A multiscale 3D network for lung nodule detection using flexible nodule modeling.

BACKGROUND: Lung cancer is the most common type of cancer. Detection of lung cancer at an early stage can reduce mortality rates. Pulmonary nodules may represent early cancer and can be identified through computed tomography (CT) scans. Malignant risk can be estimated based on attributes like size, shape, location, and density. PURPOSE: Deep learning algorithms have achieved remarkable advancements in this domain compared to traditional machine learning methods. Nevertheless, many existing anchor-based deep learning algorithms exhibit sensitivity to predefined anchor-box configurations, necessitating manual adjustments to obtain optimal outcomes. Conversely, current anchor-free deep learning-based nodule detection methods normally adopt fixed-size nodule models like cubes or spheres. METHODS: To address these technical challenges, we propose a multiscale 3D anchor-free deep learning network (M3N) for pulmonary nodule detection, leveraging adjustable nodule modeling (ANM). Within this framework, ANM empowers the representation of target objects in an anisotropic manner, with a novel point selection strategy (PSS) devised to accelerate the learning process of anisotropic representation. We further incorporate a composite loss function that combines the conventional L2 loss and cosine similarity loss, facilitating M3N to learn nodules' intensity distribution in three dimensions. RESULTS: Experiment results show that the M3N achieves 90.6% competitive performance metrics (CPM) with seven predefined false positives per scan on the LUNA 16 dataset. This performance appears to exceed that of other state-of-the-art deep learning-based networks reported in their respective publications. Individual test results also demonstrate that M3N excels in providing more accurate, adaptive bounding boxes surrounding the contours of target nodules. CONCLUSIONS: The newly developed nodule detection system reduces reliance on prior knowledge, such as the general size of objects in the dataset, thus it should enhance overall robustness and versatility. Distinct from traditional nodule modeling techniques, the ANM approach aligns more closely with the morphological characteristics of nodules. Time consumption and detection results demonstrate promising efficiency and accuracy which should be validated in clinical settings.

A Qualitative Study of the Care Experience of Mothers of Asthmatic Children from Low-Income Families.

Objectives: This study aimed to explore the experiences and caregiving perspectives of mothers from low-income families who have children with asthma,and to establish a foundation for the development of tailored nursing strategies specifically designed for families facing similar circumstances. Method: A descriptive qualitative research method was employed. Fifteen mothers of asthmatic children from low-income families receiving treatment at the Respiratory Centre of Chongqing Children's Hospital were purposefully sampled from June to December 2021. Semi-structured interviews were conducted to gather data on their caregiving experiences, and thematic analysis was utilized to analyze the interview data. Results: The interviewees were 27-42 years old (SD=32.3 years), 33.3% were full-time mothers(A woman who quit work, in order to focus on taking care of the child and the family ), 53.3% had one child, 46.6% had a college degree or higher, and 100% had health insurance. Through in-depth interviews, four main themes and eight sub-themes were identified, including (a) insufficient knowledge about the disease, (b) anxiety and uncertainty, (c) insufficient social support system, and (d) insufficient resources for medical services. The first theme describes a weak willingness to learn and medication discontinuation at will. The second theme describes financial burden and psychological stress, the third theme describes lack of family support and low social participation, and the fourth theme describes insufficient health insurance support and unequal distribution of healthcare resources. Conclusion: Mothers from low-income families with asthmatic children face substantial psychological burdens and familial pressures. It is crucial for healthcare professionals to actively engage with and deepen their understanding of these mothers' caregiving experiences and psychological well-being. By doing so, positive coping strategies can be developed, promoting the physical and mental health of these mothers and improving asthma control in their children.

An Improved Anchor-Free Nodule Detection System Using Feature Pyramid Network.

Lung cancer (LC) is the leading cause of cancer death. Detecting LC at the earliest stage facilitates curative treatment options and will improve mortality rates. Computer-aided detection (CAD) systems can help improve LC diagnostic accuracy. In this work, we propose a deep-learning-based lung nodule detection method. The proposed CAD system is a 3D anchor-free nodule detection (AFND) method based on a feature pyramid network (FPN). The deep learning-based CAD system has several novel properties: (1) It achieves region proposal and nodule classification in a single network, forming a one-step detection pipeline and reducing operation time. (2) An adaptive nodule modelling method was designed to detect nodules of various sizes. (3) The proposed AFND also establishes a novel center point selection mechanism for better classification. (4) Based on the new nodule model, a composite loss function integrating cosine similarity (CS) loss and SmoothL1loss was designed to further improve the nodule detection accuracy. Experimental results show that the AFND outperforms other similar nodule detection systems on the LUNA 16 dataset.

AF automatic classification based on different time-delay values of the recurrence plot.

Recurrence plot (RP) has been widely used to transform 1D ECG waveforms into 2D images and explore the recurrence patterns of the electrical signals generated by the cardiac system. However, selecting the critical parameter time-delay τ for creating an RP has not been well-reported. In this work, we investigated the influence of different τ values on the RP-based AF prediction. And the results illustrated that the best classification performance could be achieved at τ=1 with full characters of the dynamic system.Clinical Relevance-This work established the AF classification system based on recurrence features and found the optimal parameters of the recurrence plot improving the ECG-based classification performance.

Atrial fibrillation classification based on the 2D representation of minimal subset ECG and a non-deep neural network.

Atrial fibrillation (AF) is the most common cardiac arrhythmia, and its early detection is critical for preventing complications and optimizing treatment. In this study, a novel AF prediction method is proposed, which is based on investigating a subset of the 12-lead ECG data using a recurrent plot and ParNet-adv model. The minimal subset of ECG leads (II &V1) is determined via a forward stepwise selection procedure, and the selected 1D ECG data is transformed into 2D recurrence plot (RP) images as an input to train a shallow ParNet-adv Network for AF prediction. In this study, the proposed method achieved F1 score of 0.9763, Precision of 0.9654, Recall of 0.9875, Specificity of 0.9646, and Accuracy of 0.9760, which significantly outperformed solutions based on single leads and complete 12 leads. When studying several ECG datasets, including the CPSC and Georgia ECG databases of the PhysioNet/Computing in Cardiology Challenge 2020, the new method achieved F1 score of 0.9693 and 0.8660, respectively. The results suggested a good generalization of the proposed method. Compared with several state-of-art frameworks, the proposed model with a shallow network of only 12 depths and asymmetric convolutions achieved the highest average F1 score. Extensive experimental studies proved that the proposed method has a high potential for AF prediction in clinical and particularly wearable applications.

Corrigendum: Improved real-time semantic segmentation network model for crop vision navigation line detection.

[This corrects the article DOI: 10.3389/fpls.2022.898131.].

Cardiac Arrhythmia classification based on 3D recurrence plot analysis and deep learning.

Artificial intelligence (AI) aided cardiac arrhythmia (CA) classification has been an emerging research topic. Existing AI-based classification methods commonly analyze electrocardiogram (ECG) signals in lower dimensions, using one-dimensional (1D) temporal signals or two-dimensional (2D) images, which, however, may have limited capability in characterizing lead-wise spatiotemporal correlations, which are critical to the classification accuracy. In addition, existing methods mostly assume that the ECG data are linear temporal signals. This assumption may not accurately represent the nonlinear, nonstationary nature of the cardiac electrophysiological process. In this work, we have developed a three-dimensional (3D) recurrence plot (RP)-based deep learning algorithm to explore the nonlinear recurrent features of ECG and Vectorcardiography (VCG) signals, aiming to improve the arrhythmia classification performance. The 3D ECG/VCG images are generated from standard 12 lead ECG and 3 lead VCG signals for neural network training, validation, and testing. The superiority and effectiveness of the proposed method are validated by various experiments. Based on the PTB-XL dataset, the proposed method achieved an average F1 score of 0.9254 for the 3D ECG-based case and 0.9350 for the 3D VCG-based case. In contrast, recently published 1D and 2D ECG-based CA classification methods yielded lower average F1 scores of 0.843 and 0.9015, respectively. Thus, the improved performance and visual interpretability make the proposed 3D RP-based method appealing for practical CA classification.

Improved Real-Time Semantic Segmentation Network Model for Crop Vision Navigation Line Detection.

Field crops are generally planted in rows to improve planting efficiency and facilitate field management. Therefore, automatic detection of crop planting rows is of great significance for achieving autonomous navigation and precise spraying in intelligent agricultural machinery and is an important part of smart agricultural management. To study the visual navigation line extraction technology of unmanned aerial vehicles (UAVs) in farmland environments and realize real-time precise farmland UAV operations, we propose an improved ENet semantic segmentation network model to perform row segmentation of farmland images. Considering the lightweight and low complexity requirements of the network for crop row detection, the traditional network is compressed and replaced by convolution. Based on the residual network, we designed a network structure of the shunting process, in which low-dimensional boundary information in the feature extraction process is passed backward using the residual stream, allowing efficient extraction of low-dimensional information and significantly improving the accuracy of boundary locations and row-to-row segmentation of farmland crops. According to the characteristics of the segmented image, an improved random sampling consensus algorithm is proposed to extract the navigation line, define a new model-scoring index, find the best point set, and use the least-squares method to fit the navigation line. The experimental results showed that the proposed algorithm allows accurate and efficient extraction of farmland navigation lines, and it has the technical advantages of strong robustness and high applicability. The algorithm can provide technical support for the subsequent quasi-flight of agricultural UAVs in farmland operations.

Dense channel splitting network for MR image super-resolution.

High spatial resolution is desirable in magnetic resonance imaging (MRI) as it can provide detailed anatomical information, facilitating radiologists with accurate quantitative analysis. Super-resolution (SR) algorithms are effective approaches to enhance MR images' spatial resolution. In the past few years, convolutional neural network (CNN)-based SR methods have significantly improved and outperformed conventional ones. However, existing CNN-based SR methods usually do not explicitly consider the frequency property of images, leading to the limited representation of high-frequency components reflecting image details. To alleviate this problem, a dense channel splitting network (DCSN) algorithm is proposed to process the frequency bands for better feature detection. Specifically, a channel splitting module, a cascaded multi-branch dilation module, and a dense-in and recursive-out mechanism are designed to separate frequency bands of MR images and forward the high-frequency information to deeper layers for reconstruction. Several experiments are performed on real T2 brain and PD (proton density) knee images. The results indicate that the proposed network is superior to conventional CNN-based SR methods.

Exposure of Infants to Gradient Fields in a Baby MRI Scanner.

In pediatric magnetic resonance imaging (MRI), infants are exposed to rapid, time-varying gradient magnetic fields, leading to electric fields induced in the body of infants and potential safety risks (e.g. peripheral nerve stimulation). In this numerical study, the in situ electric fields in infants induced by small-sized gradient coils for a 1.5 T MRI scanner were evaluated. The gradient coil set was specially designed for the efficient imaging of infants within a small-bore (baby) scanner. The magnetic flux density and induced electric fields by the small x, y, z gradient coils in an infant model (8-week-old with a mass of 4.3 kg) were computed using the scalar potential finite differences method. The gradient coils were driven by a 1 kHz sinusoidal waveform and also a trapezoidal waveform with a 250 µs rise time. The model was placed at different scan positions, including the head area (position I), chest area (position II), and body center (position III). It was found that the induced electric fields in most tissues exceeded the basic restrictions of the ICNIRP 2010 guidelines for both waveforms. The electric fields were similar in the region of interest for all coil types and model positions but different outside the imaging region. The y-coil induced larger electric fields compared with the x- and z- coils. Bioelectromagnetics. 43:69-80, 2022. © 2021 Bioelectromagnetics Society.

Tripartite motif-containing protein 11 promotes hepatocellular carcinogenesis through ubiquitin-proteasome-mediated degradation of pleckstrin homology domain leucine-rich repeats protein phosphatase 1.

BACKGROUND AND AIMS: HCC is one of the main types of primary liver cancer, with high morbidity and mortality and poor treatment effect. Tripartite motif-containing protein 11 (TRIM11) has been shown to promote tumor formation in lung cancer, breast cancer, gastric cancer, and so on. However, the specific function and mechanism of TRIM11 in HCC remain open for study. APPROACH AND RESULTS: Through clinical analysis, we found that the expression of TRIM11 was up-regulated in HCC tissues and was associated with high tumor node metastasis (TNM) stages, advanced histological grade, and poor patient survival. Then, by gain- and loss-of-function investigations, we demonstrated that TRIM11 promoted cell proliferation, migration, and invasion in vitro and tumor growth in vivo. Mechanistically, RNA sequencing and mass spectrometry analysis showed that TRIM11 interacted with pleckstrin homology domain leucine-rich repeats protein phosphatase 1 (PHLPP1) and promoted K48-linked ubiquitination degradation of PHLPP1 and thus promoted activation of the protein kinase B (AKT) signaling pathway. Moreover, overexpression of PHLPP1 blocked the promotional effect of TRIM11 on HCC function. CONCLUSIONS: Our study confirmed that TRIM11 plays an oncogenic role in HCC through the PHLPP1/AKT signaling pathway, suggesting that targeting TRIM11 may be a promising target for the treatment of HCC.

Deep unregistered multi-contrast MRI reconstruction.

Multiple magnetic resonance images of different contrasts are normally acquired for clinical diagnosis. Recently, research has shown that the previously acquired multi-contrast (MC) images of the same patient can be used as anatomical prior to accelerating magnetic resonance imaging (MRI). However, current MC-MRI networks are based on the assumption that the images are perfectly registered, which is rarely the case in real-world applications. In this paper, we propose an end-to-end deep neural network to reconstruct highly accelerated images by exploiting the shareable information from potentially misaligned reference images of an arbitrary contrast. Specifically, a spatial transformation (ST) module is designed and integrated into the reconstruction network to align the pre-acquired reference images with the images to be reconstructed. The misalignment is further alleviated by maximizing the normalized cross-correlation (NCC) between the MC images. The visualization of feature maps demonstrates that the proposed method effectively reduces the misalignment between the images for shareable information extraction when applied to the publicly available brain datasets. Additionally, the experimental results on these datasets show the proposed network allows the robust exploitation of shareable information across the misaligned MC images, leading to improved reconstruction results.

A volumetric finite-difference method for the design of three-dimensional, arbitrary-structured MRI gradient coil.

A volumetric finite-difference based method is presented in this paper for the design of three-dimensional (3D), arbitrarily structured gradient coils in magnetic resonance imaging. In the proposed method, the coil space is discretized with quasi-rectangular elements, and the current density of each element is expressed by a finite-difference based numerical approximation of stream functions. The magnetic flux density at target field points can be calculated by those stream function values at all grids of the coil space. The optimization problem is constructed and solved to determine the stream function and coil patterns. This proposed method has been tested on several designs that include a shielded, ultra-short cylindrical coil, a partially shielded biplanar coil, and an asymmetric head coil with 3D geometries. The numerical results show that the proposed method is straightforward to implement and is versatile and suitable for designing complex structured gradient coils with high electromagnetic performance.

ReUINet: A fast GNL distortion correction approach on a 1.0 T MRI-Linac scanner.

PURPOSE: The hybrid system combining a magnetic resonance imaging (MRI) scanner with a linear accelerator (Linac) has become increasingly desirable for tumor treatment because of excellent soft tissue contrast and nonionizing radiation. However, image distortions caused by gradient nonlinearity (GNL) can have detrimental impacts on real-time radiotherapy using MRI-Linac systems, where accurate geometric information of tumors is essential. METHODS: In this work, we proposed a deep convolutional neural network-based method to efficiently recover undistorted images (ReUINet) for real-time image guidance. The ReUINet, based on the encoder-decoder structure, was created to learn the relationship between the undistorted images and distorted images. The ReUINet was pretrained and tested on a publically available brain MR image dataset acquired from 23 volunteers. Then, transfer learning was adopted to implement the pretrained model (i.e., network with optimal weights) on the experimental three-dimensional (3D) grid phantom and in-vivo pelvis image datasets acquired from the 1.0 T Australian MRI-Linac system. RESULTS: Evaluations on the phantom (768 slices) and pelvis data (88 slices) showed that the ReUINet achieved improvement over 15 times and 45 times on computational efficiency in comparison with standard interpolation and GNL-encoding methods, respectively. Moreover, qualitative and quantitative results demonstrated that the ReUINet provided better correction results than the standard interpolation method, and comparable performance compared to the GNL-encoding approach. CONCLUSIONS: Validated by simulation and experimental results, the proposed ReUINet showed promise in obtaining accurate MR images for the implementation of real-time MRI-guided radiotherapy.

Design of an insertable cone-shaped gradient coil matrix for head imaging with a volumetric finite-difference method.

Matrix gradient coils have received increasing interest in generating arbitrary-shaped magnetic fields for various magnetic resonance imaging applications. In this paper, a novel cone-shaped matrix gradient coil is proposed to design a multifunctional insertable system for head imaging. Using a volumetric finite-difference-based method, the matrix coil is designed to have comprised several coil elements that can implement localized imaging and control eddy current, dissipated power, and minimum wire gap. With the lowest total dissipated power, various current configurations are selected to generate multiple gradient fields within a large, spheroidal region of interest (ROI) and two small spherical sub-ROIs. The numerical computation results show that the designed matrix coil offers high flexibility in generating a local gradient field capable of improving the local resolution. In addition, with enhanced coil performance, the cone-shaped structure provides a patient-friendly solution for head imaging.

Comparison of catalytic effect on upgrading bio-oil derived from co-pyrolysis of water hyacinth and scrap tire over multilamellar MFI nanosheets and HZSM-5.

Catalytic co-pyrolysis of water hyacinth and scrap tire experiments were performed to evaluate the feasibility of improving the monocyclic aromatic hydrocarbons production. The production of monocyclic aromatic hydrocarbons increased from 5.31% (sole pyrolysis of water hyacinth) to 13.11% (co-pyrolysis with scrap tire). With use of zeolites, the highest production of monocyclic aromatic hydrocarbons can reach up to 69.18%. Comprehensive comparison on catalytic effects of HZSM-5 and multilamellar MFI nanosheets were provided. With the material to multilamellar MFI nanosheets ratios changes from 2:1 to 1:4, the production of monocyclic aromatic hydrocarbons increases significantly from 37.15-69.18%. The average production of monocyclic aromatic hydrocarbons produced by using multilamellar MFI nanosheets were 12.07% higher than that using HZSM-5, indicating the better performance of multilamellar MFI nanosheets in producing monocyclic aromatic hydrocarbons. This work provided a reference for the reuse of water hyacinth and scrap tire over multilamellar MFI nanosheets in energy field.

Geometric distortion characterization and correction for the 1.0 T Australian MRI-linac system using an inverse electromagnetic method.

PURPOSE: The magnetic resonance imaging (MRI)-Linac system combines a MRI scanner and a linear accelerator (Linac) to realize real-time localization and adaptive radiotherapy for tumors. Given that the Australian MRI-Linac system has a 30-cm diameter of spherical volume (DSV) with a shimmed homogeneity of ±4.05 parts per million (ppm), a gradient nonlinearity (GNL) of <5% can only be assured within 15 cm from the system's isocenter. GNL increases from the isocenter and escalates close to and outside of the edge of the DSV. Gradient nonlinearity can cause large geometric distortions, which may provide inaccurate tumor localization and potentially degrade the radiotherapy treatment. In this study, we aimed to characterize and correct the geometric distortions both inside and outside of the DSV. METHODS: On the basis of phantom measurements, an inverse electromagnetic (EM) method was developed to reconstitute the virtual current density distribution that could generate gradient fields. The obtained virtual EM source was capable of characterizing the GNL field both inside and outside of the DSV. With the use of this GNL field information, our recently developed "GNL-encoding" reconstruction method was applied to correct the distortions implemented in the k-space domain. RESULTS: Both phantom and in vivo human images were used to validate the proposed method. The results showed that the maximal displacements within an imaging volume of 30 cm × 30 cm × 30 cm after using the fifth-order spherical harmonic (SH) method and the proposed method were 6.1 ± 0.6 mm and 1.8 ± 0.6 mm, respectively. Compared with the fifth-order SH-based method, the new solution decreased the percentage of markers (within an imaging volume of 30 cm × 30 cm × 30 cm) with ≥1.5-mm distortions from 6.3% to 1.3%, indicating substantially improved geometric accuracy. CONCLUSIONS: The experimental results indicated that the proposed method could provide substantially improved geometric accuracy for the region outside of the DSV, when comparing with the fifth-order SH-based method.

Catalytic co-pyrolysis behaviors, product characteristics and kinetics of rural solid waste and chlorella vulgaris.

Catalytic and non-catalytic co-pyrolysis behaviors, kinetics and products distribution of rural solid waste (RSW) and chlorella vulgaris (CV) were studied by thermogravimetric analysis (TGA) and fixed bed reactor. TGA results showed that co-pyrolysis of RSW and CV presented synergism by decreasing the temperature relating to the first mass loss peak. All the additives reduced residual mass for co-pyrolysis (5.21%, 1.57% and 4.89% for CaO, MgO and HZSM-5). Addition of CaO increased activation energy while HZSM-5 and MgO reduced it. Co-pyrolysis of RSW and CV remarkably reduced carboxylic acids and nitrogenous compounds especially for 1:1 ratio. (30.85% and 25.87%). Catalytic pyrolysis with CaO showed the best results by increasing aliphatic hydrocarbons especially light fraction (5.96%-11.98%), reducing acids (0%-30.85%) and nitrogenous compounds (0.08%-17.26%), causing higher HHV of oil. Overall, catalytic co-pyrolysis of CV and RSW with CaO could obtain bio-oil of higher quality.

A cone-shaped gradient coil design for high-resolution MRI head imaging.

Insertable head gradient coils offer significant advantages such as high gradient strength and fast gradient switching speed owing to shorter distances to the target region of interest than whole-body cylindrical coils. To produce superior gradient performance, the local head coil is typically designed with an asymmetric configuration to accommodate both the shoulders and head of a patient, leading to tough dimensional constraints and practical limits to the coil implementation. In this paper, we propose a new cone-shaped model to improve the performance of the asymmetric head coils and to mitigate patient claustrophobia. The primary coils are designed with a larger diameter at the patient end for access and a smaller diameter at the service end to bring wires closer to the human head, while the secondary coils are arranged on a cylindrical former to improve coil efficiency. Two cases are studied in this paper. Case I: inner bore size at the patient end (diameter 42 cm) is fixed as the design reference. In this case, inner diameters at any other position vary with the conical tilting angles. Compared with a set of conical gradient coils designed with tilting angles ranging from 0 to 14°, it is found that the optimal coil performance is achieved at the tilting angle of 14°. The key performance parameters have been improved by 100%-200% for the transverse coils, and about 50% for the longitudinal coils compared with the cylindrical counterpart with the reference bore size (that is, the same diameter of 42 cm). The conical coils also produce less heat in the gradient structure and lower acoustic noise in the field of view. Case II: inner bore size at the iso-centre (diameter 34 cm) is set as the design reference. It is also found that, compared with 34 cm diameter cylindrical coils, the conical transverse coil performance has been improved at an angle of 14°. The key coil performance increases by 20%-50% for transverse coil but decreases by 20%-40% for the longitudinal coil. However, compared with the tight cylindrical structure (e.g. 34 cm diameter), the tilting angle will provide patient-friendly space for imaging and handling, which can be critical for fMRI and other brain studies.

Gradient Field Deviation (GFD) Correction Using a Hybrid-Norm Approach With Wavelet Sub-Band Dependent Regularization: Implementation for Radial MRI at 9.4 T.

In magnetic resonance imaging (MRI), system imperfections and eddy currents can cause gradient field deviation (GFD), leading to various image distortions, such as increased noise, ghosting artifacts, and geometric deformation. These distortions can degrade the clinical value of MR images. Generally, non-Cartesian image sequences, such as radial sampling, produce larger gradient deviations than Cartesian sampling, as a result of stronger eddy current-induced gradient delays and phase errors. In this paper, we developed a GFD encoding method to reduce image noise and artifacts for radial MRI. In the proposed method, a hybrid norm (combination of L2 and L1 norms) optimization problem was formed, which incorporated a wavelet sub-band adaptive regularization mechanism. The new approach seeks a regularized solution not only offering corrected images with reduced artifacts and geometric deformation, but also good preservation of anatomical structural details. The new method was evaluated with simulation and experiment at 9.4 T MRI. The results demonstrated that the proposed method can provide over 50% noise reduction and 15% artifact reduction compared with the traditional regridding method, suggesting substantially reduced GFD-induced distortions and improved image quality.