Synopsis

Stack of stars trajectory with golden angle ordering provides better motion robustness than Cartesian imaging for abdominal MRI. However, image reconstruction for non-Cartesian datasets is usually time-consuming, especially for datasets with high-density coil arrays. While additional motion correction methods can improve image quality for stack of stars, they often further increase the reconstruction time. In this work, we aim to reduce the reconstruction time for stack of stars using coil compression and improve motion robustness with a similar reconstruction time using soft gating.

Introduction

Methods

(1) Coil compression

Coil compression was applied to reduce the computation for stack of stars reconstruction with high-density coil arrays. A geometric-decomposition coil compression (GCC) method⁵ was used in this work, including a slice-by-slice coil compression and alignment of coil compression matrices along the slice direction. Compared with single coil compression (SCC)⁷, GCC exploits the spatially varying data redundancy of coil arrays and can compress data into fewer virtual coils. A similarly application of GCC has been demonstrated in cardiac imaging using stack of stars trajectory⁸. For standard reconstruction methods, including gridding and SENSE-type of parallel imaging⁹, the required computation is proportional to the number of coil elements, therefore the reconstruction time can be reduced by $$$\frac{n_c}{n_{vc}}$$$ using coil compression, where $$$n_c$$$ is the original number of coils and $$$n_{vc}$$$ is the number of virtual coils.

(2) Soft gating

Soft gating has been proposed recently for effective motion compensation, in which a motion-weighted data consistency is applied⁶. Soft gating can be easily combined with parallel imaging⁶ and compressed sensing¹⁰ without lengthening the reconstruction time. In this work, self-navigators were first obtained from the DC signal of the stack of stars datasets for each coil element. Coil clustering¹¹ was then applied to find the dominant motion, and the soft gating weights were generated accordingly.

To validate the proposed methods, an axial free-breathing volunteer dataset was acquired with IRB approval on a GE MR750 scanner using a 20-channel body coil (GE Healthcare, Waukesha, USA) and the following imaging parameters: TE/TR = 1.7/3.6 ms, FOV = 42 cm, slice thickness 4 mm, 54 slices, and 256 spokes per slice. First, both SCC and GCC with 8 virtual coils and gridding reconstruction were evaluated and compared. Different reconstruction methods were then compared, including gridding, parallel imaging (without and with GCC), and soft-gated parallel imaging (without and with GCC). A conjugate gradient (CG) algorithm⁹ was used in both parallel imaging and soft-gated parallel imaging. The number of CG iterations was kept constant for all methods. All reconstructions were performed in Matlab (Mathworks, MA, USA) on a 2015 MacBook Pro (Apple, Cupertino, USA).

Results

(1) Coil compression

Coil compression results are shown in Fig. 1. GCC has effectively compressed the original data into 8 virtual coils without noticeable compression loss. Compared to SCC, better performance was observed using GCC.

(2) Soft gating

Reconstructed images by all methods are shown in Fig. 2, and reformatted into the coronal plane in Fig. 3. Recorded reconstruction times are shown in Tab. 1. GCC significantly reduced the reconstruction time for parallel imaging and soft-gated parallel imaging. As shown in Fig. 2 and Fig. 3, soft gating improved motion robustness without increasing the reconstruction time.

Conclusion

Acknowledgements

References

1. Winkelmann S, Schaeffter T, Koehler T, Eggers H, Doessel O. An optimal radial profile order based on the golden ratio for time-resolved MRI. IEEE Trans Med Imaging 2007;26:68-76.

2. Feng L, Grimm R, Block KT, Chandarana H, Kim S, Xu J, Axel L, Sodickson DK, Otazo R. Golden-angle radial sparse parallel MRI: combination of compressed sensing, parallel imaging, and golden-angle radial sampling for fast and flexible dynamic volumetric MRI. Magn Reson Med 2014;72:707-717.

3. Cruz G, Atkinson D, Buerger C, Schaeffter T, Prieto C. Accelerated motion corrected three-dimensional abdominal MRI using total variation regularized SENSE reconstruction. Magn Reson Med 2016;75:1484-1498.

4. Yang HJ, Sharif B, Pang J, Kali A, Bi X, Cokic I, Li D, Dharmakumar R. Free-breathing, motion-corrected, highly efficient whole heart T2 mapping at 3T with hybrid radial-cartesian trajectory. Magn Reson Med 2016;75:126-136.

5. Zhang T, Pauly JM, Vasanawala SS, Lustig M. Coil compression for accelerated imaging with Cartesian sampling. Magn Reson Med 2013;69:571-582.

6. Johnson KM, Block WF, Reeder SB, Samsonov A. Improved least squares MR image reconstruction using estimates of k-space data consistency. Magn Reson Med 2012;67:1600-1608.

7. Huang F, Vijayakumar S, Li Y, Hertel S, Duensing GR. A software channel compression technique for faster reconstruction with many channels. Magn Reson Imaging 2008;26:133-141.

8. Kawaji K, Foppa M, Roujol S, Akcakaya M, Nezafat R. Whole heart coronary imaging with flexible acquisition window and trigger delay. PLoS One 2015; 10:e0112020.

9. Pruessmann KP, Weiger M, Börnert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med 2001;46:638-651.

10. Cheng JY, Zhang T, Ruangwattanapaisarn N, Alley MT, Uecker M, Pauly JM, Lustig M, Vasanawala SS. Free-breathing pediatric MRI with nonrigid motion correction and acceleration. J Magn Reson Imaging 2015;42:407-420.

11. Zhang T, Cheng JY, Chen Y, Nishimura DG, Pauly JM, Vasanawala SS. Robust self-navigated body MRI using dense coil arrays. Magn Reson Med 2016;76:197-205.