Oscillating-gradient spin-echo (OGSE) prototype sequences can shorten the effective diffusion time (Δ_eff) and add information regarding the internal structure of lesions. The purpose of our study was to investigate the utility of changes in diffusivity between short and long Δ_eff values in distinguishing different types of extra-axial brain tumors. Twelve patients with meningiomas, 13 with acoustic neuromas, and 11 with pituitary adenomas were scanned using a 3T magnetic resonance (MR) scanner. Diffusion tensor imaging (DTI) was performed using b-values of 0 and 1000 s/mm². OGSE imaging using a trapezoid-cosine waveform was performed with Δ_eff values of 6.5 and 35.2 ms. The diffusivity of each tumor was measured after setting a manually defined region of interest (ROI). The relative percentage change between the shorter and longer Δ_eff values was then calculated. The relative percentage changes of the three tumor types were compared using the Steel-Dwass test. The relative percentage changes of λ2, λ3, and mean diffusivity (MD) were significantly higher in pituitary adenomas than in meningiomas and acoustic neuromas, but not significantly different between meningiomas and acoustic neuromas. The dependence of λ2, λ3, and MD on Δ_eff values between 6.5 and 35.2 ms was stronger in pituitary adenomas than in meningiomas and acoustic neuromas, suggesting differences in internal tissue structure. Our results showed that the use of DTI with a shorter Δ_eff provides additional information about the microstructure of brain tumors. OGSE imaging may help in the differentiation of different types of tumors that show atypical imaging findings on conventional MRI.

　The diffusion properties in micro-restricted structures to measure the radial diffusivity for capillary phantoms using oscillating gradient spin echo (OGSE) was investigated. For small restricted structure sizes ranging from 6-12 μm, the OGSE was demonstrated to be beneficial. The diffusion time requires further adjustment for the structure size.

　As the three-dimensional fast spin echo (3D FSE) imaging uses greater than 100 refocusing pulses which generate several higher-order echoes, its Bloch simulation is challenging. The simulation was performed using a numerical phantom with 256 cube voxels and 320 subvoxels (isochromats) in the readout direction. The simulation was performed using approximately 1.58 billion isochromats and the calculation time was nearly 15 hours using an RTX 2080Ti graphics processing unit having 11 GB memory. The calculated result demonstrated that the 3D FSE was correctly simulated.

　We examined the diffusional kurtosis (K) inference by synthetic Q-space learning (synQSL), and the effect of diffusion-weighted image (DWI) denoising. In this study, we compared the results of two approaches for K inference : synQSL and least-squares fitting in denoised DWI. For synSQL, the effect of bias correction was also examined. Through experiments using 3.0T MRI data of healthy volunteers, we confirmed the superiority of the synQSL approach with bias correction.

　This retrospective study analyzed the magnetic resonance lymphography (MRL) findings of 112 lower limbs of 56 lymphedema patients, and identified characteristic patterns of lymphatic impairment that change with the location in lower limbs and clinical stage. The authors proposed an MRL staging system that enables a new patient stratification based on the severity of the lymphatic injury, thus potentially improving clinical management.

　The purpose of this phantom study was to assess the effectiveness of chemical exchange saturation transfer imaging (CEST) for detecting anticancer agents (chemoCEST) using clinical 3T MRI. The magnetization transfer ratio asymmetry (MTRasym) was used to map the chemoCEST contrast, which demonstrated that the anticancer drug concentration was proportional to chemoCEST. CEST MR was used to identify the anticancer drugs which corresponded to the amide (-NH₂) and hydroxyl (-OH) groups. chemoCEST imaging can serve as a new imaging tool to detect anticancer drugs and monitor their delivery.

　Pre-processing is essential for various quantitative diffusion analyses. In particular, noise reduction, Gibbs artifact removal, and distortion correction using FSL (TOPUP & EDDY) are performed. However, these pre-processing calculations take a long time. We aimed to simplify the distortion correction technique. A learning model for distortion correction was constructed by 3D U-net using image data before and after FSL distortion correction. Compared to Pre FSL DWI, predicted DWI improved the SSIM/PSNR and exhibited a significant difference. As a result, the learning model of FSL using 3D-U-net was able to correct the distortions in the high b-value DWI.