S Sivaram Kaushik1, Cathy Marszalkowski2, Kevin M Koch3,2. 1. Department of Biophysics, Medical College of Wisconsin, Milwaukee, Wisconsin, USA. skaushik@mcw.edu. 2. Department of Radiology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA. 3. Department of Biophysics, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.
Abstract
PURPOSE: By combining images created at distinct frequency offsets from the Larmor frequency, three-dimensional (3D) multispectral imaging (3D-MSI) sequences help overcome the large spatial frequency dispersion caused by metal implants. This frequency dispersion, however, varies with the implant size, orientation, and composition. Using a MAVRIC 3D-MSI acquisition, we sought to prospectively calibrate the spectral coverage needed for 3D-MSI scans. This calibration should offer a significant improvement to image quality, and reduce the scan time. METHODS: The 24 spectral bins from the calibration scan were used to generate a map of frequency offsets around the implant. The magnitude image was used to remove any outliers in the associated frequency offset map, and this processed map was used to determine the cutoff frequency offset and, hence, number of spectral bins. This approach was tested in 13 subjects, by retrospectively reconstructing MAVRIC-SL images with fewer spectral bins. Subsequently, the spectral coverage for MAVRIC-SL images was prospectively calibrated in six subjects, and based on the cutoff frequency offset, these images were acquired with fewer spectral bins. RESULTS: With fewer spectral bins, both retrospectively and prospectively calibrated MAVRIC-SL images adequately delineated the implant boundary. CONCLUSION: Incorporating this calibration procedure into future 3D-MSI exams will help improve image signal-to-noise ratio, reduce scan time, and significantly improve clinical workflow when imaging near orthopedic implants. Magn Reson Med 76:1494-1503, 2016.
PURPOSE: By combining images created at distinct frequency offsets from the Larmor frequency, three-dimensional (3D) multispectral imaging (3D-MSI) sequences help overcome the large spatial frequency dispersion caused by metal implants. This frequency dispersion, however, varies with the implant size, orientation, and composition. Using a MAVRIC 3D-MSI acquisition, we sought to prospectively calibrate the spectral coverage needed for 3D-MSI scans. This calibration should offer a significant improvement to image quality, and reduce the scan time. METHODS: The 24 spectral bins from the calibration scan were used to generate a map of frequency offsets around the implant. The magnitude image was used to remove any outliers in the associated frequency offset map, and this processed map was used to determine the cutoff frequency offset and, hence, number of spectral bins. This approach was tested in 13 subjects, by retrospectively reconstructing MAVRIC-SL images with fewer spectral bins. Subsequently, the spectral coverage for MAVRIC-SL images was prospectively calibrated in six subjects, and based on the cutoff frequency offset, these images were acquired with fewer spectral bins. RESULTS: With fewer spectral bins, both retrospectively and prospectively calibrated MAVRIC-SL images adequately delineated the implant boundary. CONCLUSION: Incorporating this calibration procedure into future 3D-MSI exams will help improve image signal-to-noise ratio, reduce scan time, and significantly improve clinical workflow when imaging near orthopedic implants. Magn Reson Med 76:1494-1503, 2016.
Authors: Alissa J Burge; Gabrielle P Konin; Jennifer L Berkowitz; Bin Lin; Matthew F Koff; Hollis G Potter Journal: Clin Orthop Relat Res Date: 2019-09 Impact factor: 4.176
Authors: Curtis N Wiens; Nathan S Artz; Hyungseok Jang; Alan B McMillan; Kevin M Koch; Scott B Reeder Journal: Magn Reson Med Date: 2017-08-21 Impact factor: 4.668
Authors: Kelly C Zochowski; Mauro A Miranda; Jacky Cheung; Erin C Argentieri; Bin Lin; S Sivaram Kaushik; Alissa J Burge; Hollis G Potter; Matthew F Koff Journal: AJR Am J Roentgenol Date: 2019-08-15 Impact factor: 3.959