T2-mapping has promise to evaluate kidney pathophysiology. Prior studies show a large variance in kidney T2, likely due to the differing acquisition sequences. To compare four T2-mapping sequences to investigate kidney T2. Phantom and prospective in vivo assessments. ISMRM/NIST and QASPER phantoms; 8 healthy volunteers (4 female, 30 ± 8 years). 3 T, spin echo-echo planar imaging (SE-EPI), multi-echo-spin echo (ME-SE), Gradient and Spin Echo (GraSE) vendor-provided sequences, and custom T2-prepared EPI, Dual-echo B0-mapping and DREAM B1-mapping. T2-mapping accuracy in the ISMRM/NIST phantom in the presence of B0 frequency offset and B1 + by scaling of flip angles, and in the QASPER phantom in the presence of diffusion by altering pump rate compared to being turned off. Participants underwent a single 45-min exam to collect four T2-mapping sequences, B0 and B1 maps. In vivo T2 values compared across sequences and the influence of B0 and B1 + was evaluated. Shapiro-Wilk, Wilcoxon signed-rank, Student's t-test, Coefficient of Variation, Pearson's correlation coefficient, linear mixed effect model. p < 0.05 considered statistically significant. SE-EPI, ME-SE, GraSE, and T2-prepared EPI had a mean absolute scaled error of 0.52, 0.52, 0.36, and 0.27 over the kidney T2 range of the ISMRM/NIST phantom. GraSE was most robust to perturbations in B0/B1 +. In the QASPER phantom, SE-EPI was highly sensitive to diffusion leading to T2 shortening (66%), while multi-echo sequences had lower diffusion sensitivity ordered by shortest echo spacing (ME-SE 81%, T2-prepared EPI 90%, GraSE 95% reduction in T2). In vivo, SE-EPI measured T2 was significantly lower than multi-echo sequences, and SE-EPI and T2-preparation had a -0.52 ± 0.08 and -0.57 ± 0.06 ms/% dependence on B1 +. To reduce B0, B1 +, and diffusion sensitivity for kidney T2-mapping, a multi-echo sequence spanning echo times up to the kidney T2 (~140 ms at 3 T) is recommended. Collecting data with different echo spacings can isolate the diffusion-related T2 component. 2. Stage 1.
Daniel et al. (Thu,) studied this question.