The interpretation and analysis of Magnetic resonance imaging (MRI) benefit from high spatial resolution. Unfortunately, direct acquisition of high spatial resolution MRI is time-consuming and costly, which increases the potential for motion artifact, and suffers from reduced signal-to-noise ratio (SNR). Super-resolution reconstruction (SRR) is one of the most widely used methods in MRI since it allows for the trade-off between high spatial resolution, high SNR, and reduced scan times. Deep learning has emerged for improved SRR as compared to conventional methods. However, current deep learning-based SRR methods require large-scale training datasets of high-resolution images, which are practically difficult to obtain at a suitable SNR. We sought to develop a methodology that allows for dataset-free deep learning-based SRR, through which to construct images with higher spatial resolution and of higher SNR than can be practically obtained by direct Fourier encoding. We developed a dataset-free learning method that leverages a generative neural network trained for each specific scan or set of scans, which in turn, allows for SRR tailored to the individual patient. With the SRR from three short duration scans, we achieved high quality brain MRI at an isotropic spatial resolution of 0.125 cubic mm with six minutes of imaging time for T2 contrast and an average increase of 7.2 dB (34.2%) in SNR to these short duration scans. Motion compensation was achieved by aligning the three short duration scans together. We assessed our technique on simulated MRI data and clinical data acquired from 15 subjects. Extensive experimental results demonstrate that our approach achieved superior results to state-of-the-art methods, while in parallel, performed at reduced cost as scans delivered with direct high-resolution acquisition.

Background Tools in image reconstruction, motion correction, and segmentation have enabled the accurate volumetric characterization of fetal brain growth at MRI. Purpose To evaluate the volumetric growth of intracranial structures in healthy fetuses, accounting for gestational age (GA), sex, and laterality with use of a spatiotemporal MRI atlas of fetal brain development. Materials and Methods T2-weighted 3.0-T half-Fourier acquired single-shot turbo spin-echo sequence MRI was performed in healthy fetuses from prospectively recruited pregnant volunteers from March 2013 to May 2019. A previously validated section-to-volume reconstruction algorithm was used to generate intensity-normalized superresolution three-dimensional volumes that were registered to a fetal brain MRI atlas with 28 anatomic regions of interest. Atlas-based segmentation was performed and manually refined. Labels included the bilateral hippocampus, amygdala, caudate nucleus, lentiform nucleus, thalamus, lateral ventricle, cerebellum, cortical plate, hemispheric white matter, internal capsule, ganglionic eminence, ventricular zone, corpus callosum, brainstem, hippocampal commissure, and extra-axial cerebrospinal fluid. For fetuses younger than 31 weeks of GA, the subplate and intermediate zones were delineated. A linear regression analysis was used to determine weekly age-related change adjusted for sex and laterality. Results The final analytic sample consisted of 122 MRI scans in 98 fetuses (mean GA, 29 weeks ± 5 [range, 20-38 weeks]). All structures had significant volume growth with increasing GA (P < .001). Weekly age-related change for individual structures in the brain parenchyma ranged from 2.0% (95% CI: 0.9, 3.1; P < .001) in the hippocampal commissure to 19.4% (95% CI: 18.7, 20.1; P < .001) in the cerebellum. The largest sex-related differences were 22.1% higher volume in male fetuses for the lateral ventricles (95% CI: 10.9, 34.4; P < .001). There was rightward volumetric asymmetry of 15.6% for the hippocampus (95% CI: 14.2, 17.2; P < .001) and leftward volumetric asymmetry of 8.1% for the lateral ventricles (95% CI: 3.7, 12.2; P < .001). Conclusion With use of a spatiotemporal MRI atlas, volumetric growth of the fetal brain showed complex trajectories dependent on structure, gestational age, sex, and laterality. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Rollins in this issue.

Functional MRI (fMRI) is widely used to study the functional organization of normal and pathological brains. However, the fMRI signal may be contaminated by subject motion artifacts that are only partially mitigated by motion correction strategies. These artifacts lead to distance-dependent biases in the inferred signal correlations. To mitigate these spurious effects, motion-corrupted volumes are censored from fMRI time series. Censoring can result in discontinuities in the fMRI signal, which may lead to substantial alterations in functional connectivity analysis. We propose a new approach to recover the missing entries from censoring based on structured low rank matrix completion. We formulated the artifact-reduction problem as the recovery of a super-resolved matrix from unprocessed fMRI measurements. We enforced a low rank prior on a large structured matrix, formed from the samples of the time series, to recover the missing entries. The recovered time series, in addition to being motion compensated, are also slice-time corrected at a fine temporal resolution. To achieve a fast and memory-efficient solution for our proposed optimization problem, we employed a variable splitting strategy. We validated the algorithm with simulations, data acquired under different motion conditions, and datasets from the ABCD study. Functional connectivity analysis showed that the proposed reconstruction resulted in connectivity matrices with lower errors in pair-wise correlation than non-censored and censored time series based on a standard processing pipeline. In addition, seed-based correlation analyses showed improved delineation of the default mode network. These demonstrate that the method can effectively reduce the adverse effects of motion in fMRI analysis.

BACKGROUND AND PURPOSE: Super-resolutionreconstruction (SRR) can be used to reconstruct 3-dimensional (3D) high-resolution (HR) volume from several 2-dimensional (2D) low-resolution (LR) stacks of MRI slices. The purpose is to compare lengthy 2D T2-weighted HR image acquisition of neonatal subjects with 3D SRR from several LR stacks in terms of image quality for clinical and morphometric assessments. METHODS: LR brain images were acquired from neonatal subjects to reconstruct isotropic 3D HR volumes by using SRR algorithm. Quality assessments were done by an experienced pediatric radiologist using scoring criteria adapted to newborn anatomical landmarks. The Wilcoxon signed-rank test was used to compare scoring results between HR and SRR images. For quantitative assessments, morphology-based segmentation was performed on both HR and SRR images and Dice coefficients between the results were computed. Additionally, simple linear regression was performed to compare the tissue volumes. RESULTS: No statistical difference was found between HR and SRR structural scores using Wilcoxon signed-rank test (p = .63, Z = .48). Regarding segmentation results, R2 values for the volumes of gray matter, white matter, cerebrospinal fluid, basal ganglia, cerebellum, and total brain volume including brain stem ranged between .95 and .99. Dice coefficients between the segmented regions from HR and SRR ranged between .83 ± .04 and .96 ± .01. CONCLUSION: Qualitative and quantitative assessments showed that 3D SRR of several LR images produces images that are of comparable quality to standard 2D HR image acquisition for healthy neonatal imaging without loss of anatomical details with similar edge definition allowing the detection of fine anatomical structures and permitting comparable morphometric measurement.

Accurate, quantitative segmentation of anatomical structures in radiological scans, such as Magnetic Resonance Imaging (MRI) and Computer Tomography (CT), can produce significant biomarkers and can be integrated into computer-aided assisted diagnosis (CADx) systems to support the interpretation of medical images from multi-protocol scanners. However, there are serious challenges towards developing robust automated segmentation techniques, including high variations in anatomical structure and size, the presence of edge-based artefacts, and heavy un-controlled breathing that can produce blurred motion-based artefacts. This paper presents a novel computing approach for automatic organ and muscle segmentation in medical images from multiple modalities by harnessing the advantages of deep learning techniques in a two-part process. (1) a 3D encoder-decoder, Rb-UNet, builds a localisation model and a 3D Tiramisu network generates a boundary-preserving segmentation model for each target structure; (2) the fully trained Rb-UNet predicts a 3D bounding box encapsulating the target structure of interest, after which the fully trained Tiramisu model performs segmentation to reveal detailed organ or muscle boundaries. The proposed approach is evaluated on six different datasets, including MRI, Dynamic Contrast Enhanced (DCE) MRI and CT scans targeting the pancreas, liver, kidneys and psoas-muscle and achieves quantitative measures of mean Dice similarity coefficient (DSC) that surpass or are comparable with the state-of-the-art. A qualitative evaluation performed by two independent radiologists verified the preservation of detailed organ and muscle boundaries.

Spatial resolution plays a critically important role in MRI for the precise delineation of the imaged tissues. Unfortunately, acquisitions with high spatial resolution require increased imaging time, which increases the potential of subject motion, and suffers from reduced signal-to-noise ratio (SNR). Super-resolution reconstruction (SRR) has recently emerged as a technique that allows for a trade-off between high spatial resolution, high SNR, and short scan duration. Deconvolution-based SRR has recently received significant interest due to the convenience of using the image space. The most critical factor to succeed in deconvolution is the accuracy of the estimated blur kernels that characterize how the image was degraded in the acquisition process. Current methods use handcrafted filters, such as Gaussian filters, to approximate the blur kernels, and have achieved promising SRR results. As the image degradation is complex and varies with different sequences and scanners, handcrafted filters, unfortunately, do not necessarily ensure the success of the deconvolution. We sought to develop a technique that enables accurately estimating blur kernels from the image data itself. We designed a deep architecture that utilizes an adversarial scheme with a generative neural network against its degradation counterparts. This design allows for the SRR tailored to an individual subject, as the training requires the scan-specific data only, i.e., it does not require auxiliary datasets of high-quality images, which are practically challenging to obtain. With this technique, we achieved high-quality brain MRI at an isotropic resolution of 0.125 cubic mm with six minutes of imaging time. Extensive experiments on both simulated low-resolution data and clinical data acquired from ten pediatric patients demonstrated that our approach achieved superior SRR results as compared to state-of-the-art deconvolution-based methods, while in parallel, at substantially reduced imaging time in comparison to direct high-resolution acquisitions.

There is a growing demand for fast, accurate computation of clinical markers to improve renal function and anatomy assessment with a single study. However, conventional techniques have limitations leading to overestimations of kidney function or failure to provide sufficient spatial resolution to target the disease location. In contrast, the computer-aided analysis of dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) could generate significant markers, including the glomerular filtration rate (GFR) and time-intensity curves of the cortex and medulla for determining obstruction in the urinary tract. This paper presents a dual-stage fully modular framework for automatic renal compartment segmentation in 4D DCE-MRI volumes. (1) Memory-efficient 3D deep learning is integrated to localise each kidney by harnessing residual convolutional neural networks for improved convergence; segmentation is performed by efficiently learning spatial-temporal information coupled with boundary-preserving fully convolutional dense nets. (2) Renal contextual information is enhanced via non-linear transformation to segment the cortex and medulla. The proposed framework is evaluated on a paediatric dataset containing 60 4D DCE-MRI volumes exhibiting varying conditions affecting kidney function. Our technique outperforms a state-of-the-art approach based on a GrabCut and support vector machine classifier in mean dice similarity (DSC) by 3.8% and demonstrates higher statistical stability with lower standard deviation by 12.4% and 15.7% for cortex and medulla segmentation, respectively.

The trade-off between image resolution, signal-to-noise ratio (SNR), and scan time in any magnetic resonance imaging (MRI) protocol is inevitable and unavoidable. Super-resolution reconstruction (SRR) has been shown effective in mitigating these factors, and thus, has become an important approach in addressing the current limitations of MRI. In this work, we developed a novel, image-based MRI SRR approach based on anisotropic acquisition schemes, which utilizes a new gradient guidance regularization method that guides the high-resolution (HR) reconstruction via a spatial gradient estimate. Further, we designed an analytical solution to propagate the spatial gradient fields from the low-resolution (LR) images to the HR image space and exploited these gradient fields over multiple scales with a dynamic update scheme for more accurate edge localization in the reconstruction. We also established a forward model of image formation and inverted it along with the proposed gradient guidance. The proposed SRR method allows subject motion between volumes and is able to incorporate various acquisition schemes where the LR images are acquired with arbitrary orientations and displacements, such as orthogonal and through-plane origin-shifted scans. We assessed our proposed approach on simulated data as well as on the data acquired on a Siemens 3T MRI scanner containing 45 MRI scans from 14 subjects. Our experimental results demonstrate that our approach achieved superior reconstructions compared to state-of-the-art methods, both in terms of local spatial smoothness and edge preservation, while, in parallel, at reduced, or at the same cost as scans delivered with direct HR acquisition.