Synthetic Magnetic Resonance Imaging Data Research Articles

Generation of Synthetic Rat Brain MRI Scans with a 3D Enhanced Alpha Generative Adversarial Network

Translational brain research using Magnetic Resonance Imaging (MRI) is becoming increasingly popular as animal models are an essential part of scientific studies and more ultra-high-field scanners are becoming available. Some disadvantages of MRI are the availability of MRI scanners and the time required for a full scanning session. Privacy laws and the 3Rs ethics rule also make it difficult to create large datasets for training deep learning models. To overcome these challenges, an adaptation of the alpha Generative Adversarial Networks (GANs) architecture was used to test its ability to generate realistic 3D MRI scans of the rat brain in silico. As far as the authors are aware, this was the first time a GAN-based approach was used to generate synthetic MRI data of the rat brain. The generated scans were evaluated using various quantitative metrics, a Turing test, and a segmentation test. The last two tests proved the realism and applicability of the generated scans to real problems. Therefore, by using the proposed new normalisation layer and loss functions, it was possible to improve the realism of the generated rat MRI scans, and it was shown that using the generated data improved the segmentation model more than using the conventional data augmentation.

Quantification of brown adipose tissue in vivo using synthetic magnetic resonance imaging: an experimental study with mice model.

The white adipose tissue (WAT) and brown adipose tissue (BAT) are associated with the development of several obesity-associated disorders. The use of imaging techniques to differentiate BAT from WAT and quantify BAT volume remains challenging, due to limitations such as spatial resolution and magnetic field inhomogeneity. This study aimed to investigate the feasibility for differentiating BAT from WAT, and quantify the BAT volume in vivo using synthetic magnetic resonance imaging (MRI). A total of 16 C57BL/6 mice were scanned using synthetic MRI. Quantitative longitudinal relaxation time (T1) and transverse relaxation time (T2) maps were obtained from the original synthetic MRI data using the synthetic MRI software offline. The T1 and T2 values of interscapular BAT (IBAT) and dorsal subcutaneous WAT were measured. The IBAT volume was calculated using synthetic MRI-derived T2-weighted images (T2WIs) based on its morphological characteristics and quantitative tissue values. The body weight of mice was measured, and the IBAT specimens were excised and weighted. The correlation between IBAT volume and the weight of IBAT gross specimen and between IBAT volume and mouse body weight was analyzed. The T1 values of BAT (330.3±19.57 ms) were higher than those of WAT (304.42±4.14 ms) (P<0.001), whereas the T2 values of BAT (66.06±5.06 ms) were lower than those of WAT (88.23±7.68 ms) (P<0.001). The area under the curve (AUC) values of the T1 and T2 for differentiating BAT from WAT was 0.942 and 0.995, respectively. The AUC of the T2 values was higher than that of T1 (P=0.04) using the DeLong test. The optimal cut-off value for T2 was 76 ms for differentiating BAT from WAT (100% sensitivity, 93.7% specificity). A moderate correlation was observed between IBAT volume and the weight of the IBAT gross specimen (r=0.662, P=0.014), and between IBAT volume and mouse body weight (r=0.653, P=0.016). The quantitative parameters derived using synthetic MRI may be used to detect and differentiate BAT from WAT in vivo. Synthetic MRI may help quantify BAT volume in vivo.

Quantitative synthetic MRI reveals grey matter abnormalities in children with drug-naïve attention-deficit/hyperactivity disorder.

To investigate the quantitative profiles of brain grey matter (GM) in pediatric drug-naïve ADHD patients using synthetic magnetic resonance imaging (SyMRI). A total of 37 drug-naïve pediatric ADHD and 27 age- and gender-matched healthy controls (HC) were enrolled in this study. Each subject underwent both SyMRI and conventional 3D T1-FSPGR scans. Quantitative parameters, T1 and T2 maps, were extracted from the SyMRI data. Between-group quantitative maps were compared using a general linear model analysis. Pearson correlation analysis was conducted to assess the association between significantly altered MR indices and clinical measurements in ADHD. Compared with the HC group, altered T1 and T2 relaxometry times in the ADHD group were mainly distributed in GM regions of the cerebellum, attention and execution control network, default mode network, and limbic areas. Moreover, the T1 value of the right cerebellum 8 was negatively correlated with the attention concentration level in ADHD (R = 0.140, P = 0.0225). With regards to T2 map, the associations were observed between the attention level of ADHD patients and left fusiform gyrus (R = 0.251, P = 0.0016), and right cerebellum crus2 (R = 0.142, P = 0.0214). Altered T1, T2 values found in specific regions of GM, including cerebellum, attention and execution control network, default mode network, and limbic areas, may reveal widespread micromorphology changes, i.e., brain iron deficiency, low myelin content, and enlarged vascular interstitial space in ADHD patients. Thus, T1, T2 values might be promising imaging markers for future ADHD studies.

Deep Learning Approach for Generating MRA Images From 3D Quantitative Synthetic MRI Without Additional Scans.

Quantitative synthetic magnetic resonance imaging (MRI) enables synthesis of various contrast-weighted images as well as simultaneous quantification of T1 and T2 relaxation times and proton density. However, to date, it has been challenging to generate magnetic resonance angiography (MRA) images with synthetic MRI. The purpose of this study was to develop a deep learning algorithm to generate MRA images based on 3D synthetic MRI raw data. Eleven healthy volunteers and 4 patients with intracranial aneurysms were included in this study. All participants underwent a time-of-flight (TOF) MRA sequence and a 3D-QALAS synthetic MRI sequence. The 3D-QALAS sequence acquires 5 raw images, which were used as the input for a deep learning network. The input was converted to its corresponding MRA images by a combination of a single-convolution and a U-net model with a 5-fold cross-validation, which were then compared with a simple linear combination model. Image quality was evaluated by calculating the peak signal-to-noise ratio (PSNR), structural similarity index measurements (SSIMs), and high frequency error norm (HFEN). These calculations were performed for deep learning MRA (DL-MRA) and linear combination MRA (linear-MR), relative to TOF-MRA, and compared with each other using a nonparametric Wilcoxon signed-rank test. Overall image quality and branch visualization, each scored on a 5-point Likert scale, were blindly and independently rated by 2 board-certified radiologists. Deep learning MRA was successfully obtained in all subjects. The mean PSNR, SSIM, and HFEN of the DL-MRA were significantly higher, higher, and lower, respectively, than those of the linear-MRA (PSNR, 35.3 ± 0.5 vs 34.0 ± 0.5, P < 0.001; SSIM, 0.93 ± 0.02 vs 0.82 ± 0.02, P < 0.001; HFEN, 0.61 ± 0.08 vs 0.86 ± 0.05, P < 0.001). The overall image quality of the DL-MRA was comparable to that of TOF-MRA (4.2 ± 0.7 vs 4.4 ± 0.7, P = 0.99), and both types of images were superior to that of linear-MRA (1.5 ± 0.6, for both P < 0.001). No significant differences were identified between DL-MRA and TOF-MRA in the branch visibility of intracranial arteries, except for ophthalmic artery (1.2 ± 0.5 vs 2.3 ± 1.2, P < 0.001). Magnetic resonance angiography generated by deep learning from 3D synthetic MRI data visualized major intracranial arteries as effectively as TOF-MRA, with inherently aligned quantitative maps and multiple contrast-weighted images. Our proposed algorithm may be useful as a screening tool for intracranial aneurysms without requiring additional scanning time.

3-dimensional throat region segmentation from MRI data based on Fourier interpolation and 3-dimensional level set methods.

A new algorithm for 3D throat region segmentation from magnetic resonance imaging (MRI) is presented. The proposed algorithm initially pre-processes the MRI data to increase the contrast between the throat region and its surrounding tissues and to reduce artifacts. Isotropic 3D volume is reconstructed using the Fourier interpolation. Furthermore, a cube encompassing the throat region is evolved using level set method to form a smooth 3D boundary of the throat region. The results of the proposed algorithm on real and synthetic MRI data are used to validate the robustness and accuracy of the algorithm.