Andrew David Scott¹, Malcolm Birch¹, and Marc Eric Miquel^1
1Clinical Physics, Barts and the London NHS Trust, London, United Kingdom

Real-time MR imaging of the soft-palate is challenging and gated techniques rely on sufficient similarity between repetitions of a speech task. The adaptive averaging technique selectively averages images acquired without gating, based on a similarity measure. We apply this technique to low SNR high frame-rate soft-palate imaging. Real-time mid-sagittal images were acquired while a subject repeatedly counted from 1-5 without gating or timing requirements. Adaptive averaging was retrospectively performed and improved image quality was demonstrated. The technique is also able to improve the quality of sufficiently long real-time acquisitions without repetition, due to the limited number of possible soft-palate configurations.

Liheng Guo¹, Ozan Sayin¹, J. Andrew Derbyshire², and Daniel A Herzka^1
1Biomedical Engineering, Johns Hopkins School of Medicine, Baltimore, MD, United States, ²Tornado Medical Systems, Toronto, Ontario, Canada

We propose a 2D self-gated dynamic imaging strategy that detects motion using near-center Cartesian phase encodes instead of dedicated navigator echoes, featuring 1) no motion-detection overhead to imaging data acquisition, 2) high motion-sampling rate and cine reconstruction frame rate both retrospectively and independently adjustable, and 3) minimal required knowledge about the motion prior to scan (no motion-dependent parameter to calculate prior to scan). Automatic retrospective reconstruction of self-gated cines have been successful for both breath-hold cardiac and musculoskeletal (knee) studies, showing the motion-detection capability and flexibility of our propose technique.

Johannes T. Schneider¹, Martin Blaimer², and Peter Ullmann^1
1Bruker BioSpin MRI GmbH, Ettlingen, Germany, ²Research Center Magnetic Resonance Bavaria, Würzburg, Germany

Parallel imaging reconstruction can benefit from utilizing conjugate k-space symmetry in combination with an optimal image background phase that contributes to sensitivity encoding. This study shows experimentally that parallel spatially selective excitation (PEX) is a suitable means to tailor the image background phase in a very flexible way in order to achieve optimal reconstruction performance. To this end, a procedure was developed to determine an optimal phase distribution yielding minimum g-factor values. Generating such phase distributions by PEX, g-factors and noise enhancement in SENSE and GRAPPA experiments were dramatically reduced and images with excellent quality were obtained.

Joshua D. Trzasko¹, and Armando Manduca^1
1Mayo Clinic, Rochester, MN, United States

In this work, we present a calibration-free locally low-rank encouraging reconstruction (CLEAR) strategy for accelerated parallel imaging applications. Whereas existing calibrationless parallel MRI methods operate entirely in k-space, using globally-constrained reconstruction models, our proposed strategy imposes constraints locally in the image domain. As we demonstrate, this approach offers substantial computational advantage, is very amenable to parallelized implementation, and naturally incorporates with Compressive Sensing-type sparsity constraints.

Kangrong Zhu¹, Adam Kerr¹, and John M. Pauly^1
1Stanford University, Stanford, CA, United States

In CAIPIRINHA, multiple slices are excited simultaneously and are shifted with respect to each other in the phase encoding direction. This is fulfilled by applying a particular phase cycle to each excited slice through the phase encoding steps. A separate reference scan is usually needed to reconstruct the aliased slices. In this work, the reconstruction problem in CAIPIRINHA is reformulated as the reconstruction problem of an undersampled 3D Cartesian dataset by exploiting the DFT encoding nature of the multiband pulses used for excitation. The aliased slices were reconstructed using 2D GRAPPA with autocalibrating signal acquired at the k-space center.

Steen Moeller¹, Junqian Xu¹, Edward J Auerbach¹, Essa Yacoub¹, and Kamil Ugurbil^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, Minnesota, United States

A Monte-Carlo type analysis with a fixed spectral frequency driving function is introduced as a new metric (L-factor) for determining residual signal leakage for slice separation from simultaneous multi-slice acquisitions. Compared with the geometry factor for noise amplification, the metric is shown to be a better quantifiable metric for high acceleration factors.

Manojkumar Saranathan¹, Dan Rettmann², Brian A Hargreaves¹, Sharon Clarke¹, and Shreyas S Vasanawala^1
1Radiology, Stanford University, Stanford, CA, United States, ²Global Applied Science Laboratory, GE Healthcare, Rochester, MN, United States

High spatio-temporal resolution is essential in hepatobiliary imaging to characterize lesion morphology and to assess contrast uptake. In addition, late arterial phase imaging is critical for hypervascular tumors such as hepatocellular carcinoma (HCC) and neuroendocrine metastases. We demonstrate feasibility and clinical performance of a new high spatio-temporal resolution technique called DISCO (DIfferential Sub-sampling with Cartesian Ordering) that combines a dual-echo SPGR sequence with pseudo-random variable density k-space segmentation and a view sharing reconstruction. A high spatial resolution of 1.1x1.5x3 mm over 60 slices was routinely achieved with a temporal resolution of ~4 seconds, enabling clear delineation of angiographic, hepatic arterial, hepatic venous and portal venous phases.

Philip J Beatty^1,2, James H Holmes³, Ann Shimakawa⁴, Howard A Rowley⁵, and Jean H Brittain^3
1Physical Sciences, Sunnybrook Research Institute, Toronto, Ontario, Canada, ²Global Applied Science Laboratory, GE Healthcare, Thornhill, Ontario, Canada, ³Global Applied Science Laboratory, GE Healthcare, Madison, Wisconsin, United States, ⁴Global Applied Science Laboratory, GE Healthcare, Menlo Park, California, United States, ⁵Radiology, University of Wisconsin-Madison, Madison, Wisconsin, United States

We propose a technique for achieving short echo time contrast with PROPELLER acquisitions using longer echo train lengths typically reserved for T2w imaging. The desired contrast is achieved by modifying how the oversampled PROPELLER data is weighted (i.e. density compensated). The method allows the use of fewer wider blades for image reconstruction compared to short ETL approaches, resulting in significant scan time reductions.

Noam Ben-Eliezer¹, Yoav Shrot², Daniel K. Sodickson¹, and Lucio Frydman^2
1Center for Biomedical Imaging, New-York University Medical Center, New-York, NY, United States, ²Chemical Physics, Weizmann Institute of Science, Rehovot, Israel

In recent years a conceptually different approach for collecting MR data has emerged, based on time progressive refocusing in the image spatial, rather than k-space, domain using quadratic phase functions. This novel approach, termed Spatiotemporal-Encoding (SPEN), allows users to overcome sizable field distortions by altogether freeing the spins’ evolution from T₂^* effects, thereby extending the reach of MRI into hitherto inaccessible regions. This work provides an analysis of the SPEN’s signal-to-noise ratio as compared to conventional k-space encoding, determining – theoretically, and experimentally – a parameter regime, where the respective SNR of these two encoding techniques is comparable.

Alexander Brunner¹, Lei Zheng², Alexander Zyl², Florian Maier¹, Wolfhard Semmler¹, Jürgen Hesser², and Michael Bock^1,3
1Dept. of Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Experimental Radiation Oncology, University Medical Center Mannheim, Heidelberg University, Mannheim, Germany, ³Department of Radiology - Medical Physics, University Hospital Freiburg, Freiburg, Germany

In real-time applications such as intravascular interventions the acquisition of 3D angiograms is challenging as both image acquisition and 3D reconstruction need to be performed in less than 1 s. In this work, we present a new acquisition and reconstruction concept for 3D angiograms which combines a fast projection acquisition pulse sequence with a multi-plane reconstruction. The approach was evaluated on a vessel bifurcation phantom and quantitatively compared to a high resolution dataset of the phantom. We demonstrated that the reconstruction of 3D-angiograms at frame rates suitable for real-time applications (e.g. intravascular interventions) is feasible.