Johanna Vannesjö¹, David Brunner¹, Christoph Barmet¹, and Klaas Paul Pruessmann^1
1Institute for Biomedical Technology, University and ETH Zurich, Zurich, Switzerland

Artefacts in anatomical brain images stemming from field fluctuations caused by breathing have previously been reported, and a navigator-based correction scheme has been shown. In this work, concurrent field monitoring was used to observe field shifts correlated with breathing, and to correct for breathing-related artefacts by including the monitored field evolutions in the image reconstruction. One advantage of the field monitoring approach, is that no alteration or extension of the imaging sequence is required.

Daniel Nicolas Splitthoff¹, and Maxim Zaitsev^1
1Dept. of Radiology, Medical Physics, Unversity Medical Center Freiburg, Freiburg, Germany

Projections have been suggested for estimating B0 inhomogeneities. A detailed analysis of the phase difference of projections is given and put into the framework of linear algebra. The insights gained from this analysis lead to the conclusion that for proper detection of inhomogeneities the cross talk of the different orders need to be taken into account. We here present a solution of how to measure the cross talk when no further a priori information is available and show the benefit in a phantom measurement.

Lars Kasper^1,2, Bertram Jakob Wilm¹, Christoph Barmet¹, and Klaas Paul Prüssmann^1
1University and ETH Zurich, Institute for Biomedical Engineering, Zurich, Zurich, Switzerland, ²University of Zurich, Laboratory for Social and Neural Systems Research, Zurich, Zurich, Switzerland

The point spread function (PSF) is a comprehensive concept to describe the imaging and reconstruction process in MRI. Its analysis gives insight into the signal model of MR sequences as well as their imperfections, thus revealing true resolution and typical artifacts. By definition, the PSF describes the mapping of a point source of MR signal onto pixels of an image reconstruction matrix. We take this definition literally and use a miniaturized water-filled NMR field probe to determine the PSF experimentally. Hereby, we treat the MR scanner as a black box and infer information about the reconstruction characteristics solely from probing.

Iulius Dragonu¹, Thomas Lange¹, Nicoleta Baxan¹, Jürgen Hennig¹, and Maxim Zaitsev^1
1Dept. of Radiology, Medical Physics, University Medical Center Freiburg, Freibug, Baden-Wuerttemberg, Germany

Single-shot EPI is a fast technique allowing the acquisition of an image following a single RF excitation. However, EPI is prone to geometric distortions in presence of magnetic field inhomogeneities. We propose here a new method based on the modelling of the PSF data signal to allow accelerated acquisition for accurate geometric distortion corrections. Fully sampled PSF data of healthy volunteers were acquired. Undersampling factors up to 12.8 in the PSF-encoding direction were subsequently simulated. The pixel shift map obtained with the undersampled data and fully sampled data were compared. For each experiment the maximum error was below 0.15 pixels.

Trong-Kha Truong¹, Nan-kuei Chen¹, and Allen W Song^1
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Partial Fourier imaging is typically used in DTI to increase the SNR. However, eddy currents induced by the diffusion gradients lead to: 1) signal loss if the echo is shifted outside the acquired k-space, 2) partial Fourier reconstruction errors if the echo is shifted outside the central k-space, and 3) variation of the effective TE, resulting in additional T2*-weighting. All three types of artifact vary with location and diffusion direction, causing errors in the diffusion tensor. Here, we propose a novel acquisition and post-processing method that can effectively correct for all three types of artifact while maintaining a high SNR.

Matthew F Koff¹, Catherine Lee Hayter¹, Parina Shah¹, Kevin M Koch², Theodore T Miller^1,3, and Hollis G Potter^1,3
1Department of Radiology and Imaging, Hospital for Special Surgery, New York, NY, United States, ²Applied Science Laboratory, General Electric Healthcare, Waukesha, Wisconsin, United States, ³Weill Cornell Medical College of Cornell University, New York, NY, United States

Significant susceptibility artifacts occur when performing MRI around orthopedic hardware. This study evaluated standard of care 2D FSE imaging with the multi-acquisition variable-resonance image combination (MAVRIC) technique. Patients with joint replacements (hip or shoulder) were scanned using optimized 2D FSE and MAVRIC sequences. MAVRIC scans improved the visualization of the synovium, bone and supraspinatus tendon in the shoulder. MAVRIC scans resulted in increased detection of synovitis, peri-prosthetic osteolysis and supraspinatus tendon tears, resulting in a change in diagnosis almost 50% of the cases. This study further supports the use of MAVRIC for clinical implementation.

Catherine Lee Hayter¹, Hollis G Potter^1,2, Douglas E Padgett³, Giorgio Perino⁴, and Bryan J Nestor^3
1Department of Radiology and Imaging, Hospital for Special Surgery, New York, NY, United States, ²Weill Cornell Medical College of Cornell University, New York, NY, United States, ³Department of Orthopedic Surgery, Hospital for Special Surgery, New York, NY, United States, ⁴Department of Pathology, Hospital for Special Surgery, New York, NY, United States

This study assessed the ability of MRI to detect different qualitative patterns of synovitis in symptomatic patients with hip arthroplasty compared to asymptomatic controls. We hypothesized a distinct, qualitative synovial MRI pattern would exist for metal on metal wear, metal on polymeric debris and aseptic lymphocytic vasculitis-associated lesions (ALVAL), that would be concordant with histologic findings. MRI could distinguish between tissue containing particulate debris and tissue without debris. MRI was sensitive in detecting polymeric debris, but did not detect the presence of metal in all samples. ALVAL elicited a specific synovial pattern on MRI that was highly concordant with histology.

Christina A. Chen¹, Weitian Chen², Stuart B. Goodman³, Brian A. Hargreaves³, Kevin M. Koch², Wenmiao Lu⁴, Anja C. Brau², Hillary J. Braun³, and Garry E. Gold^3
1Radiology, Stanford University, Stanford, CA, United States, ²GE Healthcare Applied Science Lab, ³Stanford University, ⁴Nanyang Technological University

Slice Encoding for Metal Artifact Correction (SEMAC) is a recently developed MRI method that corrects for the metal-induced artifact that has previously limited the diagnostic value of postoperative images. This study found SEMAC to accurately measure metallic implant rotation in the knee, as implant misalignment is an important cause of implant pain and revision surgery. In addition, SEMAC significantly reduces artifact compared to fast spin echo in subjects, allowing for visualization of knee pathology that was able to guide patient management.

Pauline Wong Worters¹, Kim Butts Pauly¹, and Brian A Hargreaves^1
1Stanford University, Stanford, CA, United States

Slice Encoding for Metal Artifact Correction (SEMAC) is a robust method for resolving spatial distortion of tissue around metal in MRI. This method, as well as an alternative hybrid MAVRIC sequence, uses View-Angle Tilting (VAT) to correct for distortion in the readout direction. However, VAT imposes timing limitations in order to reduce blurring due to the RF profile. In this work, a shear post-processing method is proposed to replace VAT to correct for readout distortions. Shear correction allows for longer readout acquisitions (e.g., by using a lower readout bandwidth) and avoids the timing limitations imposed by VAT, while maintaining effective artifact correction.

Kevin M Koch¹, Kevin F King¹, Weitian Chen², Garry Gold³, and Brian A Hargreaves^3
1Global Applied Science Laboratory, GE Healthcare, Waukesha, WI, United States, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States,³Department of Radiology, Stanford University, Stanford, CA, United States

MR capabilities when imaging in the direct vicinity of metallic devices have been substantially improved with the development of 3D-MSI methods. The MAVRIC, SEMAC, and VS-MSI techniques have shown promising clinical capabilities in diagnosing soft-tissue pathology in previously inaccessible regions. Here, we discuss some limitations on how close 3D-MSI, or any technique relying on frequency-encoding, can image near metal in the presence of substantial local induction gradients. The presented analysis and results can aid in explaining residual artifacts in 3D-MSI and in predicting the effective spectral coverage required by the techniques.