Single-voxel short-TR multi-TI multi-TE STEAM MRS for water-fat relaxometry.

PURPOSE: To propose a short-TR multi-TI multi-TE (SHORTIE, ['shȯr-te]) STEAM single-voxel MRS acquisition scheme for the simultaneous assessment of T1 relaxation, T2 relaxation, and the proton density fat fraction at reduced scan times when compared with conventional long-TR multi-TI STEAM and long-TR multi-TE STEAM single-voxel MRS. METHODS: Theoretical analysis for multi-TI (TI = 10, 100, 500, 1500 ms; scan time = 2:43 minutes), multi-TE (TE = 12, 15, 20, 25 ms; scan time = 2:24 minutes), and SHORTIE STEAM (all TI and TE combinations; scan time = 2:52 minutes) was carried out including Cramér-Rao lower bound and parameter estimation efficiency analysis for T1 (150-2000 ms) and T2 (5-150 ms) relaxation. The SHORTIE STEAM acquisition was compared with multi-TI STEAM and multi-TE STEAM in water-fat phantoms and in a human in vivo study of the adipose tissue depot in the supraclavicular fossa in 7 volunteers at 3 T. RESULTS: Cramér-Rao lower bound analysis revealed similar to increased variances for T1 and T2 estimators for SHORTIE STEAM. Parameter efficiency analysis demonstrated superior performance of SHORTIE, particularly for shorter T1 and T2 when compared with multi-TI STEAM and multi-TE STEAM. For the phantom data, linear regression and Bland-Altmann analysis yielded a slope/intercept/mean difference of 1.07/-15.40/-17.18 for T1 (in ms; r = 0.999), 0.93/+1.32/+1.09 for T2 (in ms; r = 0.995), and 0.98/-0.04/+0.78 for the fat fraction (in percent; r = 0.999); and for the in vivo data 1.08/+1.77/-62.2 for T1 (r = 0.994), 0.88/+6.69/-1.55 for T2 (r = 0.884), and 0.56/+34.40/-0.46 for the fat fraction (r = 0.673), respectively. CONCLUSION: The SHORTIE STEAM acquisition allows shorter scan times for the simultaneous probing of relaxation properties and spectral content in the water-fat environment when compared with combined long-TR multi-TI, and long-TR multi-TE STEAM.