Diffusion: Novel Acquisition

Wednesday 14 May 2014

Carl-Fredrik Westin^1,2, Markus Nilsson³, Filip Szczepankiewicz⁴, Ofer Pasternak¹, Evren Ozarslan¹, Daniel Topgaard⁵, and Hans Knutsson^2
1Radiology, Brigham and Women's, Harvard Medical School, Boston, MA, United States, ²Department of Biomedical Engineering, Linkoping University, Linkoping, Sweden, ³Lund University Bioimaging Center, Lund University, Lund, Sweden, ⁴Department of Medical Radiation Physics, Lund University, Lund, Sweden, ⁵Center for Chemistry and Chemical Engineering, Lund University, Lund, Sweden

Here we study how the choice of diffusion MRI gradient modulation schemes define the geometry and dimensionality of the diffusion encoding. By defining a diffusion encoding tensor, or diffusion measurement tensor, we will in the Gaussian approximation regime bring concepts as single, double, and triple-PFG into a common framework. Careful design of the speed and shape of the q-space trajectory can produce a planar diffusion encoding that is isotropic in the selected plane. The presented work shows that it is possible to perform in vivo circular diffusion encoding imaging of the human brain with a good SNR.

Steven Baete^1,2 and Fernando Emilio Boada^1,2
1Center for Biomedical Imaging, Dept. of Radiology, NYU Langone Medical Center, New York, New York, United States, ²CAI2R, Center for Advanced Imaging Innovation and Research, NYU Langone Medical Center, New York, New York, United States

In the conventional rectangular sampling of q-space in Diffusion Spectrum Imaging, the angular resolution attainable is proportional to the number of shells and the highest b-value acquired. Hence, increasing angular resolution results in acquiring more samples and increasing acquisition time. In this abstract, the angular resolution of the recently proposed radial q-space sampling is shown to be independent of the number of shells acquired above a certain threshold. This results in improved DSI reconstructions at a lower number of shells and at shorter acquisition times, as illustrated by in vivo brain data and computer simulations.

In this work, we propose a unified compressed sensing framework to reconstruct the diffusion signals and propagators from raw data sub-sampled from both q-space and k-space. With the proposed method, we can reduce scanning time in both k-space and q-space, and obtain good reconstruction quality with low RMSE compared to the estimation using fully sampled data.

Tim Sprenger^1,2, Jonathan I. Sperl¹, Brice Fernandez³, Vladimir Golkov¹, Ek Tsoon Tan⁴, Christopher Hardy⁴, Luca Marinelli⁴, Michael Czisch⁵, Philipp Saemann⁵, Axel Haase², and Marion I. Menzel^1
1GE Global Research, Munich, Germany, ²Technical University Munich, Munich, Germany, ³GE Healthcare, Munich, Germany, ⁴GE Global Research, Niskayuna, NY, United States, ⁵Max Planck Institute of Psychiatry, Munich, Germany

In Diffusion Kurtosis Imaging (DKI), the data is sampled in a series of concentric shells in the diffusion encoding space (q-space) and usu-ally suffers from low SNR. In this work a novel acquisition scheme for kurtosis imaging is presented, based on undersampled diffusion spectrum imaging (DSI) followed by a compressed sensing (CS) reconstruction of q-space. The undersampling thereby yields a compara-ble number of q-space sampling points as in standard DKI schemes whereas the CS-based denoising is shown to improve stability and accuracy of the kurtosis tensor estimation.

Yu-Chien Wu^1,2 and Chandana Kodiweera^1
1Dartmouth College, Hanover, New Hampshire, United States, ²Radiology and Imaging Sciences, Indiana University, Indianapolis, Indiana, United States

We propose a novel sequence, RSA, for fast and high spatial resolution advanced Diffusion-Weighted Imaging (DWI), which has been used widely to assess the integrity and directional information of white matter. Conventionally, DWI uses single-shot spin-echo EPI sequences. However, SS-EPI suffers from geometric distortion and long echo time (TE) with high spatial resolution (e.g, cubic voxel size of 1mm). Moreover, approaches using high and/or multiple b-values demand adequate SNR for accurate estimations. Because only one short-axis blade is acquired per DW direction in the RSA sequence, it reduces geometric distortion, shortens TE to increase signals, and reduces the scan time.

Rafael O'Halloran¹, Murat Aksoy¹, Eric Aboussouan¹, Eric Peterson¹, Anh Van^1,2, and Roland Bammer^1
1Radiology, Stanford University, Stanford, CA, United States, ²Zentralinstitut für Medizintechnik, Technische Universität München, Munich, Germany

While single-shot EPI remains the workhorse of clinical diffusion-weighted imaging, diffusion-weighted steady state free precession (DW-SSFP) imaging has some attractive properties as acquisitions are pushed to higher-resolution. The price of imaging in the steady state, however, is the undesirable propagation of uncorrected phase errors into the phase coherence pathway tree leading to loss of signal and contrast. The only way to combat this effect is to prospectively correct the phase. In this work an approach to prospectively correct rigid-body-motion induced phase errors in real time is presented and demonstrated to be effective in human volunteers.

Tim Schakel¹, Hans Hoogduin², and Marielle Philippens^1
1Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ²Radiology, UMC Utrecht, Utrecht, Netherlands

Single shot DW-EPI suffers from geometric distortions and is limited in resolution. A preparation phase (Diffusion sensitized driven equilibrium (DSDE)) which restores the diffusion weighted magnetization along the longitudinal axis allows for additional readout strategies such as TFE. DSDE-TFE can produce distortion free, high resolution diffusion weighted images, also in otherwise difficult areas to perform diffusion weighted imaging such as the neck.

Bertram L. Koelsch^1,2, Galen D. Reed^1,2, Kayvan R. Keshari³, Robert A. Bok¹, Daniel B. Vigneron^1,2, John Kurhanewicz^1,2, and Peder E. Z. Larson^1,2
1Department of Radiology and Biomedical Imaging, UCSF, San Francisco, CA, United States, ²UC Berkeley - UCSF Graduate Program in Bioengineering, San Francisco, CA, United States, ³Memorial Sloan-Kettering Cancer Center, NY, United States

Apparent diffusion coefficient (ADC) maps of water have become useful for cancer identification and characterization. The growing field of hyperpolarized¹³C has also proven to be useful in identifying tumors by measuring the real-time metabolism of hyperpolarized ¹³C pyruvate to lactate. In this study, we developed a novel diffusion-weighted hyperpolarized ¹³C EPI sequence on a clinical 3T MR scanner to rapidly obtain ADC maps of hyperpolarized ¹³C metabolites in vivo. Clinically, ADC mapping of hyperpolarized ¹³C lactate could allow for improved classification of tumor grade and metastatic potential by measuring both enhanced metabolic flux and differences in lactate’s microenvironment.

Ece Ercan¹, Aranee Techawiboonwong², Maarten J. Versluis¹, Andrew G. Webb¹, and Itamar Ronen^1
1C.J. Gorter Center for High Field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Electrical Engineering, Mahidol University, Nakornpathom, Thailand

2D Diffusion Weighted Chemical Shift (DW-CSI) Imaging is a challenging method which lacks robustness due to the multi-shot nature of the acquisition, combined with the low SNR and the relatively high gradient strength needed for adequate diffusion weighting of the slow diffusing metabolites. Here, we show for the first time a method that accounts for both amplitude and phase inter-shot fluctuations, and generates robust, reproducible and anatomically meaningful DW-CSI and metabolite ADC maps.

Constantin von Deuster^1,2, Christian T. Stoeck², Martin Buehrer², Jack Harmer¹, Rachel W. Chan³, David Atkinson³, and Sebastian Kozerke^1,2
1Division of Imaging Sciences & Biomedical Engineering, King's College London, London, London, United Kingdom, ²Institute for Biomedical Engineering University and ETH Zurich, University and ETH Zurich, Zurich, Zurich, Switzerland, ³Centre for Medical Imaging, University College London, London, London, United Kingdom

A navigated, simultaneous slice excitation implementation for accelerated free-breathing cardiac diffusion imaging using STEAM is proposed. By using an improved navigator gating strategy and dual-slice excitation, significant gains in scan efficiency are demonstrated. Diffusion tensor maps acquired using the proposed method compare favourably with data from conventional sequential slice breathhold acquisitions.