GSAL: Geometric structure adversarial learning for robust medical image segmentation

• A new learning mechanism for medical image segmentation. We introduce a novel Geometric Structure Learning Mechanism (GSLM) that enhances model learning "focus, path, and difficulty". It enables geometric structure attention learning to bridge image features with large differences, thus capturing the contextual dependencies of images. The image features maintain consistency and continuity along the internal and external geometry structure, which improves the integrity and boundary accuracy of the segmentation results. To the best of our knowledge, we are the first attempt to explicitly establish the target's geometric structure, which has been successfully applied to medical image segmentation. • A novel geometric structure adversarial learning for robust medical image segmentation. We present the geometric structure adversarial learning model (GSAL) that consists of a geometric structure generator, skeleton-like and boundary discriminators, and a geometric structure fusion sub-network. The generator yields the geometric structure that preserves interior characteristics consistency and external boundary structure continuity. The dual discriminators are trained simultaneously to enhance and correct the characterization of interior structure and boundary structure, respectively. The fusion sub-network aims to fuse the geometric structure that optimized by adversarial learning to refine the final segmentation results with higher credibility. • State-of-art results on widely-used benchmarks. Our GSAL achieves SOTA performance on a variety of benchmarks, including Kvasir&CVC-612 dataset, COVID-19 dataset, and LIDC-IDRI dataset. It confirms the robustness and generalizability of our framework. In addition, our method has great advantages in terms of the integrity and boundary accuracy of the segmentation target compared to other competitive methods. GSAL can also achieve a considerable trade-off in terms of accuracy, inference speed, and model complexity, which helps deploy in clinical practice systems. Automatic medical image segmentation plays a crucial role in clinical diagnosis and treatment. However, it is still a challenging task due to the complex interior characteristics (e.g. , inconsistent intensity, low contrast, texture heterogeneity) and ambiguous external boundary structures. In this paper, we introduce a novel geometric structure learning mechanism (GSLM) to overcome the limitations of existing segmentation models that lack learning "focus, path, and difficulty." The geometric structure in this mechanism is jointly characterized by the skeleton-like structure extracted by the mask distance transform (MDT) and the boundary structure extracted by the mask distance inverse transform (MDIT). Among them, the skeleton-like and boundary pay attention to the trend of interior characteristics consistency and external structure continuity, respectively. With this idea, we design GSAL, a novel end-to-end geometric structure adversarial learning for robust medical image segmentation. GSAL has four components: a geometric structure generator, which yields the geometric structure to learn the most discriminative features that preserve interior characteristics consistency and external boundary structure continuity, skeleton-like and boundary structure discriminators, which enhance and correct the characterization of internal and external geometry to mutually promote the capture of global contextual dependencies, and a geometric structure fusion sub-network, which fuses the two complementary and refined skeleton-like and boundary structures to generate the high-quality segmentation results. The proposed approach has been successfully applied to three different challenging medical image segmentation tasks, including polyp segmentation, COVID-19 lung infection segmentation, and lung nodule segmentation. Extensive experimental results demonstrate that the proposed GSAL achieves favorably against most state-of-the-art methods under different evaluation metrics. The code is available at: https://github.com/DLWK/GSAL. [ BSTRACT FROM AUTHOR] Copyright of Pattern Recognition is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full . (Copyright applies to all s.)

Immune persistence induced by two-dose BBIBP-CorV vaccine with different intervals, and immunogenicity and safety of a homologous booster dose in high-risk occupational population.

BACKGROUND: The immune persistence of neutralizing antibodies elicited by BBIBP-CorV vaccines on day 0-14, 0-21 and 0-28 schedule, and the immunogenicity and safety of a homologous booster dose after different priming vaccination regimens is scarcely reported. METHODS: : Responders (GMT≥16) at day 28, after priming with the two-dose vaccine, were followed up at 3, 6, and 10 months. Eligible participants received a homologous booster dose at month 10 and were followed-up 28 days post-booster. RESULTS: The GMT of neutralizing antibodies in 0-28d-10 m and 0-21d-10 m group were significantly higher than 0-14d-10 m group from month 3 (71.6 & 64.2 vs 46.4, p < 0.001) to month 10 (32.4 & 28.8 vs 20.3, p < 0.001) after the second dose. On day 28 post-booster, a remarkable rebound in neutralizing antibodies (246.2, 277.5, and 288.6, respectively) was observed in the three groups. All adverse reactions were mild after booster injection. CONCLUSIONS: The priming two-dose BBIBP-CorV vaccine with 0-28 days and 0-21 days schedule could lead to a longer persistence of neutralizing antibody than the 0-14 days schedule. Regardless of the priming vaccination regimens, a homologous booster dose led to a strong rebound in neutralizing antibodies and might persist for at least 18 months.

An outbreak of acute respiratory disease caused by HAdV-55 in Beijing, China, 2020.

Human adenoviruses (HAdVs) can cause acute respiratory diseases (ARDs) worldwide, and HAdV-55 is a reemergent pathogen in recent years. In the study, we investigated an outbreak of ARD at a school due to HAdV-55 in Beijing, China, during the early outbreak of coronavirus disease 2019 (COVID-19). The epidemic prevention team was dispatched to the school to collect epidemiologic data and nasopharyngeal samples. Then, real-time reverse transcription polymerase chain reaction (PCR) and multiplex PCR assays were used to detect severe acute respiratory syndrome coronavirus 2 and other respiratory pathogens, respectively. One representative HAdV-55 isolate was selected and submitted for whole-genome sequencing using a MiSeq system and the whole-genome phylogenetic tree was conducted based on the maximum likelihood method. The outbreak lasted from January 27 to February 6, 2020, and 108 students developed fever, among whom 60 (55.56%) cases were diagnosed with HAdV-55 infection in the laboratory using real-time PCR and 56 cases were hospitalized. All the confirmed cases had a fever and 11 cases (18.33%) presented with a fever above 39°C. Other main clinical symptoms included sore throat (43.33%) and headache (43.33%). We obtained and assembled the full genome of one isolate, BJ-446, with 34 761 nucleotides in length. HAdV-55 isolate BJ-446 was 99.85% identical to strain QS-DLL, which was the first HAdV-55 strain in China isolated from an ARD outbreak in Shanxi in 2006. One and four amino acid mutations were observed in the hexon gene and the coding region of L2 pV 40.1 kDa protein, respectively. We identified the first HAdV-55 infection associated with the ARD outbreak in Beijing since the emergence of COVID-19. The study suggests that improved surveillance of HAdV is needed, although COVID-19 is still prevalent in the world.

EANet: Iterative Edge Attention Network for Medical Image Segmentation

Accurate and automatic segmentation of medical images can greatly assist the clinical diagnosis and analysis. However, it remains a challenging task due to (1) the diversity of scale in the medical image targets and (2) the complex context environments of medical images, including ambiguity of structural boundaries, complexity of shapes, and the heterogeneity of textures. To comprehensively tackle these challenges, we propose a novel and effective iterative edge attention network (EANet) for medical image segmentation with steps as follows. First, we propose a dynamic scale-aware context (DSC) module, which dynamically adjusts the receptive fields to extract multi-scale contextual information efficiently. Second, an edge-attention preservation (EAP) module is employed to effectively remove noise and help the edge stream focus on processing only the boundary-related information. Finally, a multi-level pairwise regression (MPR) module is designed to combine the complementary edge and region information for refining the ambiguous structure. This iterative optimization helps to learn better representations and more accurate saliency maps. Extensive experimental results demonstrate that the proposed network achieves superior segmentation performance to state-of-the-art methods in four different challenging medical segmentation tasks, including lung nodule segmentation, COVID-19 infection segmentation, lung segmentation, and thyroid nodule segmentation. The source code of our method is available at https://github.com/DLWK/EANet

CGRNet: Contour-guided graph reasoning network for ambiguous biomedical image segmentation

In this work, we propose to address the existing problem of biomedical image segmentation that often produces results, which fail to capture the exact contours of the target and suffer from ambiguity. Most previous techniques are suboptimal because they often simply concatenate contour information to alleviate this problem, while ignoring the correlation between regions and contours. As a matter of fact, the relationship between cross-domain features is an important clue for ambiguous pixel segmentation in biomedical images. To this end, we contribute a simple yet effective framework called Contour-Guided Graph Reasoning Network (CGRNet) for more accurate segmentation against ambiguity, which is capable of capturing the semantic relations between object regions and contours through graph reasoning. Specifically, we first perform a global graph representation of the low-level and high-level features extracted by the feature extractor, where clusters of pixels with similar features are mapped to each vertex. Further, we explicitly combine contour information as the geometric prior, which can aggregate features of contour pixels to graph vertices and focus on features along the boundaries. Then, the cross-domain features propagate information through the vertices on the graph to efficiently learn and reason about the semantic relations. Finally, the learned refinement graph features are projected back to the original pixel coordinate space for the final pixel-wise segmentation task. Extensive experiments on the three publicly available Kvasir, CVC-612, and COVID19-100 datasets show the effectiveness of our CGRNet with superior performance to existing state-of-the-art methods. Our code is publicly available at: https://github.com/DLWK/CGRNet.

Safety and immunogenicity of inactivated SARS-CoV-2 vaccine in high-risk occupational population: a randomized, parallel, controlled clinical trial.

BACKGROUND: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and the resulting coronavirus disease 2019 (COVID-19) have a substantial burden on health-care systems around the world. This is a randomized parallel controlled trial for assessment of the immunogenicity and safety of an inactivated SARS-CoV-2 vaccine, aiming to determine an appropriate vaccination interval of the vaccine for high-risk occupational population. METHODS: In an ongoing randomized, parallel, controlled phase IV trial between January and May 2021 in Taiyuan City, Shanxi Province, China, we randomly assigned the airport ground staff and public security officers aged 18 to 59 years to receive two doses of inactivated SARS-CoV-2 vaccine at 14 days, 21 days, or 28 days. The serum neutralizing antibody to live SARS-CoV-2 was performed at baseline and 28 days after immunization. Long-term data are being collected. The primary immunogenicity endpoints were neutralization antibody seroconversion and geometric mean titer (GMT) at 28 days after the second dose. Analysis of variance (ANOVA), chi-square, and logistic regression analysis were used for data analysis. RESULTS: A total of 809 participants underwent randomization and received two doses of injections: 270, 270, 269 in the 0-14, 0-21, and 0-28 vaccination group, respectively. By day 28 after the second injection, SARS-CoV-2 neutralizing antibody of GMT was 98.4 (95% CI: 88.4-108.4) in the 0-14 group, which was significantly lower compared with 134.4 (95% CI: 123.1-145.7) in the 0-21 group (P < 0.001 vs 0-14 group) and 145.5 (95% CI: 131.3-159.6) in the 0-28 group (P < 0.001 vs 0-14 group), resulting in the seroconversion rates to neutralizing antibodies (GMT ≥ 16) of 100.0% for all three groups, respectively. The intention-to-treat (ITT) analysis yielded similar results. All reported adverse reactions were mild. CONCLUSIONS: Both a two-dose of inactivated SARS-CoV-2 vaccine at 0-21 days and 0-28 days regimens significantly improved SARS-CoV-2 neutralizing antibody level compared to the 0-14 days regimen in high-risk occupational population, with seroconversion rates of 100.0%. TRIAL REGISTRATION: Chinese Clinical Trial Registry, ChiCTR2100041705, ChiCTR2100041706. Registered 1 January 2021, www.chictr.org.cn .

Cationic Compounds with SARS-CoV-2 Antiviral Activity and Their Interaction with Organic Cation Transporter/Multidrug and Toxin Extruder Secretory Transporters.

In the wake of the COVID-19 pandemic, drug repurposing has been highlighted for rapid introduction of therapeutics. Proposed drugs with activity against SARS-CoV-2 include compounds with positive charges at physiologic pH, making them potential targets for the organic cation secretory transporters of kidney and liver, i.e., the basolateral organic cation transporters, OCT1 and OCT2; and the apical multidrug and toxin extruders, MATE1 and MATE2-K. We selected several compounds proposed to have in vitro activity against SARS-CoV-2 (chloroquine, hydroxychloroquine, quinacrine, tilorone, pyronaridine, cetylpyridinium, and miramistin) to test their interaction with OCT and MATE transporters. We used Bayesian machine learning models to generate predictions for each molecule with each transporter and also experimentally determined IC50 values for each compound against labeled substrate transport into CHO cells that stably expressed OCT2, MATE1, or MATE2-K using three structurally distinct substrates (atenolol, metformin and 1-methyl-4-phenylpyridinium) to assess the impact of substrate structure on inhibitory efficacy. For the OCTs substrate identity influenced IC50 values, although the effect was larger and more systematic for OCT2. In contrast, inhibition of MATE1-mediated transport was largely insensitive to substrate identity. Unlike MATE1, inhibition of MATE2-K was influenced, albeit modestly, by substrate identity. Maximum unbound plasma concentration/IC50 ratios were used to identify potential clinical DDI recommendations; all the compounds interacted with the OCT/MATE secretory pathway, most with sufficient avidity to represent potential DDI issues for secretion of cationic drugs. This should be considered when proposing cationic agents as repurposed antivirals. SIGNIFICANCE STATEMENT: Drugs proposed as potential COVID-19 therapeutics based on in vitro activity data against SARS-CoV-2 include compounds with positive charges at physiological pH, making them potential interactors with the OCT/MATE renal secretory pathway. We tested seven such molecules as inhibitors of OCT1/2 and MATE1/2-K. All the compounds blocked transport activity regardless of substrate used to monitor activity. Suggesting that plasma concentrations achieved by normal clinical application of the test agents could be expected to influence the pharmacokinetics of selected cationic drugs.

Synthesis of a novel anti-fog and high-transparent coating with high wear resistance inspired by dry rice fields.

During the outbreak of COVID-19, the fogging of goggles was a fatal problem for doctors. At present, there are many ways to prevent fogging by adjusting surface wettability. However, the mechanical properties of most super-hydrophilic antifogging coatings are poor, easy to lose their antifogging properties when encountering fingers or cloth friction. To address this issue, the Konjac Glucomannan was cross-linked with water-soluble silicone fluid to form a binder, then being combined with the modified Ecokimera to prepare an eco-friendly super-hydrophilic coating that possessed excellent super-hydrophilicity, and the water contact angle (WCA) was 2.51 ± 1°. In addition, the WCA is still about 5° after 180 times of antifogging tests. The friction resistance of the coating was as high as 24 m. Moreover, the light transmittance was only reduced by 3%. Besides, they also had the excellent self-cleaning property. After being stored in the laboratory environment for 90 days, it can still maintain the hydrophilic property (WCA is about 5°). In general, the method proposed in this study is low-cost and eco-friendly, and can be widely used in the preparation of antifogging coatings.

Clinical characteristics and laboratory features of COVID-19 in high altitude areas: A retrospective cohort study.

Coronavirus disease 2019 (COVID-19) is highly contagious and has affected the whole world. We seek to investigate the clinical and laboratory characteristics of COVID-19 patients in the high altitude areas of Sichuan, China. In this retrospective cohort study, a total of 67 patients with laboratory-confirmed SARS-CoV-2 infections in Sichuan's Ngawa Tibetan and Qiang Autonomous Prefecture were included from February 1, 2020, to March 2, 2020. Their clinical characteristics, as well as radiological and laboratory features, were extracted. Four (6.0%) patients were categorized as severe cases; 39 (58.2%) were non-severe cases, and 24 (35.8%) were asymptomatic cases. A total of 46 (68.7%) patients were associated with cluster infection events in this study. The most common symptoms were cough, sputum production, dyspnea, fatigue or myalgia, and headache. Seven (10.4%) patients showed leucopenia, and 20 (29.9%) patients showed lymphopenia. Lymphocyte counts and neutrophil-to-lymphocyte ratios (NPR) were different between the three groups. In total, 14 (20.9%) patients had thrombocytopenia, and prothrombin times (PT) and fibrinogen levels differed between groups. We also found significant differences in sodium, chloride and calcium levels between the three groups. Antiviral therapy did not lead to obvious adverse events or shortened durations from initial positive to subsequent negative nuclei acid tests. Advanced age, hypertension, high neutrophil count, the neutrophil-to-lymphocyte ratio, fibrinogen and lactate dehydrogenase levels were identified as independent risk factors for symptomatic cases of COVID-19. In conclusion, the symptoms of patients in high altitude areas were mild, and about one third were asymptomatic. We also identified several independent risk factors for symptomatic cases of COVID-19.

Epidemiological and clinical based study on four passages of COVID-19 patients: intervention at asymptomatic period contributes to early recovery.

BACKGROUND: With the worldwide spread of the 2019 novel coronavirus, scarce knowledge is available on the clinical features of more than two passages of patients. Further, in China, early intervention policy has been enacted since February. Whether early intervention contributes to swift recovery is still unknown. Hence, in this study, we focused on the patients from an isolated area, investigated the epidemiological and clinical characteristics of four serial passages of the virus. METHODS: From January 25 to February 29, 2020, all patient data on the SARS-CoV-2 passages in this isolated area were traced, and the patients were grouped according to the passaging of SARS-CoV-2. Clinical characteristics of patients, including laboratory, radiology, treatment and outcomes, were collected and analyzed. RESULTS: A total of 78 patients with four passages of virus transmission were included in this study. One patient transmitted SARS-CoV-2 to 8 patients (passage 2, P2), who next infected 23 patients (passage 3, P3), and then 46 patients (passage 4, P4). P2 received antiviral treatment when they had symptom, whereas P4 received antiviral treatment during their asymptomatic period. The incubation periods for P2, P3 and P4 patients were 7 days (IQR:2-12), 8 days (IQR:4-13) and 10 days (IQR:7-15), respectively. P2 patients showed lymphocytopenia (0.79 × 109/L), decreased lymphocyte percentages (12.15%), increased white blood cell count (6.51 × 109/L), increased total bilirubin levels (25% of P2 patients), increased C-reactive protein levels (100% of P2 patients) and abnormal liver function. By chest CT scans, all P2 patients (100%), 15 of P3 patients (65.22%) and 16 of P4 patients (34.78%) showed abnormality with typical feature of ground glass opacity. All of P2 patients (100%) received oxygen therapy, and in contrast, 19 of P4 patients (41.3%) received oxygen therapy. Further, significant decreased nucleic acid positive periods was found in P4 group (16 days, IQR: 10-23), compared with that of P2 group (22 days, IQR: 16-27). Moreover, the severity ratios were sharply decreased from 50% (P2 patients) to 4.35% (P4 patients), and the case fatality rate is zero. CONCLUSIONS: Judged from four passages of patients, early intervention contributes to the early recovery of COVID-19 patients.

Isolation and genome sequence analyses of 2019-nCoV from COVID-19 patients in Fujian Province, China

To explore genomic characteriation of 2019-nCoV, throat swab specimens from COVID-19 patients were inoculated on Vero-E6 cells. Six days post-inoculation, supernatants of cell cultures were collected, tested, and sequenced on the Ion Torrent S5 Next-Generation Sequencing system. Two strains viruses were isolated from 12 specimens (eight of them were positive in ORF1ab gene and N gene, one of them was positive in N gene and three of them were negative in ORF1ab gene and N gene), as confirmed by online BLAST search at NCBI website. Genetic similarity of more than 99.9% were observed, from genomic level to structure genes including E, M, N and S genes, between the 2 isolates and the 2019-nCoV reference strain Wuhan-Hu-1 that isolated from Wuhan, Hubei province. Therefore, these results indicated that 2019-nCoV isolates in Fujian province have not mutated significantly yet so far.

A Controllable Inflammatory Response and Temporary Abnormal Coagulation in Moderate Disease of COVID-19 in Wuhan, China.

BACKGROUND: The coronavirus disease 2019 (COVID-19) is now a worldwide challenge for public health. Among 7 million patients, about 80% present mild to moderate disease, but studies dedicate to these patients are actually scarce. The aim of our study is to clarify the characteristics of laboratory test index of COVID-19 patient with moderate symptoms during the first wave of the pandemic in Wuhan, China. METHODS: In this retrospective cohort study, we included 107 adult inpatients with confirmed moderate disease of COVID-19 from the Affiliated Hospital of Jianghan University during February and early March 2020. All of these patients were recovered from COVID-19 and discharged from hospital. Demographic, clinical, and laboratory data of admission and discharge were extracted from electronic medical records and analyzed using SPSS, as well as among young, middle age and elderly people. RESULTS: The median age of this cohort of patients was 56.0 years. And the median hospitalization time was 16 days. Common clinical manifestations included fever, cough, asthenia and shortness of breath. On admission, laboratory results showed normal or increased neutrophil ratio, low lymphocyte count, decreased hemoglobin level, and increased inflammatory indicators (erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP)); and some patients were complicated with coagulation disorder and myocardial damage. Furthermore, patients older than 60 years had statistically higher CRP, ESR and fibrinogen level. As the health condition was improved at discharge, the median level of most laboratory results were in the normal range except hemoglobin and related blood cell count, as well as inflammatory indicator ESR. And patients older than 60 years showed slower recovery on coagulation parameters when compared to younger patients. CONCLUSIONS: The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection induces a controllable inflammatory response in moderate disease of COVID-19 in Wuhan, China. Since patients older than 60 years had higher inflammatory state and more dysregulated coagulation condition, it might be essential to closely assess their illness.

Clinically Applicable AI System for Accurate Diagnosis, Quantitative Measurements and Prognosis of COVID-19 Pneumonia Using Computed Tomography

Many COVID-19 patients infected by SARS-CoV-2 virus develop pneumonia (called novel coronavirus pneumonia, NCP) and rapidly progress to respiratory failure. However, rapid diagnosis and identification of high-risk patients for early intervention are challenging. Using a large computed Tomography (CT) database from 4,154 patients, we developed an AI system that can diagnose NCP and differentiate it from other common pneumonia and normal controls. The AI system can assist radiologists and physicians in performing a quick diagnosis especially when the health system is overloaded. Significantly, our AI system identified important clinical markers that correlated with the NCP lesion properties. Together with the clinical data, our AI system was able to provide accurate clinical prognosis that can aid clinicians to consider appropriate early clinical management and allocate resources appropriately. We have made this AI system available globally to assist the clinicians to combat COVID-19.