Abstract
Background Diffusion-weighted imaging acquired using echo-planar imaging is susceptible to distortion due to its sensitivity to static magnetic field inhomogeneity and eddy currents induced by motion-probing gradients. Although "topup" and "eddy" in the FMRIB (Functional Magnetic Resonance Imaging of the Brain) Software Library (FSL) are widely used for distortion correction in research, their reliance on offline post-processing, as well as the considerable time and multiple steps required, limits their feasibility in routine clinical practice. Diffusion tensor imaging (DTI) with reverse encoding distortion correction (RDC), which can be obtained directly on the MRI scanner console, reduces post-processing time by eliminating the need for FSL. Objective This study aimed to visually assess distortions in non-RDC fractional anisotropy (FA(non-RDC)), topup- and eddy-corrected fractional anisotropy (FA) in FSL (FA(FSL)), and FA(RDC) relative to MPRAGE (Magnetization-Prepared Rapid Gradient-Echo), and to validate FA and apparent diffusion coefficient (ADC) derived from RDC DTI. Materials and methods This prospective study included 10 healthy volunteers (mean age, 40.1 ± 5.7 years). Non-RDC b(0) images, non-RDC DTI, RDC DTI, and MPRAGE were acquired on a 3T MRI scanner. Non-RDC b(0) images and non-RDC DTI were subsequently processed using topup and eddy in FSL for distortion correction. FA and ADC were registered to the standard space, and the mean values were calculated for each of the 50 regions of interest (ROIs) defined using the Johns Hopkins University-International Consortium of Brain Mapping (JHU-ICBM)-labels 1 mm white matter atlas. All statistical analyses were performed using Bonferroni-corrected Wilcoxon signed-rank tests, with the significance level set at P < 0.05. Results The processing time for RDC DTI was approximately two minutes, whereas topup and eddy in FSL required about 90 minutes per subject. Distortions observed at the cranial base and near the frontal sinuses on FA(non-RDC) were substantially mitigated in FA(RDC), comparable to FA(FSL). Although 10 ROIs in FA(non-RDC) and nine ROIs in FA(FSL) showed significant differences compared with FA(RDC), the FA differences for each ROI were minimal, with a mean ± 95% confidence interval (CI) of -0.003 ± 0.006 and -0.002 ± 0.005. Furthermore, two ROIs in ADC(non-RDC) and none in ADC(FSL) showed significant differences compared with ADC(RD) (C); however, the ADC differences for each ROI were minimal, with a mean ± 95% CI of 2 ± 16 and -1 ± 10 × 10(-6) mm(2)/s. Conclusion RDC DTI effectively reduced distortions while yielding FA and ADC values comparable to those of non-RDC DTI and FSL DTI. Given the shorter post-processing time and fewer processing steps, RDC DTI may be considered promising for clinical applications.