Search

Cerebral microbleeds (CMBs) are associated with white matter damage, various neu-rodegenerative and cerebrovascular diseases. CMBs occur as small, circular hypointense lesions on T2*-weighted gradient recalled echo (GRE) and susceptibility weighted imaging (SWI) images, and hyperintense on quantitative susceptibility mapping (QSM) images due to their paramagnetic nature. Accurate detection of CMBs would help to determine the CMB lesion count and distribution, which would be further useful to understand the clinical impact of CMBs and to obtain quantitative imaging biomarkers. In this work, we propose a fully automated, deep learning-based, 2-step algorithm, using structural and anatomical properties of CMBs from any single input image modality (e.g. GRE / SWI / QSM) for their accurate detection. Our method consists of an initial candidate detection step, that detects CMBs with high sensitivity and a candidate discrimination step using a knowledge distillation framework to classify CMB and non-CMB instances, followed by a morphological clean-up step. We used 4 datasets consisting of different modalities specified above, acquired using various protocols and with a variety of pathological and demographic characteristics. On cross-validation within datasets, our method achieved a cluster-wise true positive rate (TPR) over 90% with an average of less than 2 false positives per subject. Our method is flexible in terms of the input modality and provides comparable cluster-wise TPR and a better cluster-wise precision compared to existing state-of-the-art methods. When evaluated across different datasets, our method showed good generalisability with a cluster-wise TPR greater than 80% with different modalities.

Cerebral microbleeds (CMBs) appear as small, circular, well defined hypointense lesions of a few mm in size on T2*-weighted gradient recalled echo (T2*-GRE) images and appear enhanced on susceptibility weighted images (SWI). Due to their small size, contrast variations and other mimics (e.g. blood vessels), CMBs are highly challenging to detect automatically. In large datasets (e.g. the UK Biobank dataset), exhaustively labelling CMBs manually is difficult and time consuming. Hence it would be useful to preselect candidate CMB subjects in order to focus on those for manual labelling, which is essential for training and testing automated CMB detection tools on these datasets. In this work, we aim to detect CMB candidate subjects from a larger dataset, UK Biobank, using a machine learning-based, computationally light pipeline. For our evaluation, we used 3 different datasets, with different intensity characteristics, acquired with different scanners. They include the UK Biobank dataset and two clinical datasets with different pathological conditions. We developed and evaluated our pipelines on different types of images, consisting of SWI or GRE images. We also used the UK Biobank dataset to compare our approach with alternative CMB preselection methods using non-imaging factors and/or imaging data. Finally, we evaluated the pipeline’s generalisability across datasets. Our method provided subject-level detection accuracy > 80% on all the datasets (withindataset results), and showed good generalisability across datasets, providing a consistent accuracy of over 80%, even when evaluated across different modalities.

INTRODUCTION AND OBJECTIVE:We previously reported on the single slice T1 mapping of human bladder using a pair of spoiled gradient recalled echo (SPGR) images acquired at different flip angles. Her...

Background/Aims:In the present study, dural venous sinuses and cerebral veins were examined using 3D SPGR MRV and 2D TOF MRV. These methods were compared in terms of detectability of venous structures and their diameters. Methods:A total of 110 patients(66 female and 44 male) who had contrast-enhanced 3D SPGR MRV and 2D TOF MRV examinations using a 1.5 T MRI machine in May 2008-June 2011 period were included in the present study. Diameters of dural venous sinuses and veins were measured at three different planes at a position 1cm distal to the site they drained into, and average values were used. Both MRV methods were compared to reveal whether the diameters calculated were different in age groups and between genders.Results:SSS, bilateral TS’s, right sigmoid sinus, Galen vein and bilateral ICV’s all could be determined using 3D SPGR MRV. Not all dural venous sinuses and veins other than right sigmoid sinus could be visualized in all patients using 2D TOF MRV.There were significant differences between the two examinations for SSS and ISS, bilateral TS and sigmoid sinuses, SS, bilateral Labbe and ICV and Galen vein diameters(p

Purpose To propose and validate an efficient method, based on a biophysically motivated signal model, for removing the orientation-dependent part of R 2 * using a single gradient-recalled echo (GRE) measurement. Methods The proposed method utilized a temporal second-order approximation of the hollow-cylinder-fiber model, in which the parameter describing the linear signal decay corresponded to the orientation-independent part of R 2 * . The estimated parameters were compared to the classical, mono-exponential decay model for R 2 * in a sample of an ex vivo human optic chiasm (OC). The OC was measured at 16 distinct orientations relative to the external magnetic field using GRE at 7T. To show that the proposed signal model can remove the orientation dependence of R 2 * , it was compared to the established phenomenological method for separating R 2 * into orientation-dependent and -independent parts. Results Using the phenomenological method on the classical signal model, the well-known separation of R 2 * into orientation-dependent and -independent parts was verified. For the proposed model, no significant orientation dependence in the linear signal decay parameter was observed. Conclusions Since the proposed second-order model features orientation-dependent and -independent components at distinct temporal orders, it can be used to remove the orientation dependence of R 2 * using only a single GRE measurement.

Background Previous quantitative synthetic MRI of the brain has been solely performed in 2D. Purpose To evaluate the feasibility of the recently developed sequence 3D-QALAS for brain cortical thickness and volumetric analysis. Study type Reproducibility/repeatability study. Subjects Twenty-one healthy volunteers (35.6 ± 13.8 years). Field strength/sequence 3D T1 -weighted fast spoiled gradient recalled echo (FSPGR) sequence was performed once, and 3D-QALAS sequence was performed twice with a 3T scanner. Assessment FreeSurfer and FIRST were used to measure cortical thickness and volume of subcortical structures, respectively. Agreement with FSPGR and scan-rescan repeatability were evaluated for 3D-QALAS. Statistical tests Percent relative difference and intraclass correlation coefficient (ICC) were used to assess reproducibility and scan-rescan repeatability of the 3D-QALAS sequence-derived measurements. Results Percent relative difference compared with FSPGR in cortical thickness of the whole cortex was 3.1%, and 89% of the regional areas showed less than 10% relative difference in cortical thickness. The mean ICC across all regions was 0.65, and 74% of the structures showed substantial to almost perfect agreement. For volumes of subcortical structures, the median percent relative differences were lower than 10% across all subcortical structures, except for the accumbens area, and all structures showed ICCs of substantial to almost perfect agreement. For the scan-rescan test, percent relative difference in cortical thickness of the whole cortex was 2.3%, and 97% of the regional areas showed less than 10% relative difference in cortical thickness. The mean ICC across all regions was 0.73, and 80% showed substantial to almost perfect agreement. For volumes of subcortical structures, relative differences were less than 10% across all subcortical structures except for the accumbens area, and all structures showed ICCs of substantial to almost perfect agreement. Data conclusion 3D-QALAS could be reliably used for measuring cortical thickness and subcortical volumes in most brain regions. Level of evidence 3 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:1834-1842.

PURPOSE: To compare 2D gradient-recalled echo (GRE) and 2D spin-echo (SE) echo-planar imaging (EPI) MR elastography (MRE) for measurement of hepatic stiffness in adult patients with known or suspected liver disease at 3 Tesla. MATERIALS AND METHODS: Three hundred and eighty-seven consecutive patients underwent MRE of the liver at 3 Tesla with 2D-GRE and 2D-SE-EPI sequences. ‘Mean liver stiffness (LS)’ calculated by averaging 3 ROIs in the right lobe, ‘Maximum LS’ calculated by an ROI in the right lobe; and ‘Freehand LS’ calculated by an ROI in the entire liver were measured by two independent readers. Inter-observer and inter-class variability in stiffness measurements were assessed. Stiffness values were correlated with degree of liver fibrosis (METAVIR scores) in 97 patients who underwent biopsy. The diagnostic performance was compared by a receiver-operating characteristic analysis. RESULTS: The technical failure rate was 2.8% for 2D-SE-EPI (11/387) and 4.1% for 2D-GRE (16/387, 9 had R(2)* > 80 s(−1) indicating iron overload). There is high reproducibility for both GRE and SE-EPI variants (ICC = 0.84–0.94 for both GRE and SE-EPI MRE). The highest sensitivity, specificity, and accuracy of differentiating mild fibrosis (F0–F2) from advanced fibrosis (F3–F4) are 0.84 (GRE Freehand measurement), 0.92 (GRE Maximum stiffness measurement), and 0.88 (GRE Freehand measurement), respectively. CONCLUSIONS: High intra-class correlation and intra-reader correlation are seen on measured hepatic stiffness for both 2D-GRE and 2D-SE-EPI MRE. 2D-SE-EPI has lower failure rate. Diagnostic performance of both sequences is equivalent, with highest sensitivity for 2D-GRE Freehand stiffness measurement, and highest specificity 2D-GRE Maximum stiffness measurement.

Non-invasive evaluation of liver fibrosis has evolved over the last couple of decades. Currently, elastography techniques are the most widely used non-invasive methods for clinical evaluation of chronic liver disease (CLD). MR elastography (MRE) of the liver has been used in the clinical practice for nearly a decade and continues to be widely accepted for detection and staging of liver fibrosis. With MRE, one can directly visualize propagating shear waves through the liver and an inversion algorithm in the scanner automatically converts the shear wave properties into an elastogram (stiffness map) on which liver stiffness can be calculated. The commonly used MRE method, two-dimensional gradient recalled echo (2D-GRE) sequence has produced excellent results in the evaluation of liver fibrosis in CLD from various etiologies and newer clinical indications continue to emerge. Advances in MRE technique, including 3D MRE, automated liver elasticity calculation, improvements in shear wave delivery and patient experience, are promising to provide a faster and more reliable MRE of liver. Innovations, including evaluation of mechanical parameters, such as loss modulus, displacement, and volumetric strain, are promising for comprehensive evaluation of CLD as well as understanding pathophysiology, and in differentiating various etiologies of CLD. In this review, the current status of the MRE of liver in CLD are outlined and followed by a brief description of advanced techniques and innovations in MRE of liver.

Purpose To extend the null signal method (NSM) for B1 mapping to 3 T magnetic resonance imaging (MRI). Background The NSM operates in the steady state regime and exploits the linearity of the spoiled gradient recalled echo (SPGR) signal around the 180° flip angle (FA). Using linear regression, B1 maps are derived from three SPGR images acquired at different FAs with a short repetition time. While the conventional NSM allows accurate mapping of B1 for moderate B1 variation, we observed that this method fails for the larger B1 variations typical of high-field MRI. Methods We analyzed the effect of the FA range of the acquired SPGR images on B1 determination using the NSM for 3 T MRI through extensive numerical and in vivo analyses. B1 maps derived from the extended angle-range NSM (EA-NSM) were calculated and compared to those derived from the conventional, more restricted angle range, NSM, and to those derived from the reference, but much more time-consuming, double angle method (DAM). Furthermore, we investigated the compatibility of EA-NSM B1 mapping and the half-scan and SENSE reconstruction methods for accelerating acquisition time. Results Our results show that the use of the conventional FA range leads to substantial inaccuracies in B1 determination. Both numerical and in vivo analyses demonstrate that expanding the FA range of the acquired SPGR images substantially improves the accuracy of B1 maps. Furthermore, B1 maps derived from EA-NSM were demonstrated to be quantitatively comparable to those derived from the lengthy DAM protocol. We also found that B1 maps derived from SPGR images using the EA-NSM and imaging acceleration methods were comparable to those derived from images acquired without acceleration. Finally, the use of half scanning combined with SENSE reconstruction permits whole-brain B1 mapping in ~1 min. Conclusions The EA-NSM permits accurate, fast, and practical B1 mapping in a 3 T clinical setting.

Introduction: Cerebral Microbleeds (CMBs) are increasingly recognised as an important predictor of small vessel disease and cannot be detected on routine Computed Tomography (CT) or Magnetic Resonance Imaging (MRI) sequences. Aim: CMBs on MRI and to compare the sensitivity of Susceptibility Weighted Imaging (SWI) and T2* Gradient Recalled Echo (GRE) sequences in detection of these microbleeds. Also, to evaluate association between topography (size, number and location) of CMBs and various risk factors. Materials and Methods: This Prospective Observational analytical study comprised of 67 patients with CMBs who were evaluated with Siemens 1.5 Tesla MRI machine. T2*-weighted GRE sequence and SWI were taken in all the patients in addition to routine MRI sequences. Two observers blinded to clinical information, independently interpreted the T2*GRE and SWI sequences. Both observers recorded the topographical details of CMBs which included presence, number, size and location and their association with various risk factors (hypertension, diabetes, smoking and hyperlipidemia). Inter-rater agreement for the number of CMBs on SWI and T2*GRE was calculated using Cohen’s Kappa (κ) as measure of agreement. Intra-Class Correlation Coefficient (ICC) for reliability index in test-retest, intra-rater and inter-rater reliability was also analysed. The intra-rater reliabilities of each rater were compared. The correlations between CMBs and risk factors was performed with Pearson’s correlation Coefficient (r). The p-value

BACKGROUND AND PURPOSE: SWI is valuable for characterization of intracranial hemorrhage and mineralization but has long acquisition times. We compared a highly accelerated wave–controlled aliasing in parallel imaging (CAIPI) SWI sequence with 2 commonly used alternatives, standard SWI and T2*-weighted gradient recalled-echo (T2*W GRE), for routine clinical brain imaging at 3T. MATERIALS AND METHODS: A total of 246 consecutive adult patients were prospectively evaluated using a conventional SWI or T2*W GRE sequence and an optimized wave–CAIPI SWI sequence, which was 3–5 times faster than the standard sequence. Two blinded radiologists scored each sequence for the presence of hemorrhage, the number of microhemorrhages, and severity of motion artifacts. Wave–CAIPI SWI was then evaluated in head-to-head comparison with the conventional sequences for visualization of pathology, artifacts, and overall diagnostic quality. Forced-choice comparisons were used for all scores. Wave–CAIPI SWI was tested for superiority relative to T2*W GRE and for noninferiority relative to standard SWI using a 15% noninferiority margin. RESULTS: Compared with T2*W GRE, wave–CAIPI SWI detected hemorrhages in more cases (P

Purpose The 4th International Workshop on MRI Phase Contrast and QSM (2016, Graz, Austria) hosted the first QSM Challenge. A single-orientation gradient recalled echo acquisition was provided, along with COSMOS and the χ33 STI component as ground truths. The submitted solutions differed more than expected depending on the error metric used for optimization and were generally over-regularized. This raised (unanswered) questions about the ground truths and the metrics utilized. Methods We investigated the influence of background field remnants by applying additional filters. We also estimated the anisotropic contributions from the STI tensor to the apparent susceptibility to amend the χ33 ground truth and to investigate the impact on the reconstructions. Lastly, we used forward simulations from the COSMOS reconstruction to investigate the impact noise had on the metric scores. Results Reconstructions compared against the amended STI ground truth returned lower errors. We show that the background field remnants had a minor impact in the errors. In the absence of inconsistencies, all metrics converged to the same regularization weights, whereas structural similarity index metric was more insensitive to such inconsistencies. Conclusion There was a mismatch between the provided data and the ground truths due to the presence of unaccounted anisotropic susceptibility contributions and noise. Given the lack of reliable ground truths when using in vivo acquisitions, simulations are suggested for future QSM Challenges.

Different trends of echo time dependent gradient recalled echo MRI signals in different brain regions have been attributed to signal compartments in image voxels. It remains unclear how variations in gradient recalled echo MRI signals change as a function of MRI field strength, and how data processing may impact signal compartment parameters. We used two popular quantitative susceptibility mapping methods of processing raw phase images (Laplacian and path-based unwrapping with V-SHARP) and expressed values in the form of induced frequency shifts (in Hz) in six specific brain regions at 3T and 7T. We found the frequency shift curves to vary with echo time, and a good overlap between 3T and 7T mean frequency shift curves was present. However, the amount of variation across participants was greater at 3T, and we were able to obtain better compartment model fits of the signal at 7T. We also found the temporal trends in the signal and compartment frequency shifts to change with the method used to process images. The inter-participant averaged trends were consistent between 3T and 7T for each quantitative susceptibility pipeline. However, signal compartment frequency shifts generated using different pipelines may not be comparable.

Purpose Hyperpolarized (HP) gas MRI of the rodent lung is of great interest because of the increasing need for novel biomarkers with which to develop new therapies for respiratory diseases. The use of fast gradient-recalled echo (FGRE) for high-resolution HP gas rodent lung MRI is challenging as a result of signal loss caused by significant diffusion weighting, particularly in the larger airways. In this work, a modified FGRE approach is described for HP 3He rodent lung MRI using a centric-out readout scheme (ie, x-centric), allowing high-resolution, density-weighted imaging. Methods HP 3He x-centric imaging was performed in a phantom and compared with a conventional partial-echo FGRE acquisition for in-plane spatial resolutions varying between 39 and 312 µm. Partial-echo and x-centric acquisitions were also compared for high spatial-resolution breath-hold (1 s) imaging of rodent lungs. Results X-centric provided improved signal-to-noise ratio efficiency by a factor of up to 13/1.7 and 6.7/1.8, compared with the partial-echo FGRE for the airways/parenchyma of mouse and rat, respectively, at high spatial resolutions in vivo (

IntroductionWhile image coregistration is inevitable and quantitative parametric maps, such as R2* and QSM, are increasingly used in multi-parametric studies of the brain, there is a lack of investigations on the reliability of quantitative metrics after coregistration. The purposes of this study were 1) to evaluate the reliability of R2* and QSM at 7T and 2) to assess the statistical agreement in the quantitative metrics obtained by two different coregistration approaches.MethodsWe compared the reliability of R2* and quantitative susceptibility maps obtained from brains of eight healthy participants by two coregistration approaches: 1) transformation of pre-processed quantitative maps and 2) processing quantitative maps after transformation using pixel- and ROI-based analyses. Two-sample Kolmogorov-Smirnov, Mann-Whitney U, Paired T, Intraclass-correlation tests were performed appropriately.ResultsR2* remained invariant regardless of the cogeneration timing. However, magnetic susceptibility was significantly altered when processed in transformed space, whereas it remained invariant in all KS-tests and pixel values were only different in 2 out of 64 U-tests for direct QSM coregistration. Paired t-test revealed that ROI-based group-mean R2* was invariant to both approaches, while group-mean susceptibility was invariant to direct coregistration but differed in one structure processed in transformed space. For all pairs of measurements of R2*, ICCs were excellent. ICCs for magnetic susceptibility were excellent when processed in its native space while the ICCs were lower than 0.9 or poor when processed in transformed space. Further analysis revealed that the choice of interpolation approach affected the resultant QSM.ConclusionsOur study shows that R2* could be safely processed in a transformed space, whereas QSM was less reliable when processed in the transformed space. Hence, caution is advised when using QSM in a multi-parametric study, and it is strongly recommended to process QSM in its native space prior to any coregistration or spatial transformation.

The centromedian (CM) nucleus is an intralaminar thalamic nucleus that is considered as a potentially effective target of deep brain stimulation (DBS) and ablative surgeries for the treatment of multiple neurological and psychiatric disorders. However, the structure of CM is invisible on the standard T1- and T2-weighted (T1w and T2w) magnetic resonance images, which hamper it as a direct DBS target for clinical applications. The purpose of the current study is to demonstrate the use of quantitative susceptibility mapping (QSM) technique to image the CM within the thalamic region. Twelve patients with Parkinson's disease, dystonia, or schizophrenia were included in this study. A 3D multi-echo gradient recalled echo (GRE) sequence was acquired together with T1w and T2w images on a 3-T MR scanner. The QSM image was reconstructed from the GRE phase data. Direct visual inspection of the CM was made on T1w, T2w, and QSM images. Furthermore, the contrast-to-noise ratios (CNRs) of the CM to the adjacent posterior part of thalamus on T1w, T2w, and QSM images were compared using the one-way analysis of variance (ANOVA) test. QSM dramatically improved the visualization of the CM nucleus. Clear delineation of CM compared to the surroundings was observed on QSM but not on T1w and T2w images. Statistical analysis showed that the CNR on QSM was significantly higher than those on T1w and T2w images. Taken together, our results indicate that QSM is a promising technique for improving the visualization of CM as a direct targeting for DBS surgery.

Hemorrhagic transformation (HT) is one of the most devastating complications of reperfusion therapy in acute ischemic stroke. Prediction of an upcoming HT remains beyond current techniques in routine clinical practice. If made available, such information would benefit the management of acute ischemic stroke patients and help to tailor therapeutic strategies. This study aims at providing a machine learning framework for predicting occurrence and extent of HT from source perfusion-weighted magnetic resonance imaging (PWI) combined with diffusion weighted imaging (DWI). The model relies on a LSTM network based on PWI combined with DWI imaging features into a fully connected neural network. A retrospective comparative analysis performed on 155 acute stroke patients demonstrate the efficacy of the LSTM model (AUC-ROC: \(89.4\%\)) against state-of-the-art machine learning models. Predicted likelihood of HT at the voxel level was evaluated against HT annotations of stroke neurologists obtained from follow-up gradient recalled echo (GRE) imaging.

The objective of our study was to compare standard breath-held, free-breathing, and compressed sensing (CS) accelerated 2D gradient-recalled echo (GRE) MR elastography (MRE) techniques for measuring liver stiffness.Twenty-five adult volunteers (20 without liver disease and five with liver disease) underwent MRE with five different 2D GRE techniques: breath-held; free-breathing; and CS with acceleration factors of 1.5, 2, and 3. Four axial images were obtained through the mid liver with each technique, and liver stiffness measurements (in kilopascals) were made by three blinded readers using elastograms with 95% confidence maps. Liver stiffness measurements were compared between techniques using repeated-measures ANOVA. Absolute agreement between techniques was evaluated for each reader using single-measure intraclass correlation coefficients (ICCs). Bland-Altman analyses were performed to assess bias between techniques using breath-held MRE as the reference.Individual subject mean liver stiffness values ranged from 1.35 to 5.25 kPa for the population. There was no significant difference in mean liver stiffness values between MRE techniques for readers 1 and 3 (p0.05). A significant difference (p = 0.02) due to higher stiffness measurements using free-breathing MRE was observed for reader 2. There was excellent absolute agreement between MRE techniques for each reader (all ICCs0.940). Bias between techniques ranged from -0.102 to 0.089 kPa for reader 1, -0.119 to 0.121 kPa for reader 2, and -0.074 to 0.085 kPa for reader 3.Free-breathing and CS accelerated 2D GRE MRE techniques yield similar results to the conventional breath-held technique with only slight bias and may be useful in pediatric and adult patients with limited ability to breath-hold.

The first clinical application of magnetic resonance elastography (MRE) was in the evaluation of chronic liver disease (CLD) for detection and staging of liver fibrosis. In the past 10 years, MRE has been incorporated seamlessly into a standard magnetic resonance imaging (MRI) liver protocol worldwide. Liver MRE is a robust technique for evaluation of liver stiffness and is currently the most accurate noninvasive imaging technology for evaluation of liver fibrosis. Newer MRE sequences including spin-echo MRE and 3 dimensional MRE have helped in reducing the technical limitations of clinical liver MRE that is performed with 2D gradient recalled echo (GRE) MRE. Advances in MRE technology have led to understanding of newer mechanical parameters such as dispersion, attenuation, and viscoelasticity that may be useful in evaluating pathological processes in CLD and may prove useful in their management.This review article will describe the changes in CLD that cause an increase in stiffness followed by principle and technique of liver MRE. In the later part of the review, we will briefly discuss the advances in liver MRE.

The purpose of this study is 1) to demonstrate reproducibility of spin echo-echo planar imaging (SE-EPI) magnetic resonance elastography (MRE) to estimate kidney stiffness; and 2) to compare SE-EPI MRE and gradient recalled echo (GRE) MRE-derived stiffness estimations in various anatomical regions of the kidney. Kidney MRE was performed on 33 healthy subjects (8 for SE-EPI MRE reproducibility and 25 for comparison with GRE MRE; age range: 22-66 years) in a 3 T MRI scanner. To demonstrate SE-EPI MRE reproducibility, subjects were scanned for the first scan and then asked to leave the scan room and repositioned again for the second (repeat) scan. Similar set-up was used for GRE MRE as well. The displacement data was then processed to obtain overall stiffness estimates of the kidney. Concordance correlation analyses were performed to determine SE-EPI MRE reproducibility and agreement between GRE MRE and SE-EPI MRE derived stiffness. A high concordance correlation (ρc = 0.95; p-value

PURPOSE To accurately separate water and fat signals for bipolar multi-echo gradient-recalled echo sequence using a convolutional neural network (CNN). METHODS A CNN architecture was designed and trained using the relationship between multi-echo images from the bipolar multi-echo gradient-recalled echo sequence and artifact-free water-fat-separated images. The artifact-free water-fat-separated images for training the CNN were obtained from multiple signals with different TEs by using iterative decomposition of water and fat with echo asymmetry and the least-squares estimation method, in which multiple signals at different TEs were acquired using a single-echo gradient-recalled echo sequence. We also proposed a data augmentation method using a synthetic field inhomogeneity to generate multi-echo signals, including various bipolar multi-echo gradient-recalled echo artifacts so that the CNN could prevent overfitting and increase the separation accuracy. We trained the CNN using in vivo knee images and tested it using in vivo knee, head, and ankle images. RESULTS In vivo imaging results showed that the proposed CNN could separate water-fat images accurately. Although the proposed CNN was trained using only in vivo knee images, the proposed CNN could also separate water-fat images of different imaging regions. The proposed data augmentation method could prevent overfitting even with a limited number of training data sets and make the method robust to magnetic field inhomogeneities. CONCLUSION The proposed CNN could obtain water-fat-separated images from the multi-echo images acquired from the bipolar multi-echo gradient-recalled echo sequence, which included artifacts from the bipolar gradients.

Magnetic resonance imaging (MRI) has been considered a valuable diagnostic tool for noninvasive imaging of the vasculature in children and adults for more than two decades. While a variety of non-contrast MRI methods have been described for imaging of both arteries and veins (e.g., time-of-flight, phase contrast, and balanced steady-state free precession imaging), contrast-enhanced magnetic resonance angiography/venography are the most commonly employed vascular imaging techniques due to their high spatial and contrast resolutions and general reliability. In this technical innovation article, we describe a novel 3-D respiratory-triggered gradient recalled echo Dixon-based MR angiography/MR venography technique that provides high-resolution anatomical imaging of the vasculature of the neck, body and extremities without the need for intravenous contrast material or breath-holding.

Gradient recalled echo (GRE) sequences are nowadays routinely used in clinical applications of magnetic resonance imaging (MRI). Contrast of GRE images is determined by several different physical mechanisms, like proton density, T2 relaxation time, gradients of magnetic field, and spatial heterogeneities of spin-spin interaction. Although the combined effect of all these parameters is of great value for physicians, it is possible that their separate mapping may reveal new features, that are masked in traditional GRE images. Separate mapping of physical parameters, contributing to contrast of GRE images, is called the parameterization, and has not been entirely solved by now. The present publication reports on development of the necessary tools for the solution. Finally, results obtained on volunteers are presented.

Magnetic resonance imaging (MRI) assessment for appendicitis is limited by exam time and patient cooperation. The radially sampled 3-dimensional (3-D) T1-weighted, gradient recalled echo sequence (radial GRE) is a free-breathing, motion robust sequence that may be useful in evaluating appendicitis in children. To compare the rate of detection of the normal appendix with contrast-enhanced radial GRE versus contrast-enhanced 3-D GRE and a multi-sequence study including contrast-enhanced radial GRE. This was a retrospective study of patients ages 7–18 years undergoing abdominal-pelvic contrast-enhanced MRI between Jan. 1, 2012, and April 1, 2016. Visualization of the appendix was assessed by consensus between two pediatric radiologists. The rate of detection of the appendix for each sequence and combination of sequences was compared using a McNemar test. The rate of detection of the normal appendix on contrast-enhanced radial GRE was significantly higher than on contrast-enhanced 3-D GRE (76% vs. 57.3%, P=0.003). The rate of detection of the normal appendix with multi-sequence MRI including contrast-enhanced radial GRE was significantly higher than on contrast-enhanced 3-D GRE (81.3% vs. 57%, P

To assess the utility of the radial reformation of three-dimensional fat-suppressed multi-echo gradient-recalled-echo (3D FS me-GRE) for evaluating acetabular labral injuries and femoroacetabular impingement (FAI). A total of 25 patients with suspected acetabular labral injuries were examined using 3D FS me-GRE and radial 2D T2*-weighted imaging (T2*WI) on a 3-T magnetic resonance imaging (MRI) scanner. The range of acetabular labral injuries was evaluated by radial reformation through the center of the acetabulum perpendicular to the plane across the entire acetabular rim (type 1 radial reformation) of 3D FS me-GRE and radial 2D T2*WI. To evaluate the FAI morphology, we performed radial reformation perpendicular to the central axis of the femoral head and neck (type 2 radial reformation) of 3D FS me-GRE. Acetabular labral injuries were identified in 23 patients, and no acetabular labral injury was seen in two patients on type 1 radial reformation of 3D FS me-GRE and radial 2D T2*WI. The diagnostic concordance rate for the range of acetabular labral injuries between the two imaging methods was 76.0%, and there was excellent agreement for the injured angles (r = 0.977, p

Purpose To analyze the effect of neglecting nonzero echo times (TEs) in the conventional model of multicomponent driven equilibrium single pulse observation of T1 and T2 (mcDESPOT). Theory and Methods Formulations of the two-component spoiled gradient recalled echo (SPGR) and balanced steady state free precession (bSSFP) models that incorporate nonzero TE effects are presented in the context of mcDESPOT and compared with the conventionally used SPGR and bSSFP models which ignore nonzero TEs. Relative errors in derived parameter estimates from conventional mcDESPOT, omitting TE effects, are assessed using simulations over a wide range of experimental and sample parameters. Results The neglect of nonzero TE leads to an overestimate of the SPGR signal and an underestimate of the bSSFP signal. These effects can introduce large errors in parameter estimates derived from conventional mcDESPOT under realistic imaging conditions. Conclusion SPGR and bSSFP signal models accounting for nonzero TE effects should be incorporated into quantitative mcDESPOT analyses. Magn Reson Med, 2015. Published 2015. This article is a U.S. Government work and is in the public domain in the USA

The T2*-weighted gradient recalled echo sequence is a sensitive means to detect blood degradation products. While not a routine sequence in magnetic resonance imaging of the spine in small animals, it can provide additional valuable information in select cases. The goal of this retrospective, cross-sectional study was to describe findings when acquiring this sequence during magnetic resonance imaging examination of the spine in small animals. The University of Tennessee's veterinary radiology database was searched for dogs and cats that underwent magnetic resonance imaging for suspect spinal disease in which a T2*-weighted gradient recalled echo sequence was acquired and susceptibility artifact was identified. The following information was recorded: signalment, clinical signs, location and appearance of susceptibility artifact, and final diagnosis. Thirty-nine cases were included in the study. Extradural susceptibility artifacts were observed in cases of intervertebral disc herniation with or without associated hemorrhage (n = 28), extradural hemorrhage associated with spinal trauma (n = 2), hemophilia (n = 1), and in a cystic extradural mass (n = 1). Remaining lesions displaying susceptibility artifact were intramedullary and included presumptive acute noncompressive nucleus pulposus extrusion (n = 2), hematoma (n = 1), hemangiosarcoma metastasis (n = 1), intramedullary disc extrusion (n = 1), presumptive meningomyelitis (n = 1), and a mass of undetermined etiology (n = 1). Inclusion of a T2*-weighted gradient recalled echo sequence may be helpful in spinal magnetic resonance imaging when standard imaging sequences are ambiguous or intramedullary lesions are observed.

Background Susceptibility-weighted imaging (SWI) in neuroimaging can be challenging due to long scan times of three-dimensional (3D) gradient recalled echo (GRE), while faster techniques such as 3D interleaved echo-planar imaging (iEPI) are prone to motion artifacts. Here we outline and implement a 3D short-axis propeller echo-planar imaging (SAP-EPI) trajectory as a faster, motion-correctable approach for SWI. Methods Experiments were conducted on a 3T MRI system. The 3D SAP-EPI, 3D iEPI, and 3D GRE SWI scans were acquired on two volunteers. Controlled motion experiments were conducted to test the motion-correction capability of 3D SAP-EPI. The 3D SAP-EPI SWI data were acquired on two pediatric patients as a potential alternative to 2D GRE used clinically. Results The 3D GRE images had a better target resolution (0.47 × 0.94 × 2 mm, scan time = 5 min), iEPI and SAP-EPI images (resolution = 0.94 × 0.94 × 2 mm) were acquired in a faster scan time (1:52 min) with twice the brain coverage. SAP-EPI showed motion-correction capability and some immunity to undersampling from rejected data. Conclusion While 3D SAP-EPI suffers from some geometric distortion, its short scan time and motion-correction capability suggest that SAP-EPI may be a useful alternative to GRE and iEPI for use in SWI, particularly in uncooperative patients. J. Magn. Reson. Imaging 2015;41:1447–1453. © 2014 Wiley Periodicals, Inc.

T2*-weighted magnetic resonance imaging (MRI) has been reported to help improve detection of intracranial hemorrhage and is widely used in human neuroimaging. To assess the utility of this technique in small animals, interpretations based on this sequence were compared with those based on paired T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences in 200 dogs and cats that underwent brain MRI for suspected intracranial disease. Two sets of images (T2 + FLAIR and T2*) were reviewed separately in random order unaccompanied by patient information and were interpreted as normal or abnormal based on whether intracranial abnormalities were seen. The number and location of intracranial lesions were recorded. Eighty-five studies were considered normal and 88 were considered abnormal based on both sets of images, with good agreement (κ = 0.731) between the two. Susceptibility artifact was present in 33 cases (16.5%) on T2*-weighted images. In 12 cases (6%) a total of 69 lesions were seen on T2*-weighted images that were not seen on T2/FLAIR, all of which were associated with susceptibility artifact caused by presumed intracranial hemorrhage. Pseudolesions were seen on T2*-weighted images in five cases, none of which were associated with susceptibility artifact. Abnormalities were seen on T2/FLAIR images that were not seen on T2*-weighted images in 35 cases, confirming that T2* does not replace standard spin echo sequences. These results support inclusion of T2*-weighted sequences in small animal brain MRI studies and indicate that that a large number of abnormalities (especially hemorrhagic lesions) can go undetected if it is not performed.

Purpose To implement in-phase and out-of-phase (IP/OP) techniques with Magnetization-prepared gradient recalled echo (MP-GRE) and to evaluate the feasibility and diagnostic image quality among pre and post-optimized MP-GRE sequences, including patients unable to cooperate with breath-hold requirements. Materials and methods Institutional review board approval with waiver of informed consent was obtained for this HIPAA-compliant retrospective study. Two groups of patients were included in the study, before and after optimization of MP-GRE parameters, with seventy-three (24 noncooperative/49 cooperative) and sixty-four (22 noncooperative/42 cooperative) consecutive patients, respectively. The motion-insensitive sequence used in this study was a single-shot 2D MP-GRE. Two radiologists qualitatively evaluated the sequences to identify the presence of phase cancelation artifact in OP images and to determine image quality, extent of artifacts (respiratory ghosting, bounce-point artifact, spatial misregistration and pixel graininess) and lesion conspicuity on the various sequences. The ability to visually detect liver steatosis and fatty adrenal adenomas was evaluated. Qualitative analyses were compared using the Wilcoxon and Mann–Whitney tests. Results There were statistically significant differences between all MP-GRE sequences concerning phase cancelation artifact ( p 1 ) and post-optimized (IP 2 ) MP-GRE IP sequences, respectively, in all patients. Bounce point artifacts were significantly more pronounced in MP-GRE IP 1 ( p 2 ( p = .0027). MP-GRE OP and MP-GRE IP 2 showed significantly higher overall image quality ( p MP-GRE sequences subjectively identified hepatic steatosis (n = 20) and adrenal adenomas (n = 5) based on signal loss from IP to OP sequence. Conclusion Single shot IP/OP MP-GRE is feasible and allows motion resistant imaging with adequate diagnostic image quality. This technique is able to provide IP and OP information in patients unable to suspend respiration.

We studied the feasibility of the precise size measurement of extraocular muscles using fast magnetic resonance imaging sequences and gaze fixation targets. We recruited 20 healthy volunteers and optimized fat-suppressed fast T1-weighted gradient-recalled echo (FS T1-GRE) and single-shot fast spin-echo (SSFSE) imaging for evaluating extraocular muscles. With each eye at 40° abduction and adduction, we scanned the short-axis plane at the maximum diameter of the bilateral lateral and medial rectus muscles and measured the maximal cross-sectional area of the muscles during extension and contraction. We evaluated interobserver reproducibility and variability between the size measurements using the two MR sequences and the measurement reproducibility using FS T1-GRE imaging. The interobserver reproducibility in the muscle size measurements using FS T1-GRE and SSFSE imaging was excellent (r = 0.93–0.94) and moderate to good (r = 0.54–0.75), respectively. The interobserver variability was less with FS T1-GRE than SSFSE imaging (p

The aim of this study was to evaluate the usefulness of three-dimensional T1-weighted spoiled gradient-recalled echo (3D T1-GRE) images for the preoperative anatomical evaluation of lumbosacral lipoma, thick filum terminale, and myelomeningocele as a means of compensating for the drawbacks of 3D heavily T2-weighted (3D hT2-W) images.Nine patients with lumbosacral lipomas, one patient with tight filum terminale, and five patients with myelomeningoceles were included in this study. 3D T1-GRE images were compared with 3D hT2-W images or conventional magnetic resonance images in terms of delineation of lipomas and other structures in the patients with lipomas and tight filum terminale. For patients with myelomeningoceles, 3D T1-GRE images were compared with 3D hT2-W images in terms of artifacts in the cerebrospinal fluid (CSF) space.The 3D T1-GRE images demonstrated lipomas with good contrast to the spinal cord and CSF space and more clearly delineated the anatomical relationship between lipomas and these structures than did the 3D hT2-W images. The 3D T1-GRE images delineated dural defects through which extradural lipomas penetrated into the intradural space. The 3D T1-GRE images also demonstrated the presence or absence of lipomas in the filum terminale and the absence of artifact in the myelomeningoceles. Furthermore, they were useful for differentiating artifacts observed on the 3D hT2-W images from nerve elements.The complementary use of 3D T1-GRE and 3D hT2-W images may compensate for the drawbacks of 3D hT2-W images and may eventually improve lesion visualization and surgical decision making.

Purpose To develop a breathhold three-dimensional (3D) Dixon technique for fat suppressed imaging of myocardial infarction. Materials and Methods A pulse sequence was developed that uses a radial fan-beam k-space segmentation scheme for efficient coverage of k-space, enabling 3D scans in a single breathhold. The sequence uses a dual-echo bipolar readout to enable Dixon fat–water separation for improved visualization of epicardial and pericardial delayed enhancement. The 3D Dixon method was compared with a conventional 2D fast gradient recalled echo (FGRE) -based technique in 25 patients. Results Pericardial visualization scores and confidence were higher while overall image quality and artifacts were slightly worse for 3D Dixon compared with 2D FGRE. Robust fat suppression was achieved in 21 of 25 cases using the 3D Dixon method. Conclusion A 3D breathhold method for fat–water separated imaging of myocardial delayed enhancement was developed and validated. J. Magn. Reson. Imaging 2013;38:1362–1368. © 2013 Wiley Periodicals, Inc.

Purpose To evaluate the accuracy of an equilibrium magnetization (M0) map obtained using a two-dimensional (2D) spoiled gradient-recalled echo (SPGR) pulse sequence with variable flip angle (VFA). Materials and Methods Single-slice 2D SPGR images of 4% agar gel phantoms with different gadolinium-diethylene triamine pentaacetic acid (Gd-DTPA) concentrations (0–1 mM) were obtained with a VFA (2–30°). The 2D SPGR-VFA data were acquired with different repetition times (TRs; 7.8–117.2 ms), Gaussian and sinc RF pulses, and different field strengths (4.7, 7, and 9.4 Tesla). M0 and T1 maps were calculated from the 2D SPGR-VFA data. M0 and T1 values were compared with those calculated from free-relaxed 2D gradient-recalled echo (GRE) images and inversion recovery-prepared 2D SPGR images. The M0 and T1 slice profiles were also investigated. Results Consistent M0 values were obtained, regardless of the different Gd concentrations, TRs, and pulse sequences. The M0 slice profiles calculated from the sliced SPGR-VFA data quantitatively reproduced those calculated from the free-relaxed sliced GRE. In contrast, the T1 values calculated from the 2D SPGR-VFA data were underestimated at a high Gd concentration, short TR, and Gaussian RF pulse. Conclusion M0 values calculated from 2D SPGR-VFA images are highly quantitative. J. Magn. Reson. Imaging 2013;38:1245–1250. © 2013 Wiley Periodicals, Inc.