

The NCSG and slice-SPIRiT methods provide more accurate reconstructions for MB spiral cine imaging compared with conjugate-gradient SENSE. These methods hold promise for non-Cartesian MB imaging.

Highly accelerated vessel-selective arterial spin labeling angiography using sparsity and smoothness constraints

Purpose

To demonstrate that vessel selectivity in dynamic arterial spin labeling angiography can be achieved without any scan‐time penalty or noticeable loss of image quality compared with conventional arterial spin labeling angiography.

Methods

Simulations on a numerical phantom were used to assess whether the increased sparsity of vessel‐encoded angiograms compared with non‐vessel‐encoded angiograms alone can improve reconstruction results in a compressed‐sensing framework. Further simulations were performed to study whether the difference in relative sparsity between nonselective and vessel‐selective dynamic angiograms was sufficient to achieve similar image quality at matched scan times in the presence of noise. Finally, data were acquired from 5 healthy volunteers to validate the technique in vivo. All data, both simulated and in vivo, were sampled in 2D using a golden‐angle radial trajectory and reconstructed by enforcing image domain sparsity and temporal smoothness on the angiograms in a parallel imaging and compressed‐sensing framework.

Results

Relative sparsity was established as a primary factor governing the reconstruction fidelity. Using the proposed reconstruction scheme, differences between vessel‐selective and nonselective angiography were negligible compared with the dominant factor of total scan time in both simulations and in vivo experiments at acceleration factors up to R = 34. The reconstruction quality was not heavily dependent on hand‐tuning the parameters of the reconstruction.

Conclusion

The increase in relative sparsity of vessel‐selective angiograms compared with nonselective angiograms can be leveraged to achieve higher acceleration without loss of image quality, resulting in the acquisition of vessel‐selective information at no scan‐time cost.

Inversion recovery UTE based volumetric myelin imaging in human brain using interleaved hybrid encoding

PURPOSE

Direct myelin imaging can improve the characterization of myelin-related diseases such as multiple sclerosis. In this study, we explore a novel method to directly image myelin using inversion recovery-prepared hybrid encoding (IR-HE) UTE MRI.

METHODS

The IR-HE sequence uses an adiabatic inversion pulse to suppress the long T₂ white matter signal, followed by 3D dual-echo HE utilizing both single point imaging and radial frequency encoding, for which the subtraction image between 2 echoes reveals the myelin signal with high contrast. To reduce scan time, it is common to obtain multiple spokes per IR. Here, we invented a novel method to improve the HE, adapted for the multi-spoke IR imaging-termed interleaved HE-for which single point imaging encoding is interleaved between radial frequency encodings near nulling point to allow more efficient IR-signal suppression. To evaluate the proposed approach, a computer simulation, myelin phantom experiment, an ex vivo experiment with a cadaveric multiple sclerosis brain, and an in vivo experiment with 8 healthy volunteers and 13 multiple sclerosis patients were performed.

RESULTS

The computer simulation showed that IR-interleaved HE allows for improved contrast of myelin signal with reduced imaging artifacts. The myelin phantom experiment showed IR-interleaved HE allows direct imaging of myelin lipid with excellent suppression of water signal. In the ex vivo and in vivo experiments, the proposed method demonstrated highly specific imaging of myelin in white matter of the brain.

CONCLUSION

IR-interleaved HE allows for time-efficient, high-contrast direct myelin imaging and can detect demyelinated lesions in multiple sclerosis patients.

IPEM Topical Report: A 2018 IPEM survey of MRI use for external beam radiotherapy treatment planning in the UK

The benefits of integrating MRI into the radiotherapy pathway are well published, however there is little consensus in guidance on how to commission or implement its use. With a view to developing consensus guidelines for the use of MRI in external beam radiotherapy (EBRT) treatment planning in the UK, a survey was undertaken by an Institute of Physics and Engineering in Medicine (IPEM) working-party to assess the current landscape of MRI use in EBRT in the UK. A multi-disciplinary working-party developed a survey to understand current practice using MRI for EBRT treatment planning; investigate how MRI is currently used and managed; and identify knowledge gaps. The survey was distributed electronically to radiotherapy service managers and physics leads in 71 UK radiotherapy (RT) departments (all NHS and private groups). The survey response rate was 87% overall, with 89% of NHS and 75% of private centres responding. All responding centres include EBRT in some RT pathways: 94% using Picture Archiving and Communication System (PACS) images potentially acquired without any input from RT departments, and 69% had some form of MRI access for planning EBRT. Most centres reporting direct access use a radiology scanner within the same hospital in dedicated (26%) or non-dedicated (52%) RT scanning sessions. Only two centres reported having dedicated RT MRI scanners in the UK, lower than reported in other countries. Six percent of radiotherapy patients in England (data not publically available outside of England) have MRI as part of their treatment, which again is lower than reported elsewhere. Although a substantial number of centres acquire MRI scans for treatment planning purposes, most centres acquire less than five patient scans per month for each treatment site. Commissioning and quality assurance of both image registration and MRI scanners was found to be variable across the UK. In addition, staffing models and training given to different staff groups varied considerably across the UK, reflecting the current lack of national guidelines. The primary barriers reported to MRI implementation in EBRT planning included costs (e.g. lack of a national tariff for planning MRI), lack of MRI access and/or capacity within hospitals. Despite these challenges, significant interest remains in increasing MRI-assisted EBRT planning over the next five years.