E. Brian Welch

High-Resolution 7T MRI of the Human Hippocampus In Vivo

To describe an initial experience imaging the human hippocampus in vivo using a 7T magnetic resonance (MR) scanner and a protocol developed for very high field neuroimaging.

Development of chemical exchange saturation transfer at 7T

Chemical exchange saturation transfer (CEST) MRI is a molecular imaging method that has previously been successful at reporting variations in tissue protein and glycogen contents and pH. We have implemented amide proton transfer (APT), a specific form of chemical exchange saturation transfer imaging, at high field (7T) and used it to study healthy human subjects and patients with multiple sclerosis. The effects of static field inhomogeneities were mitigated using a water saturation shift referencing method to center each z‐spectrum on a voxel‐by‐voxel basis. Contrary to results obtained at lower fields, APT imaging at 7T revealed significant contrast between white and gray matters, with a higher APT signal apparent within the white matter. Preliminary studies of multiple sclerosis showed that the APT asymmetry varied with the type of lesion examined. An increase in APT asymmetry relative to healthy tissue was found in some lesions. These results indicate the potential utility of APT at high field as a noninvasive biomarker of white matter pathology, providing complementary information to other MRI methods in current clinical use. Magn Reson Med, 2011.

A novel AIF tracking method and comparison of DCE-MRI parameters using individual and population-based AIFs in human breast cancer.

Quantitative analysis of dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) data requires the accurate determination of the arterial input function (AIF). A novel method for obtaining the AIF is presented here and pharmacokinetic parameters derived from individual and population-based AIFs are then compared. A Philips 3.0 T Achieva MR scanner was used to obtain 20 DCE-MRI data sets from ten breast cancer patients prior to and after one cycle of chemotherapy. Using a semi-automated method to estimate the AIF from the axillary artery, we obtain the AIF for each patient, AIF(ind), and compute a population-averaged AIF, AIF(pop). The extended standard model is used to estimate the physiological parameters using the two types of AIFs. The mean concordance correlation coefficient (CCC) for the AIFs segmented manually and by the proposed AIF tracking approach is 0.96, indicating accurate and automatic tracking of an AIF in DCE-MRI data of the breast is possible. Regarding the kinetic parameters, the CCC values for K(trans), v(p) and v(e) as estimated by AIF(ind) and AIF(pop) are 0.65, 0.74 and 0.31, respectively, based on the region of interest analysis. The average CCC values for the voxel-by-voxel analysis are 0.76, 0.84 and 0.68 for K(trans), v(p) and v(e), respectively. This work indicates that K(trans) and v(p) show good agreement between AIF(pop) and AIF(ind) while there is a weak agreement on v(e).

A nonrigid registration algorithm for longitudinal breast MR images and the analysis of breast tumor response

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) can estimate parameters relating to blood flow and tissue volume fractions and therefore may be used to characterize the response of breast tumors to treatment. To assess treatment response, values of these DCE-MRI parameters are observed at different time points during the course of treatment. We propose a method whereby DCE-MRI data sets obtained in separate imaging sessions can be co-registered to a common image space, thereby retaining spatial information so that serial DCE-MRI parameter maps can be compared on a voxel-by-voxel basis. In performing inter-session breast registration, one must account for patient repositioning and breast deformation, as well as changes in tumor shape and volume relative to other imaging sessions. One challenge is to optimally register the normal tissues while simultaneously preventing tumor distortion. We accomplish this by extending the adaptive bases algorithm through adding a tumor-volume preserving constraint in the cost function. We also propose a novel method to generate the simulated breast magnetic resonance (MR) images, which can be used to evaluate the proposed registration algorithm quantitatively. The proposed nonrigid registration algorithm is applied to both simulated and real longitudinal 3D high resolution MR images and the obtained transformations are then applied to lower resolution physiological parameter maps obtained via DCE-MRI. The registration results demonstrate the proposed algorithm can successfully register breast MR images acquired at different time points and allow for analysis of the registered parameter maps.

Real-Time Compressive Sensing MRI Reconstruction Using GPU Computing and Split Bregman Methods.

Compressive sensing (CS) has been shown to enable dramatic acceleration of MRI acquisition in some applications. Being an iterative reconstruction technique, CS MRI reconstructions can be more time-consuming than traditional inverse Fourier reconstruction. We have accelerated our CS MRI reconstruction by factors of up to 27 by using a split Bregman solver combined with a graphics processing unit (GPU) computing platform. The increases in speed we find are similar to those we measure for matrix multiplication on this platform, suggesting that the split Bregman methods parallelize efficiently. We demonstrate that the combination of the rapid convergence of the split Bregman algorithm and the massively parallel strategy of GPU computing can enable real-time CS reconstruction of even acquisition data matrices of dimension 40962 or more, depending on available GPU VRAM. Reconstruction of two-dimensional data matrices of dimension 10242 and smaller took ~0.3 s or less, showing that this platform also provides very fast iterative reconstruction for small-to-moderate size images.

Quantitative effects of using compressed sensing in dynamic contrast enhanced MRI

Dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) involves the acquisition of images before, during and after the injection of a contrast agent. In order to perform quantitative modeling on the resulting signal intensity time course, data must be acquired rapidly, which compromises spatial resolution, signal to noise and/or field of view. One approach that may allow for gains in temporal or spatial resolution or signal to noise of an individual image is to use compressed sensing (CS) MRI. In this study, we demonstrate the accuracy of extracted pharmacokinetic parameters from DCE-MRI data obtained as part of pre-clinical and clinical studies in which fully sampled acquisitions have been retrospectively undersampled by factors of 2, 3 and 4 in Fourier space and then reconstructed with CS. The mean voxel-level concordance correlation coefficient for K(trans) (i.e. the volume transfer constant) obtained from the 2× accelerated and the fully sampled data is 0.92 and 0.90 for mouse and human data, respectively; for 3×, the results are 0.79 and 0.79, respectively; for 4×, the results are 0.64 and 0.70, respectively. The mean error in the tumor mean K(trans) for the mouse and human data at 2× acceleration is 1.8% and -4.2%, respectively; at 3×, 3.6% and -10%, respectively; at 4×, 7.8% and -12%, respectively. These results suggest that CS combined with appropriate reduced acquisitions may be an effective approach to improving image quality in DCE-MRI.

Dynamic B0 shimming at 7 T

Dynamic slice-wise shimming improves B0 field homogeneity by updating shim coil currents for every slice in a multislice acquisition, producing better field homogeneity over a volume than can be obtained by a single static global shim. The first aim of this work was to evaluate the performance of slice-wise field-map-based second-order dynamic shimming in a human high-field 7 T clinical scanner vis-à-vis image based second order static global shimming. Another goal was to characterize eddy currents induced by second and third order shim switching. A final aim was to compare global and dynamic shimming through shim orders to elucidate the relative benefits of going to higher orders and to dynamic shim updating from a static shimming regime. An external hardware module was used to store and dynamically update slice-optimized shim values during multislice data acquisition. High-bandwidth multislice gradient echo scans with B0 field mapping and low-bandwidth single-shot echo planar scans were performed on phantoms and humans using second-order dynamic and static global shims. For the measurement of second and third order shim induced eddy currents, step response temporal phase changes of individual shims were measured and fit to shim harmonics spatially and to multiexponential decay functions temporally. Finally, an order-wise field-map-based comparison was performed with first, second and third order global static shimming, first and second order dynamic shimming, as well as combined second or third order global and first order dynamic shim. Dynamic shimming considerably improved B0 homogeneity compared to static global shimming both in phantoms and in human subjects, reducing image distortion and signal dropout. The unshielded second and third order shims generated strong B0 and self and cross-term eddy fields, with multiple time constants ranging from milliseconds to seconds. Field homogeneity improved with increasing order of shim, with dynamic shimming performing better than global shimming. Hybrid global and dynamic shimming approach yielded field homogeneity better than global static shims but worse than dynamic shims.

Quantitative effects of inclusion of fat on muscle diffusion tensor MRI measurements.

To determine the minimum water percentage in a muscle region of interest that would allow diffusion tensor (DT−) MRI data to reflect the diffusion properties of pure muscle accurately.

Imaging body composition in obesity and weight loss: challenges and opportunities

Obesity is a threat to public health worldwide primarily due to the comorbidities related to visceral adiposity, inflammation, and insulin resistance that increase risk for type 2 diabetes and cardiovascular disease. The translational research portfolio that originally described these risk factors was significantly enhanced by imaging techniques, such as dual-energy X-ray absorptiometry (DEXA), computed tomography (CT), and magnetic resonance imaging (MRI). In this article, we briefly review the important contributions of these techniques to understand the role of body composition in the pathogenesis of obesity-related complications. Notably, these imaging techniques have contributed greatly to recent findings identifying gender and racial differences in body composition and patterns of body composition change during weight loss. Although these techniques have the ability to generate good-quality body composition data, each possesses limitations. For example, DEXA is unable to differentiate type of fat, CT has better resolution but provides greater ionizing radiation exposure, and MRI tends to require longer imaging times and specialized equipment for acquisition and analysis. With the serious need for efficacious and cost-effective therapies to appropriately identify and treat at-risk obese individuals, there is greater need for translational tools that can further elucidate the interplay between body composition and the metabolic aberrations associated with obesity. In conclusion, we will offer our perspective on the evolution toward an ideal imaging method for body composition assessment in obesity and weight loss, and the challenges remaining to achieve this goal.

Multi‐parametric MRI characterization of healthy human thigh muscles at 3.0 T – relaxation, magnetization transfer, fat/water, and diffusion tensor imaging

Muscle diseases commonly have clinical presentations of inflammation, fat infiltration, fibrosis, and atrophy. However, the results of existing laboratory tests and clinical presentations are not well correlated. Advanced quantitative MRI techniques may allow the assessment of myo‐pathological changes in a sensitive and objective manner. To progress towards this goal, an array of quantitative MRI protocols was implemented for human thigh muscles; their reproducibility was assessed; and the statistical relationships among parameters were determined. These quantitative methods included fat/water imaging, multiple spin‐echo T2 imaging (with and without fat signal suppression, FS), selective inversion recovery for T1 and quantitative magnetization transfer (qMT) imaging (with and without FS), and diffusion tensor imaging. Data were acquired at 3.0 T from nine healthy subjects. To assess the repeatability of each method, the subjects were re‐imaged an average of 35 days later. Pre‐testing lifestyle restrictions were applied to standardize physiological conditions across scans. Strong between‐day intra‐class correlations were observed in all quantitative indices except for the macromolecular‐to‐free water pool size ratio (PSR) with FS, a metric derived from qMT data. Two‐way analysis of variance revealed no significant between‐day differences in the mean values for any parameter estimate. The repeatability was further assessed with Bland–Altman plots, and low repeatability coefficients were obtained for all parameters. Among‐muscle differences in the quantitative MRI indices and inter‐class correlations among the parameters were identified. There were inverse relationships between fractional anisotropy (FA) and the second eigenvalue, the third eigenvalue, and the standard deviation of the first eigenvector. The FA was positively related to the PSR, while the other diffusion indices were inversely related to the PSR. These findings support the use of these T1, T2, fat/water, and DTI protocols for characterizing skeletal muscle using MRI. Moreover, the data support the existence of a common biophysical mechanism, water content, as a source of variation in these parameters. Copyright