View Presentation Video

Stanislav Motyka¹, Paul Weiser², Bernhard Strasser¹, Dario Goranovic¹, Lukas Hingerl¹, Gilbert Hangel^1,3, Eva Niess¹, Stephan Wampl⁴, Fabian Niess¹, Simon Robinson¹, Georg Langs², Siegfried Trattnig¹, and Wolfgang Bogner^5
1High Field MR Center, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria, ²Computational Imaging Research Lab, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria, ³Department of Neurosurgery, Medical University of Vienna, Vienna, Austria, ⁴High Field MR Center, Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria, ⁵Medical University of Vienna, Vienna, Austria

Keywords: Motion Correction, Brain

The abstract compares two approaches to dynamically estimate B0 maps at 7T. The U-net predicts B0 maps based on the initial B0 map and the movement information. The dual-echo EPI-based navigator directly measures B0 maps. Both methods yield comparable results.  However, the deep learning approach could overcome the major disadvantage of the EPI-based navigator, thus the need for the dead time in the parent sequence, if the motion can be mapped externally.

View Presentation Video

Cemre Ariyurek¹, Tess E. Wallace¹, Tobias Kober^2,3,4, Sila Kurugol¹, and Onur Afacan^1
1Quantitative Intelligent Imaging Lab (QUIN), Department of Radiology, Boston Children's Hospital and Harvard Medical School, Boston, MA, United States, ²Advanced Clinical Imaging Technology, Siemens Healthineers International AG, Lausanne, Switzerland, ³Department of Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, ⁴LTS5, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

Keywords: Motion Correction, Body

Abdominal MRI scans often require breath-holding to prevent image quality degradation, which can be challenging for patients. XD-GRASP enables generation of motion-robust images for free-breathing abdominal MRI by binning the data into respiratory phases. In this study, we compared three navigation techniques, namely k-space center, free induction decay navigators (FIDnavs) and pilot tone (PT), for XD-GRASP reconstruction. FIDnavs and PT have advantages over k-space center navigation since they are insensitive to gradient delays, independent of the imaging plane and acquired more frequently. The image quality ranking means (1=best,2=moderate,3=worst) were 1.4, 1.6 and 2 for FIDnavs, PT and k-space center, respectively.

View Presentation Video

Gonzalo Gabriel Rodriguez¹, Hector Lise de Moura¹, Lauren O'Donnell¹, Ravinder Regatte^1,2, and Guillaume Madelin^1,2
1Center for Biomedical Imaging, Radiology Department, NYU School of Medicine, New York, NY, United States, ²Vilcek Institute of Graduate Biomedical Sciences, NYU Langone Health, New York, NY, United States

Keywords: Data Processing, Non-Proton, Super-Resolution

We compared three algorithms to generate a high-resolution (HR) ²³Na image from simultaneously-acquired low-resolution (HR) ²³Na density-weighted MRI and HR ¹H density, T₁ and T₂ maps from MRF in brain at 7 T: U-net, PLS-regression, and hybrid. The multi-scale structural similarity index between generated HR ²³Na images and HR ground truth was higher than 0.95 for the three methods. Overall, the hybrid method showed better results, generating the sharpest HR image while keeping the highest similarity between the acquired and generated LR images.

View Presentation Video

Gabrio Rizzuti¹, Tim Schakel¹, Niek Huttinga¹, Tristan van Leeuwen², and Alessandro Sbrizzi^1
1Universitair Medisch Centrum Utrecht, Utrecht, Netherlands, ²Centrum Wiskunde & Informatica, Amsterdam, Netherlands

Keywords: Motion Correction, Brain, brain, clinical protocols, motion correction, retrospective correction

We discuss the application of a novel 3D retrospective rigid motion correction and reconstruction scheme to in-vivo data for clinical 3D brain protocols. The correction scheme leverages multiple scans contained in a typical MR session, of which some are corrupted by motion but at least one scan may be devoid of motion artifacts. The uncorrupted scan is then used as a reference to regularize a generalized rigid-motion registration problem. We discuss the potential of the proposed algorithm with a prospective in-vivo study.

View Presentation Video

Brian Nghiem^1,2, Zhe Wu², Lars Kasper², and Kâmil Uludağ^1,2
1Dpt. of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ²University Health Network, Toronto, ON, Canada

Keywords: Motion Correction, Motion Correction

We showed that accurate 3D retrospective motion correction of T1w MPRAGE data can be achieved with a UNet-assisted joint estimation algorithm. We compared the proposed method to using the UNet on its own and the standard joint estimation algorithm. Joint estimation (with and without the UNet) outperformed using the stand-alone UNet. The UNet-assisted joint estimation algorithm converged faster than its UNet-free counterpart. We demonstrated the importance of adapting to the changing levels of artifacts over the course of the joint estimation algorithm by sequentially employing different UNets trained for correcting different levels of motion corruption.

View Presentation Video

Jongyeon Lee¹, Wonil Lee¹, and HyunWook Park^1
1Korea Advanced Institute of Science and Technology, Daejeon, Korea, Republic of

Keywords: Motion Correction, Motion Correction, Brain Motion Correction

Motion correction in MRI has been successful with deep learning techniques to reduce motion artifacts. However, these methods were based on supervised learning, which have required a large dataset to train neural networks. To overcome this issue, we propose a new technique to correct motion artifacts without any training dataset, which optimizes a neural network only with a single motion-corrupted image. We adopt Deep Image Prior framework to capture image priors from a convolutional neural network using motion simulation based on MR physics. Using the deep image prior, our method finely reduces motion artifacts without any dataset.

View Presentation Video

Zhiqiang Li¹, James Murchison¹, Dori Shoshan¹, Melvyn B Ooi², and John P Karis^1
1Barrow Neurological Institute, Phoenix, AZ, United States, ²Philips Healthcare, Houston, TX, United States

Keywords: Motion Correction, Motion Correction, artifact, rapid scan, clinical study, high value

A motion-corrected 2D multi-shot EPI (msEPI) technique was compared to conventional 2D GRE to determine if it can be used as a faster technique for blood sensitive imaging in the emergency department (ED) setting. msEPI was found to be superior to GRE (p < 0.001) in motion artifact, overall image quality, and lesion detection. These results and reduced scan time make the motion-corrected 2D msEPI a viable alternative for blood sensitive imaging in the ED setting.

View Presentation Video

Chuanli Cheng¹, Lei Zhang², Hao Peng¹, Hairong Zheng¹, Xin Liu¹, Huimao Zhang², and Chao Zou^1
1Shenzhen institutes of advanced technology, Chinese Academy of Sciences, Shenzhen, China, ²Radiology Department, Bethune First Hospital of Jilin University, Changchun, China

Keywords: Segmentation, Fat, brown adipose tissue

The noninvasive assessment of BAT volume is fundamental for characterizing the longitudinal effects in rodents. In this work, dynamic fat fraction images of 34 rats fed under different conditions were acquired before and after noradrenaline injection for 2.5 hours. The iBAT regions were recognized and labelled automatically by identifying the regions with significant changes in FF images. Then a deep learning network was built up by training the FF images and the automatically identified mask images in all rats. The dice similarity coefficient, precision rate and recall rate of the network were found to be 0.897±0.061, 0.901±0.068 and 0.889±0.086, respectively.

View Presentation Video

Polina Emeliyanova^1,2, Laura M. Parkes^1,2, Stephen R. Williams³, and Caroline Lea-Carnall^1,2
1Division of psychology, communication and human neuroscience, Faculty of biology, medicine and health, University of Manchester, Manchester, United Kingdom, ²Geoffrey Jefferson Brain Research Centre, Manchester Academic Health Science Centre, Manchester, United Kingdom, ³Division of Informatics, Imaging and Data Science, University of Manchester, Manchester, United Kingdom

Keywords: Data Acquisition, Spectroscopy, Functional MRS, multi-echo MRS, 3T, glutamate, creatine (Cr), N-Asetyl-Aspartate (NAA)

fMRS studies have shown dynamic change in glutamate (Glu) concentrations during visual stimulation. Here, we examine the effect of TE from 9.3-280ms on the signal during fMRS acquisition and whether T2 relaxation parameters change during a visual task compared to rest. We observed a short Glu T₂-compartment between 2.5-4.5ms arising from 48% (rest) and 16% (active) of the total signal with significant increase in Glu signal at TE =140ms during stimulation. These preliminary results indicate bi-exponential decay of Glu in the visual cortex at 3T encouraging future work with more robust fMRS paradigms.  

View Presentation Video

Guodong Weng^1,2, Piotr Radojewski^1,2, Federico Turco^1,2, and Johannes Slotboom^1,2
1Support Center for Advanced Neuroimaging (SCAN), University of Bern, Bern, Switzerland, ²Translational Imaging Center, sitem-insel, Bern, Switzerland

Keywords: Pulse Sequence Design, Spectroscopy, Downfield MRSI

Many interesting metabolites, for instance adenosine triphosphate (ATP), glutathione (GSH) and homocarnosine (hCs), have resonances in the downfield range (>4.7 ppm) in 1H-MRS. Previous studies have studied the downfield spectra of metabolites in human brain using single-voxel spectroscopy (SVS) and, most recently, 2D one-slice MRSI. The study presents a 3D whole brain mapping sequence suited for downfield metabolites which does not require water suppression pulses.

View Presentation Video

Ryan K Robison^1,2,3, Justin R Haynes^2,3, Sandeep Ganji^4,5, Wellington Pham^2,3, Seth A Smith^2,3, Victoria L Morgan^2,3, Reid C Thompson⁶, Reed A Omary³, John C Gore^2,3, and Changho Choi^2,3
1Philips, Nashville, TN, United States, ²Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States, ³Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States, ⁴Philips, Rochester, MN, United States, ⁵Mayo Clinic, Rochester, MN, United States, ⁶Neurological Surgery, Vanderbilt University Medical Center, Nashville, TN, United States

Keywords: Data Acquisition, Spectroscopy

Lactate is important as a source of energy and a product of metabolism, so the capability to measure lactate accurately and noninvasively by MRS is of high value. We report a new lactate editing approach that enables detection of the lactate 1.3 ppm resonance with minimal contaminations from threonine and other compounds. Narrow-band editing 180° RF pulses were implemented in 1H MR MEGA-PRESS and the sequence was optimized with numerical and phantom analyses. With data from six healthy adult subjects, the lactate and threonine concentrations in healthy brain were estimated to be 0.5±0.1 and 1.0±0.1 mM, respectively.

Jia Yuping¹ and Dou Weiqiang^2
1Lixia District, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Jinan, China, ²MR Research, GE Healthcare, Beijing, China

Keywords: Data Analysis, Cancer

This study aimed to determine the clinical potential of intravoxel incoherent motion (IVIM) DWI in predicting perineural invasion (PNI) of rectal cancer (RC). IVIM parameters derived from different mathematical models, including apparent diffusion coefficient from mono-exponential model, true diffusion coefficientand and perfusion fraction from bi-exponential model, and the distributed diffusion coefficient from stretched-exponential model showed significant differences between 72 with PNI and 76 without PNI. With these findings, IVIM may thus be considered effective in preoperatively predicting PNI status in RC and further help clinicians make individual treatment plans.

View Presentation Video

Siqi Cai^1,2, Yang Zhou¹, Chao Zou¹, Linhai Jiang³, Qian Wan¹, Chunxiang Jiang¹, Changjun Tie¹, and Lijuan Zhang^1,2
1Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, ²University of Chinese Academy of Science, Beijing, China, ³The Instrumental Analysis Center of Shenzhen University, Shenzhen University, Shenzhen, China

Keywords: Data Analysis, CEST & MT

Chemical exchange between endogenous labile guanidinium (Guan) protons of creatine and free water and its thermal dependency were investigated with CEST imaging. The apparent offsets of Cr-CEST were linearly correlated to the experiment temperatures and was relatively insensitive to Cr concentration and pH variations at a range corresponding to physiological and pathological conditions, suggesting a potential utility of Cr-CEST for in vivo thermometry based on endogeneous creatine of the biological tissue.

View Presentation Video

Julius Åkesson¹, Katarina Steding-Ehrenborg¹, Johannes Töger¹, Björn Östenson¹, Pia Sjöberg¹, and Einar Heiberg^1
1Clinical Physiology, Department of Clinical Sciences Lund, Lund University, Skåne University Hospital, Lund University, Lund, Sweden

Keywords: Data Processing, Software Tools, Exercise, real-time

This study aimed to enable ECG-free construction of respiratory controlled 3D cardiac magnetic resonance cine series from short-axis 2D real-time images through automated detection of cardiac phases. This is important for imaging during exercise, since obtaining a reliable ECG-signal can be difficult. Visually coherent cines could be constructed for midventricular slice positions through detecting cardiac phases from automated left ventricular segmentations in real-time timeframes, but not for apical and basal slice positions. Deep learning-based regression of cardiac phases directly from images was investigated to handle this problem, but apical and basal cine construction remains an open problem.

View Presentation Video

Hwapyeong Cho¹ and Hyung Joon Cho^1
1Ulsan National Institute of Science and Technology, Ulsan, Korea, Republic of

Keywords: Data Analysis, Relaxometry, Myelin, Magnetic field perturbation

In this study, the effects of myelin structures such as myelin volume fraction (MVF) and myelin sheath thickness (MST) were investigated by measuring the intra/extracellular water signals in post-mortem aging rat brains. R₂ values calculated by intra/extracellular water signals obtained using Carr-Purcell-Meiboom-Gill (CPMG) sequence are theoretically verified by comparison with transmission electron microscopy (TEM) and simulation results. It was confirmed that MVF affects the increase of R₂ and MST affects the decrease of R₂, and the analysis of R₂ through this relationship was statistically significant. These results suggest that effects of myelin structures may affect CPMG-based intra/extracellular water signals.

View Presentation Video

Yu Chen¹, Runke Wang¹, Ruokun Li², Fuhua Yan², and Yuan Feng^1,2,3
1Shanghai JiaoTong University, School of Biomedical Engineering, Shanghai, China, ²Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China, ³Institute of Medical Robotics, Shanghai Jiao Tong University, Shanghai, China

Keywords: Elastography, Brain

Conventional gradient echo (GRE) based Magnetic Resonance Elastography (MRE) suffers from long TE and scan time at low mechanical excitation (<25Hz). Displacement Encoding with a Stimulated Echo (DENSE) can record displacement with short TR, useful for low-frequency excitation. Here, 3D excitation was used to compensate for potential signal loss. A multiphase acquisition scheme with stack-of-star radial sampling was also used along with interleaved multi-slab acquisition and Hadamard encoding for MRE acceleration. Phantom experiments were conducted for validation. 32 slices brain MRE scan at 20Hz can be achieved in 11min30s, 5 folds faster than conventional GRE-based MRE.  

View Presentation Video

Radhika Tibrewala¹, Mahesh Keerthivasan^1,2, Mary Bruno³, Christopher Collins¹, Bruno Madore⁴, and Daniel K Sodickson^1
1Center for Advanced Imaging Innovation and Research, NYU Grossman School of Medicine, New York, NY, United States, ²Siemens Medical Solutions USA, New York, NY, United States, ³NYU Grossman School of Medicine, New York, NY, United States, ⁴Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, United States

Keywords: Motion Correction, Hybrid & Novel Systems Technology

Deep-seated motion due to digestion and bladder filling can be difficult to monitor during MR acquisition. In this study, we use a small ultrasound-based sensor to detect bladder and pelvic floor motion, with a real-time MRI acquisition and a pilot-tone motion-sensitive signal as a reference. A neural network is used to predict the bladder area at various times during the MRI scan by using the US signal features. Our results indicate that the correlations between the ultrasound sensor and the MRI are beneficial for detecting bladder motion over time. 

View Presentation Video

Nelson Homero Gil¹, Wei-Ching Lo², Bryan Clifford², Min Lang¹, Komal Awan¹, Daniel Nicolas Splitthoff³, Daniel Polak³, Stephen Cauley¹, and Susie Huang^1
1Radiology, Massachusetts General Hospital, Boston, MA, United States, ²Siemens Medical Solutions USA, Boston, MA, United States, ³Siemens Healthcare GmbH, Erlangen, Germany

Keywords: Motion Correction, Neurodegeneration

Volumetric brain MRI using the T1-weighted-MPRAGE sequence is an important component of the clinical evaluation of dementia. However, MPRAGE has a long acquisition time and is especially prone to patient motion artifact. A recent development to address this issue is a technique called Scout Accelerated Motion Estimation and Reduction (SAMER). In this work, we used a set of 90 MPRAGE scans derived from 10 healthy volunteers to demonstrate that SAMER is effective at correcting various degrees of motion, including severe motion with non-diagnostic image quality, and greatly increases the accuracy of volumetric brain measurements.

View Presentation Video

Siyuan Hu¹, Yong Chen¹, Mark Griswold¹, and Dan Ma^1
1Case Western Reserve University, Cleveland, OH, United States

Keywords: Motion Correction, MR Fingerprinting

MR Fingerprinting (MRF) simultaneously measures multiple tissue properties within a single scan. Although 2D MRF has been shown to be less sensitivie to motion than conventional imaging, bulk motion is still problematic for longer 3D MR Fingerprinting scans. We propose to integrate 3D fat navigators with 3D MRF acquisitions for motion correction in neuroimaging, especially for non-sedated pediatric imaging. The proposed method was validated in in vivo scans of healthy subjects with various types of motions and in non-sedated infants. We showed that the fat-navigated 3D MRF framework could resolve and correct bulk motions of both healthy volunteers and infants.