Deep Transfer Learning for COVID-19 Detection and Lesion Recognition Using Chest CT Images.

Starting from December 2019, the global pandemic of coronavirus disease 2019 (COVID-19) is continuously expanding and has caused several millions of deaths worldwide. Fast and accurate diagnostic methods for COVID-19 detection play a vital role in containing the plague. Chest computed tomography (CT) is one of the most commonly used diagnosis methods. However, a complete CT-scan has hundreds of slices, and it is time-consuming for radiologists to check each slice to diagnose COVID-19. This study introduces a novel method for fast and automated COVID-19 diagnosis using the chest CT scans. The proposed models are based on the state-of-the-art deep convolutional neural network (CNN) architecture, and a 2D global max pooling (globalMaxPool2D) layer is used to improve the performance. We compare the proposed models to the existing state-of-the-art deep learning models such as CNN based models and vision transformer (ViT) models. Based off of metric such as area under curve (AUC), sensitivity, specificity, accuracy, and false discovery rate (FDR), experimental results show that the proposed models outperform the previous methods, and the best model achieves an area under curve of 0.9744 and accuracy 94.12% on our test datasets. It is also shown that the accuracy is improved by around 1% by using the 2D global max pooling layer. Moreover, a heatmap method to highlight the lesion area on COVID-19 chest CT images is introduced in the paper. This heatmap method is helpful for a radiologist to identify the abnormal pattern of COVID-19 on chest CT images. In addition, we also developed a freely accessible online simulation software for automated COVID-19 detection using CT images. The proposed deep learning models and software tool can be used by radiologist to diagnose COVID-19 more accurately and efficiently.

Are Older People Really More Susceptible to SARS-CoV-2?

Since the outbreak, COVID-19 has spread rapidly across the globe due to its high infectivity and lethality. Age appears to be one of the key factors influencing the status and progression of SARS-CoV-2 infection, as multiple reports indicated that the majority of COVID-19 infections and severe cases are elderly. Most people simply assume that the elderly are more susceptible to SARS-CoV-2 than the young, but the mechanism behind it is still open to question. The older and younger people are at similar risk of infection because their infection process is the same and they must be exposed to the virus first. However, whether they will get sick after exposure to the virus and how their disease progresses depend on their immune mechanisms. In older populations, inflammation and immune aging reduce their ability to resist SARS-CoV-2 infection. Meanwhile, under the influence of comorbidities, ACE2 receptor and various cytokines undergo corresponding changes, thus accelerating the entry, replication, and transmission of SARS-CoV-2 in the body, promoting disease progression, and leading to severe illness and even death. In addition, the relatively fragile mental state of the elderly can also affect their timely recovery from COVID-19. Therefore, once older people are infected with SARS-CoV-2, they are more prone to severe illness and death with a poor prognosis, and they should strengthen protection to avoid exposure to the virus.

Deep Transfer Learning for COVID-19 Detection and Lesion Recognition Using Chest CT Images

Starting from December 2019, the global pandemic of coronavirus disease 2019 (COVID-19) is continuously expanding and has caused several millions of deaths worldwide. Fast and accurate diagnostic methods for COVID-19 detection play a vital role in containing the plague. Chest computed tomography (CT) is one of the most commonly used diagnosis methods. However, a complete CT-scan has hundreds of slices, and it is time-consuming for radiologists to check each slice to diagnose COVID-19. This study introduces a novel method for fast and automated COVID-19 diagnosis using the chest CT scans. The proposed models are based on the state-of-the-art deep convolutional neural network (CNN) architecture, and a 2D global max pooling (globalMaxPool2D) layer is used to improve the performance. We compare the proposed models to the existing state-of-the-art deep learning models such as CNN based models and vision transformer (ViT) models. Based off of metric such as area under curve (AUC), sensitivity, specificity, accuracy, and false discovery rate (FDR), experimental results show that the proposed models outperform the previous methods, and the best model achieves an area under curve of 0.9744 and accuracy 94.12% on our test datasets. It is also shown that the accuracy is improved by around 1% by using the 2D global max pooling layer. Moreover, a heatmap method to highlight the lesion area on COVID-19 chest CT images is introduced in the paper. This heatmap method is helpful for a radiologist to identify the abnormal pattern of COVID-19 on chest CT images. In addition, we also developed a freely accessible online simulation software for automated COVID-19 detection using CT images. The proposed deep learning models and software tool can be used by radiologist to diagnose COVID-19 more accurately and efficiently.

Low expression of EXOSC2 protects against clinical COVID-19 and impedes SARS-CoV-2 replication.

New therapeutic targets are a valuable resource for treatment of SARS-CoV-2 viral infection. Genome-wide association studies have identified risk loci associated with COVID-19, but many loci are associated with comorbidities and are not specific to host-virus interactions. Here, we identify and experimentally validate a link between reduced expression of EXOSC2 and reduced SARS-CoV-2 replication. EXOSC2 was one of the 332 host proteins examined, all of which interact directly with SARS-CoV-2 proteins. Aggregating COVID-19 genome-wide association studies statistics for gene-specific eQTLs revealed an association between increased expression of EXOSC2 and higher risk of clinical COVID-19. EXOSC2 interacts with Nsp8 which forms part of the viral RNA polymerase. EXOSC2 is a component of the RNA exosome, and here, LC-MS/MS analysis of protein pulldowns demonstrated interaction between the SARS-CoV-2 RNA polymerase and most of the human RNA exosome components. CRISPR/Cas9 introduction of nonsense mutations within EXOSC2 in Calu-3 cells reduced EXOSC2 protein expression and impeded SARS-CoV-2 replication without impacting cellular viability. Targeted depletion of EXOSC2 may be a safe and effective strategy to protect against clinical COVID-19.

Multiomic analysis reveals cell-type-specific molecular determinants of COVID-19 severity.

The determinants of severe COVID-19 in healthy adults are poorly understood, which limits the opportunity for early intervention. We present a multiomic analysis using machine learning to characterize the genomic basis of COVID-19 severity. We use single-cell multiome profiling of human lungs to link genetic signals to cell-type-specific functions. We discover >1,000 risk genes across 19 cell types, which account for 77% of the SNP-based heritability for severe disease. Genetic risk is particularly focused within natural killer (NK) cells and T cells, placing the dysfunction of these cells upstream of severe disease. Mendelian randomization and single-cell profiling of human NK cells support the role of NK cells and further localize genetic risk to CD56bright NK cells, which are key cytokine producers during the innate immune response. Rare variant analysis confirms the enrichment of severe-disease-associated genetic variation within NK-cell risk genes. Our study provides insights into the pathogenesis of severe COVID-19 with potential therapeutic targets.

Efficacy and Safety of Lianhua Qingwen for Patients with COVID-19: A Systematic Review and Meta-Analysis.

BACKGROUND: Corona virus disease 2019 (COVID-19) has spread around the world since its outbreak, and there is no ascertained effective drug up to now. Lianhua Qingwen (LHQW) has been widely used in China and overseas Chinese, which had some advantages in the treatment of COVID-19. OBJECTIVE: To evaluate the efficacy and safety of LHQW for COVID-19 by conducting a systematic review with meta-analysis. METHODS: A comprehensive literature search was conducted in 12 electronic databases from their establishment to October 30, 2021. Note Express 3.2.0 was used for screening of trials, and the data was independently extracted in duplicate by 2 researchers. The risk of bias of randomized controlled trials (RCTs) and retrospective studies were assessed by using the Cochrane collaboration tool and Newcastle Ottawa Scale, respectively, followed by data analysis using RevMan 5.3. The RCTs or retrospective studies to treat COVID-19 using LHQW were included. The intervention measures in the experimental group were LHQW alone or combined with chemical drugs (LCWC), and that in the control group were chemical drugs (CDs). Outcome measures included computed tomography (CT) recovery rate, disappearance rates of primary (fever, cough, fatigue), respiratory, gastrointestinal and other symptoms, exacerbation rate and adverse reaction. Subgroup analysis was conducted according to whether LHQW was combined with CDs and the different treatment methods in the control group. RESULTS: Nine trials with 1,152 participants with COVID-19 were included. The CT recovery rates of LHQW and LCWC were 1.36 and 1.32 times of CDs, respectively (P<0.05). Compared with CDs, LCWC remarkably increased the disappearance rates of fever, cough, fatigue, expectoration, shortness of breath, and muscle soreness (P<0.05). LHQW also obviously decreased the exacerbation rate, which was 0.45 times of CDs alone (P<0.05). There was no obvious difference between LCWC and CDs in adverse reaction (P>0.05). CONCLUSIONS: LHQW was more suitable for treating COVID-19 patients with obvious expectoration, shortness of breath and muscle soreness. LHQW had advantages in treating COVID-19 with no obvious exacerbation. (PROSPERO No. CRD42021235937).

Clinical Characteristics, Transmissibility, Pathogenicity, Susceptible Populations, and Re-infectivity of Prominent COVID-19 Variants.

In addition to the rapid, global spread of SARS-CoV-2, new and comparatively more contagious variants are of considerable concern. These emerging mutations have become a threat to the global public health, creating COVID-19 surges in different countries. However, information on these emerging variants is limited and scattered. In this review, we discuss new variants that have emerged worldwide and identify several variants of concern, such as B.1.1.7, B.1.351, P.1, B.1.617.2 and B.1.1.529, and their basic characteristics. Other significant variants such as C.37, B.1.621, B.1.525, B.1.526, AZ.5, C.1.2, and B.1.617.1 are also discussed. This review highlights the clinical characteristics of these variants, including transmissibility, pathogenicity, susceptible population, and re-infectivity. It provides the latest information on the recent variants of SARS-CoV-2. The summary of this information will help researchers formulate reasonable strategies to curb the ongoing COVID-19 pandemic.