Optimizing spatiotemporal sampling for k-t BLAST and k-t SENSE: application to high-resolution real-time cardiac steady-state free precession.

In k-t BLAST and k-t SENSE, data acquisition is accelerated by sparsely sampling k-space over time. This undersampling in k-t space causes the object signals to be convolved with a point spread function in x-f space (x = spatial position, f = temporal frequency). The resulting aliasing is resolved by exploiting spatiotemporal correlations within the data. In general, reconstruction accuracy can be improved by controlling the k-t sampling pattern to minimize signal overlap in x-f space. In this work, we describe an approach to obtain generally favorable patterns for typical image series without specific knowledge of the image series itself. These optimized sampling patterns were applied to free-breathing, untriggered (i.e., real-time) cardiac imaging with steady-state free precession (SSFP). Eddy-current artifacts, which are otherwise increased drastically in SSFP by the undersampling, were minimized using alternating k-space sweeps. With the synergistic combination of the k-t approach with optimized sampling and SSFP with alternating k-space sweeps, it was possible to achieve a high signal-to-noise ratio, high contrast, and high spatiotemporal resolutions, while achieving substantial immunity against eddy currents. Cardiac images are shown, demonstrating excellent image quality and an in-plane resolution of approximately 2.0 mm at >25 frames/s, using one or more receiver coils.     

On the influence of training data quality in k-t BLAST reconstruction.

This work investigated how the quality of prior information (i.e., data acquired during the training stage) influences k-t BLAST reconstruction. The impact of several factors, such as the amount of training data, the presence of spatial misregistration in the training data, and the effects of filtering, was investigated with simulations and in vivo data. It is shown that k-t BLAST outperforms sliding window reconstruction, even with very limited training data. By increasing the amount of training data, reconstruction error continues to decrease, albeit by a diminishing amount. However, an increased amount of training data also increases susceptibility to misregistration of the training data. Filtering of the training data with the goal of reducing truncation artifacts had only minor impact on reconstruction errors. Considering the balance among obtaining the most benefit from the training data, minimizing susceptibility to misregistration, and keeping data acquisition to a minimum, it is concluded that in cardiac imaging the training datasets should be limited to 10-20 profiles in k-space for a typical field of view. The training data may be acquired in a separate breathhold without much penalty, if care is taken to minimize misregistration, such as with a navigator.     

Accelerating cine phase-contrast flow measurements using k-t BLAST and k-t SENSE.

Conventional phase-contrast velocity mapping in the ascending aorta was combined with k-t BLAST and k-t SENSE. Up to 5.3-fold net acceleration was achieved, enabling single breath-hold acquisitions. A standard phase-contrast (PC) sequence with interleaved acquisition of the velocity-encoded segments was modified to collect data in 2 stages, a high-resolution under sampled and a low-resolution fully sampled training stage. In addition, a modification of the k-t reconstruction strategy was tested. This strategy, denoted as "plug-in," incorporates data acquired in the training stage into the final reconstruction for improved data consistency, similar to conventional keyhole. "k-t SENSE plug-in" was found to provide best image quality and most accurate flow quantification. For this strategy, at least 10 training profiles are required to yield accurate stroke volumes (relative deviation <5%) and good image quality. In vivo 2D cine velocity mapping was performed in 6 healthy volunteers with 30-32 cardiac phases (spatial resolution 1.3 x 1.3 x 8-10 mm(3), temporal resolution of 18-38 ms), yielding relative stroke volumes of 106 +/- 18% (mean +/- 2*SD) and 112 +/- 15% for 3.8 x and 5.3 x net accelerations, respectively. In summary, k-t BLAST and k-t SENSE are promising approaches that permit significant scan-time reduction in PC velocity mapping, thus making high-resolution breath-held flow quantification possible.     

Comparison of k‐t SENSE/k‐t BLAST with conventional SENSE applied to BOLD fMRI

Purpose:

To compare k‐t BLAST (broad‐use linear‐acquisition speedup technique)/k‐t SENSE (sensitivity encoding) with conventional SENSE applied to a simple fMRI paradigm.

Materials and Methods:

Blood oxygen level‐dependent (BOLD) functional magnetic resonance imaging (fMRI) was performed at 3 T using a displaced ultra‐fast low‐angle refocused echo (UFLARE) pulse sequence with a visual stimulus in a block paradigm. Conventional SENSE and k‐t BLAST/k‐t SENSE data were acquired. Also, k‐t BLAST/k‐t SENSE was simulated at different undersampling factors from fully sampled data after removal of lines of k‐space data. Analysis was performed using SPM5.

Results:

Sensitivity to the BOLD response in k‐t BLAST/k‐t SENSE was comparable with that of SENSE in images acquired at an undersampling factor of 2.3. Simulated k‐t BLAST/k‐t SENSE yielded reliable detection of activation‐induced BOLD contrast at undersampling factors of 5 or less. Sensitivity increased significantly when training data were included in k‐space before Fourier transformation (known as “plug‐in”).

Conclusion:

k‐t BLAST/k‐t SENSE performs at least as well as conventional SENSE for BOLD fMRI at a modest undersampling factor. Results suggest that sufficient sensitivity to BOLD contrast may be achievable at higher undersampling factors with k‐t BLAST/k‐t SENSE than with conventional parallel imaging approaches, offering particular advantages at the highest magnetic field strengths. J. Magn. Reson. Imaging 2010;32:235–241. © 2010 Wiley‐Liss, Inc.     

Cartesian SENSE and k-t SENSE reconstruction using commodity graphics hardware.

This study demonstrates that modern commodity graphics cards (GPUs) can be used to perform fast Cartesian SENSE and k-t SENSE reconstruction. Specifically, the SENSE inversion is accelerated by up to two orders of magnitude and is no longer the time-limiting step. The achieved reconstruction times are now well below the acquisition times, thus enabling real-time, interactive SENSE imaging, even with a large number of receive coils. The fast GPU reconstruction is also beneficial for datasets that are not acquired in real time. We demonstrate that it can be used for interactive adjustment of regularization parameters for k-t SENSE in the same way that one would adjust window and level settings. This enables a new way of performing imaging reconstruction, where the user chooses the setting of tunable reconstruction parameters, in real time, depending on the context in which the images are interpreted.     

Accelerating cardiac cine 3D imaging using k-t BLAST.

By exploiting spatiotemporal correlations in cardiac acquisitions using k-t BLAST, gated cine 3D acquisitions of the heart were accelerated by a net factor of 4.3, making single breathhold acquisitions possible. Sparse sampling of k-t space along a sheared grid pattern was implemented into a cine 3D SSFP sequence. The acquisition of low-resolution training data, which was required to resolve aliasing in the k-t BLAST method, was either interleaved into the sampling process or obtained in a separate prescan to allow for shorter breathhold durations in patients with heart disease. Volumetric datasets covering the heart with 20 slices at a spatial resolution of 2 x 2 x 5 mm3 were recorded with 20 cardiac phases in a total breathhold duration of 25-27 sec, or 18 sec if partial Fourier sampling was additionally employed. The feasibility of the method was demonstrated on healthy volunteers and on patients. The comparison of endocardial area derived from single slices of the 3D dataset with values extracted from separate single-slice acquisitions showed no significant differences. By shortening the acquisition substantially, k-t BLAST may greatly facilitate volumetric imaging of the heart for evaluation of regional wall motion and the assessment of ventricular volume and ejection fraction.     

k-t BLAST reconstruction from non-Cartesian k-t space sampling.

Current implementations of k-t Broad-use Linear Acqusition Speed-up Technique (BLAST) require the sampling in k-t space to conform to a lattice. To permit the use of k-t BLAST with non-Cartesian sampling, an iterative reconstruction approach is proposed in this work. This method, which is based on the conjugate gradient (CG) method and gridding reconstruction principles, can efficiently handle data that are sampled along non-Cartesian trajectories in k-t space. The approach is demonstrated on prospectively gated radial and retrospectively gated Cartesian imaging. Compared to a sliding window (SW) reconstruction, the resulting image series exhibit lower artifact levels and improved temporal fidelity. The proposed approach thus allows investigators to combine the specific advantages of non-Cartesian imaging or retrospective gating with the acceleration provided by k-t BLAST.     

k-t2 BLAST: exploiting spatiotemporal structure in simultaneously cardiac and respiratory time-resolved volumetric imaging.

Multidimensional imaging resolving both the cardiac and respiratory cycles simultaneously has the potential to describe important physiological interdependences between the heart and pulmonary processes. A fully five-dimensional acquisition with three spatial and two temporal dimensions is hampered, however, by the long acquisition time and low spatial resolution. A technique is proposed to reduce the scan time substantially by extending the k-t BLAST framework to two temporal dimensions. By sampling the k-t space sparsely in a lattice grid, the signal in the transform domain, x-f space, can be densely packed, exploiting the fact that large regions in the field of view have low temporal bandwidth. A volumetric online prospective triggering approach with full cardiac and respiratory cycle coverage was implemented. Retrospective temporal interpolation was used to refine the timing estimates for the center of k-space, which is sampled for all cardiac and respiratory time frames. This resulted in reduced reconstruction error compared with conventional k-t BLAST reconstruction. The k-t(2) BLAST technique was evaluated by decimating a fully sampled five-dimensional data set, and feasibility was further demonstrated by performing sparsely sampled acquisitions. Compared to the fully sampled data, a fourfold improvement in spatial resolution was accomplished in approximately half the scan time.     

Fast dynamic contrast-enhanced lung MR imaging using k-t BLAST: a spatiotemporal perspective

Dynamic contrast-enhanced MR imaging has long been an attractive alternative to measure pulmonary perfusion as it offers simultaneous acquisition of high-resolution anatomical images and various functional information without exposing to ionizing radiation. As higher temporal resolution in addition to simultaneous acquisition of more slices from different positions favors more precise diagnosis, rapid acquisition of multiple images during bolus contrast administration remains essential to pulmonary perfusion imaging. Nevertheless, the branching morphology together with asynchronization of contrast-enhanced pulmonary perfusion scattered among distinct blood vessels imposes difficulties to faster imaging. This work demonstrates that k-t broad-use linear acquisition speed-up technique (k-t BLAST), having substantial performance on accelerating cardiac cine imaging, can be applied to accelerate dynamic contrast-enhanced lung imaging up to a factor of 5 with errors less than 6% on five healthy subjects and less than 10% on 13 patients, respectively, in the overall signal intensity. Perfusion parameter estimates show somewhat less errors than those in overall signal intensity. Results from healthy subjects and two groups of patients with various diseases show high consistency between fully sampled datasets and their accelerated counterparts. These suggest feasibility of accelerated contrast-enhanced lung images in clinical examinations and potential of extending k-t BLAST into related applications.     

Accelerated cardiac perfusion imaging using k‐t SENSE with SENSE training

In k‐t sensitivity encoding (SENSE), MR data acquisition performed in parallel by multiple coils is accelerated by sparsely sampling the k‐space over time. The resulting aliasing is resolved by exploiting spatiotemporal correlations inherent in dynamic images of natural objects. In this article, a modified k‐t SENSE reconstruction approach is presented, which aims at improving the temporal fidelity of first‐pass, contrast‐enhanced myocardial perfusion images at high accelerations. The proposed technique is based on applying parallel imaging on the training data in order to increase their spatial resolution. At a net acceleration of 5.8 (k‐t factor = 8, training profiles = 11) accurate representations of dynamic signal‐intensities were achieved. The efficacy of this approach as well as limitations due to noise amplification were investigated in computer simulations and in vivo experiments. Magn Reson Med, 2009. © 2009 Wiley‐Liss, Inc.     

k-t GRAPPA: a k-space implementation for dynamic MRI with high reduction factor.

A novel technique called "k-t GRAPPA" is introduced for the acceleration of dynamic magnetic resonance imaging. Dynamic magnetic resonance images have significant signal correlations in k-space and time dimension. Hence, it is feasible to acquire only a reduced amount of data and recover the missing portion afterward. Generalized autocalibrating partially parallel acquisitions (GRAPPA), as an important parallel imaging technique, linearly interpolates the missing data in k-space. In this work, it is shown that the idea of GRAPPA can also be applied in k-t space to take advantage of the correlations and interpolate the missing data in k-t space. For this method, no training data, filters, additional parameters, or sensitivity maps are necessary, and it is applicable for either single or multiple receiver coils. The signal correlation is locally derived from the acquired data. In this work, the k-t GRAPPA technique is compared with our implementation of GRAPPA, TGRAPPA, and sliding window reconstructions, as described in Methods. The experimental results manifest that k-t GRAPPA generates high spatial resolution reconstruction without significant loss of temporal resolution when the reduction factor is as high as 4. When the reduction factor becomes higher, there might be a noticeable loss of temporal resolution since k-t GRAPPA uses temporal interpolation. Images reconstructed using k-t GRAPPA have less residue/folding artifacts than those reconstructed by sliding window, much less noise than those reconstructed by GRAPPA, and wider temporal bandwidth than those reconstructed by GRAPPA with residual k-space. k-t GRAPPA is applicable to a wide range of dynamic imaging applications and is not limited to imaging parts with quasi-periodic motion. Since only local information is used for reconstruction, k-t GRAPPA is also preferred for applications requiring real time reconstruction, such as monitoring interventional MRI.     

k‐t PCA: Temporally constrained k‐t BLAST reconstruction using principal component analysis

The k‐t broad‐use linear acquisition speed‐up technique (BLAST) has become widespread for reducing image acquisition time in dynamic MRI. In its basic form k‐t BLAST speeds up the data acquisition by undersampling k‐space over time (referred to as k‐t space). The resulting aliasing is resolved in the Fourier reciprocal x‐f space (x = spatial position, f = temporal frequency) using an adaptive filter derived from a low‐resolution estimate of the signal covariance. However, this filtering process tends to increase the reconstruction error or lower the achievable acceleration factor. This is problematic in applications exhibiting a broad range of temporal frequencies such as free‐breathing myocardial perfusion imaging. We show that temporal basis functions calculated by subjecting the training data to principal component analysis (PCA) can be used to constrain the reconstruction such that the temporal resolution is improved. The presented method is called k‐t PCA. Magn Reson Med, 2009. © 2009 Wiley‐Liss, Inc.     

Temporal filtering effects in dynamic parallel MRI

Autocalibrated parallel MRI methods such as TSENSE or k‐t SENSE have been presented for dynamic imaging studies as they are able to provide images with high temporal resolution. One key element of these techniques is the temporal averaging of the undersampled raw data to obtain an unaliased image. This image represents the temporal average (also known as direct current, DC) and is used to derive the reconstruction parameters. In this work, we show that aliasing artifacts can be introduced in the DC signal obtained from the undersampled raw data. These artifacts lead to undesired temporal filtering effects when the DC signal is used for coil sensitivity calibration or when the DC signal is subtracted from the raw data. It is demonstrated that the temporal filtering effects can be reduced significantly by filtering the DC signal. Magn Reson Med, 2011. © 2011 Wiley‐Liss, Inc.     

Dynamic contrast-enhanced myocardial perfusion MRI accelerated with k-t sense.

In the k-t sensitivity encoding (k-t SENSE) method spatiotemporal data correlations are exploited to accelerate data acquisition in dynamic MRI studies. The present study demonstrates the feasibility of applying k-t SENSE to contrast-enhanced myocardial perfusion MRI and using the speed-up to increase spatial resolution. At a net acceleration factor of 3.9 (k-t factor of 5 with 11 training profiles) accurate representations of dynamic signal intensity (SI) changes were achieved in computer simulations. In vivo, 5x k-t SENSE was compared with 2x SENSE (identical acquisition parameters except for in-plane spatial resolution = 1.48 x 1.48 mm(2) vs. 2.64 x 2.64 mm(2), respectively). In 10 volunteers no differences in myocardial SI profiles were found (relative peak enhancement = 151% vs. 149.7%, maximal upslope = 12.9%/s vs. 13.3%/s for 2x SENSE and 5x k-t SENSE, respectively, all P > 0.05). Overall image quality was similar, but endocardial dark rim artifacts were reduced with k-t SENSE. Signal-to-noise ratio (SNR) in the myocardium was greater with 5x k-t SENSE by a factor of 1.36 +/- 0.23 at peak contrast enhancement with the relative yield decreasing with increasing dynamics in the object in accordance to theory. Higher nominal acceleration factors of up to 10-fold were shown to be feasible in computer simulations and in vivo.     

Accelerated phase‐contrast MR imaging: Comparison of k‐t BLAST with SENSE and doppler ultrasound for velocity and flow measurements in the aorta

Purpose

To evaluate differences in velocity and flow measurements in the aorta between accelerated phase‐contrast (PC) magnetic resonance imaging (MRI) using SENSE and k‐t BLAST and in peak velocity to Doppler ultrasound.

Materials and Methods

Two‐dimensional PC‐MRI perpendicular to the ascending and descending aorta was performed in 11 volunteers using SENSE (R = 2) and k‐t BLAST (2‐, 4‐, 6‐, and 8‐fold). Peak velocity, mean velocity, and stroke volume of the accelerated PC‐MRI experiments were correlated. Peak velocities were compared to Doppler ultrasound.

Results

All acceleration techniques showed significant correlations for peak velocity with Doppler ultrasound. However, k‐t BLAST 6 and 8 showed a significant underestimation. Strong correlations between SENSE and k‐t BLAST were found for all three parameters. Significant differences in peak velocity were found between SENSE and all k‐t BLAST experiments, but not for 2‐fold k‐t BLAST in the ascending aorta, and 2‐ and 4‐fold k‐t BLAST in the descending aorta. For mean velocity no significant differences were found. Stroke volume showed significant differences for all k‐t BLAST experiments in the ascending and for 6‐ and 8‐fold k‐t BLAST in the descending aorta.

Conclusion

Peak velocity of accelerated PC‐MRI correlated with CW Doppler measurements, but high k‐t BLAST acceleration factors lead to a significant underestimation. SENSE with R = 2 and 2‐fold k‐t BLAST are most highly correlated in phase‐contrast flow measurements. J. Magn. Reson. Imaging 2009;29:817–824. © 2009 Wiley‐Liss, Inc.     

Feasibility of k-t BLAST technique for measuring "seven-dimensional" fluid flow

PURPOSE: To investigate the feasibility of rapid MR measurement of "seven-dimensional" (three velocity components, three dimensions, and time) fluid flow using the k-t Broad-use Linear Acquisition Speed-Up Technique (BLAST). MATERIALS AND METHODS: Complete k-space data were acquired for pulsatile fluid flow in a model of a stenosed carotid bifurcation. The data was subsampled to simulate "training" and "accelerated acquisition" data for reconstruction using k-t BLAST. RESULTS: Flow waveforms estimated from k-t BLAST reconstructions were in good agreement with those measured from the full data set for overall speedup factors up to approximately four times when slice-by-slice undersampling in k(y) was used. Accuracy was better than 25 mm/second or 7% (root-mean-square error) for individual time frames under these conditions. Flow patterns in the plane of symmetry, near the bifurcation, and in the stenosis were also in good agreement with those reconstructed from the full data set. Improved performance was obtained from undersampling in both k(y) and k(z), when acceleration factors up to 12 times gave acceptable results. CONCLUSION: The k-t BLAST technique can be applied to flow quantification, and may make feasible the acquisition of time-resolved blood flow from extended arterial regions within acceptable examination times.     

Single-breathhold four-dimensional assessment of left ventricular volumes and function using k-t BLAST after application of extracellular contrast agent at 3 Tesla

PURPOSE: To prospectively determine the feasibility and accuracy of a four-dimensional (4D) k-space over time broad-use linear acquisition speed-up technique (k-t BLAST) for the evaluation of left ventricular (LV) volumes in comparison to standard multiple-breathhold cine imaging, using a 3.0 Tesla (3T) MR system. MATERIALS AND METHODS: In 23 subjects, short-axis cine loops completely covering the LV were acquired using conventional turbo gradient echo (GRE) imaging. Immediately after administration of gadobenate dimeglumine, a rapid single-breathhold k-t BLAST 4D dataset with the same coverage was acquired and reconstructed to short-axis views. Quantitative aortic flow measurement for LV stroke volume (LVSV) was used to calibrate both techniques. For GRE and k-t BLAST cine imaging: LV volumes, ejection fraction (EF), and blood-to-myocardium-contrast (BMC) were determined. RESULTS: k-t BLAST and GRE sequences showed a strong correlation for LV volumes and EF (r = 0.97-0.99; P < 0.001). Excellent agreement was also found between the LVSV determined by aortic flow measurements and LVSV assessed using GRE sequence and k-t BLAST sequence. BMC of GRE was similar to that of k-t BLAST cine imaging. CONCLUSION: The use of the single-breathhold 4D k-t BLAST technique for the assessment of LV volume is feasible and accurate in 3T MRI.     

Self-calibration method for radial GRAPPA/k-t GRAPPA.

Generalized autocalibrating partially parallel acquisitions (GRAPPA), an important parallel imaging technique, can be easily applied to radial k-space data by segmenting the k-space. The previously reported radial GRAPPA method requires extra calibration data to determine the relative shift operators. In this work it is shown that pseudo-full k-space data can be generated from the partially acquired radial data by filtering in image space followed by inverse gridding. The relative shift operators can then be approximated from the pseudo-full k-space data. The self-calibration method using pseudo-full k-space data can be applied in both k and k-t space. This technique avoids the prescans and hence improves the applicability of radial GRAPPA to image static tissue, and makes k-t GRAPPA applicable to radial trajectory. Experiments show that radial GRAPPA calibrated with pseudo-full calibration data generates results similar to radial GRAPPA calibrated with the true full k-space data for that image. If motion occurs during acquisition, self-calibrated radial GRAPPA protects structural information better than externally calibrated GRAPPA. However, radial GRAPPA calibrated with pseudo-full calibration data suffers from residual streaking artifacts when the reduction factor is high. Radial k-t GRAPPA calibrated with pseudo-full calibration data generates reduced errors compared to the sliding-window method and temporal GRAPPA (TGRAPPA).     

Assessment of left ventricular volumes and mass with fast 3D cine steady-state free precession k-t space broad-use linear acquisition speed-up technique (k-t BLAST)

PURPOSE: To compare left ventricular (LV) volume and mass assessment using two-dimensional (2D) cine steady-state free precession (SSFP) and k-t space broad-use linear acquisition speed-up technique (k-t BLAST) accelerated 3D magnetic resonance imaging (MRI). MATERIALS AND METHODS: On a commercially available 1.5T MR scanner, 2D cine SSFP, six- and eight-fold accelerated 3D k-t BLAST were performed to evaluate LV volumes and mass in 17 volunteers. After semiautomatic segmentation of the different MR data sets, the resulting volumes and mass were compared according to the mean difference, 95% confidence interval, standard deviation (SD), Pearson's correlation coefficient, Bland-Altman analysis, and the Pitman-Morgan test. RESULTS: Data acquisition was successful in all subjects. The number of required breathholds was reduced from a maximal of five for the 2D cine SSFP sequence to two for 3D k-t BLAST sequences. Comparing LV volumes, there was excellent agreement between 2D and 3D cine 8x k-t BLAST SSFP volumes (mean difference +/- 2SD end-diastolic volume [EDV] = 5 +/- 8 mL, end-systolic volume [ESV] = 1 +/-12 mL, and stroke volume [SV] = 3 +/- 8 mL), and mass (-1.8 +/- 9 g). CONCLUSION: k-t BLAST-accelerated 3D sequences allow accurate assessment of LV volumes and mass compared to 2D cine SSFP. This method may reduce costs and increase patient comfort due to shortened data acquisition time and reduced number of breathholds.     

Parallel MRI with extended and averaged GRAPPA kernels (PEAK-GRAPPA): optimized spatiotemporal dynamic imaging

Purpose

To evaluate an optimized k‐t‐space related reconstruction method for dynamic magnetic resonance imaging (MRI), a method called PEAK‐GRAPPA (Parallel MRI with Extended and Averaged GRAPPA Kernels) is presented which is based on an extended spatiotemporal GRAPPA kernel in combination with temporal averaging of coil weights.

Materials and Methods

The PEAK‐GRAPPA kernel consists of a uniform geometry with several spatial and temporal source points from acquired k‐space lines and several target points from missing k‐space lines. In order to improve the quality of coil weight estimation sets of coil weights are averaged over the temporal dimension.

Results

The kernel geometry leads to strongly decreased reconstruction times compared to the recently introduced k‐t‐GRAPPA using different kernel geometries with only one target point per kernel to fit. Improved results were obtained in terms of the root mean square error and the signal‐to‐noise ratio as demonstrated by in vivo cardiac imaging.

Conclusion

Using a uniform kernel geometry for weight estimation with the properties of uncorrelated noise of different acquired timeframes, optimized results were achieved in terms of error level, signal‐to‐noise ratio, and reconstruction time. J. Magn. Reson. Imaging 2008;28:1226–1232. © 2008 Wiley‐Liss, Inc.