1161
Highly Accelerated Prospectively Gated Free Breathing Dynamic Abdominal Imaging with Cartesian Sampling1Diagnostic Radiology, University of Texas MD Anderson Cancer Center, Houston, TX, United States, 2Global MR Applications & Workflow, GE Healthcare, Madison, WI, United States, 3Department of Imaging Physics, University of Texas MD Anderson Cancer Center, Houston, TX, United States, 4Global MR Applications & Workflow, GE Healthcare, Houston, TX, United States
Synopsis
Dynamic abdominal imaging has the requirements of high spatial-temporal resolution, large spatial coverage, and acquisition timing with the contrast injection. Golden angle radial sampling in combination with compressed sensing and parallel imaging has been reported for free breathing dynamic volumetric imaging. However, radial sampling suffers from streaking artifacts, inflexible FOV prescription, and long reconstruction time. In this study, we report a Cartesian sampling technique that combines compressed sensing, parallel imaging, temporal view sharing, and navigator for prospective gating. The potential of the technique for free breathing dynamic abdominal imaging is demonstrated in both volunteers and patients.
Introduction
Recent work on radial sampling with sparse reconstruction allowed high spatial-temporal free breathing dynamic imaging of the liver1. However, streaking artifacts and restrictions (long reconstruction times, square field of view and square pixels) from radial sparse sampling can be undesirable. We propose to address these limitations by incorporating into DISCO2 (Differential Subsampling with Cartesian Ordering, which is a high spatial-temporal Cartesian imaging technique utilizing view sharing and parallel imaging) the following enhancements: 1) spectrally selective inversion for fat suppression3, 2) automatically placed cylindrical navigator pulse4 for free breathing acquisition and 3) Compressed Sensing5 (CS) to compensate the scan time cost from navigators. The high efficiency of our proposed technique could enable a simplified and consistent workflow with multiple arterial phase capture, and potentially replacing the need for contrast monitoring with a fixed delay of 20 seconds after contrast injection. We demonstrate the feasibility of our technique in healthy volunteers and patients.Methods
In our proposed technique, randomly undersampled k-space points are segmented and each segment shares a spectrally selective fat suppression pulse and a navigator pulse for motion monitoring as shown in Fig 1. In the reconstruction, temporal view sharing is utilized first, followed by CS5 reconstruction of randomly undersampled data using L1-norm minimization of total variation and imposing sparsity iteratively to recover uniform k-space in the parallel imaging domain. Finally, data driven parallel imaging6 with auto-calibration signal (ACS) points produces the complete k-space data. Scan workflow is shown in Figure 2. A prospectively gated mask phase is collected in 18-24s. The scanner pauses for contrast injection and allows the technologist to inspect and confirm the image quality. Contrast injection is followed by a programmed delay of 20s and four to five phases of free breathing dynamic imaging with 8-10s of scan time per phase that can cover both arterial and venous phases. Finally, a delayed phase is collected. Temporal acceleration via view sharing is restricted only to the continuous scanning part of the acquisition to minimize temporal blurring. Reconstruction is online and recon lag per phase is less than 5 seconds, which allows nearly instant review of the image quality. Volunteers and patients were scanned on 3.0T scanners (Discovery MR 750w & Signa Premier, GE Healthcare, Waukesha, WI) using high density body receiver array coils under IRB approval. The following scan parameters were used: Axial scan plane, 3D SPGR sequence, TR/TE = 3.0/1.0 msec, rBW = 83 kHz, image resolution of 1.5x1.7 mm in-plane and 5 mm slice thickness interpolated to 2.5 mm, FOV = 40x32 cm, flip angle = 10o.Results
Figure 3 shows an example healthy volunteer image demonstrating the benefit of prospective navigation. Despite the substantial hit on scan efficiency (3s vs. 10s per phase), the use of navigator substantially minimized the breathing artifacts while allowing at least dual arterial phases in free breathing. Figure 4 shows example images from free breathing patient scanning. Although the images were all acquired post gadolinium contrast injection, the quality demonstrates the potential motion robustness if they were acquired for imaging contrast injection.Discussion
With a combination of CS, parallel imaging, and DISCO view sharing, coverage of the entire abdomen can be achieved with Cartesian sampling in about 3 to 4 seconds of acquisition time or only a couple of respiratory cycles in free breathing. Our results show that such a high acceleration can be exploited for free breathing dynamic imaging of the abdomen for better and more reliable capture of the first pass of contrast. Upon further validation, this technique may enable improved MRI of the liver, such as for the assessment of liver cancer or metastases, where high spatial resolution and high image quality over the whole liver is a significant unmet need.Conclusion
Free breathing dynamic abdominal imaging with prospectively gated highly accelerated Cartesian sampling is demonstrated in both volunteers and patients with streamlined workflow. Upon further validation, this technique may provide an alternative for patients who have limited breath-hold capacity or are under sedation.Acknowledgements
The authors would like to acknowledge Brandy Reed and Stacy Hash for their help in collecting the MR scans for this project.References
1. Feng L, Axel L, Chandarana H, Block KT, Sodickson DK, Otazo R. XD-GRASP: Golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing. Magn Reson Med. 2016;75(2):775-88.
2. Saranathan M, Rettmann DW, Hargreaves BA, Clarke SE, Vasanawala SS. Differential subsampling with cartesian ordering (DISCO): a high spatio-temporal resolution dixon imaging sequence for multiphasic contrast enhanced abdominal imaging. J Magn Reson Imaging 2012;35:1484–1492
3. Wang K, Takei N, Morrison C, Bancroft LH, Wang P, Holmes JH, Bayram E, Strigel RM, Kelcz F, and Korosec FR. Dynamic Contrast-Enhanced Breast MRI using A Chemically Fat-Suppressed View-Sharing Technique. In Proceedings of 25th Annual Meeting of ISMRM, Honolulu, HI, USA, 2017. #2120.
4. Goto, Takao, & Kabasawa, Hiroyuki (2010). Automated positioning of scan plane and navigator tracker in MRI liver scan. Medical Imaging Technology, (suppl), 8.
5. King K, Xu D, Brau AC, Lai P, Beatty PJ, Marinelli L. A New Combination of Compressed Sensing and Data Driven Parallel Imaging. In Proceedings of 18th Annual Meeting of ISMRM, Stockholm, Sweden, 2010. p. 4881.
6. Brau ACS, Beatty PJ, Skare S, Bammer R. Comparison of reconstruction accuracy and efficiency among autocalibrating data-driven parallel imaging methods. Mag Reson Med 2008;59:382–395.
Figures
