To assess the quantitative performance of three-dimensional (3D) fast spin-echo (FSE) Dixon amide proton transfer (APT) imaging of brain tumors compared with B0 correction with separate mapping methods.Methods:
Twenty-two patients with brain tumors (54.2 ± 18.7 years old, 12 males and 10 females) were scanned at 3 Tesla (T). Z-spectra were obtained at seven different frequency offsets at ±3.1 ppm, ± 3.5 ppm, ± 3.9 ppm, and −1560 ppm. The scan was repeated three times at +3.5 ppm with echo shifts for Dixon B0 mapping. The APT image corrected by a three-point Dixon-type B0 map from the same scan (3D-Dixon) or a separate B0 map (2D-separate and 3D-separate), and an uncorrected APT image (3D-uncorrected) were generated. We compared the APT-weighted signals within a tumor obtained with each 3D method with those obtained with 2D-separate as a reference standard.Results:
Excellent agreements and correlations with the 2D-separate were obtained by the 3D-Dixon method for both mean (ICC = 0.964, r = 0.93, P < 0.0001) and 90th-percentile (ICC = 0.972, r = 0.95, P < 0.0001) APT-weighted signals. These agreements and correlations for 3D-Dixon were better than those obtained by the 3D-uncorrected and 3D-separate methods.Conclusion:
The 3D FSE Dixon APT method with intrinsic B0 correction offers a quantitative performance that is similar to that of established two-dimensional (2D) methods.