Comparison of Phase Estimation Methods for Quantitative Susceptibility Mapping Using a Rotating-Tube Phantom

Figure 4

Each row in this figure shows five different examples of sequence (at 1 mm resolution, TR = 45 ms) and phase estimation method pairs for three different angles of rotation in different tubes. These examples were selected to illustrate the spatial nature of different phase estimations errors and artifacts. The corresponding plots in the last column show the resulting frequency measurements as a function of angle of rotation (modulo 180°), for a voxel inside the tube, after frequency referencing (blue circle). The predicted frequency offset as obtained from equation (1) is also shown in solid red line. The background ROI (#9) used is shown with a square overlaid on the frequency maps. (a) SEGE + GBC shows a phase wrapping error in tube 3 in the frequency reference ROI for the 160-degree angle. (b) SEGE + Laplace shows smoothly varying frequency maps; however, the values deviate from the expected result at every angle. (c) MEGE + Slope, without 3D phase unwrapping, shows the presence of phase wrapping in areas with large frequency values. The frequency reference ROI is within a phase wrapped region across all angles. (d, e) MEGE + MAGPI-unopt and MAGPI’s phase estimation show frequency maps consistent with values predicted from the model.

(a)

(b)

(c)

(d)

(e)