Computed Tomography Reconstruction Research Articles

Initial Clinical Images From a Second-Generation Prototype Silicon-Based Photon-Counting Computed Tomography System

To demonstrate the feasibility andpotential of using a second-generation prototype photon-counting computed tomography (CT) system to provide simultaneous high spatial resolution images and high spectral resolution material information across a range of routine imaging tasks using clinical patient exposure levels. The photon-counting system employs an innovative silicon-based photon-counting detector to provide a balanced approach to ultra-high-resolution spectral CT imaging. An initial cohort of volunteer subjects was imaged using the prototype photon-counting system. Acquisition technique parameters and radiation dose exposures were guided by routine clinical exposure levels used at the institution. Images were reconstructed in native slice thickness using an early version of a spectral CT reconstruction algorithm Samples of images across a range of clinical tasks were selected and presented for review. Clinical cases are presented across inner ear, carotid angiography, chest, and musculoskeletal imaging tasks. Initial reconstructed images illustrate ultra-high spatial resolution imaging. The fine detail of small structures and pathologies is clearly visualized, and structural boundaries are well delineated. The prototype system additionally provides concomitant spectral information with high spatial resolution. This initial study demonstrates that routine imaging at clinically appropriate patient exposure levels is feasible using a novel deep-silicon photon-counting detector CT system. Furthermore, a deep-silicon detector may provide a balanced approach to photon-counting CT, providing high spatial resolution imaging with simultaneous high-fidelity spectral information.

A novel radiographic analysis system for subaxial cervical spine pedicle screw placement

BackgroundPrecise pedicle screw placement of the subaxial cervical spine is difficult. Not every hospital is equipped with a guidance system that can provide effective help. Computed tomography (CT) scanning is almost a routine preoperative examination for cervical spine surgery in all hospitals. Appropriate measurement and analysis of the CT images could assist optimal cervical pedicle screw placement. The purpose of this study is to propose a new and universal method using computed tomography (CT) morphological parameters analysis to assist optimal cervical pedicle screw placement from C3 to C7.MethodsA localization system with six parameters was designed based on preoperative CT reconstruction to guide subaxial cervical spine pedicle screw placement. The six parameters were distance from the starting point to the midline [D1], distance from the starting point to the lower edge of the inferior articular process [D2], transverse section angle [TSA], sagittal section angle [SSA], pedicle width [PW], and pedicle height [PH]. The six parameters were analyzed in 53 participants.ResultsCombining D1 and D2 could localize the entrance of the pedicle screw, and we concluded that D1 and TSA and D2 and SSA could be a new standard for determination of the transverse and sagittal orientation of the pedicle screw. The six parameters were closely related to the patient’s gender, height, and weight. PH and PW were linearly correlated and could guide selection of the appropriate pedicle screw. SSA was an independent parameter of the relevant vertebral body, and changes in SSA had nothing to do with the curvature or posture of the cervical spine.ConclusionsUnderstanding and applying the six-parameter localization system are essential for achieving accurate and optimal pedicle screw placement in subaxial cervical spine, regardless of cervical sagittal alignment.

Feasibility Analysis and Surgical Details of Vestibular Schwannoma Resection via Retrolabyrinthine Approach With Preservation of Endolymphatic Sac.

Objectives: To study the feasibility of reserving the endolymphatic sac in the cerebellopontine angle (CPA) and the fundus of the internal auditory canal (IAC) by the retrolabyrinthine approach. Design: Single-center retrospective study. Methods: Through 3-dimensional preoperative computed tomography reconstruction, vestibular schwannoma (VS) resection was performed using a retrolabyrinthine approach with preservation of the endolymphatic sac in selected patients, and hearing and facial nerve functions were followed to assess the feasibility of this operation and the effectiveness of function preservation. Results: VS was completely removed in all cases and the postoperative detectable hearing retention rate (AAO-HNS hearing rating grade A, B, and C) was 80% (4/5). Postoperative facial nerve function was well preserved (HB grade I), no leakage of cerebrospinal fluid or other cranial complications was observed in all patients, and no recurrence was observed during follow-up. Conclusions: With less trauma and a promising route for hearing preservation and facial nerve functions, the retrolabyrinthine approach is a potential choice for the treatment of VS located in the IAC-CPA.

Melon and rostral muscle morphology of Gervais' beaked whale (Mesoplodon europaeus): Alternating patterns of bilateral asymmetry.

Toothed whales utilize specialized nasal structures such as the lipid-rich melon to produce sound and propagate it into the aquatic environment. Very little nasal morphology of mesoplodont beaked whales has been described in the literature, and the anatomy of the melon and associated musculature of Gervais' beaked whale (Mesoplodon europaeus) remains undescribed. Heads of three (n = 3) Gervais' beaked whales were examined in detail via dissection as well as computed tomography (CT) and magnetic resonance imaging (MRI). Two additional Gervais' beaked whale individuals (n = 2) were studied via archived CT and MRI scans. Representative transverse dissection sections of the melon were processed for polarized light imaging to verify the presence of tendons inserting into the melon tissue. Three-dimensional (3D) CT reconstructions of the melon, rostral muscles, and associated structures were performed to assess morphology and spatial relationships. In all individuals, the melon's main body demonstrated a bilaterally asymmetrical, curvilinear geometry. This curvilinear shape was defined by a pattern of alternating asymmetry in the medial rostral muscles that projected into the melon's tissue. In transverse polarized light imaging, a network of tendons originating from these asymmetrical rostral muscle projections was observed permeating the melon's lipid tissue. This curvilinear melon morphology and associated asymmetrical musculature suggest a means of lengthening the lipid pathway within a relatively short dimensional footprint. In addition, the species-specific arrangement of muscular projections suggests complex fine-tuning of the melon's geometry during echolocation. Further studies may lend additional insight into the function of this unusual melon morphology.

Multi-layer clustering-based residual sparsifying transform for low-dose CT image reconstruction.

The recently proposed sparsifying transform (ST) models incur low computational cost and have been applied to medical imaging. Meanwhile, deep models with nested network structure reveal great potential for learning features in different layers. In this study, we propose a network-structured ST learning approach for X-ray computed tomography (CT), which we refer to as multi-layer clustering-based residual sparsifying transform (MCST) learning. The proposed MCST scheme learns multiple different unitary transforms in each layer by dividing each layer's input into several classes. We apply the MCST model to low-dose CT (LDCT) reconstruction by deploying the learned MCST model into the regularizer in penalized weighted least squares (PWLS) reconstruction. The proposed MCST model combines a multi-layer sparse representation structure with multiple clusters for the features in each layer that are modeled by a rich collection of transforms. We train the MCST model in an unsupervised manner via a block coordinate descent (BCD) algorithm. Since our method is patch-based, the training can be performed with a limited set of images. For CT image reconstruction, we devise a novel algorithm called PWLS-MCST by integrating the pre-learned MCST signal model with PWLSoptimization. We conducted LDCT reconstruction experiments on XCAT phantom data, Numerical Mayo Clinical CT dataset and "LDCT image and projection dataset" (Clinical LDCT dataset). We trained the MCST model with two (or three) layers and with five clusters in each layer. The learned transforms in the same layer showed rich features while additional information is extracted from representation residuals. Our simulation results and clinical results demonstrate that PWLS-MCST achieves better image reconstruction quality than the conventional filtered back-projection (FBP) method and PWLS with edge-preserving (EP) regularizer. It also outperformed recent advanced methods like PWLS with a learned multi-layer residual sparsifying transform (MARS) prior and PWLS with a union of learned transforms (ULTRA), especially for displaying clear edges and preserving subtledetails. In this work, a multi-layer sparse signal model with a nested network structure is proposed. We refer this novel model as the MCST model that exploits multi-layer residual maps to sparsify the underlying image and clusters the inputs in each layer for accurate sparsification. We presented a new PWLS framework with a learned MCST regularizer for LDCT reconstruction. Experimental results show that the proposed PWLS-MCST provides clearer reconstructions than several baseline methods. The code for PWLS-MCST is released at https://github.com/Xikai97/PWLS-MCST.

The Reliability of Measurements for Tibial Torsion: A Comparison of CT, MRI, Biplanar Radiography, and 3D Reconstructions With and Without Standardized Measurement Training

Background: Accurate and reliable assessment of tibial torsion is important for the identification and treatment of tibial rotation malalignment; however, the ideal rotational measurement modality and technique are controversial. This study compares rotational measurements between computed tomography (CT), magnetic resonance imaging (MRI), biplanar erect radiograph (BER) reconstructions, and three-dimensional (3D) reconstructed CT before and after standardized training of observers to evaluate the reliability of each assessment method.Methods: Eight adult cadaveric specimens underwent CT, MRI, and BER imaging. Tibial torsion was measured by three independent observers (one resident and two experienced orthopaedic surgeons) both before and after standardized measurement instruction. Reliability for inter-observer measurement was evaluated using the Intraclass Correlation Coefficient (ICC). Measurement values for CT, MRI, and BER reconstructions were compared to 3D CT reconstructions analyzed using custom software to align and measure tibial torsion (used as the reference standard).Results: Before training, there was poor inter-observer reliability for CT (ICC=0.492, p=0.014) and moderate interobserver reliability for MRI (ICC=0.633, p=0.002). There was no inter-method reliability between 3D CT and MRI for two of the three surgeons and moderate to good reliability between 3D CT and CT. After training, the interobserver reliability for CT improved to 0.536 and the inter-observer reliability for MRI improved to 0.701. The BER measurements (no observer involvement) had moderate reliability compared to the 3D CT reconstructions (ICC=0.69, p=0.026). Measurement error was 4 degrees for CT pre-training and 7 degrees post-training, and 7 degrees for MRI pre-training and 8 degrees post-training.Conclusions: A standardized training regimen for MRI measurements improved both inter-observer and intraobserver reliability. Inter-method reliability between CT, MRI, and BER compared to reference 3D CT reconstructions demonstrated that all imaging modalities are a valid means to measure tibial rotation but differ in reliability from moderate to good. When assessing tibial torsional deformities, it is important to consider these variations from true rotation and feel comfortable using them for pre-operative planning purposes.Level of Evidence: N/A for a cadaveric imaging study Key Concepts•A standardized training regimen for MRI measurements of tibial torsion significantly improved inter-observer and intra-observer reliability.•CT, MRI, and BER measurements are valid imaging modalities compared to 3D CT reference standards for measuring tibial rotation.•Established manual measurements of tibial torsion from MRI may be improved with adequate training.•Measurement error of up to 8 degrees can be seen from true tibial torsion during the imaging assessment.

Gamma radiation detector selection for CT scanner

Three types of gamma radiation detectors associated with distributed electronics namely, NaI (Tl), HPGe and LaBr3(Ce) are compared primarily focusing on electronic noise and scattering noise. Additionally, detectors of same make, material, size and electronics are also compared. A methodology is proposed to select the most suitable detector for computed tomography (CT) among the available options. Standard deviation parameter is employed to estimate electronic noise without performing CT experiment. Kanpur theorem-1(KT-1) is used to estimate the scattering noise quantitatively after verifying its sensitivity to scattering noise. The impact of scattering noise on CT profiles is evaluated using dice similarity dice coefficient. A good resemblance between KT-1 and dice coefficient is observed. A maximum difference of 56% in scattering noise is observed when five detectors used simultaneously instead of single detector whereas a discrepancy of 85% is observed between different types of radiation detectors. As far as ease of handling, operational and capital cost is concern one has to compromise minimum 12% of accuracy in CT reconstruction if NaI (Tl) detector is used with respect to best alternative available.The proposed methodology can be applied to measurement that require minimal scattering interference data other than CT experiments. The manufacturer can add noise level of detector as a characteristic parameter in the data sheet.

Imaging Findings in Pediatric Accessory Cranial Sutures using 3D CT Reconstruction: Fracture or Rudimentary Suture.

Objective:Although accessory sutures are considered to be relatively rare, the consequences of a missed diagnosis are profound. Distinguishing between accessory sutures and cranial fractures can be difficult, especially in cases of suspected non-accidental trauma. High-resolution imaging is the best way to discern between two- and three-dimensional computerized tomography (3D CT) is considered the preferred method for evaluation. The goal of this study was to determine the impact of 3D CT scans in distinguishing between accessory sutures and cranial fractures in suspected child abuse cases in a rural community and the importance of early detection in such cases, as well as call attention to the consequences of initial misinterpretation. Materials and methods:The researchers conducted a retrospective chart review of all pediatric patients diagnosed with cranial fractures (265 in total) at University Medical Center between May 30, 2016, and May 30, 2021. Initial computed tomography (CT) scans and subsequent 3D CT scans were evaluated for each patient that fit the inclusion criteria, 13 in total. Patients were then categorized into two groups based on the final diagnosis on the radiology report: accessory cranial suture or cranial fracture. Once these patients were identified, the etiology and structural components of each were evaluated, and the key differences were highlighted. Results:Our results showed that, of the 11 cases of suspected non-accidental trauma, six were finally diagnosed with accessory sutures with the use of 3D CT scans, and of those six, four were diagnosed with cranial fractures from the initial CT scan report due to a similar presentation and asymmetric nature. Conclusion:Discerning between fracture and accessory suture is essential in evaluating pediatric patients presenting with signs ofcranial fracture due to the increased risk of misinterpretation that can lead to severe legal consequences considering that cranial suture variants may mimic intentional injury and be mistaken for child abuse, causing significant distress for patients and their families.

Body composition assessment with limited field-of-view computed tomography: A semantic image extension perspective.

Field-of-view (FOV) tissue truncation beyond the lungs is common in routine lung screening computed tomography (CT). This poses limitations for opportunistic CT-based body composition (BC) assessment as key anatomical structures are missing. Traditionally, extending the FOV of CT is considered as a CT reconstruction problem using limited data. However, this approach relies on the projection domain data which might not be available in application. In this work, we formulate the problem from the semantic image extension perspective which only requires image data as inputs. The proposed two-stage method identifies a new FOV border based on the estimated extent of the complete body and imputes missing tissues in the truncated region. The training samples are simulated using CT slices with complete body in FOV, making the model development self-supervised. We evaluate the validity of the proposed method in automatic BC assessment using lung screening CT with limited FOV. The proposed method effectively restores the missing tissues and reduces BC assessment error introduced by FOV tissue truncation. In the BC assessment for large-scale lung screening CT datasets, this correction improves both the intra-subject consistency and the correlation with anthropometric approximations. The developed method is available at https://github.com/MASILab/S-EFOV.

PSDP: Pseudo-supervised dual-processing for low-dose cone-beam computed tomography reconstruction

PSDP: Pseudo-supervised dual-processing for low-dose cone-beam computed tomography reconstruction

Characterization of breast tissues in density and effective atomic number basis via spectral X-ray computed tomography

Objective. Differentiation of breast tissues is challenging in X-ray imaging because tissues might share similar or even the same linear attenuation coefficients μ. Spectral computed tomography (CT) allows for more quantitative characterization in terms of tissue density (ρ) and effective atomic number (Z eff) by exploiting the energy dependence of μ. The objective of this study was to examine the potential of ρ/Z eff decomposition in spectral breast CT so as to explore the benefits of tissue characterization and improve the diagnostic accuracy of this emerging 3D imaging technique. Approach. In this work, 5 mastectomy samples and a phantom with inserts mimicking breast soft tissues were evaluated in a retrospective study. The samples were imaged at three monochromatic energy levels in the range of 24–38 keV at 5 mGy per scan using a propagation-based phase-contrast setup at SYRMEP beamline at the Italian national synchrotron Elettra. Main results. A custom-made algorithm incorporating CT reconstructions of an arbitrary number of spectral energy channels was developed to extract the density and effective atomic number of adipose, fibro-glandular, pure glandular, tumor, and skin from regions selected by a radiologist. Significance. Preliminary results suggest that, via spectral CT, it is possible to enhance tissue differentiation. It was found that adipose, fibro-glandular and tumorous tissues have average effective atomic numbers (5.94 ± 0.09, 7.03 ± 0.012, and 7.40 ± 0.10) and densities (0.90 ± 0.02, 0.96 ± 0.02, and 1.07 ± 0.03 g cm−3) and can be better distinguished if both quantitative values are observed together.

Photon-counting spectral CT reconstruction with sparse and double low-rank components fusion

ObjectivePhoton-counting CT (computed tomography) has aroused more attention. The relatively high dose of X-ray increases concerns about radiation, dose reduction can mitigate radiation risk, however, projection data within energy channel suffers from strong noise, leading to degraded results. Besides, there is high correlation among channels since the multi-channel data is captured from the same object, which should be utilized to improve reconstruction quality. To suppress noise while preserving details in reconstruction and improve material decomposition accuracy, we proposed a novel optimization-based spectral CT reconstruction method with a fusion framework. MethodsTwo results were obtained by implementing two iterative reconstructions with different constraints and can be treated as complementary components. The first constraint is sparsity, the second constraint is double low-rank, moreover, a heuristic strategy is designed to solve double low-rank constrained optimization. We developed a guided filter based fusion framework to fuse sparse component and double low-rank component. ResultsExperimental results showed that our method achieved satisfying visualization, higher values of PSNR (peak signal-to-noise ratio) and SSIM (structure similarity), moreover, lower material decomposition error in this work.Conclusion.Our method can well perform spectral CT reconstruction under low-signal-to-ratio situation and it has decent performance in noise suppression, edge preserving and material discrimination.

Manejo multimodal de aneurisma cirsoideo. Reporte de un caso y revisión de la literatura

Scalp arteriovenous malformations (AVMs), commonly known as cirsoid aneurysms, are complex collections of arteries and veins that communicate directly. It has been postulated that AVMs occur both through the activation of inactive fistulous lesions and through the formation of new fistulous vascular channels through the canalization of injured vessels. Depending on their size and complexity, a variety of strategies can be used to treat them. The case of a 53-year-old male is presented who came for evaluation due to an increase in the volume of a pulsating lesion that extended from the right parietotemporal region accompanied by holocraneal headache and tinnitus. A cerebral angiotomography was performed in which a lesion composed of ectatic serpentine images, with a right extracranial location, subsequently diagnostic cerebral angiography was performed demonstrating a right frontoparietal fistulous and plexiform arteriovenous shunt fed by afferents from the right superficial temporal, posterior auricular, occipital and middle meningeal arteries, with drainage through a varicose and ectatic vein towards the external jugular vein. An endovascular embolization of the arteriovenous shunt was performed via the femoral route in two stages, achieving a “pressure cooker” construction to perform the injection of embolizing fluid and have a subtotal embolization of 93%, with this it was possible to perform a microsurgical resection with a right fronto-parietal-temporal approach resecting the deepest part, leaving the scalp with embolized vessels without resecting. To close the surgical wound, a simple flap rotation was performed to cover the skin defect derived from its resection. The patient underwent post-resection cerebral panangiography, identifying the absence of arteriovenous shunt. Depending on the extent and anatomical complexity of the malformation, clinical manifestations may vary. More than 80% of AVMs will be present at birth and the rest may develop after head trauma or surgical interventions. Appropriate therapeutic intervention is needed through selective angiography, during which endovascular embolization of the major feeding vessels can be performed before open surgical intervention. Volume-based three-dimensional computed tomography reconstructions can be used for further planning. Surgical reduction of post-embolization arterial supply will reduce intraoperative hemorrhage, facilitating total surgical resection. Removal can result in large, full-thickness skin defects that may be difficult to reconstruct without compromising functional and cosmetic results. Keywords: Cirsoid aneurysm; multimodal management; treatment; endovascular; giant

Effect of the Innominate Bone Horizontal Rotation on Acetabular Version: A Retrospective Radiological Study on a Middle Eastern Population

Background: The impact of acetabular horizontal rotation on the development of femoroacetabular impingement and subsequently osteoarthritis is well-studied in the literature. However, there is not a clear relationship between the rotation of the hemipelvis and the version of the acetabulum. Purpose: The purpose of this study was to evaluate the influence of the rotation of the hemipelvis on the version of the acetabulum. Methods: Through a retrospective study, three-dimensional reconstructions of high-resolution CT (computed tomography) scans of 154 patients receiving pelvic scans for non-orthopedic causes were selected from our institution’s database. The horizontal rotation of the different parts of the hemipelvis was evaluated using the following parameters: superior iliac spine angle (SIS), inferior iliac spine angle (IIS), roof edge angle (REA), equatorial edge angle (EEA) and ischiopubic angle (IPA). Results: The results showed a significant positive correlation between the different angles of the innominate bone and the version of the acetabulum such as when the proximal innominate bone rotates, the cranial part of the acetabulum rotates in the opposite direction. Increased anteversion angles in females compared to males were also observed. Conclusion: The observations suggest that, in an asymptomatic population, the acetabulum should not be considered a separate entity independent from the rest of the innominate bone and that the version of the acetabulum correlates with the rotation of the hemipelvis.

Cerebrovascular segmentation in phase-contrast magnetic resonance angiography by a Radon projection composition network.

Cerebrovascular segmentation based on phase-contrast magnetic resonance angiography (PC-MRA) provides patient-specific intracranial vascular structures for neurosurgery planning. However, the vascular complex topology and spatial sparsity make the task challenging. Inspired by the computed tomography reconstruction, this paper proposes a Radon Projection Composition Network (RPC-Net) for cerebrovascular segmentation in PC-MRA, aiming to enhance distribution probability of vessels and fully obtain the vascular topological information. Multi-directional Radon projections of the images are introduced and a two-stream network is used to learn the features of the 3D images and projections. The projection domain features are remapped to the 3D image domain by filtered back-projection transform to obtain the image-projection joint features for predicting vessel voxels. A four-fold cross-validation experiment was performed on a local dataset containing 128 PC-MRA scans. The average Dice similarity coefficient, precision and recall of the RPC-Net achieved 86.12%, 85.91% and 86.50%, respectively, while the average completeness and validity of the vessel structure were 85.50% and 92.38%, respectively. The proposed method outperformed the existing methods, especially with significant improvement on the extraction of small and low-intensity vessels. Moreover, the applicability of the segmentation for electrode trajectory planning was also validated. The results demonstrate that the RPC-Net realizes an accurate and complete cerebrovascular segmentation and has potential applications in assisting neurosurgery preoperative planning.

Structural Gaussian priors for Bayesian CT reconstruction of subsea pipes

A non-destructive testing (NDT) application of X-ray computed tomography (CT) is inspection of subsea pipes in operation via 2D cross-sectional scans. Data acquisition is time-consuming and costly due to the challenging subsea environment. While reducing the number of projections in a CT scan yields time and cost savings, this compromises the reconstruction quality when conventional reconstruction methods are used. To address this issue we take a Bayesian approach to CT reconstruction and focus on designing an effective prior that introduces additional information to the problem. We propose a new class of structural Gaussian priors to enforce expected material properties in different regions of the reconstructed image based on available a-priori information about the pipe's layered structure. The prior is composed from Gaussian distributions, and we utilize this to propose an efficient computational strategy to sample the resulting posterior distribution. This is essential in practical NDT applications for fast processing of the large-scale data we see in CT. Numerical experiments with synthetic and real data show that the proposed structural Gaussian prior reduces artifacts and enhances contrast in the reconstruction, compared to using only a conventional Gaussian Markov random field prior or no prior information at all. Furthermore, we obtain uncertainty estimates that indicate the certainty of the reconstructions increase when more prior information is added.

A novel hybrid generative adversarial network for CT and MRI super-resolution reconstruction

Objective. Computed tomography (CT) and magnetic resonance imaging (MRI) are widely used in medical imaging modalities, and provide valuable information for clinical diagnosis and treatment. However, due to hardware limitations and radiation safety concerns, the acquired images are often limited in resolution. Super-resolution reconstruction (SR) techniques have been developed to enhance the resolution of CT and MRI slices, which can potentially improve diagnostic accuracy. To capture more useful feature information and reconstruct higher quality super-resolution images, we proposed a novel hybrid framework SR model based on generative adversarial networks. Approach. The proposed SR model combines frequency domain and perceptual loss functions, which can work in both frequency domain and image domain (spatial domain). The proposed SR model consists of 4 parts: (i) the discrete Fourier transform (DFT) operation transforms the image from the image domain to frequency domain; (ii) a complex residual U-net performs SR in the frequency domain; (iii) the inverse discrete Fourier transform (iDFT) operation based on data fusion transforms the image from the frequency domain to image domain; (iv) an enhanced residual U-net network is used for SR of image domain. Main results. Experimental results on bladder MRI slices, abdomen CT slices, and brain MRI slices show that the proposed SR model outperforms state-of-the-art SR methods in terms of visual quality and objective quality metric such as the structural similarity (SSIM) and the peak signal-to-noise ratio (PSNR), which proves that the proposed model has better generalization and robustness. (Bladder dataset: upscaling factor of 2: SSIM = 0.913, PSNR = 31.203; upscaling factor of 4: SSIM = 0.821, PSNR = 28.604. Abdomen dataset: upscaling factor of 2: SSIM = 0.929, PSNR = 32.594; upscaling factor of 4: SSIM = 0.834, PSNR = 27.050. Brain dataset: SSIM = 0.861, PSNR = 26.945). Significance. Our proposed SR model is capable of SR for CT and MRI slices. The SR results provide a reliable and effective foundation for clinical diagnosis and treatment.

Change of Femoral Anteversion Angle in Children With Intoeing Gait Measured by Three-Dimensional Computed Tomography Reconstruction: 3-Year Follow-Up Study.

To investigate long-term changes in femoral anteversion angle (FAA) in children with intoeing gait and to identify factors that affect FAA changes. We retrospectively analyzed three-dimensional computed tomography data from 2006 to 2022 of children with intoeing gait with ≥3 years of follow-up without active treatment. The study examined the mean changes in FAA, the effects of sex, age, and initial FAA on FAA change, and mean FAAs by age. Changes in FAA severity up to eight years of age were also observed and analyzed by sex. A total of 126 lower limbs of 63 children (30 males, 33 females) with intoeing gait were included, with a mean age of 5.11±1.05 years and a mean follow-up period of 43.59±7.74 months. The initial FAA was 41.42°±8.29° and the follow-up FAA was 33.25°±9.19°, indicating a significant decrease (p<0.001). Significant correlations were observed between age and changes in FAA, as well as between initial FAA and changes in FAA (r=0.248, p=0.005; r=-0.333, p<0.001). At age 8 years, only 22 limbs were classified as having mild FAA severity. During the follow-up period, children with intoeing gait had a significant decreased in FAA. No significant difference in FAA change was found between sex, but younger children and those with greater initial FAA were more likely to have decreased FAA. However, most children retained moderate to severe severity of increased FAA. Further studies are required to validate these findings.

TRCT-GAN: CT reconstruction from biplane X-rays using transformer and generative adversarial networks

Computed tomography (CT) provides a three-dimensional view of a patient's internal organs. Compared to CT volumes, X-ray imaging can significantly reduce the patient's exposure to ionizing radiation. Moreover, X-ray images are more economical and widely applied in surgical procedures. However, X-ray images can only provide two-dimensional information. In this paper, an end-to-end GAN network, named TRCT-GAN, is proposed to reconstruct chest CT volumes from biplane X-ray images. In the GAN network, the Transformer network module is employed to enhance the feature representation of X-ray images. Moreover, a dynamic attention module is added to exploit some 2D feature maps and 3D feature maps to enhance the contextual association. The experimental results demonstrate that the proposed network can effectively produce high-quality CT reconstructions from X-ray images.

Sparsier2Sparse: Self-supervised convolutional neural network-based streak artifacts reduction in sparse-view CT images.

Sparse-view computed tomography (CT) has attracted a lot of attention for reducing both scanning time and radiation dose. However, sparsely-sampled projection data generate severe streak artifacts in the reconstructed images. In recent decades, many sparse-view CT reconstruction techniques based on fully-supervised learning have been proposed and have shown promising results. However, it is not feasible to acquire pairs of full-view and sparse-view CT images in real clinical practice. In this study, we propose a novel self-supervised convolutional neural network (CNN) method to reduce streak artifacts in sparse-view CTimages. We generate the training dataset using only sparse-view CT data and train CNN based on self-supervised learning. Since the streak artifacts can be estimated using prior images under the same CT geometry system, we acquire prior images by iteratively applying the trained network to given sparse-view CT images. We then subtract the estimated steak artifacts from given sparse-view CT images to produce the finalresults. We validated the imaging performance of the proposed method using extended cardiac-torso (XCAT) and the 2016 AAPM Low-Dose CT Grand Challenge dataset from Mayo Clinic. From the results of visual inspection and modulation transfer function (MTF), the proposed method preserved the anatomical structures effectively and showed higher image resolution compared to the various streak artifacts reduction methods for all projection views. We propose a new framework for streak artifacts reduction when only the sparse-view CT data are given. Although we do not use any information of full-view CT data for CNN training, the proposed method achieved the highest performance in preserving fine details. By overcoming the limitation of dataset requirements on fully-supervised-based methods, we expect that our framework can be utilized in the medical imagingfield.