Magnetic Resonance in Medicine

PurposeTo optimize a 100 msec pulse for producing CEST MRI contrast and evaluate in mice. MethodsA gradient ascent algorithm was employed to generate a family of 100 point, 100 msec pulses for use in CEST pulse trains ( PRECISE). Gradient ascent optimizations were performed for exchange rates (kca) = 500 s-1, 1,500 s-1, 2,500 s-1, 3,500 s-1 and 4,500 s-1 and offsets ({Delta}{omega}) = 9.6, 7.8, 4.2 and 2.0 ppm. 7 PRECISE pulse shapes were tested on an 11.7 T scanner using a phantom containing three representative CEST agents with peak saturation B1 = 4 T. The pulse producing the most contrast in phantoms was then evaluated for CEST MRI pH mapping of the kidneys in healthy mice after iopamidol administration. ResultsThe most promising pulse in terms of contrast performance across all three phantoms was the 9.6 ppm, 2500 s-1 optimized pulse with [~]2.7 x improvement over Gaussian and [~]1.3xs over Fermi pulses. This pulse also displayed a large improvement in contrast over the Gaussian pulse after administration of iopamidol in live mice. ConclusionA new 100 msec pulse was developed based on gradient ascent optimizations which produced better contrast compared to standard Gaussian and Fermi pulses in phantoms. This shape also showed a substantial improvement for CEST MRI pH mapping in live mice over the Gaussian shape and appears promising for a wide range of CEST applications.

Physiological brain pulsations, primarily driven by cardiac and respiratory activity, play a key role in driving neurofluid circulation and waste clearance. Capturing the temporal dynamics of cardiac- and respiratory-driven brain pulsations (0.2-1.5 Hz) requires fast imaging with TRs near 100 ms, which is often unachievable in functional MRI or dynamic diffusion MRI. As a result, valuable physiological information remains hidden in these datasets. Here, we introduce TRACC-PHYSIO, a time-domain analytical framework designed to quantify physiological coupling and pulse time delays in dynamic MRI without requiring a fast acquisition. TRACC-PHYSIO uses cross-correlation to detect co-fluctuations between slowly sampled dynamic MRI data and simultaneously recorded physiological waveforms. It measures two key metrics: the peak Coupling Coefficient (peak CorrCoeff), quantifying the strength of co-fluctuations, and the TimeDelay, reflecting the relative arrival time of the physiological impulse in the brain with millisecond-level temporal resolution. The primary aim of this study is to validate TRACC-PHYSIO through systematic simulations that model realistic dynamic MR signals with mixed physiological components. We comprehensively evaluate TRACC-PHYSIOs performance under a wide range of conditions, including varying cardiac-to-respiratory composition ratios, TRs, and acquisition times. Results demonstrate that TRACC-PHYSIO can robustly assess coupling strengths and time delays for both cardiac (TRACC-Cardiac) and respiratory (TRACC-Respiratory) components, even in datasets with long TRs up to 3 seconds. By enabling a reliable time-domain coupling analysis, TRACC-PHYSIO opens new avenues for revealing brain pulsation mechanisms and elucidating the physiological drivers of neurofluid dynamics in health and disease. This stimulation study provides a valuable reference for interpreting TRACC-PHYSIO results and understanding associated uncertainties in future applications. Highlights- TRACC-PHYSIO is a time-domain method developed to estimate cardiac and respiratory coupling and their pulsation time delays in dynamic MRI without requiring high temporal resolution. - TRACC-PHYSIO was validated through systematic simulations across varying physiological compositions and MR acquisition parameters. - Results demonstrated that TRACC-PHYSIO reliably quantifies cardiac and respiratory components in dynamic MR signals. - The stimulation study provides a useful reference for interpreting TRACC-PHYSIO results and understanding associated uncertainties in future applications.

PurposeTo introduce the OmniShim Toolbox, a software package designed to calibrate and perform B0 shimming in user-defined regions of interest across systems from different vendors, supporting various shim orders. MethodsIn this study, we systematically compared the performance of vendor-implemented B0 shim routines with our custom developed OmniShim toolbox, designed to improve static magnetic field homogeneity. Single-voxel magnetic resonance spectroscopy (MRS) data were acquired from the prefrontal cortex and occipital lobe, while magnetic resonance spectroscopic imaging (MRSI) data were collected from regions above and below the corpus callosum in healthy volunteers. Measurements were conducted on both 3T and 7T MR systems to evaluate the robustness and scalability of each B0 shimming strategy across different field strengths. Additionally, functional MRI (fMRI) data were acquired at 7T to assess the impact of improved shimming on EPI data quality and BOLD contrast. ResultsThe OmniShim Toolbox demonstrated superior B0 homogeneity across all applications and field strength, leading to reduced signal dropout in fMRI and MRSI data, significantly improved spectral linewidths in SV MRS and MRSI data as well improved detection of neural networks by resting state fMRI. ConclusionThe proposed OmniShim Toolbox offers a robust and flexible approach to control B0 inhomogeneity, resulting in substantial improvements in image quality and spectral resolution. These enhancements benefit a broad range of MR applications.

PurposeB0 inhomogeneity within the brain-feeding arteries is a major issue for pseudo-continuous arterial spin labeling (PCASL) at 7T because it reduces labeling efficiency and leads to a loss of perfusion signal. This study aimed to develop a vessel-specific dynamic B0 field shimming method for 7T PCASL to enhance labeling efficiency by correcting off-resonance in the arteries within the labeling region. MethodsWe implemented a PCASL sequence with dynamic B shimming at 7T that compensates for B0 field offsets at the brain-feeding arteries by updating linear shimming terms and adding a phase increment to the PCASL RF pulses. Rapidly acquired vessel-specific B field maps were used to calculate dynamic shimming parameters. We evaluated both 2D and 3D variants of our method, comparing their performance against established global frequency offset and optimal-encoding-scheme (OES)-based corrections. Cerebral blood flow (CBF) maps were quantified before and after corrections. CBF values from different methods in the whole brain, white matter, and grey matter regions were compared. ResultsAll off-resonance correction methods significantly enhanced perfusion signals across the brain. The proposed vessel-specific dynamic B shimming method improved labeling efficiency while maintaining optimal static shimming in the imaging region. Perfusion-weighted images demonstrated the superiority of 3D dynamic B shimming method compared to global or 2D-based correction approaches. CBF analysis revealed that 3D dynamic B shimming significantly increased CBF values relative to the other methods. ConclusionOur proposed dynamic B0 shimming method offers a significant advancement in PCASL robustness and effectiveness, enabling full utilization of 7T ASLs high sensitivity and spatial resolution.

IntroductionBrain tissue is exceptionally soft. Recent in vivo work, including intrinsic MRE using naturally occurring cardiac pulsations, has shown stiffness values far below those obtained from post-mortem testing or externally actuated MRE. Such extreme softness allows a balance of elastic and inertial forces at even low frequencies, restoring uniqueness in viscoelastic iMRE inversion. Here, we demonstrate that viscoelastic iMRE can provide unique and stable stiffness estimates across frequencies, enabled by the brains ultra-soft nature. MethodiMRE data was obtained for 8 healthy subjects from a previous 7T MRI study, and MRE data was obtained at 50 Hz for 38 healthy subjects from two prior studies. The elastic-to-inertial force ratio was calculated for all subjects and compared between intrinsic and extrinsic datasets. The convergence of the viscoelastic iMRE was evaluated across a range of initial conditions and was examined at approximately 1, 2, and 3 Hz. ResultsAt a brain stiffness of 2600 Pa, the iMRE force ratio exceeded the MRE value by nearly four orders of magnitude, whereas at 4-35 Pa it fell to within approximately 0.5-2.5 orders of magnitude. Convergence was consistent across initial stiffness estimates. The mean storage modulus across all subjects was 5.29{+/-}0.95 Pa at 1 Hz, 34{+/-}17 Pa at 2 Hz, and 160{+/-}98 Pa at 3 Hz, consistent with previously reported frequency dependency. ConclusionBalanced force ratios, consistent convergence, and physiologically plausible results support uniqueness of the viscoelastic inversion. These findings resolve a key limitation in iMRE modeling and provide further evidence for the brains ultra-soft nature at low frequencies.

The pulsatility of blood flow through cerebral arteries is clinically important, as it is intrinsically associated with cerebrovascular health. In this study we outline a new MRI approach to measuring the real-time pulsatile flow in cerebral arteries, which is based on the inflow phenomenon associated with fast gradient-recalled-echo acquisitions. Unlike traditional phase-contrast techniques, this new method, which we dub Dynamic Inflow Magnitude Contrast (DIMAC), does not require velocity-encoding gradients as sensitivity to flow velocity is derived purely from the inflow effect. We achieved this using a highly accelerated single slice EPI acquisition with a very short TR (15 ms) and a 90{degrees} flip angle, thus maximizing inflow contrast. We simulate the spoiled GRE signal in the presence of large arteries and perform a sensitivity analysis to demonstrate that in the regime of high inflow contrast it shows much greater sensitivity to flow velocity over blood volume changes. We support this theoretical prediction with in-vivo data collected in two separate experiments designed to demonstrate the utility of the DIMAC signal contrast. We perform a hypercapnia challenge experiment in order to experimentally modulate arterial tone within subjects, and thus modulate the arterial pulsatile flow waveform. We also perform a thigh-cuff release challenge, designed to induce a transient drop in blood pressure, and demonstrate that the continuous DIMAC signal captures the complex transient change in the pulsatile and non-pulsatile components of flow. In summary, this study proposes a new role for a well-established source of MR image contrast and demonstrates its potential for measuring both steady-state and dynamic changes in arterial tone. HighlightsO_LIWe present a novel method for measuring pulsatility of cerebral arteries. C_LIO_LIThe inflow effect on fast GRE imaging can be exploited to yield a flow velocity dependent signal. C_LIO_LIWe measure pulsatile flow through cerebral arteries dynamically on a beat-to-beat basis. C_LIO_LIWe use physiological challenges to demonstrate sensitivity to dynamic and steady-state changes in vascular tone. C_LI

PurposeTo create a realistic in-silico brain phantom for positive and negative magnetic susceptibility that incorporates susceptibility anisotropy, enabling the evaluation of how susceptibility anisotropy influences susceptibility separation algorithm performance. MethodsWe expanded an existing QSM validation phantom by creating separate maps for positive and negative susceptibility, with the option of modeling susceptibility anisotropy. Multi-echo gradient echo data were simulated to evaluate four susceptibility separation techniques ({chi}-separation, DECOMPOSE-QSM, APART-QSM, and [Formula]). To assess the impact of noise, simulations were performed at different SNR levels (50, 100, 200, 300). ResultsOur findings showed that the error in negative susceptibility estimates increased by up to 53% when susceptibility anisotropy was present, compared to the case without susceptibility anisotropy, with {chi}-separation being the algorithm that was most sensitive to anisotropy. Robustness to noise varied across the assessed algorithms, with APART-QSM and {chi}-separation having the highest and lowest sensitivity to noise, respectively. ConclusionThe modified phantom is open-source and can serve as a numerical ground truth for evaluating susceptibility separation methods. Our findings emphasize the importance of incorporating susceptibility anisotropy into susceptibility separation models to improve their accuracy.

PurposeNon-water-suppressed proton spectroscopy, 1H-MRS, is desirable, as retaining the strong water resonance can facilitate automated online data corrections, internal concentration referencing, and monitoring of line narrowing effects in functional MRS. Removal of the water suppression module can also mitigate magnetization transfer effects and slightly reduce the minimum achievable TR and total RF power deposition. However, water suppression is typically considered essential due to eddy current-induced antisymmetric sidebands on the water resonance that distort the spectral baseline and obscure metabolite signals. Theory and MethodsThe Gradient Impulse Response Function (GIRF) was used to predict time-dependent magnetic field perturbations during the FID that generate the artefactual sidebands. The GIRF was measured in a one-time calibration, independent of spectroscopy acquisitions, enabling post-processing correction of the sidebands without sequence modification or additional dedicated hardware. GIRF-corrected non-water-suppressed single-voxel-spectroscopy (SVS) was compared to otherwise identical water-suppressed acquisitions in eight participants at 3T using semi-LASER and MEGA-PRESS sequences. ResultsAcross participants, GIRF correction reduced sideband amplitudes to levels comparable with the spectral baseline, enabling recovery of the underlying metabolite signals. Systematic increases in quantified metabolite concentrations were observed relative to water-suppressed acquisitions, consistent with water-suppression-induced magnetization transfer effects. Total creatine exhibited the largest increase, with enhancement ratios of 1.069{+/-}0.039 for MEGA-PRESS and 1.535{+/-}0.160 for semi-LASER acquisitions. ConclusionGradient-induced artefactual sidebands in non-water-suppressed MR spectroscopy can be effectively corrected using the GIRF to predict time-dependent magnetic field perturbations during the FID. In principle, the approach extends to other SVS sequences and field strengths following appropriate GIRF calibration.

PurposeHigh-dimensional and dynamic MRI are often limited by thermal noise, particularly in accelerated acquisitions. Although image-domain low-rank denoising methods (e.g., MP-PCA and NORDIC) are effective, reliance on a stationary noise distribution limits applicability to non-Cartesian sampling and advanced reconstruction methods. This work introduces MR KLEAN, a k-space low-rank denoising framework agnostic to acquisition trajectory and reconstruction strategy. Theory and MethodsMR KLEAN exploits locally low-rank structure in multichannel, high-dimensional k-space. Data are prewhitened using a noise-only calibration scan to enforce independent and identically distributed, zero-mean, unit-variance noise. Casorati matrices are generated through local k-space patches and denoised via singular-value thresholding, with thresholds from Monte-Carlo simulations under known noise statistics. MR KLEAN was evaluated in (1) a phantom study using Cartesian 3D FLASH, (2) an ASL study with spiral readout and compressed sensing reconstruction to assess generaliz-ability and preservation of temporal information via resting-state connectivity analysis, and (3) an accelerated cardiac cine study to assess performance under rapid temporal dynamics. ResultsMR KLEAN increased SNR and CNR in phantom study. In vivo ASL showed reduced noise for perfusion images, improved relative SNR, and substantially enhanced resting-state network detectability and functional connectivity sensitivity. In cardiac imaging, MR KLEAN reduced noise and improved delin-eation of fine anatomical features while preserving temporal fidelity across cardiac phases. ConclusionMR KLEAN provides robust, acquisition-and reconstruction-agnostic k-space denoising, improving image quality and allowing flexible spatial-temporal trade-offs. Results further support that high-dimensional k-space data retain intrinsic low-rank structure analogous to image space despite temporal signal variations.

PurposeTo demonstrate the advantages of spatiotemporal magnetic field monitoring to correct eddy current-induced artifacts (ghosting and geometric distortions) in high gradient strength diffusion MRI (dMRI). MethodsA dynamic field camera with 16 NMR field probes was used to characterize eddy current fields induced from diffusion gradients for different gradients strengths (up to 300 mT/m), diffusion directions, and shots in a 3D multi-shot EPI sequence on a 3T Connectom scanner. The efficacy of dynamic field monitoring-based image reconstruction was demonstrated on high-resolution whole brain ex vivo dMRI. A 3D multi-shot image reconstruction framework was informed with the actual nonlinear phase evolution measured with the dynamic field camera, thereby accounting for high-order eddy currents fields on top of the image encoding gradients in the image formation model. ResultsEddy current fields from diffusion gradients at high gradient strength in a 3T Connectom scanner are highly nonlinear in space and time, inducing high-order spatial phase modulations between odd/even echoes and shots that are not static during the readout. Superior reduction of ghosting and geometric distortion was achieved with dynamic field monitoring compared to ghosting approaches such as navigator- and structured low-rank-based methods or MUSE, followed by image-based distortion correction with eddy. Improved dMRI analysis is demonstrated with diffusion tensor imaging and high-angular resolution diffusion imaging. ConclusionStrong eddy current artifacts characteristic of high gradient strength dMRI can be well corrected with dynamic field monitoring-based image reconstruction, unlike the two-step approach consisting of ghosting correction followed by geometric distortion reduction with eddy.

PurposeMyelin water fraction (MWF) mapping in the central nervous system is a topic of intense research activity. One framework for this requires parameter estimation from a decaying biexponential signal. However, this is often an ill-posed nonlinear problem resulting in unreliable parameter estimates. For linear least-squares (LLS) problems, the ridge regression theorem (RRT) shows that a Tikhonov regularization parameter exists that will reduce mean square error (MSE) in parameter estimates. We present and apply a nonlinear version of the RRT,{lambda} -NL-RR, to MWF mapping. MethodsFor simulated and experimental data, we estimated parameter values with conventional nonlinear least-squares (NLLS) and compared these with values obtained from{lambda} -NL-RR, with the regularization parameter value defined by generalized cross validation. We applied regularization only to signals identified as biexponential according to the Bayesian information criterion. ResultsUnder conditions of modest SNR and closely spaced exponential time constants in which conventional biexponential analysis methods yield particularly inaccurate results,{lambda} -NL-RR decreases MSE by ~10-15%. ConclusionRegularization of the NLLS parameter estimation problem for the biexponential model decreased MSE for simulated and in vivo MRI brain data. In addition, this work provides a general framework for regularization of a broad class of NLLS problems.

PurposeTo demonstrate the feasibility and performance of mesoscopic whole brain T2*-weighted (T2*w) MRI at 10.5 T by combining a motion-robust multi-echo gradient-echo (GRE) method with high-density RF receive arrays. MethodsMulti-echo GRE data were collected in healthy adults at isotropic 0.5 mm resolution using a custom-built 16-channel transmit/80-channel receive (16Tx/80Rx) RF coil. Whole brain images were reconstructed with navigator-guided joint motion and field correction and were used for quantitative R2* and susceptibility ({chi}) mapping. Intrinsic signal-to-noise ratio (iSNR) and quantification precision for R2* and{chi} were also estimated. The results were compared with those obtained in the same subjects with matched resolution at 7 T using the commercial Nova 1Tx/32Rx coil, to demonstrate the iSNR and quantification precision gains at 10.5 T. G-factors were also calculated at each field strength to evaluate parallel imaging performances. To demonstrate the benefit of increased parallel imaging performances at 10.5 T, whole brain images with higher acceleration were also obtained using a custom-built 16Tx/128Rx coil. Resultsthe utilized motion robust GRE sequence and reconstruction effectively reduced artifacts from motion and field changes during scans, producing high-quality whole-brain T2*w images and multi-parametric maps at 10.5 T with delineation of fine-scale brain structures. Compared to 7 T, the 10.5 T approach led to gains in both iSNR and quantification precision of R2* and{chi} . Quantitatively, iSNR estimated in the peripheral cortical gray matter increased by 42%. Parallel imaging performances were also improved at 10.5 T owing to the utilized high-density coils compared to the commonly used commercially available coil at 7 T, allowing high-quality images with up to 12-fold combined acceleration when using the 128Rx coil. ConclusionIt is feasible to perform motion-robust whole-brain mesoscopic multi-echo gradient echo imaging of the human brain at 10.5 T. Intrinsic SNR and quantification precision of R2* and{chi} were estimated and compared with 7 T results. The results presented here may shed light on future optimal implementation of anatomic T2*w brain MRI at ultrahigh field beyond 7 T.

AbstractO_ST_ABSPurposeC_ST_ABSMaps of the MRI parameters R2* and magnetic susceptibility () enable the investigation of microscopic tissue changes in brain disease. However, cardiac-induced signal instabilities increase the variability of brain maps of R2* and . In this study, we introduce ISME - a sampling strategy that minimizes the level of cardiac-induced instabilities in brain maps of R2* and . MethodsISME uses phase-encoding gradients to shift the k-space frequency of the acquired data between consecutive readouts of a multi-echo train. As a result, the multi-echo data at a given k-space index is acquired at different phases of the cardiac cycle. We compare the variability of R2* and maps acquired with ISME and with standard multi-echo trajectories in N=10 healthy volunteers. We investigate the effect of both trajectories on the spatial aliasing of pulsating MR signals and propose a weighted-least squares (NWLS) approach for the estimation of R2* that accounts for the increase of the residuals with echo time. ResultsISME reduces the variability of R2* and maps across repetitions by 25/26/21% and 24/32/23% in the cerebellum/brainstem/whole brain, respectively. With ISME, the spatial aliasing of pulsating MR signals is incoherent between raw echo images, leading to visually sharper R2* maps. The proposed NWLS approach for the estimation of R2* reduces the dependence of the fitting residuals on echo time and the variability of R2* by an additional 3/2/1% in the cerebellum/brainstem/whole brain. ConclusionISME allows the mitigation of cardiac-induced signal instabilities in brain maps of R2* and , improving reproducibility.

PurposeIntervertebral disc degeneration (IVDD), resulting in the depletion of hydrophilic glycosaminoglycans (GAGs) located in the nucleus pulposus (NP), can lead to debilitating neck and back pain. Magnetic Resonance Imaging (MRI) is a promising means of IVD assessment due to the sensitivity of MRI tissue relaxation properties to matrix composition. Furthermore, anomalous (i.e. non-monoexponential) relaxation models have shown higher sensitivity to specific matrix components compared to conventional monoexponential models. Here, we extend the use of the stretched exponential model, an anomalous relaxation model, to IVD relaxometry data. Theory and MethodsT1{rho} and T2 relaxation data were measured in the cervical IVDs of healthy volunteers and IVDs adjacent to cervical fusion, and analyzed using both conventional and stretched-exponential (SE) models. Model differences were evaluated via goodness of fit in the healthy data. Normalized histograms of the resultant quantitative MRI (qMRI) maps were described using stable distributions, and data were compared across adjacent disc segments. ResultsIn the healthy IVDs, we found lower mean squared error in the SE relaxation model fitting behavior compared to monoexponential models, supporting anomalous relaxation behavior in healthy IVDs. SE model parameter T1{rho} increased level-wise in the caudal direction, especially in the nucleus pulposus, while conventional T1{rho} and T2 monoexponential measures did not vary. For IVDs adjacent to cervical fusion, SE parameters deviated near the fusion site compared with those in the healthy population. ConclusionSE modeling of T1{rho} relaxation provides greater sensitivity to level-wise variation in IVD matrix properties compared with conventional relaxation modeling, and could provide improved sensitivity to early stages of IVD degeneration. The improved model fit and correlation between the SE T1{rho} parameter with IVD level suggests SE modeling may be a more sensitive method for detection of GAG content variation.

PurposeTo develop a SNR enhancement method for chemical exchange saturation transfer (CEST) imaging using a denoising convolutional autoencoder (DCAE), and compare its performance with state-of-the-art denoising methods. MethodThe DCAE-CEST model encompasses an encoder and a decoder network. The encoder learns features from the input CEST Z-spectrum via a series of 1D convolutions, nonlinearity applications and pooling. Subsequently, the decoder reconstructs an output denoised Z-spectrum using a series of up-sampling and convolution layers. The DCAE-CEST model underwent multistage training in an environment constrained by Kullback-Leibler divergence, while ensuring data adaptability through context learning using Principal Component Analysis processed Z-spectrum as a reference. The model was trained using simulated Z-spectra, and its performance was evaluated using both simulated data and in-vivo data from an animal tumor model. Maps of amide proton transfer (APT) and nuclear Overhauser enhancement (NOE) effects were quantified using the multiple-pool Lorentzian fit, along with an apparent exchange-dependent relaxation metric. ResultsIn digital phantom experiments, the DCAE-CEST method exhibited superior performance, surpassing existing denoising techniques, as indicated by the peak SNR and Structural Similarity Index. Additionally, in vivo data further confirms the effectiveness of the DCAE-CEST in denoising the APT and NOE maps when compared to other methods. While no significant difference was observed in APT between tumors and normal tissues, there was a significant difference in NOE, consistent with previous findings. ConclusionThe DCAE-CEST can learn the most important features of the CEST Z-spectrum and provide the most effective denoising solution compared to other methods.

Massively multidimensional diffusion-relaxation correlation MRI (MMD-MRI) provides information beyond the traditional voxel-averaged metric that may better characterize the microstructural characteristics of biological tissues. MMD-MRI reproducibility has been established in clinical settings, but has yet to be thoroughly evaluated under preclinical conditions, where superior hardware and modulated gradient waveforms enhance its performance. In this study, we investigate the reproducibility of MMD-MRI on a micro-imaging system using ex vivo mouse brains. Notably, the estimated signal fractions of intra-voxel spectral components in the MD-MRI distribution, corresponding to white and gray matter, along with the frequency-dependent parameters, demonstrated high reproducibility. We identified bias between scan and rescan in some of the metrics, which we attribute to the time gap between repeated scans pointing to a long-time progressive fixation effect. We compare our results with in vivo results from clinical scanners and show the reproducibility of diffusion frequency-dependent metrics to benefit from the improved gradient hardware on our preclinical setup. Our results inform future micro-imaging ex vivo MMD-MRI studies of the reproducibility of MMD-MRI metrics and their dependence on fixation time.

PurposeDynamic T1-weighted lung oxygen enhanced MRI (OE-MRI) is challenging at 3 T due to decreased longitudinal relaxivity of oxygen and increased magnetic susceptibility difference between air and tissue interfaces relative to 1.5 T, leading to poor signal quality. In this work, we evaluate the robustness of an alternative T2*-sensitised lung dynamic OE-MRI protocol in humans at 3 T. MethodsSimulations were performed to predict OE contrast behaviour and optimise the MRI protocol. Sixteen healthy subjects underwent dynamic free-breathing OE-MRI acquisitions using a dual echo RF-spoiled gradient echo acquisition at 3 T on two MRI scanners at different institutions. Non-linear registration and tissue density variation correction were applied. Percent signal enhancement (PSE) maps and {Delta}R2* were derived. Intra-class correlation coefficient (ICC) and Bland-Altman analyses were used to evaluate reproducibility of the OE indices across two sites and vendors as well as scan-rescan repeatability. ResultsSimulations and experimental data show negative contrast on oxygen inhalation due to substantial dominance of {Delta}R2* at TE longer than 0.2 ms when using our chosen flip angle and TR. Mean PSE values were TE dependent and mean {Delta}R2* was 0.14 ms-1 {+/-} 0.03 ms-1, ICC values for intra-scanner (ICCintra) and inter-scanner (ICCinter) variability for OE indices were high (ICCintra > 0.74; ICCinter = 0.70) and 95% limits of agreement showed strong agreement of repeated measures. ConclusionOur results demonstrate excellent scan-rescan repeatability for the PSE indices and good reproducibility for {Delta}R2* across two sites and vendors, suggesting potential utility in multi-centre clinical studies.

PurposeTo measure magnetic susceptibility ({chi}) with Quantitative Susceptibility Mapping (QSM) and evaluate its repeatability using four phantom doping materials relevant to QSM applications. MethodsA cylindrical phantom was constructed containing vials of agarose gel doped with two paramagnetic materials (ferritin, USPIO) and two diamagnetic materials (CaCl2, CaCO3) at five concentrations each. Single orientation QSM measurements (MEDI+0) were carried out on the phantom at 3T and 7T. We measured molar susceptibility ({chi}mol) from QSM and evaluated the test-retest repeatability of {chi} using the standard error of the measurement (SEM). We evaluated material lifespan by conducting a t-test of {chi}mol at various timepoints. Results{chi}mol (ppm{middle dot}L{middle dot}mmol-1) were measured as 1.67 {+/-} 0.24 and 0.74 {+/-} 0.09 (USPIO: 3T and 7T, respectively), 10-2x(8.13 {+/-} 1.35; 8.13 {+/-} 1.19) (Ferritin: 3T; 7T), 10-4x(-2.68 {+/-} 0.24; -2.71 {+/-} 0.37) (CaCl2: 3T; 7T), and 10-5x(-9.52 {+/-} 1.44; -9.53 {+/-} 1.18) (CaCO3: 3T; 7T). The USPIO SEM (1.5 {+/-} 2.0; 5.1 {+/-} 2.0 ppb at 3T; 7T) was greater than the ferritin SEM (1.2 {+/-} 1.0; 2.2 {+/-} 1.3 ppb at 3T; 7T). The CaCl2 SEM (7.5 {+/-} 5.5; 1.2 {+/-} 0.6 ppb at 3T; 7T) was greater than the CaCO3 SEM (1.2 {+/-} 0.6; 0.9 {+/-} 0.7 ppb at 3T; 7T). We observed no significant changes in molar susceptibility for ferritin and CaCO3 over the measured timeframes (24 months and 15 months, respectively). ConclusionWe recommend using ferritin and CaCO3 in the construction of susceptibility phantoms, removing later echo times for CaCO3 QSM reconstructions.

PurposeMotivated by the need to improve GABA-edited magnetic resonance spectroscopy (MRS) quality, we developed a three-module framework to improve transient averaging based on quality. We hypothesized that training a deep learning (DL) to differentiate spectrum quality could improve transient averaging compared to traditional averaging. MethodsThe transient averaging framework was approached through three modules: (1) a continuous-valued automated quality labeling algorithm using both traditional and recently developed MRS quality metrics, (2) a dual-domain (time and frequency) DL model that learns from these quality labels to assess quality scores for new data, and (3) a transient weighting algorithm informed by DL quality scores. The labeling algorithm was used to produce quality labels focused on retaining GABA peak shape in difference spectra (1) to train the DL model (2). The DL model quality scores were used to assign weights (3) for transient pairs within the final average difference spectrum. Results were compared to an existing software weighting algorithm for transient averaging and traditional transient averaging. ResultsRetaining only GABA-edited transient pairs with quality labels above zero, defined by metrics evaluating peak shapes, resulted in overall better traditional and recently developed mean metric values as well as better visual assessment of GABA and Glx peaks. Applying the trained DL model to in vivo scans, the average difference spectra calculated from the DL quality scores and weighting algorithm resulted in lower fit errors than averaging all transients with equal weights. ConclusionThe proposed framework can optimize transient averaging based on quality for edited-MRS.

Reproducible and accurate fitting of the magnetic resonance spectrum is critical for estimating metabolite concentrations. We have previously developed a fitting software called ProFit-1D which was shown to fit 9.4T semi-LASER data from the human brain with high accuracy and precision. In this study, we adapted ProFit-1D to fit J-difference edited spectra acquired at a clinical field strength of 3T and to assessed its performance in simulated and in vivo data. ProFit-1D was adapted to fit J-difference edited data with alterations to the fitting range to exclude the 1.3 ppm lipid resonance, starting T2 relaxation constants, initial fit parameters, and adaptive spectral baseline determination. The accuracy of ProFit-1D was systematically evaluated on simulated GABA-edited and 2-hydroxyglutarate-edited (2HG-edited) data with different types of in vivo parameter variations and compared to that of LCModel and Gannet, two software commonly used to fit J-difference edited data. The precision of ProFit-1D was also evaluated in GABA-edited spectra acquired in vivo in the occipital cortex (OCC) and medial prefrontal cortex (mPFC) of healthy participants at 3T using subsets of averages and compared to that of LCModel and Gannet. The 2HG fit error was also evaluated for ProFit-1D in 2HG-edited spectra acquired in glioma patients and compared to that of LCModel. Overall, it was found that ProFit-1D generally produced fits with low parameter fit errors across a variety of parameter variations. GABA, glutamate plus glutamine (Glx), and 2HG levels were also more accurately estimated with ProFit-1D than with LCModel and Gannet across different spectral disturbances and simulated concentrations. ProFit-1D was found to be as precise as LCModel and more precise than Gannet in estimating GABA and Glx. 2HG fit errors were 45% lower with ProFit-1D than with LCModel. Thus, ProFit-1D was found to produce high-quality fits to J- difference edited data with high accuracy and precision.