Free-breathing high isotropic resolution quantitative susceptibility mapping (QSM) of liver using 3D multi-echo UTE cones acquisition and respiratory motion-resolved image reconstruction

Abstract

Purpose: To enable free-breathing and high isotropic resolution liver quantitative susceptibility mapping (QSM) using 3D multi-echo UTE cones acquisition and respiratory motion-resolved image reconstruction.

Methods: Using 3D multi-echo UTE cones MRI, a respiratory motion was estimated from the k-space center of the imaging data. After sorting the k-space data with estimated motion, respiratory motion state-resolved reconstruction was performed for multi-echo data followed by nonlinear least-squares fitting for proton density fat fraction (PDFF), $R_2^{*}$ , and fat-corrected B₀ field maps. PDFF and B₀ field maps were subsequently used for QSM reconstruction. The proposed method was compared with motion-averaged (gridding) reconstruction and conventional 3D multi-echo Cartesian MRI in moving gadolinium phantom and in vivo studies. Region of interest (ROI)-based linear regression analysis was performed on these methods to investigate correlations between gadolinium concentration and QSM in the phantom study and between $R_2^{*}$ and QSM in in vivo study.

Results: Cones with motion-resolved reconstruction showed sharper image quality compared to motion-averaged reconstruction with a substantial reduction of motion artifacts in both moving phantom and in vivo studies. For ROI-based linear regression analysis of the phantom study, susceptibility values from cones with motion-resolved reconstruction ( ${\mathrm{QSM}}_{\mathrm{ppm}}$ = 0.31 $\times {\text{gadolinium}}_{\mathrm{mM}}+$ 0.05, $R^2$ = 0.999) and Cartesian without motion ( ${\mathrm{QSM}}_{\mathrm{ppm}}$ = 0.32 $\times {\text{gadolinium}}_{\mathrm{mM}}+$ 0.04, $R^2$ = 1.000) showed linear relationships with gadolinium concentrations and showed good agreement with each other. For in vivo, motion-resolved reconstruction showed higher goodness of fit ( ${\mathrm{QSM}}_{\mathrm{ppm}}$ = 0.00261 $\times R_{2_{s^{-1}}}^{*}-$ 0.524, $R^2$ = 0.977) compared to motion-averaged reconstruction ( ${\mathrm{QSM}}_{\mathrm{ppm}}$ = 0.0021 $\times R_{2_{s^{-1}}}^{*}-$ 0.572, $R^2$ = 0.723) in ROI-based linear regression analysis between $R_2^{*}$ and QSM.

Conclusion: Feasibility of free-breathing liver QSM was demonstrated with motion-resolved 3D multi-echo UTE cones MRI, achieving high isotropic resolution currently unachievable in conventional Cartesian MRI.

Keywords: 3D multi-echo UTE cones k-space sampling trajectory; free-breathing liver QSM; liver iron overload; motion-resolved image reconstruction.

FIGURE 1

3D multi-echo UTE cones k-space sampling trajectory with pseudo-random view ordering and gradient waveforms. 3D multi-echo UTE cones k-space showing 300 cone interleaves with pseudo-random view ordering in a 3D $k_x-k_y-k_z$ view (A), in a $k_x-k_z$ plane, and in a $k_x-k_y$ plane (B). (B) One exemplary cone interleaf is shown as thick black lines. (C) The gradient waveforms for cone interleaves acquisition.

FIGURE 2

Pipeline for respiratory motion-resolved reconstruction and QSM reconstruction. After the acquisition of 3D multi-echo UTE cones MRI data, SVD of direct current (DC, k-space center) components is performed along coil dimension to extract respiratory signals for each echo. (A) Additional SVD is applied to previously extracted respiratory signals to estimate improved respiratory signal. Based on the estimated respiratory signal, the number of motion states is chosen such that each of the motion state has the same number of cone interleaves. (B) Then, motion-resolved reconstruction is performed. After choosing one motion state data, water/fat separation and background field removal are performed from multi-echo complex-valued signals to acquired local (tissue) field map, proton density fat fraction (PDFF) map, and ${\mathrm{R}}_2^{\ast }$ map. (C) Then, dipole inversion is performed to estimate the susceptibility map.

FIGURE 3

Image quality comparison of gadolinium phantom. (A) Magnitude images, local field maps, and susceptibility maps from Cartesian without periodic table motion, cones without periodic table motion, cones with periodic table motion using motion-averaged, and motion-resolved reconstructions. Conventional gridding reconstruction was performed for cones without periodic table motion. (B) Detailed information regarding periodic table motion. (C) Gadolinium phantom configuration. (A) The yellow arrow on the top row indicates the $B_0$ direction.

FIGURE 4

Result of linear regression analysis. (A) ROI-based linear regression between gadolinium (Gd) concentrations and susceptibility values from Cartesian without periodic table motion (red dashed line), cones without periodic table motion (green dashed line), cones with motion-resolved (blue dashed line), and the reference (orange solid line). The reference susceptibility values are 0.41, 0.82, 1.63, and 3.26 ppm for gadolinium concentrations of 1.25, 2.5, 5, and 10 mM, respectively. (B) ROI-based linear regression between susceptibility values from Cartesian without periodic table motion and two cones without motion + gridding (green dashed line) and with motion + motion-resolved (blue dashed line) with the line of unity (red solid line).

FIGURE 5

Image quality comparison on a representative healthy volunteer. Axial (A) and coronal (B) views of magnitude images, ${\mathrm{R}}_2^{\ast }$, and susceptibility maps from BH Cartesian, FB Cones with motion-averaged, and motion-resolved reconstructions. ROI-based mean ± SD ${\mathrm{R}}_2^{\ast }$ and susceptibility values are shown at the bottom of each map. Circular ROIs are drawn in red. (C,D) Zoomed-in views of the green boxes in susceptibility maps from FB Cones with motion-averaged and motion-resolved reconstructions.

FIGURE 6

Image quality comparison on one patient with liver iron overload. Axial (A) and coronal (B) views of magnitude images, ${\mathrm{R}}_2^{\ast }$, and susceptibility maps from BH Cartesian, FB Cones with motion-averaged, and motion-resolved reconstructions. ROI-based mean ± SD ${\mathrm{R}}_2^{\ast }$ and susceptibility values are shown at the bottom of each map. Circular ROIs are drawn in red. (C) Zoomed-in view of the green boxes in susceptibility maps from FB Cones with motion-averaged and motion-resolved reconstructions.

FIGURE 7

Image quality comparison on a healthy volunteer with deep breathing (severe motion artifacts). Axial (A) and coronal (B) views of magnitude images, ${\mathrm{R}}_2^{\ast }$, and susceptibility maps from BH Cartesian, FB Cones with motion-averaged, and motion-resolved reconstructions. ROI-based mean ± SD ${\mathrm{R}}_2^{\ast }$ and susceptibility values are shown at the bottom of each map. Circular ROIs are drawn in red. (C,D) Zoomed-in views of the green boxes in susceptibility maps from FB Cones with motion-averaged and motion-resolved reconstructions.

FIGURE 8

Linear regression analysis result. ROI-based linear regression results between median susceptibility values and corresponding ${\mathrm{R}}_2^{\ast }$ values of 12 human subjects from BH Cartesian (orange solid line, top row), FB Cones with motion-averaged reconstruction (blue solid line, top row), and FB Cones with motion-resolved reconstruction (red solid line, top row) are shown. Within the ROIs, the entire susceptibility values versus the corresponding ${\mathrm{R}}_2^{\ast }$ values from BH Cartesian (orange solid line, bottom row), FB Cones with motion-averaged reconstruction (blue solid line, bottom row), and FB Cones with motion-resolved reconstruction (red solid line, bottom row) are shown.