Magnetic Resonance Materials in Physics, Biology and Medicine最新文献

Objectives: Brain tumor detection, classification and segmentation are challenging due to the heterogeneous nature of brain tumors. Different deep learning-based algorithms are available for object detection; however, the performance of detection algorithms on brain tumor data has not been widely explored. Therefore, we aim to compare different object detection algorithms (Faster R-CNN, YOLO & SSD) for brain tumor detection on MRI data. Furthermore, the best-performing detection network is paired with a 2D U-Net for pixel-wise segmentation of abnormal tumor cells.

Materials and methods: The proposed model was evaluated on the Brain Tumor Figshare (BTF) dataset, and the best-performing detection network cascaded with 2D U-Net for pixel-wise segmentation of tumors. The best-performing detection network was also fine-tuned on BRATS 2018 data to detect and classify the glioma tumor.

Results: For the detection of three tumor types, YOLOv5 achieved the highest mAP of 89.5% on test data compared to other networks. For segmentation, YOLOv5 combined with 2D U-Net achieved a higher DSC compared to the 2D U-Net alone (DSC: YOLOv5 + 2D U-Net = 88.1%; 2D U-Net = 80.5%). The proposed method was compared with the existing detection and segmentation network i.e. Mask R-CNN and achieved a higher mAP (YOLOv5 + 2D U-Net = 89.5%; Mask R-CNN = 67%) and DSC (YOLOv5 + 2D U-Net = 88.1%; Mask R-CNN = 44.2%).

Conclusion: In this work, we propose a deep-learning-based method for multi-class tumor detection, classification and segmentation that combines YOLOv5 with 2D U-Net. The results show that the proposed method not only detects different types of brain tumors accurately but also delineates the tumor region precisely within the detected bounding box.

Objective: To compare compressed sensing (CS) and the Cascades of Independently Recurrent Inference Machines (CIRIM) with respect to image quality and reconstruction times when 12-fold accelerated scans of patients with neurological deficits are reconstructed.

Materials and methods: Twelve-fold accelerated 3D T2-FLAIR images were obtained from a cohort of 62 patients with neurological deficits on 3 T MRI. Images were reconstructed offline via CS and the CIRIM. Image quality was assessed in a blinded and randomized manner by two experienced interventional neuroradiologists and one experienced pediatric neuroradiologist on imaging artifacts, perceived spatial resolution (sharpness), anatomic conspicuity, diagnostic confidence, and contrast. The methods were also compared in terms of self-referenced quality metrics, image resolution, patient groups and reconstruction time. In ten scans, the contrast ratio (CR) was determined between lesions and white matter. The effect of acceleration factor was assessed in a publicly available fully sampled dataset, since ground truth data are not available in prospectively accelerated clinical scans. Specifically, 451 FLAIR scans, including scans with white matter lesions, were adopted from the FastMRI database to evaluate structural similarity (SSIM) and the CR of lesions and white matter on ranging acceleration factors from four-fold up to 12-fold.

Results: Interventional neuroradiologists significantly preferred the CIRIM for imaging artifacts, anatomic conspicuity, and contrast. One rater significantly preferred the CIRIM in terms of sharpness and diagnostic confidence. The pediatric neuroradiologist preferred CS for imaging artifacts and sharpness. Compared to CS, the CIRIM reconstructions significantly improved in terms of imaging artifacts and anatomic conspicuity (p < 0.01) for higher resolution scans while yielding a 28% higher SNR (p = 0.001) and a 5.8% lower CR (p = 0.04). There were no differences between patient groups. Additionally, CIRIM was five times faster than CS was. An increasing acceleration factor did not lead to changes in CR (p = 0.92), but led to lower SSIM (p = 0.002).

Discussion: Patients with neurological deficits can undergo MRI at a range of moderate to high acceleration. DL reconstruction outperforms CS in terms of image resolution, efficient denoising with a modest reduction in contrast and reduced reconstruction times.

Objective: To optimize high-resolution 7 T MRI of the cochlea and measure normal cochlea and the cochlear nerve morphometry in vivo.

Materials and methods: Eight volunteers with normal hearing were scanned at 7 T using an optimized protocol. Two neuroradiologists independently scored image quality. The basal turn lumen diameter (BTLD), height, width, length and volume of the cochlear, long (LD) and short (SD) diameter the calculated cross-sectional area (CSA) of the cochlear nerve were measured. Intra and inter-observer reliability was assessed using intraclass correlation (ICC).

Results: 3D T2W DRIVE combined with dielectric pads, allowed acquisition of high-resolution images showing detailed structures, such as the crista ampullaris in the semicircular canals. The overall grading scores from neuroradiologists were excellent. In the left ear, averaging over all subjects gave BTLD of 2.6 ± 0.05 mm, height of 4.9 ± 0.1 mm, width of 4.4 ± 0.2 mm, length of 36.5 ± 0.4 mm, volume of 0.16 ± 0.02 ml, LD of 1.31 ± 0.1 mm, SD of 1.06 ± 0.1 mm, and CSA of 1.1 ± 0.1 mm². The right ear gave BTLD of 2.6 ± 0.04 mm, height of 4.9 ± 0.1 mm, width of 4.4 ± 0.3 mm, length of 35.5 ± 0.4 mm, volume of 0.16 ± 0.02 ml, LD of 1.29 ± 0.1 mm, SD of 1.07 ± 0.1 mm, and CSA of 1.10 ± 0.2 mm². No statistically significant difference was found between the sides of the head (p-value > 0.05). The intra-observer reliability was high (0.77-0.94), while the inter-observer reliability varied from moderate to high (0.55-0.81).

Conclusion: 7 T MRI can provide excellent visualization of the internal structure of the cochlear and of the vestibulocochlear nerve in vivo.

Standard quantitative abdominal MRI techniques are time consuming, require breath-holds, and are susceptible to patient motion artifacts. Magnetic resonance fingerprinting (MRF) is naturally multi-parametric and quantifies multiple tissue properties, including T₁ and T₂. This work includes T₂* and off-resonance mapping into a free-breathing MRF framework utilizing a pilot tone navigator. The new acquisition and reconstruction are compared to current clinical standards. Prospective. Ten volunteers. 3 T scanner, Quadratic-RF MRF, Balanced SSFP, Inversion recovery spin-echo, LiverLab. MRI ROIs were evaluated in the liver, spleen, pancreas, kidney (cortex and medulla), and paravertebral muscle by two abdominal imaging investigators for ten healthy adult volunteers for clinical standard, breath-Hold (BH) qRF-MRF, and free-breathing qRF-MRF with pilot-tone (PT) acquisitions. Bland-Altman analysis as well as Student's T tests were used to evaluate and compare the respective ROI analyses. Quantitative values between breath-Hold (BH) and free-breathing qRF-MRF with pilot-tone (PT) results show good agreement with clinical standard T1 and T2 quantitative mapping, and Dixon q-VIBE (acquired using the Siemens LiverLAB). In this work, we show free-breathing abdominal MRF (T₁, T₂) with T₂* results that are quantitatively comparable to current breath-hold MRF and clinical techniques.

Objective: Computing the trajectories of mandibular condyles directly from MRI could provide a comprehensive examination, providing both anatomical and kinematic details. This study aimed to investigate the feasibility of extracting 3D condylar trajectories from 2D real-time MRI.

Materials and methods: Twenty healthy subjects underwent real-time MRI while performing jaw opening and closing movements. One axial and two sagittal slices were segmented using a U-Net-based algorithm. After motion compensation, the centers of mass of the resulting masks were projected onto the coordinate system based on anatomical markers and temporally adjusted. The quality of the computed trajectories was evaluated using metrics designed to estimate movement reproducibility, head motion, and slice placement symmetry.

Results: The segmentation of the axial slices demonstrated good-to-excellent quality; however, the segmentation of the sagittal slices required some fine-tuning. On average, the intercuspal position shifted by 0.6 mm after an opening-closing cycle. The difference in the superior-inferior coordinate of the condyles in the intercuspal position was 1.5 mm on average. Some subjects demonstrated a significant discrepancy between the axial and the sagittal trajectories.

Discussion: Real-time MRI enables the extraction of condylar trajectories for evaluating some clinically relevant parameters. However, attention is required during patient installation and image acquisition.

Object: This study aimed to evaluate the relaxivity and uniformity of agarose gel phantoms added with relaxation modifiers. It is hypothesized that the modifiers could manipulate the T1 and T2 relaxations as well as the signal uniformity.

Materials and methods: Twenty agarose gel phantoms with different GdCl₃ and FeCl₃ volume fractions were prepared. The phantoms were scanned using a 3-T scanner implementing a turbo spin echo sequence to acquire T1 and T2 images. The SNR of the images were computed using Image-J software from 1, 3, and 25 regions-of-interest (ROIs) and were inverted as T1 and T2 curves.

Results: With the increase in relaxation modifier content, T1 SNR increased at a faster rate at very short TR and reached saturation at TR well below 400 ms. Agarose gel phantoms containing GdCl₃ showed a higher saturation value as compared to phantoms containing FeCl₃. For T2 SNR, differences between plots are observed at low TE. As TE gets larger, the SNR between plots is incomparable. The SNR for both groups was uniform among 1, 3, and 25 ROIs.

Discussions: It can be concluded that GdCl₃ and FeCl₃ solutions can be used as effective relaxation modifiers to reduce T1 but not T2 relaxation times.

Objective: To investigate the trade-off between magnet bore diameter and the distance between the conductive Faraday shield and RF head coil for low-field point-of-care neuroimaging systems.

Methods: Electromagnetic simulations were performed for three different Faraday shield geometries and two commonly used RF coil designs (spiral and solenoid) to assess the effects of a close-fitting shield on the RF coil's transmit and receive efficiencies. Experimental measurements were performed to confirm the accuracy of the simulations. Parallel simulations were performed to assess the static magnet ( $B_0$ ) field as a function of the magnet bore diameter. The obtainable SNR was then calculated as a function of these two related variables.

Results: Simulations of the RF coil characteristics and $B_1^{+}$ transmit efficiencies agreed well with corresponding experimentally determined parameters. Overall, the RF coil transmit efficiency was, as expected, higher when the gap between the shield and coil increased. The calculated intrinsic SNR showed that maximum SNR would be obtained for a cylindrical shield of diameter 310 mm with an inner diameter of the magnet of 320 mm (assuming 10 mm for the gradient coils).

Conclusion: This work presents an overview of the trade-offs in transmit efficiencies for RF coils used for POC MRI neuroimaging as a function of coil-to-shield distance and inner diameter of the Halbach magnet. Results show that there is a relatively shallow optimum between a magnet diameter of 290 and 330 mm, with values falling more than 10% if either smaller or larger magnets are used.

Deep-learning-based MR image reconstruction in settings where large fully sampled dataset collection is infeasible requires methods that effectively use both under-sampled and fully sampled datasets. This paper evaluates a weakly supervised, multi-coil, physics-guided approach to MR image reconstruction, leveraging both dataset types, to improve both the quality and robustness of reconstruction. A physics-guided end-to-end variational network (VarNet) is pretrained in a self-supervised manner using a 4 $\times$ under-sampled dataset following the self-supervised learning via data undersampling (SSDU) methodology. The pre-trained weights are transferred to another VarNet, which is fine-tuned using a smaller, fully sampled dataset by optimizing multi-scale structural similarity (MS-SSIM) loss in image space. The proposed methodology is compared with fully self-supervised and fully supervised training. Reconstruction quality improvements in SSIM, PSNR, and NRMSE when abundant training data is available (the high-data regime), and enhanced robustness when training data is scarce (the low-data regime) are demonstrated using weak supervision for knee and brain MR image reconstructions at 8 $\times$ and 10 $\times$ acceleration, respectively. Multi-coil physics-guided MR image reconstruction using both under-sampled and fully sampled datasets is achievable with transfer learning and fine-tuning. This methodology can provide improved reconstruction quality in the high-data regime and improved robustness in the low-data regime at high acceleration rates.

Objective: This study aimed to assess the reproducibility of MRI-derived radiomic features across multiple gray-level discretization levels for classifying Parkinson's disease (PD) subtypes, and to evaluate the impact of ComBat harmonization on feature stability and machine learning performance.

Methods: T1-weighted MRI scans from 140 PD patients (70 tremor-dominant, 70 postural instability gait difficulty) and 70 healthy controls were obtained from the Parkinson's progression markers initiative (PPMI) database. Radiomic features were extracted from 16 brain regions using 6 discretization levels (8, 16, 32, 64, 128, and 256 bins). ComBat harmonization was applied using a combined batch variable incorporating both scanner models and discretization levels. Intraclass correlation coefficients (ICC) and Kruskal-Wallis tests assessed feature reproducibility before and after harmonization. Support vector machine classifiers were used for PD subtype classification.

Results: ComBat harmonization significantly improved feature reproducibility across all feature groups. The percentage of features showing excellent robustness (ICC ≥ 0.90) increased substantially after harmonization. The proportion of features significantly affected by discretization levels was reduced following harmonization. Classification accuracy improved dramatically, from a range of 0.42-0.49 before harmonization to 0.86-0.96 after harmonization across most discretization levels. AUC values similarly increased from 0.60-0.67 to 0.93-0.99 after harmonization.

Conclusions: ComBat harmonization significantly enhanced the reproducibility of radiomic features across discretization levels and improved PD subtype classification performance. This study highlights the importance of harmonization in radiomics research for PD and suggests potential clinical applications in personalized treatment planning.

Object: Spatio-temporal MRI methods offer rapid whole-brain multi-parametric mapping, yet they are often hindered by prolonged reconstruction times or prohibitively burdensome hardware requirements. The aim of this project is to reduce reconstruction time using deep learning.

Materials and methods: This study focuses on accelerating the reconstruction of volumetric multi-axis spiral projection MRF, aiming for whole-brain T1 and T2 mapping, while ensuring a streamlined approach compatible with clinical requirements. To optimize reconstruction time, the traditional method is first revamped with a memory-efficient GPU implementation. Deep Learning Initialized Compressed Sensing (Deli-CS) is then introduced, which initiates iterative reconstruction with a DL-generated seed point, reducing the number of iterations needed for convergence.

Results: The full reconstruction process for volumetric multi-axis spiral projection MRF is completed in just 20 min compared to over 2 h for the previously published implementation. Comparative analysis demonstrates Deli-CS's efficiency in expediting iterative reconstruction while maintaining high-quality results.

Discussion: By offering a rapid warm start to the iterative reconstruction algorithm, this method substantially reduces processing time while preserving reconstruction quality. Its successful implementation paves the way for advanced spatio-temporal MRI techniques, addressing the challenge of extensive reconstruction times and ensuring efficient, high-quality imaging in a streamlined manner.