A multi-output network with U-net enhanced class activation map and robust classification performance for medical imaging analysis

Computer vision in medical diagnosis has achieved a high level of success in diagnosing diseases with high accuracy. However, conventional classifiers that produce an image-to-label result provide insufficient information for medical professionals to judge and raise concerns over the trust and reliability of a model with results that cannot be explained. To gain local insight of cancerous regions, separate tasks such as imaging segmentation needs to be implemented to aid the doctors in treating patients which doubles the training time and costs which renders the diagnosis system inefficient and difficult to be accepted by the public. To tackle this issue and drive the AI-first medical solutions further, this paper proposes a multi-output network which follows a U-Net architecture for image segmentation output and features an additional CNN module for auxiliary classification output. Class Activation Maps or CAMs are a method of providing insight into a convolutional neural network's feature maps that lead to its classification but in the case of lung diseases, the region of interest is enhanced by U-net assisted Class Activation Mapping (CAM) visualization. Therefore, our proposed model combines image segmentation models and classifiers to crop out only the lung region of a chest X-ray's class activation map to provide a visualization that improves the explainability and can generate classification results simultaneously which builds trust for AI-led diagnosis system. The proposed U-Net model achieves 97.72% accuracy and a dice coefficient of 0.9691 on a testing data from the COVID-QU-Ex Dataset which includes both diseased and healthy lungs.

Advances in the application of CRISPR-Cas system for SARS-CoV-2 diagnostic test

The outbreak of coronavirus disease 2019（COVID-19）, caused by severe acute respiratory syndrome coronavirus（SARS-CoV-2） at the end of 2019, results in a global rapid pandemic.The emerging infectious disease is life threaten and profoundly undermines the normal operation all over the world.Rapid and accurate detection of SARS-CoV-2 is an essential component of efforts to combat SARS-CoV-2 spread.Although many technologies for SARS-CoV-2 detection have been commercialized and have played a role in the control of COVID-19 epidemic to some extent yet, each of them is still with certain limitations.Recently, increasing number of teams attempt to detect SARS-CoV-2 by CRISPR-Cas（clustered regularly interspaced short palindromic repeats-Cas） system.With excellent specificity and sensitivity, it has been believed to be a potential technology in COVID-19 diagnosis and therapy.The review provides an overview of CRISPR-Cas system for SARS-CoV-2 detection and COVID-19 therapy, along with its clinical translation potential.Hope it has some referential significance for the control of SARS-CoV-2 spread and COVID-19 epidemic.

How the COVID-19 Pandemic Impacts Green Inventions: Evidence from a Quasi-Natural Experiment in China

This paper investigates whether the COVID-19 (coronavirus disease 2019) pandemic affects the green inventions of firms, universities, and firm–university collaborations (FUCs) differently. Our identification used provincial-level monthly data from China. Results from the difference-in-differences (DID) model showed that the COVID-19 pandemic has prompted the output of three types of green invention patents. After the parallel-trend test, placebo test, and triple-difference estimation, our conclusion has good robustness. However, the COVID-19 pandemic also influences the role of other policies, such as the SO2-emissions-trading pilot policy for universities' green inventions. There has been a slight change in the effect of dual carbon targets on green inventions since the start of the pandemic. The positive effect of the COVID-19 pandemic has been weaker for provinces where the pandemic has been more severe than in other provinces. The results of this study are compared with the results and empirical evidence of other related studies and the theoretical logic of COVID-19 crisis-promoted green inventions are discussed.

Dexamethasone Ameliorates Severe Pneumonia But Slightly Enhances Viral Replication in Lung of SARS-CoV-2-Infected Syrian Hamster (preprint)

Background: The pandemic of SARS-CoV-2 has turned into a global public health crisis. Acute SARS-CoV-2 infection is associated with severe pneumonia, multiple-organ failures and deaths. Currently, treatment for SARS-CoV-2 infection and severe pneumonia is largely lacking. Several clinical trials demonstrated that glucocorticoid dexamethasone is effective to reduce disease severity and mortality. However, whether dexamethasone is clinically sufficient to treat COVID-19 is unknown.Methods: We tested the therapeutic effect of dexamethasone on SARS-CoV-2 infection and pneumonia in a Syrian hamster model. Survival rate, body weight loss, viral RNA, antibody responses, severity of lung inflammation and injury were measured in a 7-day acute infection course.Findings: Dexamethasone reduces body weight loss and relieves the diffusion of lung injury in SARS-CoV-2-infected hamster by inhibiting the excessive proinflammatory cytokines including IL-4, IL-6, IL-10, IL-13, TNF-α and IFN-γ. Dexamethasone rescues hamsters from the lethal infection of SARS-CoV-2 variant D614G. Dexamethasone attenuates serum neutralizing antibody and RBD-specific antibody titers, and slightly increases viral RNA level in lung tissues.Interpretation: Overall, using the hamster model, this study improves our understanding of the therapeutic mechanisms and drawbacks of dexamethasone treatment of COVID-19, and suggests that an antiviral is needed to accompany the dexamethasone treatment regimen.Funding: National Science Key Research and Development Project of China, National Natural Science Foundation of China, the CAMS Innovation Fund for Medical Sciences and China Postdoctoral Science Foundation.Declaration of Interest: The authors declare no competing interests.Ethical Approval: All the animal experiments were approved by the Medical Ethics Committee(SUCM2021-112).

Adaptive Epidemic Forecasting and Community Risk Evaluation of COVID-19 (preprint)

Pandemic control measures like lock-down, restrictions on restaurants and gatherings, social-distancing have shown to be effective in curtailing the spread of COVID-19. However, their sustained enforcement has negative economic effects. To craft strategies and policies that reduce the hardship on the people and the economy while being effective against the pandemic, authorities need to understand the disease dynamics at the right geo-spatial granularity. Considering factors like the hospitals' ability to handle the fluctuating demands, evaluating various reopening scenarios, and accurate forecasting of cases are vital to decision making. Towards this end, we present a flexible end-to-end solution that seamlessly integrates public health data with tertiary client data to accurately estimate the risk of reopening a community. At its core lies a state-of-the-art prediction model that auto-captures changing trends in transmission and mobility. Benchmarking against various published baselines confirm the superiority of our forecasting algorithm. Combined with the ability to extend to multiple client-specific requirements and perform deductive reasoning through counter-factual analysis, this solution provides actionable insights to multiple client domains ranging from government to educational institutions, hospitals, and commercial establishments.

Difference in Risk Factors Associated with Acute Respiratory Distress Syndrome and Intubation/Death in Patients with Coronavirus Disease 2019: A Two-hospital Retrospective Cohort Study in Hubei, China (preprint)

Background: The reviews on the risk factors with ARDS and the worse outcomes concluded lacking robust data of risk factors to prevent COVID-19 and identified an urgent need for large sample and high-quality research in this area, as well as the features of the ARDS.Methods: This retrospective cohort study included 333 COVID-19 inpatients at two hospitals in Hubei of China in 2020. The COVID-19-related ARDS was diagnosed according to the Berlin criteria. The outcomes were ARDS development and the intubation or in-hospital death. The cox proportional hazard ratio (HR) models were employed to determine the significant risk factors. Results: The median number of days from symptom onset to ARDS diagnosis was 11.0 (IQR, 8.0–13.0). Up to 84.1% COVID-19-related ARDS patients demonstrated multiple organ injuries. The mortality rates were 41.9% and 85.7% in moderate and severe ARDS. The survival patients on invasive mechanical ventilation (IMV) had been intubated earlier since ARDS diagnosis than those who had not survived (5.5 median days, IQR 4.0-7.0 days versus 11.5 median days, IQR 6.0-14.0 days, P < 0.001). Males and all abnormal laboratory indices associated with the higher risk of ARDS (P<0.05) but were not linked with the risk of intubation or death (P>0.05). The sensitivity analyses found that lymphocyte count of < 1000 per mm3 at hospital admission were still significantly associated with developing ARDS when adjusting for age and male gender (HR, 4.10; 95% CI, 2.40-7.10), and oxygenation index (OI) ratio < 150 were more likely to predict the intubation/death after age adjustment (HR, 2.50; 95% CI, 1.17-5.30). Conclusion: The SARS-CoV-2-caused ARDS was not the typical ARDS according to Berlin criteria. The alive patients with IMV had been intubated earlier since ARDS diagnosis than those who had not survived. We identified male gender and abnormal laboratory indices associated with the ARDS but were not linked with the intubation/death. Sensitivity analysis concluded lymphocyte count of < 1000 per mm3 could predict ARDS while OI ratio less than 150 could predict intubation/death.

Clinical Characteristics and Outcomes of Patients with Coronavirus Disease 2019-associated ARDS and the Early Intubation: A Two-Hospital Retrospective Cohort Study in Hubei, China (preprint)

Background: More evidence in understanding the heterogeneity of COVID-19-associated acute respiratory distress syndrome (ARDS) and in improving strategy to increase the survival from the critical patients intubated is always needed. The study aimed to comprehensively explore the features of COVID-19-associated ARDS and the features and outcomes between the early and late intubation groups. Methods: This retrospective cohort included 65 adult COVID-19 inpatients with ARDS at two hospitals in Hubei, China. The ARDS in these patients was diagnosed according to the Berlin criteria. We defined intubation within 7 days of ARDS diagnosis as ‘early’ intubation and that performed from the eighth day as ‘late’ intubation based on literatures. The outcomes were invasive mechanical ventilation and in-hospital death. The log-binomial regression models were used to explore the risk factors and the Kaplan-Meier statistic was used to estimate the risk of mortality. Results: The median number of days from symptom onset to ARDS diagnosis was 11.0 (IQR, 8.0–13.0). Up to 84.1% COVID-19-related ARDS patients demonstrated multiple organ injuries. The mortality rates were 41.9% and 85.7% in moderate and severe ARDS. The early intubation and the late intubation had the differences in days from symptom onset/hospital admission/ARDS diagnosis to intubation (P = 0.023, P = 0.011, P < 0.001). Compared with the early-intubation group, the late-intubation group showed less severity at admission (median oxygenation index 159.0 95% CI 134.0-203.0 vs. 133.9 95% CI 98.3-183.2), but required more aggressive therapies (ICU 80% vs. 70%, CRRT 50% vs. 10%, prone-position 50% vs. 30%, and ECMO 50% vs. 10%) and had higher risk to die at hospital (RR, 3.18; 95% CI 1.98-5.12). Conclusion: The ARDS caused by COVID-19 was not typical ARDS due to prolonged onset time, multiple organ injuries, and higher mortalities. The late-intubation group showed less severity at admission but higher risk of in-hospital death than the early-intubation group. 

Blockchain for Decentralized Multi-Drone to Combat COVID-19 (preprint)

Currently, drones represent a promising technology for combating Coronavirus disease 2019 (COVID-19) due to the transport of goods, medical supplies to a given target location in the quarantine areas experiencing an epidemic outbreak. Drone missions will increasingly rely on drone collaboration, which requires the drones to reduce communication complexity and be controlled in a decentralized fashion. Blockchain technology becomes a must in industrial applications because it provides decentralized data, accessibility, immutability, and irreversibility. Therefore, Blockchain makes data public for all drones and enables drones to log information concerning world states, time, location, resources, delivery data, and drone relation to all neighbors drones. This paper introduces decentralized independent multi-drones to accomplish the task collaboratively. Improving blockchain with a consensus algorithm can improve network partitioning and scalability in order to combat COVID-19. The multi-drones task is to combat COVID-19 via monitoring and detecting, social distancing, sanitization, data analysis, delivering goods and medical supplies, and announcement while avoiding collisions with one another. We discuss End to End (E2E) delivery application of combination blockchain and multi-drone in combating COVID-19 and beyond future pandemics. Furthermore, the challenges and opportunities of our proposed framework are highlighted.

Globally Local: Hyper-local Modeling for Accurate Forecast of COVID-19 (preprint)

Multiple efforts to model the epidemiology of SARS-CoV-2 have recently been launched in support of public health response at the national, state, and county levels. While the pandemic is global, the dynamics of this infectious disease varies with geography, local policies, and local variations in demographics. An underlying assumption of most infectious disease compartment modeling is that of a well mixed population at the resolution of the areas being modeled. The implicit need to model at fine spatial resolution is impeded by the quality of ground truth data for fine scale administrative subdivisions. To understand the trade-offs and benefits of such modeling as a function of scale, we compare the predictive performance of a SARS-CoV-2 modeling at the county, county cluster, and state level for the entire United States. Our results demonstrate that accurate prediction at the county level requires hyper-local modeling with county resolution. State level modeling does not accurately predict community spread in smaller sub-regions because state populations are not well mixed, resulting in large prediction errors. As an important use case, leveraging high resolution modeling with public health data and admissions data from Hillsborough County Florida, we performed weekly forecasts of both hospital admission and ICU bed demand for the county. The repeated forecasts between March and August 2020 were used to develop accurate resource allocation plans for Tampa General Hospital. 2010 MSC92-D30, 91-C20

On Machine Learning-Based Short-Term Adjustment of Epidemiological Projections of COVID-19 in US (preprint)

Epidemiological models have provided valuable information for the outlook of COVID-19 pandemic and relative impact of different mitigation scenarios. However, more accurate forecasts are often needed at near term for planning and staffing. We present our early results from a systemic analysis of short-term adjustment of epidemiological modeling of COVID 19 pandemic in US during March-April 2020. Our analysis includes the importance of various types of features for short term adjustment of the predictions. In addition, we explore the potential of data augmentation to address the data limitation for an emerging pandemic. Following published literature, we employ data augmentation via clustering of regions and evaluate a number of clustering strategies to identify early patterns from the data. From our early analysis, we used CovidActNow as our underlying epidemiological model and found that the most impactful features for the one-day prediction horizon are population density, workers in commuting flow, number of deaths in the day prior to prediction date, and the autoregressive features of new COVID-19 cases from three previous dates of the prediction. Interestingly, we also found that counties clustered with New York County resulted in best preforming model with maximum of R2= 0.90 and minimum of R2= 0.85 for state-based and COVID-based clustering strategy, respectively.

Prevalence and Prognostic Value of Elevated Troponins in Patients Hospitalised for Coronavirus Disease 2019: A Systematic Review and Meta-Analysis (preprint)

Background: The clinical significance of cardiac troponin measurement in patients hospitalised for coronavirus disease-2019 (covid-19) is uncertain. We investigated the prevalence of elevated troponins in these patients and its prognostic value for predicting mortality. Methods: : Studies were identified by searching electronic databases and preprint servers. We included studies of hospitalised covid-19 patients that reported the frequency of troponin elevations above the upper reference limit and/or the association between troponins and mortality. Meta-analyses were performed using random-effects models. Results: : Forty-four studies were included. Elevated troponins were found in 21.3% (95% confidence interval [CI] 18.0-24.9 %) of patients who received troponin test on hospital admission. Elevated troponins on admission were associated with a higher risk of subsequent death (risk ratio 2.81, 95% CI 2.01-3.93) after adjusting for confounders in multivariable analysis. The pooled sensitivity of elevated admission troponins for predicting death was 0.64 (95% CI 0.58-0.70), and the specificity was 0.88 (0.82-0.92). The post-test probability of death was about 50% for patients with elevated admission troponins, and was about 7% for those with non-elevated troponins on admission. There were significant heterogeneity and publication bias in the analyses, and many included studies were at risk of selection bias due to the lack of systematic troponin measurement and inadequate follow-up. Conclusion: Elevated troponins were relatively common in patients hospitalised for covid-19. Troponin measurement on admission might help in risk stratification, especially in identifying patients at high risk of death when troponin levels are elevated. High-quality prospective studies are needed to validate these findings. Systematic Review Registration: PROSPERO (CRD42020176747).

Clinical characteristics of 106 patients with neurological diseases and co-morbid coronavirus disease 2019: a retrospective study (preprint)

Objectives:To describe the clinical characteristics of patients with coronavirus disease 2019 (COVID-19) with co-morbid neurological symptoms. Design:Retrospective case series. Setting:Huoshenshan Hospital in Wuhan, China. Participants:From 4 February to 14 April 2020, 106 patients with neurological diseases were enrolled from all patients in the hospital with confirmed COVID-19 and divided into a severe group and a nonsevere group according to their COVID-19 diagnosis. Main outcome measures:Clinical characteristics, laboratory results, imaging findings, and treatment methods were all retrieved through an electronic medical records system and recorded in spreadsheets. Results:The mean (standard deviation, SD) age of patients was 72.7 (11.8) years, and 64 patients were male (60.4%). Among patients with co-morbid neurological diseases, 81 had a previous cerebral infarction (76.4%), 20 had dementia (18.9%), 10 had acute cerebral infarction (9.4%), 5 had sequelae of cerebral haemorrhage (4.7%), 4 had intracranial mass lesions (3.8%), 3 had epilepsy (2.8%), 2 had Parkinsons disease (1.9%), and 1 had myelopathy (0.9%). Fever (n = 62, 58.5%) was the most common symptom. The most common neurological symptoms were myalgia (n = 26, 24.5%), followed by extremity paralysis (n = 20, 18.9%), impaired consciousness (n = 17, 16%), and positive focal neurological signs (n = 42, 39.6%). Eight patients (7.5%) died. There were more patients with altered mental status in the severe group than in the non-severe group (6 [10.2%] vs. 0, P = 0.033). The inflammatory response in the severe group was more significant than that in the non-severe group. There were more patients taking anticoagulant drugs (25 [42.4%] vs. 4 [8.5%], P < 0.001) and sedative drugs (22 [37.3%] vs. 9 [19.1%], P = 0.041) in the severe group than in the non-severe group. Amid all 93 patients with cerebrovascular diseases, only 32 (34.4%) were taking aspirin, 13 (14%) taking clopidogrel, and 33 (35.5%) taking statins. Conclusions:Patients with COVID-19 with co-morbid neurological diseases had an advanced age, a high rate of severe illness, and a high mortality rate. Among the neurological symptoms, altered mental status was more common in patients with severe COVID-19 with co-morbid neurological diseases.

Heparin-induced thrombocytopenia is associated with a high risk of mortality in critical COVID-19 patients receiving heparin-involved treatment (preprint)

Background Coronavirus infectious disease 2019 (COVID-19) has developed into a global pandemic. It is essential to investigate the clinical characteristics of COVID-19 and uncover potential risk factors for severe disease to reduce the overall mortality rate of COVID-19. Methods Sixty-one critical COVID-19 patients admitted to the intensive care unit (ICU) and 93 severe non-ICU patients at Huoshenshan Hospital (Wuhan, China) were included in this study. Medical records, including demographic, platelet counts, heparin-involved treatments, heparin-induced thrombocytopenia-(HIT) related laboratory tests, and fatal outcomes of COVID-19 patients were analyzed and compared between survivors and nonsurvivors. Findings Sixty-one critical COVID-19 patients treated in ICU included 15 survivors and 46 nonsurvivors. Forty-one percent of them (25/61) had severe thrombocytopenia, with a platelet count (PLT) less than 50x109/L, of whom 76% (19/25) had a platelet decrease of >50% compared to baseline; 96% of these patients (24/25) had a fatal outcome. Among the 46 nonsurvivors, 52.2% (24/46) had severe thrombocytopenia, compared to 6.7% (1/15) among survivors. Moreover, continuous renal replacement therapy (CRRT) could induce a significant decrease in PLT in 81.3% of critical CRRT patients (13/16), resulting in a fatal outcome. In addition, a high level of anti-heparin-PF4 antibodies, a marker of HIT, was observed in most ICU patients. Surprisingly, HIT occurred not only in patients with heparin exposure, such as CRRT, but also in heparin-naive patients, suggesting that spontaneous HIT may occur in COVID-19. Interpretation Anti-heparin-PF4 antibodies are induced in critical COVID-19 patients, resulting in a progressive platelet decrease. Exposure to a high dose of heparin may trigger further severe thrombocytopenia with a fatal outcome. An alternative anticoagulant other than heparin should be used to treat COVID-19 patients in critical condition.

Highly pathogenic coronavirus N protein aggravates lung injury by MASP-2-mediated complement over-activation (preprint)

An excessive immune response contributes to SARS-CoV, MERS-CoV and SARS-CoV-2 pathogenesis and lethality, but the mechanism remains unclear. In this study, the N proteins of SARS-CoV, MERS-CoV and SARS-CoV-2 were found to bind to MASP-2, the key serine protease in the lectin pathway of complement activation, resulting in aberrant complement activation and aggravated inflammatory lung injury. Either blocking the N protein:MASP-2 interaction or suppressing complement activation can significantly alleviate N protein-induced complement hyper-activation and lung injury in vitro and in vivo. Complement hyper-activation was also observed in COVID-19 patients, and a promising suppressive effect was observed when the deteriorating patients were treated with anti-C5a monoclonal antibody. Complement suppression may represent a common therapeutic approach for pneumonia induced by these highly pathogenic coronaviruses.