Comparison of spin-echo echo planar imaging and gradient-recalled echo sequences in magnetic resonance elastography of liver at 1.5T same MRI scanner.

Abstract

Magnetic resonance elastography (MRE) is used to measure liver stiffness with gradient-recalled echo (GRE)-based and spin-echo echo planar imaging (SE-EPI)-based sequences. We compared the liver stiffness (LS) values of the two sequences on a 1.5-T MR imaging scanner. This is a retrospective study. An MRE imaging section was obtained from a horizontal section of the liver. Region of interest was drawn on the elastogram, and the mean LS and pixel values were measured and compared. The correlations between proton density fat fraction, R2* values, and biochemical data from electronic medical records were confirmed, and multivariate analysis was performed. The mean LS values were 3.01±1.78kPa for GRE and 3.13±1.57kPa for SE-EPI, showing excellent agreement and a strong correlation between the two sequences (correlation coefficient r=0.96). The mean pixel values were 369.5±142.7 pixels for GRE and 490.1±197.9 pixels for SE-EPI, showing a significant difference by the Wilcoxon rank sum test (p<0.01). There were no LS unmeasurable cases in SE-EPI, but seven (2.5%) were unmeasurable in GRE, and multivariate analysis showed a significant difference with p<0.001 in R2* values (mean, 92.7Hz) for the GRE method. The GRE and SE-EPI methods were comparable for LS measurements in 1.5-T liver MRE, indicating that SE-EPI MRE is more useful because GRE MRE may not measure cases with high R2* values and the region of interest tends to be smaller.