The main parallel imaging techniques have been implemented under MATLAB (SENSE, PILS and SPACE-RIP for image domain, and SMASH and GRAPPA for k-space domain), evaluating its characteristics about image quality and signal to noise ratio. PILS and SMASH, with an specific coil configuration, provided the best results. SENSE and SPACE-RIP were more versatile with any coil configuration. All methods arrived to its acceleration factor theoretical limit (R=4 for 4 coils) with images without noise. Only those methods that depend on a matrix reconstruction inversion (SENSE and SPACE-RIP) did not arrive to the theoretical limit for noisy images due to its instability problems. PILS provided the best results although SENSE and SPACE-RIP can be applied under more general conditions. For the k-space domain methods, SMASH gave the best results under an specific coil array configuration

Reconstruction of magnetic resonance images from data not falling on a Cartesian grid is widely used for fast acquisitions, and it is a Fourier inversion problem typically solved using convolution interpolation, also known as gridding. This work presents a comparison between two gridding reconstruction methods to reconstruct magnetic resonance images from acquisitions using spiral trajectories through k-space. One method (grid-driven) is not based on a density compensation function while the other one (Direct Summation) uses Voronoi cells for the determination of the necessary areas to estimate the corresponding density compensation function. Both methods have been applied to the same image to see the reconstruction quality of each method. Both methods have correctly reconstructed the original image using only 13.73% of the original full-grid data from a Cartesian trajectory.

M. E. Brummer, M. Sanz-Blasco, S. Krishnan, L. H. Hamilton, S. Ramamurthy, and D. Moratal Pediatrics and Radiology, Emory University, Atlanta, GA, United States, Universitat Politècnica de València, València, Spain, Pediatrics, Emory University, Atlanta, GA, United States, Bioengineering, Georgia Institute of Technology, Atlanta, GA, United States, Center for Biomaterials and Tissue Engineering, Universitat Politècnica de València, València, Spain

Introduction A well-known reconstruction technique developed by Gerchberg, based on “error energy” reduction [1], is extended in this study to sparsely sampled dynamic cardiac magnetic resonance imaging (CMRI). A-priori knowledge of static and dynamic regions in the FOV is used to sample a subset of phase-encoding views on a regular Cartesian grid, allowing a reduction in overall imaging time. Similar to the direct-inversion Noquist method [2], the iterative reconstruction does not use either data-substitution or temporal interpolation. Instead, the inherent temporally band-limited properties of the spatially bounded object, the static FOV, are used to recover additional resolution from information embedded in sparse k-t samples. The algorithm iteratively “corrects” the data applying the band-limited constraint in the image domain and the acquired data in the Fourier domain. The proposed method is compared to a full-grid reconstruction (“truth”), the original Noquist reconstruction and to a method that uses temporal interpolation (UNFOLD). Convergence properties and noise amplification due to undersampled data are investigated. Methods The cardiac phantom introduced in [2] was used in simulation (Fig 3). Typically observed cardiac dynamics, such as concentric contraction (1) from systole (3b) to diastole (3c), vertical (2), horizontal (3) translational motion and sinusoidal (4) and transient (5) intensity events are simulated in k-space. Temporal changes were sampled uniformly at 16 phases during the cardiac cycle with a nominal full-grid spatial resolution of 256 x 256. Noisy datasets had normally distributed white noise with 10 dB SNR added to k-space data. Figure 1. shows acquisition patterns as per the Stairwell and UNFOLD algorithms (49.6% and 50 % data reduction respectively). Iterative reconstruction (Figure 2) begins with first producing estimated images from the sparse data using temporal nearest-neighbor substitution to generate a full dataset. This result is transformed to the Fourier domain and corrected by replacing all phase-encoding views that were sampled with the corresponding original data. These k-space data are returned to image domain via inverse Fourier transform. The error in the known static region due to residual dynamic content is minimized by temporal low-pass filtering in this region. This process continues until convergence, (no further change in error energy) is achieved. Results and Discussion The convergence patterns shown in Figure 3f. suggest ideal reconstruction, similar to direct inversion [2] for the Stairwell algorithm as seen (Fig 3a,b), while the UNFOLD reconstruction (Fig 3c) shows residual artifact due to incomplete estimation of the dynamic spectrum. The proposed iterative process however, yields improvements relative to the original UNFOLD temporal interpolation method (Fig 3g). Diastolic (Fig 3d) and systolic (Fig 3h) iterative reconstructions of a cardiac-gated CMRI acquisition (normal volunteer) show anticipated noise amplification in dynamic regions relative to full-matrix (Fig 3e). Table 1. quantifies this SNR reduction in a sample dynamic region in 10 dB phantom reconstructions.

Introduction A 2D analytical cardiac magnetic resonance imaging (MRI) phantom in the Fourier domain was introduced by Brummer et al. [1] to compare acceleration strategies based on a reduction of the field of view. In related work by our group and others the value of a standardized simulation phantom to test and compare reconstruction methods for cardiac imaging has become evident. In this work, a 4D analytical phantom in the Fourier domain is proposed, aimed to serve as a flexible, objective, standardized benchmark for evaluation and comparison of different image reconstruction techniques in dynamic 3D MRI.

Introduction The Noquist method for accelerated cardiac imaging takes advantage of the spatiotemporal redundancy whenever the field of view contains static regions [1]. Feasibility of the method for reconstruction of dynamic SSFP and velocity-encoded flow imaging has been demonstrated by reconstruction of originally conventional full-grid data sets from retrospectively selected sparse subsets [1, 2]. Notable characteristics of Noquist compared to alternative methods include full preservation of spatiotemporal resolution. The method was shown to further improve acceleration, by integrating naturally with parallel imaging [3], combining individual acceleration factors at accordingly accumulated SNR penalty. This study reports first results of an implementation of Noquist for a prospectively gated dynamic cardiac imaging method (GE FastCARD). The objective of this investigation is to demonstrate actual improvement of image quality by accelerated acquisition using prior knowledge of spatiotemporal redundancy.

Introduction Effective strategies for sparse sampling of kt-space are important for fast dynamic imaging. Following observations in early landmark papers [1-3] the temporal sampling rate of each k-space view is often linked to the signal energy represented by that view in k-space for a certain class of images, in general considering only the spatial or spatiotemporal spectrum of the entire image. Based on resulting criteria, sparse sampling schemes typically sample low spatial frequencies more frequently than high spatial frequencies, implicitly considering signal-to-noise ratio SNR as the dominant factor in image quality [1-4]. For cardiac imaging this concept is represented by the BRISK formalism [2]. Sampling schemes for angiography and dynamic lesion enhancement imaging were established by TRICKS [3] and STBB [4]. In this work we have studied analysis of temporal spectral content as a function of the k-space frequency, illustrated with simulated and actual ciné cardiac imaging data. Dynamic spectral content was observed both globally in the image, and spatially-selectively in relevant portions of the image.

L. H. Hamilton, J. A. Fabregat, D. Moratal, S. Ramamurthy, S. Kozerke, and M. E. Brummer School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States, Universidad Politécnica de Valencia, Valencia, Spain, Children’s Healthcare of Atlanta, Atlanta, GA, United States, University and ETH Zurich, Zurich, Switzerland, School of Medicine, Emory University, Atlanta, GA, United States

-1 contains identical blocks of inverse Fourier coefficients. Noquist reduces this model by representing the static part of the FOV only once in the entire sequence, since it is identical in all frames. The total size of the ciné image is thus reduced from T(NS+ND) to NS+TND. This propagates into accordingly reduced size requirements for each of the equally-sized temporal k-space sample vectors F for each time point to ND+NS/T. Selection of a subset of k-space views in each frame is constrained by the requirement to keep matrix M invertible. SENSE: An image f is observed by a receiver coil c with sensitivity Sc as the product fc=f.Sc, with corresponding Fourier data Fc=Mfc. With complete prior knowledge of Sc at all image locations for each of C independent receiver coils, complete conventional sampling of Fc yields C-fold redundant data. Accordingly, reduction of the number of acquired k-space views by a factor R up to C may present a feasible image reconstruction problem. Recovery of f from the reduced Fourier data sets Fc, simultaneously acquired from all C coils, may be achieved through different strategies, with different constraints in flexibility in sampling patterns and computational complexity.

An algorithm is presented for fully automated detection of brain contours from single-echo 3-D coronal MRI data. The technique detects structures in a head data volume in a hierarchical fashion. Detections consist of histogram-based thresholding operation, followed by a morphological cleanup procedure of the binary threshold mask images. Anatomic knowledge, essential for the discrimination between desired and undesired structures, is implemented through a sequence of conventional and new morphological operations. Innovative use of 3-D distance transformations allows implicit evaluation of anatomic relationships for structure recognition. Overlap tests between neighbouring slice images are used to propagate coherent 2-D brain masks through the third dimension. A summary of results of testing the algorithm on 23 test data sets is presented, with a discussion of potential for clinical application and generalization to other problems, and of limitations of the technique.

Functional magnetic resonance imaging (fMRI) for therapeutic monitoring of transcranial magnetic stimulation (TMS) F. L. Giesel1, A. Hempel2, E. Hempel3, T. Wuestenberg4, U. Seidl5, J. Schroeder2, M. Essig1; 1Department of Radiology, German Cancer Research Center, Heidelberg, GERMANY, 2Department of Psychiatry, University of Heidelberg, Heidelberg, GERMANY, 3Department of Medical Engineering, Forschungszentrum Karlsruhe, Karslruhe, GERMANY, 4Department of Neurology, Humboldt-University of Berlin, Berlin, GERMANY, 5Department of Psychiatry, Universtiy of Heidelberg, Heidelberg, GERMANY.

BACKGROUND:Autosomal-dominant polycystic kidney disease (ADPKD) is characterized by gradual renal enlargement and cyst growth prior to loss of renal function. Standard radiographic imaging has not provided the resolution and accuracy necessary to detect small changes in renal volume or to reliably measure renal cyst volumes. The Consortium for Radiologic Imaging Studies in Polycystic Kidney Disease (CRISP) is longitudinally observing ADPKD individuals using high-resolution magnetic resonance (MR) imaging to determine if change in renal and cyst volumes can be detected over a short period of time, and if they correlate with decline in renal function early in disease.METHODS:Standardization studies were conducted in phantoms and four subjects at each participating clinical center. After, in the full-scale protocol, healthy ADPKD individuals 15 to 45 years old with creatinine clearance>70 mL/min underwent standardized MR renal imaging, renal iothalamate clearance, comprehensive clinical evaluation, and determination of 24-hour urinary albumin and electrolyte excretion. Stereology was used from T1-weighted images to quantify renal volume, and region-growing thresholding was used from T2-weighted images to determine cyst volume. Renal structures were evaluated in relation to demographic, clinical, and biochemical variables using means/medians, standard deviations, and Pearson correlations.RESULTS:Reliability coefficients for MR renal and cyst volume measurements in phantoms were 99.9% and 89.2%, respectively. In the full-scale protocol, 241 ADPKD individuals (145 women and 96 men) were enrolled. Total renal, cyst, and % cyst volume were significantly greater in each decade group. Hypertensive individuals demonstrated greater renal, cyst, and % cyst volume than normotensive subjects. Age-adjusted renal (r = -0.31, P < 0.0001), cyst (r = -0.36, P < 0.0001), and % cyst volume (r = -0.35, P < 0.0001) were inversely related to glomerular filtration rate (GFR). Age-adjusted renal volume (r = 0.42, P < 0.0001), cystic (r = 0.39, P < 0.0001, and % cyst volume (r = 0.41, P < 0.0001) were related with urinary albumin excretion.CONCLUSION:MR measures of renal and cyst volume are reliable and accurate in patients with ADPKD. ADPKD is characterized by significant cystic involvement that increases with age. Structure (renal and cyst volume) and function (GFR) are inversely related and directly related with the presence of hypertension and urinary albumin excretion in individuals with normal renal function.

Eisner, R. L. Ph.D.; Brummer, M. E. M.S.; Hoffman, J. C. M.D.; Coumans, J. C. Ph.D. Author Information