Prantik Kundu¹, Sadegh M. Salehi¹, Bradley A. Cahn², Mercy H. Mazurek², Matthew M. Yuen², Jo Schlemper¹, Barbara Gordon-Kundu², Rafael O'Halloran¹, Michal Sofka¹, and Kevin N. Sheth^2
1Hyperfine Research Inc., Guilford, CT, United States, ²Yale University School of Medicine, New Haven, CT, United States

Low-field (64 mT) point-of-care (POC)-MRI was acquired at the bedside from patients with ischemic and hemorrhagic stroke in the neurointensive care unit at a major academic medical center (n=128). An AI system was trained to quantify brain midline shift (MLS), a standard neuroradiological marker of brain injury from POC-MRI, using anatomical annotations from independent neuroradiologists. A cross-validation experiment showed that AI estimates of MLS from POC-MRI were associated with stroke severity and disability at subsequent discharge. AI estimates of MLS greater than 1.5 mm were positively predictive of poor discharge outcome in the full sample and in ischemic stroke.
 

Thomas Campbell Arnold¹, Ramya Muthukrishnan², Steven N Baldassano¹, Samantha By³, Brian Welch³, Brian Litt^1,4, and Joel M. Stein^5
1Bioengineering, University of Pennsylvania, Philadelphia, PA, United States, ²Computer Science, University of Pennsylvania, Philadelphia, PA, United States, ³Hyperfine Research, Guilford, CT, United States, ⁴Neurology, Perelman School of Medicine, Philadelphia, PA, United States, ⁵Radiology, Perelman School of Medicine, Philadelphia, PA, United States

A growing body of literature demonstrates the value of MRI-based volumetric measures in diagnosing and treating neurodegenerative disorders. However, clinicians cannot offer biomarker screening for at-risk patients due to MRI systems’ high cost and limited access. Low-field MRI scanners offer a potential method for collecting low-cost images that could be analyzed for longitudinal biomarker changes or used in population-level studies. Here we examine the reliability of volumetric measurements made on a low-field MRI system. We compare the variability of 6 tissue volumes collected over 40 scans for 3T and 64mT systems, as well as the overlap between volume segmentations.

Sophie M. Peereboom¹, Christian Guenthner¹, Mohammed M. Albannay¹, and Sebastian Kozerke^1
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

Proton MR spectroscopy can assess cardiac triglyceride levels. Although spectral separation and thermal polarization decrease at lower field strength, increased T₂^* and T₂ values, decreased T₁ values, decreased echo times and the possibility to decrease receiver bandwidth are clear advantages compared to higher fields. In this work the feasibility of in vivo skeletal muscle and cardiac proton spectroscopy at 0.75T is demonstrated. For this purpose, a clinical 3T Philips Achieva scanner was ramped down to a field strength of 0.75T. Results are compared to spectra acquired at 1.5T and 3T and simulations confirm experimental findings.

Neeraj Kulkarni¹, John Greenhalgh¹, Robert Wolf¹, David Chu¹, Rachel Caffrey², Brianna Elizabeth Damadian¹, and Raymond Damadian^1
1FONAR Corporation, Melville, NY, United States, ²Lehigh University, Bethlehem, PA, United States

Despite decades of imaging research focused on the knee, little attention has been paid to dynamic, weight-bearing knee MRI emulating the physiologic condition of the joint. Here, we utilize dynamic imaging in alternating axial and sagittal planes to capture the patellofemoral joint of one volunteer performing one cycle of knee flexion/extension. Our findings indicate that it is possible to characterize patellar tracking in both axial and sagittal planes utilizing fast imaging sequences (<2 minutes) on mid-field upright MRI. Additionally, it is possible to quantify the patellofemoral kinematics on these images using patellar flexion angle and anteroposterior patellar translation. 

Mao Sheng¹, ZhongChang Miao², Jian Bao³, Renjie Zong³, SiSeung Kim³, Huiyao Zhang³, and Bing Keong Li^3
1Department of Radiology, Children’s Hospital of Soochow University, Suzhou, China, ²Department of Radiology, The First People’s Hospital of Lianyungang, Jiangsu Province, China, ³Jiangsu LiCi Medical Device Co., Ltd, Lianyungang, China

A dedicated 0.35T Neonatal-Infant Brain MRI system is developed to investigate the effectiveness and safeness for neonatal and infant patients. 66 volunteers from 0-12 months are recruited to undergo a clinical trial and it is found that low field brain images displayed high grey-white tissue contrast, which is notably helpful for clinical diagnosis. The newly developed system also has very low acoustic noise and that most of the volunteer showed no sign of temperature raise. Low field MRI system therefore has the potential to be a effective and safer alternative MRI system for neonatal and infant patients.

Sai Abitha Srinivas^1,2, Stephen Cauley^3,4, Jason P Stockmann^3,4, Charlotte R Sappo^1,2, Christopher E Vaughn^1,2, Lawrence L Wald^3,4,5, William A Grissom^1,2,6,7, and Clarissa Zimmermann Cooley^3,4
1Vanderbilt University Institute of imaging science, Nashville, TN, United States, ²Department of Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, ³Harvard Medical School, Boston, MA, United States, ⁴Dept. of Radiology, Massachusetts General Hospital, Athinoula A Martinos Center for Biomedical Imaging, Boston, MA, United States, ⁵Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States, ⁶Department of Electrical Engineering, Vanderbilt University, Nashville, TN, United States, ⁷Department of Radiology, Vanderbilt University, Nashville, TN, United States

Low-field MRI scanners can operate outside MR safe rooms, but their image quality is adversely affected due to the presence of electromagnetic interference signals which can severely obscure the images. We demonstrate a generalized dynamic model that can handle time-varying external interference sources by using simultaneously acquired data from multiple EMI detectors - specifically from an electrode, traditional RF pick up coils and the primary MR coil - in-vivo, in real world EMI settings on a 47.5mT permanent magnet open MRI system.

Sydney Sherman^1,2 and Michael Cima^2,3
1Health Science and Technology, Massachusetts Institute of Technology, Cambridge, MA, United States, ²Koch Institute for Integrative Cancer Research, Cambridge, MA, United States, ³Massachusetts Institute of Technology, Cambridge, MA, United States

A low-field portable MR-based sensor was designed for the acquisition of clinical T2 relaxometry measurements in skeletal muscle. A range of low-field permanent magnet array configurations were modeled with varying sensitive region depths and homogeneous volumes; an optimization score for each design was calculated. The optimal design has a sensitive region 15-20mm from the surface of the magnet, making it capable of acquiring measurements deep into the leg such that the measurement is fully localized to skeletal muscle. The exclusion of subcutaneous fat tissue in the sensitive region will improve sensitivity to fluid shifts within the skeletal muscle.

Dion G Thomas¹, Freya G Harrison², Paul D Teal³, Petrik Galvosas^1,4, Mary J Berry^2,5, Sergei Obruchkov⁶, and Yu-Chieh Tzeng^2
1School of Chemical and Physical Sciences, Victoria University of Wellington, Wellington, New Zealand, ²Centre for Translational Physiology, University of Otago, Wellington, New Zealand, ³School of Engineering and Computer Science, Victoria University of Wellington, Wellington, New Zealand, ⁴MacDiarmid Institute for Advanced Materials and Nanotechnology, Wellington, New Zealand, ⁵Department of Paediatrics and Child Health, University of Otago, Wellington, New Zealand, ⁶Robinson Research Institute, Victoria University of Wellington, Wellington, New Zealand

In this study, we investigate how a low field single-sided NMR device can be used to monitor brain tissue properties in-vivo. This device produces a B₀ field with a sweet spot in the brain, which defines the region of tissue that is measured. An ovine model of brain hypoxia was developed, to allow tissue oxygenation to be controlled. We observed that the T₂ decreased during hypoxia, recovering once normal oxygenation levels were re-established. These results shows that single-sided NMR devices have the potential to be used for real-time monitoring applications.

Victor Han¹ and Chunlei Liu^1,2
1Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, United States, ²Helen Wills Neuroscience Institute, University of California, Berkeley, Berkeley, CA, United States

Two-photon excitation allows for excitation at half of the Larmor frequency. In Earth’s magnetic field, this corresponds to a Larmor frequency of about 2 kHz and an excitation frequency of about 1 kHz. By transmitting at 1 kHz and receiving at 2 kHz, we could transmit and receive at the same time by using filters to protect the receiver. While some harmonic distortion prevented us from fully separating the transmitted from the received signal during excitation, the always connected and unsaturated receive circuitry allowed us to eliminate receiver dead time, which can be quite long at low frequencies. 

Bart de Vos¹, Thomas O'Reilly¹, Wouter Teeuwisse¹, Rob Remis², and Andrew Webb^1
1C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, ²Circuits and Systems, Delft University of Technology, Delft, Netherlands

We reduced image artifacts and increased the attainable axial field-of-view by designing a highly linear x-gradient coil for Halbach array-based MR systems. A truncated sum of sinusoidal basis function is used for the current density. Higher order modes combined with specifying the target field inside a volume lead to a stable solution of the inverse source problem. The coil was installed on our 50 mT system and resulted in a 150% increase in linear DSV with respect to our previously designed gradient coil. Whole-brain three-dimensional images have been acquired using turbo spin echo sequences in less than 10 minutes.

Matthew Wilcox^1,2, Sai Abitha Srinivas^1,2, Christopher E Vaughn^1,2, Charlotte R. Sappo^1,2, and William A. Grissom^1,2,3,4
1Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, ²Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ³Radiology, Vanderbilt University, Nashville, TN, United States, ⁴Electrical Engineering, Vanderbilt University, Nashville, TN, United States

Despite drawbacks including production of acoustic noise and peripheral nerve simulation, long switching times, bulkiness, and high costs, use of B₀ gradient coils is ubiquitous in conventional MR imaging. Replacement of B₀ encoding with B₁⁺ encoding alleviates these concerns and allows the use of hardware which is lower-cost, more compact, and silent. As a first step towards realizing a full B₁⁺ encoding coil system, this work demonstrates an RF z-gradient solenoid coil developed for brain imaging on a 47.5 mT system. The coil performance is evaluated in simulation, bench, and scanner experiments and slice-selection using B₁⁺-selective pulses is demonstrated.  

Bart de Vos¹, Thomas O'Reilly¹, Wouter Teeuwisse¹, Rob Remis², and Andrew Webb^1
1C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, ²Circuits and Systems, Delft University of Technology, Delft, Netherlands

We describe a design method for quadrature RF coils for Halbach based magnets using a target field approach. The resulting current densities corresponding to these coils make them inherently decoupled. We constructed a coil pair for a newly built 76 mT system and imaged the wrist of a volunteer. The images showed a 33% SNR improvement over a linear coil, in close agreement with simulations.

Javad Parsa¹, Thomas O'Reilly¹, Bart de Vos¹, and Andrew Webb^1
1Leiden University Medical Center, Leiden, Netherlands

An integrated transmit coil and four-element receive array has been simulated, constructed, characterized and tested on a low field MRI system operating at 2.15 MHz. Using a combination of loops and butterfly coils neighbouring inter-element coupling is below -18 dB, with directly opposite coils ~-9 dB, and <-17 dB coupling for all receive coils to the transmit coil. Images of a phantom have been acquired with a simple sum-of-squares reconstruction.

Ed Boskamp¹, Mike Twieg¹, and Rafael O'Halloran^1
1Hyperfine, Guilford, CT, United States

MRI is sensitive to patient motion. There are a number of pulse sequence techniques like navigators and propeller to prevent motion artifacts. We are introducing a sensitive sensor based technique to detect motion and pose without touching the patient. The sensors are a number of miniaturized dipole resonators integrated into the receiver array. Motion corrupted lines in k space are flagged and rescanned, or when motion is between a limited number of poses, the data can be binned and combined before the recon process.

Konstantin Wenzel¹, Hazem Alhamwey¹, Tom O'Reilly², Layla Riemann¹, and Lukas Winter^1
1Physikalisch-Technische Bundesanstalt (PTB), Braunschweig and Berlin, Germany, ²Leiden University Medical Center (LUMC), Leiden, Netherlands

In this work, B₀-shimming techniques are investigated allowing the construction of simple, low-cost, and homogeneous Halbach-based low-field MR magnets. The techniques are applied to build a desktop MR magnet at B₀=0.1T and can be easily scaled to magnet designs of larger diameter. The presented shimming approach improved B₀ homogeneity by a factor of ~8 from 5448ppm to 682ppm in a 2D target region. All design files and code concerning this work will be made available open source on www.opensourceimaging.org.

Shao Ying Huang¹, Yi-Dan Chen¹, Ting-Ou Liang², Yan Hao Koh¹, and Wenwei Yu^3
1Singapore University of Technology and Design, Singapore, Singapore, ²Zhejiang University, Hangzhou, China, ³Chiba University, Chiba, Japan

This abstract presents the development of a digital-twin of a permanent-magnet-array-based portable MRI system. It is to guide the design, optimization, hardware debug and calibration, and evaluation of such a system at both a sub-system level and a system level. Meanwhile, it facilitates the development of a sub-system without building the whole system. It consists of the simulators for magnet array, RF coil, gradient coils, pulse sequence, k-space analysis, image reconstruction, and image quality evaluation.

Thomas O'Reilly¹ and Andrew Webb^1
1C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands

Arrays of permanent magnets have shown promise as a relatively light weight, low cost, sustainable way of generating magnetic fields suitable for MRI. In this work we show the design of a new homogenous helmet shaped, Halbach-array based magnet optimized for imaging the adult head. The magnet has a mean magnetic field strength of 59.7 mT and a homogeneity of 1313 ppm over a brain sized 25x20x15 cm³ ellipsoid. By truncating one side of the magnet the weight of the system has been reduced compared to symmetric designs while increasing the obtained field strength.

Thomas O'Reilly¹ and Andrew Webb^1
1C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands

Arrays of permanent magnets are an attractive approach to designing magnets for low field MRI systems as they are affordable, easy to handle and offer great flexibility in magnet designs. However, translating designs from simulations to realised systems is challenging. The difference between the two is often attributed to imperfections in the individual magnets that make up the permanent magnet array. Here we show that non-random errors in the magnet rotations due to imperfect magnet holders can be the dominant cause of this discrepancy and if the imperfections are known they can easily be corrected for during manufacturing.

Yang Gao^1,2, Alex T. L. Leong^1,2, Yilong Liu^1,2, and Ed X. Wu^1,2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Hong Kong SAR, China, ²Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China

MRI has impacted modern healthcare tremendously. However, MRI accessibility is low and extremely inhomogeneous globally. The recent shift in prioritizing MRI’s value has seen the emergence in developing point-of-care systems at ultra-low-field (<0.1T), particularly the use of permanent magnets for its light-weight, low power and low cost benefits. However, fundamentally, a relatively strong and uniform magnetic field is still desired for ultra-low-field MRI systems. Here, we propose a triple-ring permanent magnet design that aims to preserve and improve upon the fundamental field homogeneity requirements of MR imaging at ultra-low-field, without sacrificing the flexibility to perform both body and head imaging.

Marcus Prier^1,2, David Schote^1,2, Ivan Fomin^1,2, Thomas Witzel³, Georg Rose^1,2, and Oliver Speck^1,2
1Otto-von-Guericke University, Magdeburg, Germany, ²Research Campus STIMULATE, Magdeburg, Germany, ³Q Bio Inc, San Carlos, CA, United States

A Tabletop MRI system was developed that merges the Marinos Center Tabletop and OCRA projects, and is improved in many aspects compared to the original system. The proposed system includes a fully functional MRI console with a customizable GUI and microcontroller server. The hardware was extended with an optimized magnet design with passive shimming, active TR switches with reverse bias and RFPA noise blanking, and a low cost and fast response 4 channel gradient amplifier. It can be used for educational purposes and student courses in medical engineering.