Tobias Spronk^1,2,3, Oliver Kraff¹, Gregor Schaefers^3,4, and Harald H Quick^1,2
1Erwin L. Hahn Institute for MR Imaging, University of Duisburg-Essen, Essen, Germany, ²High-Field and Hybrid MR Imaging, University Hospital Essen, Essen, Germany, ³MRI-STaR Magnetic Resonance Institute for Safety, Technology and Research GmbH, Gelsenkirchen, Germany, ⁴MR:comp GmbH, Testing Services for MR Safety & Compatibility, Gelsenkirchen, Germany

This study presents a numerical approach to simulate artifacts of metallic implants in the MR environment, which can be applied to improve the MR image artifacts testing procedure for medical implants according to ASTM F2119. The numerical approach is validated by comparing simulations and measurements of two metallic test objects made of titanium and stainless steel at three different field strengths (1.5 T, 3 T, and 7 T).

Kübra Keskin¹, Brian Hargreaves^2,3,4, and Krishna S. Nayak^1,5
1Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, United States, ²Radiology, Stanford University, Stanford, CA, United States, ³Electrical Engineering, Stanford University, Stanford, CA, United States, ⁴Bioengineering, Stanford University, Stanford, CA, United States, ⁵Biomedical Engineering, University of Southern California, Los Angeles, CA, United States

MRI near metallic implants suffers from severe artifacts due to magnetic susceptibility, and these artifacts scale with the B₀ field strength. Techniques like “view angle tilting” and “multi spectral imaging” partially mitigate these distortions and are widely used at 1.5T and 3T. Here, we provide a pipeline for realistic simulation of MRI metal artifacts and noise using anatomic digital phantoms and susceptibility calculations. For one example application, total hip arthroplasty, we demonstrate the impact of field strength (0.2T, 0.55T, 1.5T, 3T), and MSI parameters such as receiver bandwidth, RF bandwidth, and number of spectral bins.

Alexander R Toews^1,2, Daehyun Yoon¹, and Brian A Hargreaves^1,2,3
1Radiology, Stanford University, Stanford, CA, United States, ²Electrical Engineering, Stanford University, Stanford, CA, United States, ³Bioengineering, Stanford University, Stanford, CA, United States

Multispectral imaging sequences mitigate geometric distortion near metallic implants at the expense of readout blur and residual intensity artifacts. To address these limitations, a method for jointly estimating the image content and field inhomogeneity near metal is presented. A dual polarity readout acquisition is used to provide sufficient conditioning for the estimation problem. The method is studied in simulation as well as in a physical phantom consisting of a shoulder implant embedded in a resolution grid. Results indicate that the method can greatly reduce readout blur and in some cases eliminate residual intensity artifacts.

Yuheng Huang^1,2, Xinheng Zhang^1,2, Serry Fradad¹, Lu Meng¹, Ghazal Yoosefian¹, Linda Azab¹, Xinqi Li³, Alan Kwan⁴, Rohan Dharmakumar¹, Hui Han¹, and Hsin-Jung Yang^1
1Biomedical Imaging Research Institute, Cedars Sinai Medical Center, Los Angeles, CA, United States, ²Bioengineering, UCLA, Los Angeles, CA, United States, ³Columbia University, New York, NY, United States, ⁴Cardiology, Cedars Sinai Medical Center, Los Angeles, CA, United States

A strong off-resonance field around metallic implants induces unreliable phase maps and limits the application of phase-based MRI techniques. State-of-the-art MR phase mapping and unwrapping techniques perform poorly under the influence of metallic implants because of their limited dynamic range and susceptibility to noise. In this study, we developed a fully automatic phase mapping technique based on a multi-echo GRE sequence with unevenly spaced echoes and a high-dynamic-range phase reconstruction algorithm. We tested the proposed method with numerical simulations and human subjects with metallic implants.

Xu Lulu¹, Shen Yong², Dou Weiqiang³, and Qi Liang^1
1Radiology, Jiangsu Province Hospital, Nanjing, China, ²GE Healthcare, MR Enhanced Application China, Beijing, China, ³GE Healthcare, MR Research China, Beijing, China

This study aimed to determine the optimal direction of the frequency-encoding gradient when acquiring T2WI-STIR images with titanium alloy. To investigate this, we made in vitro MRI experiment. The water-pork phantom with ten different-length titanium alloys was constructed, and the short- and long-axis ratios calculated in different frequency-encoding directions were compared with each other, respectively. We found that when the frequency-encoding axis was perpendicular to the screw, the metallic artifact was smaller in whatever long axis or short axis. This result may improve observing more anatomical structures around the embedded metal.

Anuj Sharma¹ and Andrew J Wheaton^1
1Magnetic Resonance, Canon Medical Research USA, Inc., Mayfield Village, OH, United States

Deep learning based denoising methods can significantly reduce image noise at the cost of producing unnaturally smooth images. Typically, natural-looking images are produced by blending a fraction of the acquired noisy image with the denoised image. The blending ratio is set based on visual inspection. This approach impedes workflow and is prone to inter-operator variability. We propose a method to analytically calculate the blending ratio based on a desired SNR value. The proposed method is demonstrated to produce natural-looking denoised images with consistent SNR across head, spine and knee applications.

Boris Mailhe¹, Zahra Hosseini², Bing Ji³, Hui Mao³, and Mariappan S. Nadar^1
1Digital Technology and Innovation, Siemens Healthineers, Princeton, NJ, United States, ²MR R&D Collaboration, Siemens Medical Solutions USA Inc., Atlanta, GA, United States, ³Department of Radiology and Imaging Sciences, Emory University, Atlanta, GA, United States

This work introduces a new approach to denoise 2D COSY data that typically are collected with low number of signal averages. Our method of a joint sparse parametric Matching Pursuit uses high-resolution parametric models to model the behavior of 2D COSY data in the F2 dimension and add nonparametric joint sparse constraints to regularize the peak shape in the F1 dimension. The effect and performance of denoising were demonstrated using a phantom.

William T Clarke¹ and Mark Chiew^1
1Wellcome Centre for Integrative Neuroimaging, NDCN, University of Oxford, Oxford, United Kingdom

Low rank denoising methods can reduce the noise present in spectroscopic data, a typically signal-to-noise limited technique. Low rank denoising is a data driven technique, typically exploiting the linear predictability or spatio-temporal separability of the data. Despite high levels of apparent denoising it is not clear whether denoising decreases the final uncertainty in the measurement, as the denoised data can be biased. In this work we assess the uncertainty reduction using Monte Carlo simulation and by measuring reproducibility of noisy and denoised 1H MRSI. We find low rank denoising reduces uncertainty but by much less than is apparent.

Francesco Padormo¹, Joe Martin¹, Jane Ansell¹, Elizabeth Gabriel¹, Laurence H. Jackson¹, Caitlin O'Brien¹, Simon Shah¹, David Price¹, and Geoff Charles-Edwards^1
1Guy's & St Thomas' NHS Foundation Trust, London, United Kingdom

Manufacturers typically specify new MRI systems to have a minimum Faraday Cage (FC) attenuation of 90-100dB at the resonant frequency. Anecdotal observations have suggested that systems can operate successfully with FC attenuation levels below this specification. We present a survey of attenuation measurements across eleven systems in order to gain insight into realistic values on systems in routine clinical use. 

Takeshi Nakaura¹, Hiroyuki Uetani¹, Kousuke Morita¹, Kentaro Haraoka², Akira Sasao¹, Masahiro Hatemura¹, and Toshinori Hirai^1
1Diagnostic Radiology, Kumamoto University, Kumamoto, Japan, ²Cannon Medical Systems Japan, Tochigi, Japan

We evaluated image quality of hybrid type deep learning reconstruction (hybrid-DLR) with wavelet based denoising method in T2-weighted images (T2WI) of the pituitary with various denoising level (1-5). There was a progressive increase in SNR with hybrid-DLR with increase of the denoising level. On the other hand, the SNR of conventional wavelet-based method was not increased at high denoising levels (4-5). All qualitative scores of hybrid-DLR in any denoising levels are higher than that of wavelet based denoising method, and the difference became more noticeable at higher denoising levels.

Hitoshi Kubo¹, Yuya Abe², Tomoya Yokokawa², Seira Yokoyama², and Koji Hoshi^2
1Fukushima Medical University, Fukushima, Japan, ²Hoshi General Hospital, Koriyama, Japan

We aim to assess fundamental noise reduction performance using deep learning reconstruction with a phantom at a 1.5 T MR scanner. In this study, the relationship among parameters for noise reduction, signal-to-noise ratio, image quality, and spatial resolution of images was evaluated. SNRs were increased higher significantly by DLR in all SNR ranges. Increasing ratio of SNR was varied by means of parameter settings. Combination of the DLR parameters affected varies to SNR, SSIM, and spatial resolution of the images. We should exercise caution to select DLR parameters when this technique applies to clinical images.

Hsin-Ju Lee^1,2, Hsiang-Yu Yu^3,4,5, Cheng-Chia Lee^4,5,6, Chien-Chen Chou^3,4, Chien Chen^3,4, Wen-Jui Kuo^5,7, and Fa-Hsuan Lin^1,2,8
1Physical Sciences Platform, Sunnybrook Research Institute, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Department of Epilepsy, Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan, ⁴School of Medicine, National Yang-Ming University, Taipei, Taiwan, ⁵Brain Research Center, National Yang-Ming University, Taipei, Taiwan, ⁶Department of Neurosurgery, Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan, ⁷Institute of Neuroscience, National Yang-Ming University, Taipei, Taiwan, ⁸Department of Neuroscience and Biomedical Engineering, Aalto University, Espoo, Finland

We developed the dynamic modeling of heartbeats (DMH) method to suppress the ballistocardiography (BCG) artifacts on the electroencephalography (EEG) data collected inside MRI. DMH estimates the instantaneous EEG signals at specific phases in the cardiac cycle by combining EEG signals at those phases in other cardiac cycles showing similar dynamic features. Using both simulations and empirical data at 3T, we demonstrated that the DMH approach can suppress the BCG artifacts more efficiently than Optimal Basis Set (OBS) method in both epilepsy and steady-state visual evoked potential data.

Alex Ensworth¹, Véronique Fortier^1,2, Jorge Campos Pazmino¹, and Ives R Levesque^1,2,3,4
1Medical Physics Unit, McGill University, Montreal, QC, Canada, ²Biomedical Engineering, McGill University, Montreal, QC, Canada, ³Research Institute of the McGill University Health Centre, Montreal, QC, Canada, ⁴Gerald Bronfman Department of Oncology, McGill University, Montreal, QC, Canada

Phase data for quantitative susceptibility mapping must be unwrapped to remove ambiguities due to limitations in the collected data. This work demonstrates the advantages of quality-guided region-growing over conventionally used methods and introduces a general method to correct for temporal discontinuities introduced by path-based spatial unwrapping methods. The unwrapping approaches were evaluated in 11 in vivo datasets presenting a variety of phase profiles. 

Guillaume Daval-Frérot^1,2,3, Aurélien Massire¹, Mathilde Ripart², Boris Mailhe⁴, Mariappan Nadar⁴, Alexandre Vignaud², and Philippe Ciuciu^2,3
1Siemens Healthcare SAS, Saint-Denis, France, ²CEA, NeuroSpin, CNRS, Université Paris-Saclay, Gif-sur-Yvette, France, ³Inria, Parietal, Palaiseau, France, ⁴Digital Technology and Innovation, Siemens Healthineers, Princeton, NJ, United States

Patient-induced inhomogeneities in the magnetic field cause distortions and blurring during acquisitions with long echo times, as in susceptibility-weighted imaging. Most correction methods require collecting an additional $$$\Delta{B_0}$$$ field map. To avoid that, we propose a method to approximate this field map using the single echo acquisition only. The main component of the observed phase is linearly related to $$$\Delta{B_0}$$$ and TE, and the relative impact of non-$$$\Delta{B_0}$$$ terms becomes insignificant with TE>20ms at 3T. The estimated 3D field maps, produced at 0.6 mm isotropic under 3 minutes, provide corrections equivalent to acquired ones.

Divya Varadarajan^1,2, Mukund Balasubramanian^2,3, Daniel J. Park¹, Thomas Witzel⁴, Jason P. Stockmann^1,2, and Jonathan R. Polimeni^1,2,5
1Martinos Center for Biomedical Imaging, Charlestown, MA, United States, ²Department of Radiology, Harvard Medical School, Boston, MA, United States, ³Boston Children’s Hospital, Boston, MA, United States, ⁴Qbio Inc., San Carlos, CA, United States, ⁵Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA, United States

Maps of the B0 field are routinely used for MRI scanner calibration and for post-processing corrections of geometric distortions. However, several sources of bias are present in conventional field estimates, which will result in uncorrected image artifacts, yet it is unclear what the magnitude of these biases are, whether these are large enough to warrant concern, and how to reduce these errors to extract more accurate field estimates. Here we investigate the accuracy of the standard B0 field map acquisition, demonstrate that the estimated fields vary with several acquisition parameters, and investigate sources of these errors.

seyedeh nasim adnani¹, Thomas Denney Jr.¹, Alexander Sukstansky², Dmitriy Yablonskiy², and adil bashir^1
1electrical and computer engineering, auburn university, auburn, AL, United States, ²radiology, Washington university school of medicine in St. Louis, St. Louis, MO, United States

Imaging in ultra-high field (7T and above) has pros and cons as compared with 3T. The pros are increases in SNR and frequency shifts that helps in separating the available biophysical compartments in the brain. Con is the increased effect of magnetic field inhomogeneity on T2* that renders the quantification unreliable. Therefore, the field correction methods are critical for ultra-high field quantitative T2* mapping. We have demonstrated the application of Voxel Spread Method, a post-processing technique, to addresses this issue at 7T. F-term correction significantly reduces magnetic field inhomogeneity effects for quantitative T2* mapping.

Jan Ole Pedersen¹, Vincent O. Boer², Oula Puonti², Jaco M. Zwanenburg³, and Esben Thade Petersen^2,4
1Philips Healthcare, Copenhagen, Denmark, ²Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital, Hvidovre, Denmark, ³Department of Radiology, University Medical Center Utrecht, Utrcecht, Netherlands, ⁴Section for Magnetic Resonance, DTU Health Tech, Technical University of Denmark, Kgs Lyngby, Denmark

In order to ease radiological assessment of 7T MP-FLAIR images, we have developed a fully automated bias-field correction that reduces apparent signal loss caused by inhomogeneities in the RF transmit field. Using simulations and measurements of the transmit field, the bias-field can be corrected for without relying on the simplifying assumptions used in a typical bias-field correction. The algorithm is expandable to other TSE-based sequences and adds limited additional scan time.

Sofia Chavez^1,2
1Centre for Addiction and Mental Health (CAMH), Toronto, ON, Canada, ²Psychiatry, University of Toronto, Toronto, ON, Canada

Quantitative MRI requires accurate knowledge of the spatially varying B1 and B0 fields in order to accurately account for their effects on relevant parameters included in the signal models. B1 maps in the human brain are commonly produced from a double-angle method (DAM) with  many variations in the implementations. B0 maps are usually estimated from the distortions in two 2D axial EPI scans acquired with opposing phase encode directions (topup). Here, we propose to integrate the SE-EPI scan requirements for B0 mapping with topup and B1 mapping with the DAM, for simultaneous B1 and B0 mapping with reduced scan time.

Kain Kyle^1,2 and Chenyu Wang^1,2
1Sydney Neuroimaging Analysis Centre, Sydney, Australia, ²University of Sydney, Sydney, Australia

Correction of bias fields is a common component in MRI analysis pipelines.  We present a novel optimisation method to improve the white matter segmentation used for white matter weighted N4 correction, and validate using atrophy in a healthy control cohort.