High In-plane Resolution Research Articles

A flexible 2.5D medical image segmentation approach with in-slice and cross-slice attention

Deep learning has become the de facto method for medical image segmentation, with 3D segmentation models excelling in capturing complex 3D structures and 2D models offering high computational efficiency. However, segmenting 2.5D images, characterized by high in-plane resolution but lower through-plane resolution, presents significant challenges. While applying 2D models to individual slices of a 2.5D image is feasible, it fails to capture the spatial relationships between slices. On the other hand, 3D models face challenges such as resolution inconsistencies in 2.5D images, along with computational complexity and susceptibility to overfitting when trained with limited data. In this context, 2.5D models, which capture inter-slice correlations using only 2D neural networks, emerge as a promising solution due to their reduced computational demand and simplicity in implementation. In this paper, we introduce CSA-Net, a flexible 2.5D segmentation model capable of processing 2.5D images with an arbitrary number of slices. CSA-Net features an innovative Cross-Slice Attention (CSA) module that effectively captures 3D spatial information by learning long-range dependencies between the center slice (for segmentation) and its neighboring slices. Moreover, CSA-Net utilizes the self-attention mechanism to learn correlations among pixels within the center slice. We evaluated CSA-Net on three 2.5D segmentation tasks: (1) multi-class brain MR image segmentation, (2) binary prostate MR image segmentation, and (3) multi-class prostate MR image segmentation. CSA-Net outperformed leading 2D, 2.5D, and 3D segmentation methods across all three tasks, achieving average Dice coefficients and HD95 values of 0.897 and 1.40 mm for the brain dataset, 0.921 and 1.06 mm for the prostate dataset, and 0.659 and 2.70 mm for the ProstateX dataset, demonstrating its efficacy and superiority. Our code is publicly available at: https://github.com/mirthAI/CSA-Net.

Full-field optical coherence microscopy enables high-resolution label-free imaging of the dynamics of live mouse oocytes and early embryos

High quality label-free imaging of oocytes and early embryos is essential for accurate assessment of their developmental potential, a key element of assisted reproduction procedures. To achieve this goal, we propose full-field optical coherence microscopy (FF-OCM), constructed as a compact module fully integrated with a commercial wide-field fluorescence microscope. Our system achieves optical sectioning in wide-field, high in-plane resolution of 0.5 µm, and high sensitivity to backscattered light. To demonstrate its imaging capabilities, we study live mouse oocytes and embryos at all important stages of meiotic maturation and early embryogenesis. Our system enables visualization of intracellular structures, which are not visible in common bright-field microscopy, i.e., internal structure of nuclear apparatus, cytoskeletal filaments, cellular cortex, cytoplasmic protrusions, or zona pellucida features. Additionally, we visualize and quantify intracellular dynamics like cytoplasmic stirring motion, nuclear envelope fluctuations and nucleolus mobility. Altogether, we demonstrate that FF-OCM is a powerful tool for research in developmental biology that also holds great potential for non-invasive time-lapse monitoring of oocyte and embryo quality in assisted reproduction.

3D whole heart k-space-based super-resolution cardiac T1 mapping using rotated stacks

Objective. To provide three-dimensional (3D) whole-heart high-resolution isotropic cardiac T1 maps using a k-space-based through-plane super-resolution reconstruction (SRR) with rotated multi-slice stacks. Approach. Due to limited SNR and cardiac motion, often only 2D T1 maps with low through-plane resolution (4–8 mm) can be obtained. Previous approaches used SRR to calculate 3D high-resolution isotropic cardiac T1 maps. However, they were limited to the ventricles. The proposed approach acquires rotated stacks in long-axis orientation with high in-plane resolution but low through-plane resolution. This results in radially overlapping stacks from which high-resolution T1 maps of the whole heart are reconstructed using a k-space-based SRR framework considering the complete acquisition model. Cardiac and residual respiratory motion between different breath holds is estimated and incorporated into the reconstruction. The proposed approach was evaluated in simulations and phantom experiments and successfully applied to ten healthy subjects.Main results. 3D T1 maps of the whole heart were obtained in the same acquisition time as previous methods covering only the ventricles. T1 measurements were possible even for small structures, such as the atrial wall. The proposed approach provided accurate (P > 0.4; R 2 > 0.99) and precise T1 values (SD of 64.32 ± 22.77 ms in the proposed approach, 44.73 ± 31.9 ms in the reference). The edge sharpness of the T1 maps was increased by 6.20% and 4.73% in simulation and phantom experiments, respectively. Contrast-to-noise ratios between the septum and blood pool increased by 14.50% in in vivo measurements with a k-space compared to an image-space-based SRR. Significance. The proposed approach provided whole-heart high-resolution 1.3 mm isotropic T1 maps in an overall acquisition time of approximately three minutes. Small structures, such as the atrial and right ventricular walls, could be visualized in the T1 maps.

Spatial attention-based implicit neural representation for arbitrary reduction of MRI slice spacing

Magnetic resonance (MR) images collected in 2D clinical protocols typically have large inter-slice spacing, resulting in high in-plane resolution and reduced through-plane resolution. Super-resolution technique can enhance the through-plane resolution of MR images to facilitate downstream visualization and computer-aided diagnosis. However, most existing works train the super-resolution network at a fixed scaling factor, which is not friendly to clinical scenes of varying inter-slice spacing in MR scanning. Inspired by the recent progress in implicit neural representation, we propose a Spatial Attention-based Implicit Neural Representation (SA-INR) network for arbitrary reduction of MR inter-slice spacing. The SA-INR aims to represent an MR image as a continuous implicit function of 3D coordinates. In this way, the SA-INR can reconstruct the MR image with arbitrary inter-slice spacing by continuously sampling the coordinates in 3D space. In particular, a local-aware spatial attention operation is introduced to model nearby voxels and their affinity more accurately in a larger receptive field. Meanwhile, to improve the computational efficiency, a gradient-guided gating mask is proposed for applying the local-aware spatial attention to selected areas only. We evaluate our method on the public HCP-1200 dataset and the clinical knee MR dataset to demonstrate its superiority over other existing methods.

Simultaneous two-plane flame front detection using PIV based on defocusing.

This Letter presents a simultaneous two-plane flame front detection method. It is based on a standard single-camera single-plane particle image velocimetry (PIV) system in combination with an inexpensive and compact image splitting device. The image splitting optics places images from two depth-offset planes onto the two halves of a camera sensor. A shallow depth of field ensures only one plane is in focus on each half of the sensor. By using a high-pass filter and a novel two-step filter we have devised, the out-of-focus particle images are effectively removed, while the in-focus particle images remain, allowing the turbulent flame fronts on two planes to be detected simultaneously. Our approach could be combined with conventional polarization/wavelength discrimination methods to achieve simultaneous multi-plane flame front reconstruction with similarly high in-plane spatial resolution to single-plane measurement and is suitable for practical combustion devices with limited optical access.

I3Net: Inter-Intra-slice Interpolation Network for Medical Slice Synthesis.

Medical imaging is limited by acquisition time and scanning equipment. CT and MR volumes, reconstructed with thicker slices, are anisotropic with high in-plane resolution and low through-plane resolution. We reveal an intriguing phenomenon that due to the mentioned nature of data, performing slice-wise interpolation from the axial view can yield greater benefits than performing super-resolution from other views. Based on this observation, we propose an Inter-Intra-slice Interpolation Network (I3Net), which fully explores information from high in-plane resolution and compensates for low through-plane resolution. The through-plane branch supplements the limited information contained in low through-plane resolution from high in-plane resolution and enables continual and diverse feature learning. In-plane branch transforms features to the frequency domain and enforces an equal learning opportunity for all frequency bands in a global context learning paradigm. We further propose a cross-view block to take advantage of the information from all three views online. Extensive experiments on two public datasets demonstrate the effectiveness of I3Net, and noticeably outperforms state-of-the-art super-resolution, video frame interpolation and slice interpolation methods by a large margin. We achieve 43.90dB in PSNR, with at least 1.14dB improvement under the upscale factor of ×2 on MSD dataset with faster inference.

Simultaneous T1, T2 and T2⁎ mapping of the liver with multi-shot MI-SAGE

PurposeTo apply multi-shot high-resolution multi inversion spin and gradient echo (MI-SAGE) acquisition for simultaneous liver T1, T2 and T2* mapping. MethodsInversion prepared spin- and gradient-echo EPI was developed with ascending slice order across measurements for efficient acquisition with T1, T2, and T2⁎ weighting. Multi-shot EPI was also implemented to minimize distortion and blurring while enabling high in-plane resolution. A dictionary-matching approach was used to fit the images to quantitative parameter maps, which were compared to T1 measured by modified Look-Locker (MOLLI), T1 measured by variable flip angle (VFA), T2 measured by multiple echo time-based Half Fourier Single-shot Turbo spin-Echo (HASTE), T2 measured by radial turbo-spin-echo (rTSE) and T2⁎ measured by multiple gradient echo (MGRE) sequences. ResultsThe multi-shot variant of the sequence achieved higher in-plane resolution of 1.7 × 1.7 mm2 with good image quality in 28 s. Derived quantitative maps showed comparable values to conventional mapping methods. As measured in phantom and in vivo, MOLLI, MESE and MGRE give closest values to MISAGE. VFA, HASTE and rTSE show obvious overestimation. ConclusionsThe proposed multi-shot inversion prepared spin- and gradient-echo EPI sequence allows for high-resolution quantitative T1, T2 and T2 liver tissue characterization in a single breath-hold scan.

2D sodium MRI of the human calf using half-sinc excitation pulses and compressed sensing.

Sodium MRI can be used to quantify tissue sodium concentration (TSC) in vivo; however, UTE sequences are required to capture the rapidly decaying signal. 2D MRI enables high in-plane resolution but typically has long TEs. Half-sinc excitation may enable UTE; however, twice as many readouts are necessary. Scan time can be minimized by reducing the number of signal averages (NSAs), but at a cost to SNR. We propose using compressed sensing (CS) to accelerate 2D half-sinc acquisitions while maintaining SNR and TSC. Ex vivo and in vivo TSC were compared between 2D spiral sequences with full-sinc (TE = 0.73 ms, scan time ≈ 5 min) and half-sinc excitation (TE = 0.23 ms, scan time ≈ 10 min), with 150 NSAs. Ex vivo, these were compared to a reference 3D sequence (TE = 0.22 ms, scan time ≈ 24 min). To investigate shortening 2D scan times, half-sinc data was retrospectively reconstructed with fewer NSAs, comparing a nonuniform fast Fourier transform to CS. Resultant TSC and image quality were compared to reference 150 NSAs nonuniform fast Fourier transform images. TSC was significantly higher from half-sinc than from full-sinc acquisitions, ex vivo and in vivo. Ex vivo, half-sinc data more closely matched the reference 3D sequence, indicating improved accuracy. In silico modeling confirmed this was due to shorter TEs minimizing bias caused by relaxation differences between phantoms and tissue. CS was successfully applied to in vivo, half-sinc data, maintaining TSC and image quality (estimated SNR, edge sharpness, and qualitative metrics) with ≥50 NSAs. 2D sodium MRI with half-sinc excitation and CS was validated, enabling TSC quantification with 2.25 × 2.25 mm2 resolution and scan times of ≤5 mins.

Inexpensive multi-plane particle image velocimetry based on defocusing: Proof of concept on two-component measurement

This paper presents a method for simultaneous particle image velocimetry (PIV) on parallel planes offset in depth. The method places images from two planes onto a different half of a camera sensor by using image splitting optics with variable optical path lengths. A shallow depth of field is achieved to ensure only one plane is in focus on each half of the sensor. Without needing additional lasers, the method is designed as an inexpensive means to increase the number of measurement plane(s) of single/multi-plane PIV setups and can be combined with existing plane discrimination approaches such as polarization and wavelength. The method is useful for studying instantaneous flow correlations on different planes while retaining high in-plane spatial resolution of typical planar PIV measurement. The measurement uncertainty caused by crosstalk from out-of-focus images is discussed. Experimental results from a laminar flow rig test indicate that the average measurement error of each velocity component is lower than 0.1 pixels per time step, with a 20 mm plane separation in depth and a 35 × 54 mm2 field of view. As an application with varying background scatter and out-of-plane flow motions, in-cylinder flow measurements in an optically accessible internal combustion engine were performed on two swirl planes simultaneously. Characteristics of the proposed method performing stereoscopic PIV measurements will be studied in future.

Deep learning architecture for 3D image super-resolution of late gadolinium enhanced cardiac MRI.

High-resolution late gadolinium enhanced (LGE) cardiac magnetic resonance imaging (MRI) volumes are difficult to acquire due to the limitations of the maximal breath-hold time achievable by the patient. This results in anisotropic 3D volumes of the heart with high in-plane resolution, but low-through-plane resolution. Thus, we propose a 3D convolutional neural network (CNN) approach to improve the through-plane resolution of the cardiac LGE-MRI volumes. We present a 3D CNN-based framework with two branches: a super-resolution branch to learn the mapping between low-resolution and high-resolution LGE-MRI volumes, and a gradient branch that learns the mapping between the gradient map of low-resolution LGE-MRI volumes and the gradient map of high-resolution LGE-MRI volumes. The gradient branch provides structural guidance to the CNN-based super-resolution framework. To assess the performance of the proposed CNN-based framework, we train two CNN models with and without gradient guidance, namely, dense deep back-projection network (DBPN) and enhanced deep super-resolution network. We train and evaluate our method on the 2018 atrial segmentation challenge dataset. Additionally, we also evaluate these trained models on the left atrial and scar quantification and segmentation challenge 2022 dataset to assess their generalization ability. Finally, we investigate the effect of the proposed CNN-based super-resolution framework on the 3D segmentation of the left atrium (LA) from these cardiac LGE-MRI image volumes. Experimental results demonstrate that our proposed CNN method with gradient guidance consistently outperforms bicubic interpolation and the CNN models without gradient guidance. Furthermore, the segmentation results, evaluated using Dice score, obtained using the super-resolved images generated by our proposed method are superior to the segmentation results obtained using the images generated by bicubic interpolation () and the CNN models without gradient guidance (). The presented CNN-based super-resolution method with gradient guidance improves the through-plane resolution of the LGE-MRI volumes and the structure guidance provided by the gradient branch can be useful to aid the 3D segmentation of cardiac chambers, such as LA, from the 3D LGE-MRI images.

Cervical Spinal Cord Gray Matter Area is Associated with Motor Unit Number Index in Spinal Muscular Atrophy (S33.008)

<h3>Objective:</h3> To evaluate the association between spinal cord (SC) gray matter (GM) atrophy and Motor Unit Number Index (MUNIX) in patients with 5q-Spinal Muscular Atrophy (SMA). <h3>Background:</h3> There is an unmet need for valid imaging markers to monitor the disease course and therapeutic outcome in SMA. The novel rAMIRA (radially sampled Averaged Magnetization Inversion Recovery Acquisitions) method enables high in-plane resolution MR-imaging with improved contrast of SC GM in clinically feasible acquisition times. MUNIX is an established quantitative neurophysiological measure that provides an index of the number of functioning lower motor neurons that supply a muscle. The association between SC GM area and MUNIX has not been investigated in SMA yet. <h3>Design/Methods:</h3> Using axial 2D rAMIRA imaging, we prospectively investigated 21 patients with 5q-SMA, types 2 and 3 (mean age/SD 41.3/11.6y, 9 women) and 21 age- and sex-matched healthy controls (HC) (mean age/SD 41.7/11.4y, 9 women) at the intervertebral disc level C5/C6 perpendicular to the spinal cord. SC GM areas were determined using a semi-automated approach. MUNIX of the biceps brachii (BB) was measured following the ENCALS protocol. The association between SC GM area and MUNIX was assessed using multivariable regression analysis covarying for age and sex. <h3>Results:</h3> SC GM area at C5/6 was significantly reduced in patients compared to HC (mean GM area in mm² (SD): SMA 17.5 (2.2); HC 21.1 (2.3); relative reduction: 17.1% (p&lt;0.0001)). In multivariable regression analysis SC GM area at C5/6 explained 53% of MUNIX variance of the BB in SMA patients. <h3>Conclusions:</h3> In patients with SMA, SC GM area is strongly associated with motor neuron function quantitated by MUNIX in the corresponding myotome, indicating the functional relevance of this structural marker. Further longitudinal studies are necessary to evaluate the potential of this novel imaging marker for monitoring the disease course and therapeutic outcome. <b>Disclosure:</b> Dr. Wendebourg has nothing to disclose. Dr. Kesenheimer has nothing to disclose. An immediate family member of Dr. Weidensteiner has received personal compensation for serving as an employee of Siemens Healthineers. Dr. Sander has nothing to disclose. The institution of Matthias Weigel, PhD has received research support from Biogen. Matthias Weigel, PhD has received intellectual property interests from a discovery or technology relating to health care. Tanja Haas has nothing to disclose. The institution of Dirk Fischer has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Hoffmann La Roche AG. The institution of Dirk Fischer has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Hoffmann La Roche AG. Christoph Neuwirth has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Biogen USA. Christoph Neuwirth has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Biogen Switzerland. Nathalie Braun has nothing to disclose. Dr. Weber has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Biogen Idec. Dr. Weber has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Mitsubishi Tanabe. Dr. Granziera has nothing to disclose. Prof. Sinnreich has nothing to disclose. The institution of Oliver Bieri has received research support from Swiss National Science Foundation. The institution of Oliver Bieri has received research support from MIAC Foundation. The institution of Oliver Bieri has received research support from Biogen. Oliver Bieri has received intellectual property interests from a discovery or technology relating to health care. Regina Schlaeger has nothing to disclose.

Cervical Spinal Cord Gray Matter Atrophy as an Emerging Imaging Marker in Spinal Muscular Atrophy (P2-8.014)

<h3>Objective:</h3> To prospectively evaluate cervical spinal cord (SC) gray matter (GM) area and its association with upper limb function in patients with 5q-Spinal Muscular Atrophy (SMA) using rAMIRA (radially sampled Magnetization Inversion Recovery Acquisitions) MR-imaging. <h3>Background:</h3> With the approval of the first disease modifying treatments for SMA, there is an increasing need for biomarkers that allow easy to perform, reliable, and valid disease course- and therapeutic response monitoring. The novel rAMIRA method enables high in-plane resolution MR-imaging with improved contrast of SC GM in clinically feasible acquisition times at 3 Tesla. <h3>Design/Methods:</h3> Using axial 2D rAMIRA imaging, we prospectively investigated 21 patients with 5q-SMA, types 2 and 3 (mean age/SD 41.3/11.6y, 9 women) and 21 age- and sex-matched healthy controls (HC) (mean age/SD 41.7/11.4y, 9 women) at the intervertebral disc levels C2/C3-C5/C6 perpendicular to the cord. SC GM areas were determined using a semi-automated approach. The associations between SC GM area and Revised Upper Limb Module (RULM), an established measure of upper limb disability in SMA, were assessed using multivariable regression analysis covarying for age. <h3>Results:</h3> Compared to HC cervical SC GM areas were significantly reduced in patients with SMA at C3/C4 with a relative reduction (RR) of 13.6% (p&lt;0.0001), at C4/5 RR = 16.7%, (p&lt;0.0001), at C5/6 RR = 17.1% (p&lt;0.0001), but not significantly at C2/C3 RR = 5.3% (p=0.071). In multivariable regression analysis covarying for age, GM area at C3/C4 explained 28% of RULM variance in patients with SMA. <h3>Conclusions:</h3> Cervical SC GM atrophy is detectable in patients with 5q SMA compared to HC and correlates with clinical measures of upper limb function, namely RULM. Further longitudinal investigations are necessary next steps to evaluate the potential of this novel and easy to assess imaging marker for monitoring the disease course and therapeutic response. <b>Disclosure:</b> Dr. Kesenheimer has nothing to disclose. Dr. Wendebourg has nothing to disclose. An immediate family member of Dr. Weidensteiner has received personal compensation for serving as an employee of Siemens Healthineers. Dr. Sander has nothing to disclose. The institution of Matthias Weigel, PhD has received research support from Biogen. Matthias Weigel, PhD has received intellectual property interests from a discovery or technology relating to health care. Tanja Haas has nothing to disclose. The institution of Dirk Fischer has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Hoffmann La Roche AG. The institution of Dirk Fischer has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Hoffmann La Roche AG. Christoph Neuwirth has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Biogen USA. Christoph Neuwirth has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Biogen Switzerland. Nathalie Braun has nothing to disclose. Dr. Weber has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Biogen Idec. Dr. Weber has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Mitsubishi Tanabe. Dr. Granziera has nothing to disclose. Prof. Sinnreich has nothing to disclose. The institution of Oliver Bieri has received research support from Swiss National Science Foundation. The institution of Oliver Bieri has received research support from MIAC Foundation. The institution of Oliver Bieri has received research support from Biogen. Oliver Bieri has received intellectual property interests from a discovery or technology relating to health care. Regina Schlaeger has nothing to disclose.

Image-quality characteristics in the longitudinal direction from different image-reconstruction algorithms during single-rotation volume acquisition on head computed tomography: A phantom study

A multi detector computed tomography (CT) scanner with wide-area coverage enables whole-brain volumetric scanning in a single rotation. To investigate variations in image-quality characteristics in the longitudinal direction for different image-reconstruction algorithms and strengths with phantoms. Single-rotation volume scans were performed on a 320-row multidetector CT volume scanner using three types of phantoms. Tube current was set to 200mA (standard dose) and 50mA (low dose). All images were reconstructed with filtered back projection (FBP), mild and strong levels with hybrid iterative reconstruction (HIR), and model-based IR (MBIR). Computed tomography numbers, image noise, noise power spectrum (NPS), task-based transfer function (TTF), and visual spatial resolution were used to evaluate uniformity of image quality in the longitudinal direction (Z-axis). The MBIR images showed smaller variation in CT numbers in the Z-axis. The difference in the highest and lowest CT numbers was smaller (5.0 Hounsfield units [HU]) for MBIR than for FBP (6.6 HU) and HIR (6.8 HU). The variations in image noise were the smallest for strong MBIR and the largest for FBP. The low-frequency component at NPS0.2 was lower for strong MBIR than for other algorithms. The high-frequency component at NPS0.8 was low in all reconstructions. For MBIR, the image resolution and TTFs were higher in the outer portion than in the center. Model-based IR is the optimal image-reconstruction algorithm for single-volume scan of spherical subjects owing to its high in-plane resolution and uniformity of CT numbers, image noise, and NPS in the Z-axis.

High-Resolution 3D MRI With Deep Generative Networks via Novel Slice-Profile Transformation Super-Resolution.

High-resolution magnetic resonance imaging (MRI) sequences, such as 3D turbo or fast spin-echo (TSE/FSE) imaging, are clinically desirable but suffer from long scanning time-related blurring when reformatted into preferred orientations. Instead, multi-slice two-dimensional (2D) TSE imaging is commonly used because of its high in-plane resolution but is limited clinically by poor through-plane resolution due to elongated voxels and the inability to generate multi-planar reformations due to staircase artifacts. Therefore, multiple 2D TSE scans are acquired in various orthogonal imaging planes, increasing the overall MRI scan time. In this study, we propose a novel slice-profile transformation super-resolution (SPTSR) framework with deep generative learning for through-plane super-resolution (SR) of multi-slice 2D TSE imaging. The deep generative networks were trained by synthesized low-resolution training input via slice-profile downsampling (SP-DS), and the trained networks inferred on the slice profile convolved (SP-conv) testing input for 5.5x through-plane SR. The network output was further slice-profile deconvolved (SP-deconv) to achieve an isotropic super-resolution. Compared to SMORE SR method and the networks trained by conventional downsampling, our SPTSR framework demonstrated the best overall image quality from 50 testing cases, evaluated by two abdominal radiologists. The quantitative analysis cross-validated the expert reader study results. 3D simulation experiments confirmed the quantitative improvement of the proposed SPTSR and the effectiveness of the SP-deconv step, compared to 3D ground-truths. Ablation studies were conducted on the individual contributions of SP-DS and SP-conv, networks structure, training dataset size, and different slice profiles.

3D model-based super-resolution motion-corrected cardiac T1 mapping

Objective. To provide 3D high-resolution cardiac T1 maps using model-based super-resolution reconstruction (SRR). Approach. Due to signal-to-noise ratio limitations and the motion of the heart during imaging, often 2D T1 maps with only low through-plane resolution (i.e. slice thickness of 6–8 mm) can be obtained. Here, a model-based SRR approach is presented, which combines multiple stacks of 2D acquisitions with 6–8 mm slice thickness and generates 3D high-resolution T1 maps with a slice thickness of 1.5–2 mm. Every stack was acquired in a different breath hold (BH) and any misalignment between BH was corrected retrospectively. The novelty of the proposed approach is the BH correction and the application of model-based SRR on cardiac T1 Mapping. The proposed approach was evaluated in numerical simulations and phantom experiments and demonstrated in four healthy subjects. Main results. Alignment of BH states was essential for SRR even in healthy volunteers. In simulations, respiratory motion could be estimated with an RMS error of 0.18 ± 0.28 mm. SRR improved the visualization of small structures. High accuracy and precision (average standard deviation of 69.62 ms) of the T1 values was ensured by SRR while the detectability of small structures increased by 40%. Significance. The proposed SRR approach provided T1 maps with high in-plane and high through-plane resolution (1.3 × 1.3 × 1.5–2 mm3). The approach led to improvements in the visualization of small structures and precise T1 values.

&lt;I&gt;In vivo&lt;/I&gt; mapping of hippocampal venous vasculature and oxygen saturation using dual-echo SWI/QSM at 7 T: a potential marker for neurodegeneration

Background: The current understanding of the venous system in the hippocampus is mostly based on histological and autopsy studies.1 However, the main disadvantage is that it only reveals the anatomy of the vascular system at the post-mortem stage and lacks physiological aspects associated with neuronal metabolism. In vivo characterization of the venous system using susceptibility weighted imaging (SWI) at 7 T could provide valuable information on both venous anatomy and blood oxygen saturation, through high-resolution SWI venography2 and quantitative susceptibility mapping (QSM).3 In this study, we aim to elucidate the hierarchical network of the hippocampal venous system and then test the feasibility of using venous susceptibility to characterize venous oxygenation level changes related to neurodegeneration.&#x0D; Methods: Seven healthy volunteers were recruited for this study. We used high in-plane resolution of flow-compensated dual-echo gradient echo sequence (TE1/TE2/TR=7.5/15/22 ms, voxel size: 0.25*0.25*1 mm). SWI and QSM were then reconstructed using the iterative SWI and mapping (iterative SWIM) algorithm,3 as shown in Figure 1. Hippocampus masks were extracted from the T1-MPRAGE image, which was transformed to SWI space afterwards. To reduce the partial volume effect from the tissue-vessel boundary, we extract the venous susceptibility value from each voxel along the centerline of the vessels.&#x0D; Results: High-resolution in vivo mapping of hippocampal venous vasculature exhibits a high analogy to Duvernoy’s reference4 for hippocampal vascularization. As shown in Figure 1, there is a shape of venous arch near the fimbria of the hippocampus, and small veins extending through the arch are possibly the intrahippocampal veins. The intrahippocampal veins will eventually reach the inferior ventricular vein (IVV) (anteriorly) and medial atrial vein (MAV) (posteriorly), before joining the basal vein of Rosenthal (BVR). For venous susceptibility quantification, Figure 1 shows the representative color-coded QSM for centerline extraction on BVR.&#x0D; Conclusions: Our results showed improved visualization of the micro venous system in the hippocampus using high-resolution 7 T SWI data without the contrast agent.5 In summary, the characterization of venous QSM in major tributaries related to the hippocampus offers a novel perspective on oxygen utilization in the hippocampus, which may be useful for studying age-related dementia. We delineated the hierarchical network of the hippocampus venous system using SWI/QSM at 7 T and extract the venous density and venous susceptibility value in hippocampus-related small veins and major venous tributaries, as an overall measure for venous oxygenation level related to the hippocampus, which may be used as an early marker for hippocampal atrophy in Alzheimer’s disease.

Synthesizing high-resolution magnetic resonance imaging using parallel cycle-consistent generative adversarial networks for fast magnetic resonance imaging.

The common practice in acquiring the magnetic resonance (MR) images is to obtain two-dimensional (2D) slices at coarse locations while keeping the high in-plane resolution in order to ensure enough body coverage while shortening the MR scan time. The aim of this study is to propose a novel method to generate HR MR images from low-resolution MR images along the longitudinal direction. In order to address the difficulty of collecting paired low- and high-resolution MR images in clinical settings and to gain the advantage of parallel cycle consistent generative adversarial networks (CycleGANs) in synthesizing realistic medical images, we developed a parallel CycleGANs based method using a self-supervised strategy. The proposed workflow consists of two parallely trained CycleGANs to independently predict the HR MR images in the two planes along the directions that are orthogonal to the longitudinal MR scan direction. Then, the final synthetic HR MR images are generated by fusing the two predicted images. MR images, including T1-weighted (T1), contrast enhanced T1-weighted (T1CE), T2-weighted (T2), and T2 Fluid Attenuated Inversion Recovery (FLAIR), of the multimodal brain tumor segmentation challenge 2020 (BraTS2020) dataset were processed to evaluate the proposed workflow along the cranial-caudal (CC), lateral, and anterior-posterior directions. Institutional collected MR images were also processed for evaluation of the proposed method. The performance of the proposed method was investigated via both qualitative and quantitative evaluations. Metrics of normalized mean absolute error (NMAE), peak signal-to-noise ratio (PSNR), edge keeping index (EKI), structural similarity index measurement (SSIM), information fidelity criterion (IFC), and visual information fidelity in pixel domain (VIFP) were calculated. It is shown that the proposed method can generate HR MR images visually indistinguishable from the ground truth in the investigations on the BraTS2020 dataset. In addition, the intensity profiles, difference images and SSIM maps can also confirm the feasibility of the proposed method for synthesizing HR MR images. Quantitative evaluations on the BraTS2020 dataset shows that the calculated metrics of synthetic HR MR images can all be enhanced for the T1, T1CE, T2, and FLAIR images. The enhancements in the numerical metrics over the low-resolution and bi-cubic interpolated MR images, as well as those genearted with a comparative deep learning method, are statistically significant. Qualitative evaluation of the synthetic HR MR images of the clinical collected dataset could also confirm the feasibility of the proposed method. The proposed method is feasible to synthesize HR MR images using self-supervised parallel CycleGANs, which can be expected to shorten MR acquisition time in clinical practices.

Use of Digital Tomosynthesis in Assessing Accurate Medial Epicondyle Fracture Displacement as Compared With Conventional Radiography and Computed Tomography.

Introduction:Medial epicondyle fracture displacement is notoriously difficult to determine on conventional radiography, and follow-up computed tomography (CT) is often obtained to measure precise displacement. Another option for fracture characterization is digital tomosynthesis (DT), a technology providing high in-plane resolution of bony anatomy by acquiring multiple low-dose images in a linear arc. Advantages of DT include lower radiation exposure and lower cost than CT, rapid image acquisition, and a similar patient experience to conventional radiography. The digital application of tomosynthesis is relatively new and is integrated as an add-on feature with modern radiography equipment. This study compares DT, CT and conventional radiography for measurement accuracy in medial epicondyle fractures with the goal of determining relative accuracy in measuring medial epicondyle fracture displacement.Methods:Medial epicondyle fractures were created in 5 cadaveric elbow specimens. Each specimen was imaged with conventional radiography, DT, and CT. True displacement measured by digital calipers was compared with “measured” displacement for each image acquisition. CT images included axial, sagittal, and coronal reformats. DT images of the elbow included anteroposterior (AP) longitudinal and transverse, lateral longitudinal and transverse, and axial longitudinal and transverse. Conventional radiographs included AP, lateral, and axial distal humerus images. Four physicians reviewed all images 3 months later. Each reviewer independently measured maximum apparent fracture displacement to the nearest 0.1 mm. Measurement accuracy was calculated as percent difference [(measured displacement−actual displacement)/actual displacement] for each acquisition. Mean, median, and SD for measurement accuracy were calculated. Two-tailed paired t tests were performed on each acquisition to compare the measurement accuracy.Results:Compared with conventional radiographs, accuracy of DT was superior in AP longitudinal (P=0.03), AP transverse (P=0.01), axial longitudinal (P=0.0001), and axial transverse projections (P=0.001). Accuracy of CT was superior to conventional radiography in the AP projection (P=0.03), but was equivalent in the axial projection (P=0.9). Accuracy of CT was similar to DT in AP longitudinal (P=0.6), AP transverse (P=0.5), and axial longitudinal projections (P=0.07). Accuracy of DT in the axial transverse projection was superior to CT (P=0.03).Conclusion:DT is more accurate than conventional radiography (both AP and axial views) and as accurate as CT in assessing millimeters of displacement of medial epicondyle fracture fragments.Level of Evidence:Level IV—diagnostic study.

Mono-planar T-Hex: Speed and flexibility for high-resolution 3D imaging.

PurposeThe aim of this work is the reconciliation of high spatial and temporal resolution for MRI. For this purpose, a novel sampling strategy for 3D encoding is proposed, which provides flexible k‐space segmentation along with uniform sampling density and benign filtering effects related to signal decay.MethodsFor time‐critical MRI applications such as functional MRI (fMRI), 3D k‐space is usually sampled by stacking together 2D trajectories such as echo planar imaging (EPI) or spiral readouts, where each shot covers one k‐space plane. For very high temporal and medium to low spatial resolution, tilted hexagonal sampling (T‐Hex) was recently proposed, which allows the acquisition of a larger k‐space volume per excitation than can be covered with a planar readout. Here, T‐Hex is described in a modified version where it instead acquires a smaller k‐space volume per shot for use with medium temporal and high spatial resolution.ResultsMono‐planar T‐Hex sampling provides flexibility in the choice of speed, signal‐to‐noise ratio (SNR), and contrast for rapid MRI acquisitions. For use with a conventional gradient system, it offers the greatest benefit in a regime of high in‐plane resolution <1 mm. The sampling scheme is combined with spirals for high sampling speed as well as with more conventional EPI trajectories.ConclusionMono‐planar T‐Hex sampling combines fast 3D encoding with SNR efficiency and favorable depiction characteristics regarding noise amplification and filtering effects from T2∗ decay, thereby providing flexibility in the choice of imaging parameters. It is attractive both for high‐resolution time series such as fMRI and for applications that require rapid anatomical imaging.

In-plane simultaneous multisegment imaging using a 2D RF pulse.

To develop an in-plane simultaneous multisegment (IP-SMS) imaging technique using a 2D-RF pulse and to demonstrate its ability to achieve high spatial resolution in EPI while reducing image distortion. The proposed IP-SMS technique takes advantage of periodic replicates of the excitation profile of a 2D-RF pulse to simultaneously excite multiple segments within a slice. These segments were acquired over a reduced FOV and separated using a joint GRAPPA reconstruction by leveraging virtual coils that combined the physical coil sensitivity and 2D-RF pulse spatial response. Two excitations were used with complementary spatial response profiles to adequately cover a full FOV, producing a full-FOV image that had the benefits of reduced FOV with high spatial resolution and reduced distortion. The IP-SMS technique was implemented in a diffusion-weighted single-shot EPI sequence. Experimental demonstrations were performed on a phantom and healthy human brain. In the phantom experiment, IP-SMS enabled a four-fold acceleration using an eight-channel coil without causing residual aliasing artifacts. In the human brain experiment, diffusion-weighted images with high in-plane resolution (1 × 1 mm2 ) and substantially reduced image distortion were obtained in all imaging planes in comparison with a commercial diffusion-weighted EPI sequence. The capability of IP-SMS for contiguous whole-brain coverage was also demonstrated. The proposed IP-SMS technique can realize the benefits of reduced-FOV imaging while achieving a full-FOV coverage with good image quality and time efficiency.