Motion-robust proton density fat fraction and T₂^* mapping in supraclavicular adipose tissue using radial stack-of-stars imaging

Figures

• 

  Fig. 1. Anatomy of the supraclavicular fossa The underlying image depicts an exemplary coronal slice of a gradient echo image (first echo of a radial scan with scan parameters listed in Table 1) of the upper thorax and neck region. The red segmented region encircles the fat depot of the supraclavicular fossa with the attached axillary fat depot

• 

  Fig. 2. Water-fat separation of the digital phantom. The images on the left depict the simulated water and fat images of the anatomical body phantom from the superior part of the lungs to the neck. The SCV fossa and axillary fat depot are encircled by the red boundaries. The PDFF map on the top right is computed from the fat and water images. The ground truth \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ map on the bottom right expresses a sharp contrast between low-fat tissues such as muscle and the high-fat containing adipose tissues

• 

  Fig. 3. Simulated and measured \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 fluctuations. a The image depicts a simulated gradient echo image of the neck region. Two voxels are marked on the coronal slice of the simulated magnitude image. The signal inside the blue voxel placed within the SCV fossa experiences \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 fluctuations in the order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm {15\,\mathrm{\text {Hz}}}$$\end{document}±15Hz while the signal fluctuations inside the red voxel in the humerus bone are less pronounced. b A similar analysis was…

• 

  Fig. 4. Simulations: only \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 fluctuations without local tissue displacement. a The PDFF and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ maps from simulated data with additionally simulated \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 fluctuations are shown on the left and right side, respectively. The PDFF maps do not suffer from obvious artifacts with both simulated trajectories. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy}…

• 

  Fig. 5. Simulations: respiratory motion including \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 fluctuations and local tissue displacement. a The PDFF maps from simulated data with simulated respiratory motion on the left show visible artifacts in the adipose tissue around the lungs in the Cartesian map, while the radial map does not exhibit visible artifacts. The \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ maps on the right exhibit visual artifacts localized to the areas around the lungs and superior to the lungs in the Cartesian map, while there are artifacts visible over the whole FOV in the radial map, albeit of a smaller extent compared to Cartesian. b The PDFF differences to the ground truth on the left show the Cartesian errors affecting the tissues around the lungs…

• 

  Fig. 6. Repeatability of in vivo PDFF mapping in a volunteer. The example Cartesian PDFF maps of five repetitions are shown in the top row and radial SoS maps below. It is worth noting that the range of visualized PDFF values is decreased to highlight the variations in the Cartesian maps. As indicated by the arrows, there are large visual variations of PDFF in the Cartesian case inside the SCV fossa. There are no similar variations observable in the radial SoS maps. The analysis of an ROI indicated by a black circle in the maps on the left shows increased repeatability and reduced intra-ROI standard deviation using radial SoS over Cartesian sampling

• 

  Fig. 7. Repeatability of in vivo \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ mapping in a volunteer. The example slices of the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ maps show that Cartesian reconstructions in the top row overestimate \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ and appear noisy, while radial SoS \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt}…

• 

  Fig. 8. Example \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ improvement with temporal \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 correction. The two top rows show the uncorrected and corrected \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ maps, respectively. The arrows point to regions of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ overestimation in the…

• 

  Fig. 9. Median PDFF of all volunteers. The left column represents the difference in performance in PDFF mapping with Cartesian compared to radial SoS for all volunteers. The median value is less consistent with a higher variance using the Cartesian acquisitions. The comparison between uncorrected and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$B_0$$\end{document}B0 corrected radial reconstructions in the right column shows that the median gets slightly more consistent after correction

• 

  Fig. 10. Median \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ of all volunteers. The left column represents the difference in performance in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ mapping for Cartesian compared to radial SoS for all volunteers. The difference between Cartesian and radial SoS in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T_2}^*$$\end{document}T2∗ is even more striking compared to PDFF and radial sampling consistently improves repeatability. The comparison between uncorrected and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym}…