Magnetic resonance in medical sciences : MRMS : an official journal of Japan Society of Magnetic Resonance in Medicine最新文献

Purpose: We present a novel algorithm for the automated detection of cerebral microbleeds (CMBs) on 2D gradient-recalled echo T2* weighted images (T2*WIs). This approach combines a morphology filter bank with a convolutional neural network (CNN) to improve the efficiency of CMB detection. A technical evaluation was performed to ascertain the algorithm's accuracy.

Methods: In this retrospective study, 60 patients with CMBs on T2*WIs were included. The gold standard was set by three neuroradiologists based on the Microbleed Anatomic Rating Scale guidelines. Images with CMBs were extracted from the training dataset comprising 30 cases using a morphology filter bank, and false positives (FPs) were removed based on the threshold of size and signal intensity. The extracted images were used to train the CNN (Vgg16). To determine the effectiveness of the morphology filter bank, the outcomes of the following two methods for detecting CMBs from the 30-case test dataset were compared: (a) employing the morphology filter bank and additional FP removal and (b) comprehensive detection without filters. The trained CNN processed both sets of initial CMB candidates, and the final CMB candidates were compared with the gold standard. The sensitivity and FPs per patient of both methods were compared.

Results: After CNN processing, the morphology-filter-bank-based method had a 95.0% sensitivity with 4.37 FPs per patient. In contrast, the comprehensive method had a 97.5% sensitivity with 25.87 FPs per patient.

Conclusion: Through effective CMB candidate refinement with a morphology filter bank and FP removal with a CNN, we achieved a high CMB detection rate and low FP count. Combining a CNN and morphology filter bank may facilitate the accurate automated detection of CMBs on T2*WIs.

Purpose: The 8th edition of the American Joint Committee on Cancer staging system included the depth of invasion (DOI) for the T classification of oral cancer. However, no standardized method has been established to clinically measure the DOI. This study aimed to investigate the accuracy of MRI-based DOI for oral tongue squamous cell carcinoma (OTSCC) in each MRI sequence.

Methods: We enrolled 49 patients with histologically proven OTSCC, treated surgically between April 2017 and February 2021. We divided the DOI into three groups using 5 and 10 mm, the thresholds for determining the T stage, and retrospectively evaluated the agreement between MRI-based DOI and pathological DOI (pDOI) for each MRI sequence, axial T1-weighted imaging (T1WI), T2-weighted imaging with fat suppression (FS-T2WI), contrast-enhanced T1WI with fat suppression (CE-T1WI), diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) maps. We also divided the DOI into two groups using 3 mm, the threshold for considering elective neck dissection, and evaluated the overestimation rate of MRI-based DOI in lesions with pDOI ≤ 3 mm.

Results: With 5-mm and 10-mm divisions, the accuracy of the DOI assessment was highest on DWI (0.82, weighted kappa = 0.85). With a 3-mm division, the accuracy was also highest on DWI (0.87, kappa = 0.73). The overestimation rate of the MRI-based DOI in lesions with pDOI ≤ 3 mm was lowest on DWI (27.8%).

Conclusion: DOI on DWI exhibits a comparatively higher rate of concordance with pDOI. DWI may be more useful than other MRI sequences in evaluating the DOI of OTSCC.

A woman in her sixties with portosystemic shunt and hepatic encephalopathy underwent open mesenteric vein ligation, resulting in improved portal flow and blood ammonia. In this case, 4D flow MRI was a valuable diagnostic and follow-up tool, visualizing and quantifying physiological portal hemodynamics with features distinct from those of contrast-enhanced CT and digital subtraction angiography. Our case study highlights the value of 4D flow MRI for managing portosystemic shunts.

A chemically fixed Carnegie stage 23 (approximately 56 days of gestation) human embryo specimen was imaged using 3D spin-echo and gradient-echo sequences in a static magnetic field strength of 4.74T, and a quantitative susceptibility map was calculated using the 3D gradient-echo image. The acquired 3D microscopic images (90 μm cube voxel size) clarified the relationship between R₂ (transverse relaxation rate), R₂* (apparent transverse relaxation rate), and magnetic susceptibility in the heart, liver, kidney, and spinal cord. The results suggested that the R₂* and magnetic susceptibility in each tissue were probably due to paramagnetic iron ions originating from erythrocytes. The large R₂* (~130 s^-1) and magnetic susceptibility (~0.122 ppm) in the liver were attributed to its hemopoietic function. A large magnetic susceptibility (~0.116 ppm) was also observed in the spinal cord, but we conclude that more detailed future studies are needed.

Purpose: Neuromelanin is visualized by optimizing the conditions of longitudinal relaxation (T1)-weighted imaging (T1WI). Although it was originally developed in 2D imaging, 3D imaging has been also reported, and T1WI sequences with magnetization transfer (MT) pulses are now widely used in 3D gradient echo (GRE) sequences. In this study, we assert that the use of spectral presaturation with inversion recovery (SPIR) may also be useful as an alternative to MT pulses, and we optimize SPIR and compare it with MT.

Methods: Neuromelanin images with MT pulse and SPIR (flip angles [FAs] = 19º, 22º, and 25º) were acquired from 30 healthy volunteers. To achieve the same acquisition time of 5 min, the slab thickness of the MT images was less than 1/3 of those of the SPIR images; the acquisition areas for MT and SPIR were the brainstem and the whole brain, respectively. Visual and quantitative evaluation was performed and compared on the four sequences acquired for the substantia nigra pars compacta (SNc) and the locus coeruleus (LC). For visual assessment, we used the mean score from a 3-point scale by two evaluators. For quantitative evaluation, the contrast ratios of SNc and LC were calculated in comparison with the background tissue signal.

Results: In visual assessments, the mean scores of the SPIR FA19º and FA22º images were better than others in the SNc. Regarding LC, the SPIR FA22º image yielded the best mean score. In quantitative evaluations, the MT image was significantly lower than the other three images in SNc. Regarding LC, there were no significant differences among the four acquired images (MT and SPIR FA19º, FA22º, and FA25º).

Conclusions: Detection of neuromelanin in SNc and LC was improved by the use of SPIR compared to MT pulse in 3D neuromelanin imaging.

Purpose: The kidney iron deposition can cause kidney damage and renal insufficiency in paroxysmal nocturnal hemoglobinuria (PNH) patients. Assessment of iron deposition in the kidney is essential for the early diagnosis of renal damage in PNH patients. The purpose of this study was to evaluate kidney R2* (T2* reciprocals) values in PNH patients using the iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL-IQ).

Methods: Two radiologists measured the R2* values of the renal cortex in 14 PNH patients and 13 healthy volunteers using IDEAL-IQ. Lactate dehydrogenase (LDH), a reliable marker of intravascular hemolysis, was also measured in all participants.

Results: The kidney R2* values were significantly higher in PNH patients compared with those in healthy volunteers (P < 0.001). High inter-operator reproducibility of the measurements was also acquired using IDEAL-IQ. LDH levels were also significantly higher in PNH patients compared with those in healthy volunteers (P < 0.001). Kidney R2* values strongly correlated with LDH levels in PNH patients.

Conclusion: IDEAL-IQ has a possibility of becoming a useful method for the noninvasive evaluation of renal iron overload in PNH patients.

Purpose: To delineate brain microstructures in human embryos during the formation of the various major primordia by MR microscopy, with different contrasts appropriate for each target.

Methods: We focused mainly on the internal structures in the cerebral cortex and the accessory nerves of the brain. To find appropriate sequence parameters, we measured nuclear magnetic resonance (NMR) parameters and created kernel density plots of T1 and T2 values. We performed T1-weighted gradient echo imaging with parameters similar to those used in the previous studies. We performed T2*-weighted gradient echo imaging to delineate the target structures with the appropriate sequence parameters according to the NMR parameter and flip angle measurements. We also performed high-resolution imaging with both T1- and T2*-weighted sequences.

Results: T1, T2, and T2* values of the target tissues were positively correlated and shorter than those of the surrounding tissues. In T1-weighted images with a voxel size of (30 µm)³ and (20 µm)³, various organs and tissues and the agarose gel were differentiated as in previous studies, and the structure of approximately 40 µm in size was depicted, but the detailed structures within the cerebral cortex and the accessory nerves were not delineated. In T2*-weighted images with a voxel size of (30 µm)³, the layered structure within the cerebral cortex and the accessory nerves were clearly visualized. Overall, T1-weighted images provided more information than T2*-weighted images, but important internal brain structures of interest were visible only in T2*-weighted images. Therefore, it is essential to perform MR microscopy with different contrasts.

Conclusion: We have visualized brain structures in a human embryo that had not previously been delineated by MR microscopy. We discussed pulse sequences appropriate for the structures of interest. This methodology would provide a way to visualize crucial embryological information about the anatomical structure of human embryos.

Purpose: The present study aimed to investigate whether diffusion-weighted imaging (DWI) can qualify and quantify cerebrospinal fluid (CSF) dynamics in the brains of healthy subjects. For this purpose, we developed new DWI-based fluidography and compared the CSF dynamics seen on the fluidography with two apparent diffusion coefficients obtained with different DWI signal models at anatomical spaces filled by CSF.

Methods: DWI with multiple b values was performed for 10 subjects using a 7T MRI scanner. DWI-fluidography based on the DWI signal variations in different motion probing gradient directions was developed for visualizing the CSF dynamics voxel-by-voxel. DWI signals were measured using an ROI in the representative CSF-filled anatomical spaces in the brain. For the multiple DWI signals, the mono-exponential and kurtosis models were fitted and two kinds of apparent diffusion coefficients (ADC_C and ADC_K) were estimated in each space using the Gaussian and non-Gaussian diffusion models, respectively.

Results: DWI-fluidography could qualitatively represent the features of CSF dynamics in each anatomical space. ADCs indicated that the motions at the foramen of Monro, the cistern of the velum interpositum, the quadrigeminal cistern, the Sylvian cisterns, and the fourth ventricle were more drastic than those at the subarachnoid space and anterior horns of the lateral ventricle. Those results seen in ADCs were identical to the findings on DWI-fluidography.

Conclusion: DWI-fluidography based on the features of DWI signals could show differences of CSF dynamics among anatomical spaces.

Purpose: The purpose of this study was to do a feasibility study on a microstrip transmission line (MS) RF coil for a positron emission tomography (PET) insert in a 7 Tesla human MRI system. The proposed MS coil integrated the RF shield of the PET detector as the ground conductor of the coil. We called the integrated module "MS PET coil."

Methods: A single-channel MS PET coil was developed with an integrated RF-shielded PET detector module. For comparison, we also studied a conventional MS coil with a single-layer ground conductor. A lutetium fine silicate (LFS) scintillation crystal block (14 × 14 × 4-layer) with a silicon photomultiplier (Hamamatsu Photonics K.K., Shizuoka, Japan) and a front-end readout circuit board were mounted inside the shield cage of the MS PET coil. The MS PET coil was studied with and without PET detectors. All three coil configurations were studied with a homogeneous phantom in a 7T MRI system (Siemens Healthineers, Erlangen, Germany). PET data measurements were conducted using a Cesium-137 radiation point source.

Results: The MR images were similar for the MS coil and the empty MS PET coil, as well as for the cases of MS PET coil with and without PET measurements. Compared to the empty MS PET coil (without PET detector and cable RF shield), decreases in SNR, increases in image noise and RF power, and a slight decrease in resonance frequency were seen for the case of the MS PET coil with the detector and cable shield. Differences in the PET energy histograms or in the crystal identification maps with and without MRI measurements were negligible.

Conclusions: Both the MRI and PET performances of the MS PET coil showed responses that matched the MS coil responses. The performance variations of MRI data with and without PET measurement and PET data with and without MR imaging were negligible.