Quantitative muscle water T₂ mapping using RF phase-modulated 3D gradient echo imaging

Figure S1. EPG sequence differentiation.

Figure S2: Sequence optimization script.

Figure S3: Estimation of T2fat in the subcutaneous fat from the calves of six subjects using the proposed sequence. For each subject, T2fat was estimated in the subcutaneous fat with a single-component model (assuming a FF of 1). The mean T2fat value across the 8 subjects obtained was 133 ± 4 ms. This value was then used in all simulations and for dictionary generation.

Figure S4: SNR estimated in the leg muscles of the same 6 subjects as in Figure S4, using the proposed sequence. Noise was defined as the signal's standard deviation in a circular ROI (of 5 pixels of diameter) within a single muscle and signal as the average amplitude in the same ROI.

Figure S5: A: number of excitations required to reach a steady state for $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ = 25 ms and FF = 0.3. A steady state is defined as the point at which both magnitude and phase differences between two successive excitations are below 0.1%. B/C: signal's magnitude and phase evolution over 1000 excitation for the 10 α, φ pairs used at $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ = 25 ms and FF = 0.3.

Figure S6: pSPGR signal magnitude and phase over the range of small phase increments ϕ for various tissues and field parameters. For all simulations, TR/TE = 5.5/2.25. Rows 1 and 2 (in blue): flip angle = 25°, rows 3 and 4 (in brown): flip angle = 5°.

Figure S7: $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ maps and Bland–Altman between two $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ measurements on the multi-vials phantom using a GRAPPA acceleration factor of 2 or by disabling GRAPPA.

Figure S8: Bland–Altman plot between two $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ measurements on a 15-vials phantom using 150 or 406 radial spokes. There was a small bias present (0.38 ± 1.81 ms).

Figure S9: Bland–Altman plot between two $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ measurements on the muscle legs of 6 volunteers.

Table S1: $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ and FF measurements of the 15-vial phantom obtained using the reference methods (MRS-STEAM and 3-point Dixon) and both pSPGR implementations (Cartesian and radial pSPGR). The table also reports the relative error in $$$ \mathrm{T}{2}_{{\mathrm{H}}_2\mathrm{O}} $$$ and the FF difference of the pSPGR acquisitions compared to the references. Values are represented as mean ± standard deviation for imaging methods and MRS with its 95% confidence interval.