An Ensemble Feature Selection Approach-Based Machine Learning Classifiers for Prediction of COVID-19 Disease

The respiratory disease of coronavirus disease 2019 (COVID-19) has wreaked havoc on the economy of every nation by infecting and killing millions of people. This deadly disease has taken a toll on the life of the entire human race, and an exact cure for it is still not developed. Thus, the control and cure of this disease mainly depend on restricting its transmission rate through early detection. The detection of coronavirus infection facilitates the isolation and exclusive care of infected patients. This research paper proposes a novel data mining system that combines the ensemble feature selection method and machine learning classifier for the effective identification of COVID-19 infection. Different feature selection approaches including chi-square test, recursive feature elimination (RFE), genetic algorithm (GA), particle swarm optimization (PSO), and random forest are evaluated for their effectiveness in enhancing the classification accuracy of the machine learning classifiers. The classifiers that are considered in this research work are decision tree, naïve Bayes, K-nearest neighbor (KNN), multilayer perceptron (MLP), and support vector machine (SVM). Two COVID-19 datasets were used for testing from which the best features supporting the dataset were extracted by the proposed system. The performance of the machine learning classifiers based on the ensemble feature selection methods is analyzed.

Facile construction of nanocubic Mn3[Fe(CN)6]2@Pt based electrochemical DNA sensors for ultrafast precise determination of SARS-CoV-2

Owing to the high mortality and strong infection ability of COVID-19, the early rapid diagnosis is essential to reduce the risk of severe symptoms and the loss of lung function. In clinic, the commonly used detection methods, including the computed tomography (CT) and reverse transcription-polymerase chain reaction (RT-PCR), are often time-consuming with bulky instruments, which normally require more than one hour to report the results. To shorten the analytical period for testing the COVID-19 virus (SARS-CoV-2), we proposed an ultrafast and ultrasensitive DNA sensors to achieve an accurate determination of the DNA sequence by the RNA reverse transcription (rtDNA) of the SARS-CoV-2. A nanocubic architecture of the MnFe@Pt crystals was constructed to integrate both electrocatalysis and conductivity to greatly improve the biosensing performance. After the immobilization of a specific capture and report DNA on above nanocomposite, the rtDNA can be rapidly caught to the DNA sensor to form a double-helix structure, thus generating the current signal change. Within only 10 min, the as-prepared DNA sensors exhibited ultralow detection limit (1 × 10-20 M) and wide linear detection range, together with an outstanding selectivity among various interfering substances.

Evaluation of Binding and Neutralizing Antibodies for Inactivated SARS-CoV-2 Vaccine Immunization

The circulating severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) variant presents an ongoing challenge for surveillance and detection. It is important to establish an assay for SARS-CoV-2 antibodies in vaccinated individuals. Numerous studies have demonstrated that binding antibodies (such as S-IgG and N-IgG) and neutralizing antibodies (Nabs) can be detected in vaccinated individuals. However, it is still unclear how to evaluate the consistency and correlation between binding antibodies and Nabs induced by inactivated SARS-CoV-2 vaccines. In this study, serum samples from humans, rhesus macaques, and hamsters immunized with inactivated SARS-CoV-2 vaccines were analyzed for S-IgG, N-IgG, and Nabs. The results showed that the titer and seroconversion rate of S-IgG were significantly higher than those of N-IgG. The correlation between S-IgG and Nabs was higher compared to that of N-IgG. Based on this analysis, we further investigated the titer thresholds of S-IgG and N-IgG in predicting the seroconversion of Nabs. According to the threshold, we can quickly determine the positive and negative effects of the SARS-CoV-2 variant neutralizing antibody in individuals. These findings suggest that the S-IgG antibody is a better supplement to and confirmation of SARS-CoV-2 vaccine immunization.

Sandwich ELISA for the Quantification of Nucleocapsid Protein of SARS-CoV-2 Based on Polyclonal Antibodies from Two Different Species

In this study, a cost-effective sandwich ELISA test, based on polyclonal antibodies, for routine quantification SARS-CoV-2 nucleocapsid (N) protein was developed. The recombinant N protein was produced and used for the production of mice and rabbit antisera. Polyclonal N protein-specific antibodies served as capture and detection antibodies. The prototype ELISA has LOD 0.93 ng/mL and LOQ 5.3 ng/mL, with a linear range of 1.52-48.83 ng/mL. N protein heat pretreatment (56 °C, 1 h) decreased, while pretreatment with 1% Triton X-100 increased analytical ELISA sensitivity. The diagnostic specificity of ELISA was 100% (95% CI, 91.19-100.00%) and sensitivity was 52.94% (95% CI, 35.13-70.22%) compared to rtRT-PCR (Ct < 40). Profoundly higher sensitivity was obtained using patient samples mostly containing Wuhan-similar variants (Wuhan, alpha, and delta), 62.50% (95% CI, 40.59 to 81.20%), in comparison to samples mostly containing Wuhan-distant variants (Omicron) 30.00% (6.67-65.25%). The developed product has relatively high diagnostic sensitivity in relation to its analytical sensitivity due to the usage of polyclonal antibodies from two species, providing a wide repertoire of antibodies against multiple N protein epitopes. Moreover, the fast, simple, and inexpensive production of polyclonal antibodies, as the most expensive assay components, would result in affordable antigen tests.

Momentum contrast transformer for COVID-19 diagnosis with knowledge distillation

Intelligent diagnosis has been widely studied in diagnosing novel corona virus disease (COVID-19). Existing deep models typically do not make full use of the global features such as large areas of ground glass opacities, and the local features such as local bronchiolectasis from the COVID-19 chest CT images, leading to unsatisfying recognition accuracy. To address this challenge, this paper proposes a novel method to diagnose COVID-19 using momentum contrast and knowledge distillation, termed MCT-KD. Our method takes advantage of Vision Transformer to design a momentum contrastive learning task to effectively extract global features from COVID-19 chest CT images. Moreover, in transfer and fine-tuning process, we integrate the locality of convolution into Vision Transformer via special knowledge distillation. These strategies enable the final Vision Transformer simultaneously focuses on global and local features from COVID-19 chest CT images. In addition, momentum contrastive learning is self-supervised learning, solving the problem that Vision Transformer is challenging to train on small datasets. Extensive experiments confirm the effectiveness of the proposed MCT-KD. In particular, our MCT-KD is able to achieve 87.43% and 96.94% accuracy on two publicly available datasets, respectively.

FP-CNN: Fuzzy pooling-based convolutional neural network for lung ultrasound image classification with explainable AI

The COVID-19 pandemic wreaks havoc on healthcare systems all across the world. In pandemic scenarios like COVID-19, the applicability of diagnostic modalities is crucial in medical diagnosis, where non-invasive ultrasound imaging has the potential to be a useful biomarker. This research develops a computer-assisted intelligent methodology for ultrasound lung image classification by utilizing a fuzzy pooling-based convolutional neural network FP-CNN with underlying evidence of particular decisions. The fuzzy-pooling method finds better representative features for ultrasound image classification. The FPCNN model categorizes ultrasound images into one of three classes: covid, disease-free (normal), and pneumonia. Explanations of diagnostic decisions are crucial to ensure the fairness of an intelligent system. This research has used Shapley Additive Explanation (SHAP) to explain the prediction of the FP-CNN models. The prediction of the black-box model is illustrated using the SHAP explanation of the intermediate layers of the black-box model. To determine the most effective model, we have tested different state-of-the-art convolutional neural network architectures with various training strategies, including fine-tuned models, single-layer fuzzy pooling models, and fuzzy pooling at all pooling layers. Among different architectures, the Xception model with all pooling layers having fuzzy pooling achieves the best classification results of 97.2% accuracy. We hope our proposed method will be helpful for the clinical diagnosis of covid-19 from lung ultrasound (LUS) images.

GIONet: Global information optimized network for multi-center COVID-19 diagnosis via COVID-GAN and domain adversarial strategy

The outbreak of coronavirus disease (COVID-19) in 2019 has highlighted the need for automatic diagnosis of the disease, which can develop rapidly into a severe condition. Nevertheless, distinguishing between COVID-19 pneumonia and community-acquired pneumonia (CAP) through computed tomography scans can be challenging due to their similar characteristics. The existing methods often perform poorly in the 3-class classification task of healthy, CAP, and COVID-19 pneumonia, and they have poor ability to handle the heterogeneity of multi-centers data. To address these challenges, we design a COVID-19 classification model using global information optimized network (GIONet) and cross-centers domain adversarial learning strategy. Our approach includes proposing a 3D convolutional neural network with graph enhanced aggregation unit and multi-scale self-attention fusion unit to improve the global feature extraction capability. We also verified that domain adversarial training can effectively reduce feature distance between different centers to address the heterogeneity of multi-center data, and used specialized generative adversarial networks to balance data distribution and improve diagnostic performance. Our experiments demonstrate satisfying diagnosis results, with a mixed dataset accuracy of 99.17% and cross-centers task accuracies of 86.73% and 89.61%.

Household conditions, COVID-19, and equity: Insight from two nationally representative surveys

Background

With people across the United States spending increased time at home since the emergence of COVID-19, housing characteristics may have an even greater impact on health. Therefore, we assessed associations between household conditions and COVID-19 experiences.

Methods

We used data from two nationally representative surveys: the Tufts Equity Study (TES; n = 1449 in 2021; n = 1831 in 2022) and the Household Pulse Survey (HPS; n = 147,380 in 2021; n = 62,826 in 2022). In the TES, housing conditions were characterized by heating/cooling methods; smoking inside the home; visible water damage/mold; age of housing unit; and self-reported concern about various environmental factors. In TES and HPS, household size was assessed. Accounting for sampling weights, we examined associations between each housing exposure and COVID-19 outcomes (diagnosis, vaccination) using separate logistic regression models with covariates selected based on an evidence-based directed acyclic graph.

Results

Having had COVID-19 was more likely among people who reported poor physical housing condition (odds ratio [OR] = 2.32; 95% confidence interval [CI] = 1.17-4.59; 2021), visible water damage or mold/musty smells (OR = 1.50; 95% CI = 1.10-2.03; 2022), and larger household size (5+ versus 1-2 people; OR = 1.53, 95% CI = 1.34-1.75, HPS 2022). COVID-19 vaccination was less likely among participants who reported smoke exposure inside the home (OR = 0.53; 95% CI = 0.31-0.90; 2022), poor water quality (OR = 0.42; 95% CI = 0.21-0.85; 2021), noise from industrial activity/construction (OR = 0.44; 95% CI = 0.19-0.99; 2022), and larger household size (OR = 0.57; 95% CI = 0.46-0.71; HPS 2022). Vaccination was also positively associated with poor indoor air quality (OR = 1.96; 95% CI = 1.02-3.72; 2022) and poor physical housing condition (OR = 2.27; 95% CI = 1.01-5.13; 2022). Certain heating/cooling sources were associated with COVID-19 outcomes.

Conclusions

Our study found poor housing conditions associated with increased COVID-19 burden, which may be driven by systemic disparities in housing, healthcare, and financial access to resources during the COVID-19 pandemic.

COV-MobNets: a mobile networks ensemble model for diagnosis of COVID-19 based on chest X-ray images

BACKGROUND

The medical profession is facing an excessive workload, which has led to the development of various Computer-Aided Diagnosis (CAD) systems as well as Mobile-Aid Diagnosis (MAD) systems. These technologies enhance the speed and accuracy of diagnoses, particularly in areas with limited resources or remote regions during the pandemic. The primary purpose of this research is to predict and diagnose COVID-19 infection from chest X-ray images by developing a mobile-friendly deep learning framework, which has the potential for deployment in portable devices such as mobile or tablet, especially in situations where the workload of radiology specialists may be high. Moreover, this could improve the accuracy and transparency of population screening to assist radiologists during the pandemic.

METHODS

In this study, the Mobile Networks ensemble model called COV-MobNets is proposed to classify positive COVID-19 X-ray images from negative ones and can have an assistant role in diagnosing COVID-19. The proposed model is an ensemble model, combining two lightweight and mobile-friendly models: MobileViT based on transformer structure and MobileNetV3 based on Convolutional Neural Network. Hence, COV-MobNets can extract the features of chest X-ray images in two different methods to achieve better and more accurate results. In addition, data augmentation techniques were applied to the dataset to avoid overfitting during the training process. The COVIDx-CXR-3 benchmark dataset was used for training and evaluation.

RESULTS

The classification accuracy of the improved MobileViT and MobileNetV3 models on the test set has reached 92.5% and 97%, respectively, while the accuracy of the proposed model (COV-MobNets) has reached 97.75%. The sensitivity and specificity of the proposed model have also reached 98.5% and 97%, respectively. Experimental comparison proves the result is more accurate and balanced than other methods.

CONCLUSION

The proposed method can distinguish between positive and negative COVID-19 cases more accurately and quickly. The proposed method proves that utilizing two automatic feature extractors with different structures as an overall framework of COVID-19 diagnosis can lead to improved performance, enhanced accuracy, and better generalization to new or unseen data. As a result, the proposed framework in this study can be used as an effective method for computer-aided diagnosis and mobile-aided diagnosis of COVID-19. The code is available publicly for open access at https://github.com/MAmirEshraghi/COV-MobNets .

Paper microfluidics with deep learning for portable intelligent nucleic acid amplification tests

During global outbreaks such as COVID-19, regular nucleic acid amplification tests (NAATs) have posed unprecedented burden on hospital resources. Data of traditional NAATs are manually analyzed post assay. Integration of artificial intelligence (AI) with on-chip assays give rise to novel analytical platforms via data-driven models. Here, we combined paper microfluidics, portable optoelectronic system with deep learning for SARS-CoV-2 detection. The system was quite streamlined with low power dissipation. Pixel by pixel signals reflecting amplification of synthesized SARS-CoV-2 templates (containing ORF1ab, N and E genes) can be real-time processed. Then, the data were synchronously fed to the neural networks for early prediction analysis. Instead of the quantification cycle (Cq) based analytics, reaction dynamics hidden at the early stage of amplification curve were utilized by neural networks for predicting subsequent data. Qualitative and quantitative analysis of the 40-cycle NAATs can be achieved at the end of 22nd cycle, reducing time cost by 45%. In particular, the attention mechanism based deep learning model trained by microfluidics-generated data can be seamlessly adapted to multiple clinical datasets including readouts of SARS-CoV-2 detection. Accuracy, sensitivity and specificity of the prediction can reach up to 98.1%, 97.6% and 98.6%, respectively. The approach can be compatible with the most advanced sensing technologies and AI algorithms to inspire ample innovations in fields of fundamental research and clinical settings.

COVID-19 diagnosis utilizing wavelet-based contrastive learning with chest CT images

The pandemic caused by the coronavirus disease 2019 (COVID-19) has continuously wreaked havoc on human health. Computer-aided diagnosis (CAD) system based on chest computed tomography (CT) has been a hotspot option for COVID-19 diagnosis. However, due to the high cost of data annotation in the medical field, it happens that the number of unannotated data is much larger than the annotated data. Meanwhile, having a highly accurate CAD system always requires a large amount of labeled data training. To solve this problem while meeting the needs, this paper presents an automated and accurate COVID-19 diagnosis system using few labeled CT images. The overall framework of this system is based on the self-supervised contrastive learning (SSCL). Based on the framework, our enhancement of our system can be summarized as follows. 1) We integrated a two-dimensional discrete wavelet transform with contrastive learning to fully use all the features from the images. 2) We use the recently proposed COVID-Net as the encoder, with a redesign to target the specificity of the task and learning efficiency. 3) A new pretraining strategy based on contrastive learning is applied for broader generalization ability. 4) An additional auxiliary task is exerted to promote performance during classification. The final experimental result of our system attained 93.55%, 91.59%, 96.92% and 94.18% for accuracy, recall, precision, and F1-score respectively. By comparing results with the existing schemes, we demonstrate the performance enhancement and superiority of our proposed system.

Broad learning solution for rapid diagnosis of COVID-19

COVID-19 has put all of humanity in a health dilemma as it spreads rapidly. For many infectious diseases, the delay of detection results leads to the spread of infection and an increase in healthcare costs. COVID-19 diagnostic methods rely on a large number of redundant labeled data and time-consuming data training processes to obtain satisfactory results. However, as a new epidemic, obtaining large clinical datasets is still challenging, which will inhibit the training of deep models. And a model that can really rapidly diagnose COVID-19 at all stages of the model has still not been proposed. To address these limitations, we combine feature attention and broad learning to propose a diagnostic system (FA-BLS) for COVID-19 pulmonary infection, which introduces a broad learning structure to address the slow diagnosis speed of existing deep learning methods. In our network, transfer learning is performed with ResNet50 convolutional modules with fixed weights to extract image features, and the attention mechanism is used to enhance feature representation. After that, feature nodes and enhancement nodes are generated by broad learning with random weights to adaptly select features for diagnosis. Finally, three publicly accessible datasets were used to evaluate our optimization model. It was determined that the FA-BLS model had a 26-130 times faster training speed than deep learning with a similar level of accuracy, which can achieve a fast and accurate diagnosis, achieve effective isolation from COVID-19 and the proposed method also opens up a new method for other types of chest CT image recognition problems.

Artificial Intelligence Based COVID-19 Detection and Classification Model on Chest X-ray Images

Diagnostic and predictive models of disease have been growing rapidly due to developments in the field of healthcare. Accurate and early diagnosis of COVID-19 is an underlying process for controlling the spread of this deadly disease and its death rates. The chest radiology (CT) scan is an effective device for the diagnosis and earlier management of COVID-19, meanwhile, the virus mainly targets the respiratory system. Chest X-ray (CXR) images are extremely helpful in the effective diagnosis of COVID-19 due to their rapid outcomes, cost-effectiveness, and availability. Although the radiological image-based diagnosis method seems faster and accomplishes a better recognition rate in the early phase of the epidemic, it requires healthcare experts to interpret the images. Thus, Artificial Intelligence (AI) technologies, such as the deep learning (DL) model, play an integral part in developing automated diagnosis process using CXR images. Therefore, this study designs a sine cosine optimization with DL-based disease detection and classification (SCODL-DDC) for COVID-19 on CXR images. The proposed SCODL-DDC technique examines the CXR images to identify and classify the occurrence of COVID-19. In particular, the SCODL-DDC technique uses the EfficientNet model for feature vector generation, and its hyperparameters can be adjusted by the SCO algorithm. Furthermore, the quantum neural network (QNN) model can be employed for an accurate COVID-19 classification process. Finally, the equilibrium optimizer (EO) is exploited for optimum parameter selection of the QNN model, showing the novelty of the work. The experimental results of the SCODL-DDC method exhibit the superior performance of the SCODL-DDC technique over other approaches.

COVID-19 Diagnosis in Computerized Tomography (CT) and X-ray Scans Using Capsule Neural Network

This study proposes a deep-learning-based solution (named CapsNetCovid) for COVID-19 diagnosis using a capsule neural network (CapsNet). CapsNets are robust for image rotations and affine transformations, which is advantageous when processing medical imaging datasets. This study presents a performance analysis of CapsNets on standard images and their augmented variants for binary and multi-class classification. CapsNetCovid was trained and evaluated on two COVID-19 datasets of CT images and X-ray images. It was also evaluated on eight augmented datasets. The results show that the proposed model achieved classification accuracy, precision, sensitivity, and F1-score of 99.929%, 99.887%, 100%, and 99.319%, respectively, for the CT images. It also achieved a classification accuracy, precision, sensitivity, and F1-score of 94.721%, 93.864%, 92.947%, and 93.386%, respectively, for the X-ray images. This study presents a comparative analysis between CapsNetCovid, CNN, DenseNet121, and ResNet50 in terms of their ability to correctly identify randomly transformed and rotated CT and X-ray images without the use of data augmentation techniques. The analysis shows that CapsNetCovid outperforms CNN, DenseNet121, and ResNet50 when trained and evaluated on CT and X-ray images without data augmentation. We hope that this research will aid in improving decision making and diagnostic accuracy of medical professionals when diagnosing COVID-19.

Discovery of DNA aptamers targeting SARS-CoV-2 nucleocapsid protein and protein-binding epitopes for label-free COVID-19 diagnostics

The spread of COVID-19 has affected billions of people across the globe, and the diagnosis of viral infection still needs improvement. Because of high immunogenicity and abundant expression during viral infection, SARS-CoV-2 nucleocapsid (N) protein could be an important diagnostic marker. This study aimed to develop a label-free optical aptasensor fabricated with a novel single-stranded DNA aptamer to detect the N protein. The N-binding aptamers selected using asymmetric-emulsion PCR-SELEX and their binding affinity and cross-reactivity were characterized by biolayer interferometry. The tNSP3 aptamer (44 nt) was identified to bind the N protein of wild type and Delta and Omicron variants with high affinity (K_D in the range of 0.6-3.5 nM). Utilizing tNSP3 to detect the N protein spiked in human saliva evinced the potential of this aptamer with a limit of detection of 4.5 nM. Mass spectrometry analysis was performed along with molecular dynamics simulation to obtain an insight into how tNSP3 binds to the N protein. The identified epitope peptides are localized within the RNA-binding domain and C terminus of the N protein. Hence, we confirmed the performance of this aptamer as an analytical tool for COVID-19 diagnosis.

Comparison of SARS-CoV-2 Detection in Nasopharyngeal Swab and Saliva Samples from Patients Infected with Omicron Variant

To compare the detection of the SARS-CoV-2 Omicron variant in nasopharyngeal-swab (NPS) and oral saliva samples. 255 samples were obtained from 85 Omicron-infected patients. SARS-CoV-2 load was measured in the NPS and saliva samples by using Simplexa™ COVID-19 direct and Alinity m SARS-CoV-2 AMP assays. Results obtained with the two diagnostic platforms showed very good inter-assay concordance (91.4 and 82.4% for saliva and NPS samples, respectively) and a significant correlation among cycle threshold (Ct) values. Both platforms revealed a highly significant correlation among Ct obtained in the two matrices. Although the median Ct value was lower in NPS than in saliva samples, the Ct drop was comparable in size for both types of samples after 7 days of antiviral treatment of the Omicron-infected patients. Our result demonstrates that the detection of the SARS-CoV-2 Omicron variant is not influenced by the type of sample used for PCR analysis, and that saliva can be used as an alternative specimen for detection and follow-up of Omicron-infected patients.

Omnidirectional 2.5D representation for COVID-19 diagnosis using chest CTs

The Coronavirus Disease 2019 (COVID-19) has drastically overwhelmed most countries in the last two years, and image-based approaches using computerized tomography (CT) have been used to identify pulmonary infections. Recent methods based on deep learning either require time-consuming per-slice annotations (2D) or are highly data- and hardware-demanding (3D). This work proposes a novel omnidirectional 2.5D representation of volumetric chest CTs that allows exploring efficient 2D deep learning architectures while requiring volume-level annotations only. Our learning approach uses a siamese feature extraction backbone applied to each lung. It combines these features into a classification head that explores a novel combination of Squeeze-and-Excite strategies with Class Activation Maps. We experimented with public and in-house datasets and compared our results with state-of-the-art techniques. Our analyses show that our method provides better or comparable prediction quality and accurately distinguishes COVID-19 infections from other kinds of pneumonia and healthy lungs.

Improved deep convolutional neural networks using chimp optimization algorithm for Covid19 diagnosis from the X-ray images

Applying Deep Learning (DL) in radiological images (i.e., chest X-rays) is emerging because of the necessity of having accurate and fast COVID-19 detectors. Deep Convolutional Neural Networks (DCNN) have been typically used as robust COVID-19 positive case detectors in these approaches. Such DCCNs tend to utilize Gradient Descent-Based (GDB) algorithms as the last fully-connected layers' trainers. Although GDB training algorithms have simple structures and fast convergence rates for cases with large training samples, they suffer from the manual tuning of numerous parameters, getting stuck in local minima, large training samples set requirements, and inherently sequential procedures. It is exceedingly challenging to parallelize them with Graphics Processing Units (GPU). Consequently, the Chimp Optimization Algorithm (ChOA) is presented for training the DCNN's fully connected layers in light of the scarcity of a big COVID-19 training dataset and for the purpose of developing a fast COVID-19 detector with the capability of parallel implementation. In addition, two publicly accessible datasets termed COVID-Xray-5 k and COVIDetectioNet are used to benchmark the proposed detector known as DCCN-Chimp. In order to make a fair comparison, two structures are proposed: i-6c-2 s-12c-2 s and i-8c-2 s-16c-2 s, all of which have had their hyperparameters fine-tuned. The outcomes are evaluated in comparison to standard DCNN, Hybrid DCNN plus Genetic Algorithm (DCNN-GA), and Matched Subspace classifier with Adaptive Dictionaries (MSAD). Due to the large variation in results, we employ a weighted average of the ensemble of ten trained DCNN-ChOA, with the validation accuracy of the weights being used to determine the final weights. The validation accuracy for the mixed ensemble DCNN-ChOA is 99.11%. LeNet-5 DCNN's ensemble detection accuracy on COVID-19 is 84.58%. Comparatively, the suggested DCNN-ChOA yields over 99.11% accurate detection with a false alarm rate of less than 0.89%. The outcomes show that the DCCN-Chimp can deliver noticeably superior results than the comparable detectors. The Class Activation Map (CAM) is another tool used in this study to identify probable COVID-19-infected areas. Results show that highlighted regions are completely connected with clinical outcomes, which has been verified by experts.

Real-life evaluation of a rapid antigen test (DPP SARS-CoV-2 Antigen) for COVID-19 diagnosis of primary healthcare patients, in the context of the Omicron-dominant wave in Brazil

OBJECTIVES

We aimed to investigate the real-life performance of the rapid antigen test in the context of a primary healthcare setting, including symptomatic and asymptomatic individuals that sought diagnosis during an Omicron infection wave.

METHODS

We prospectively accessed the performance of the DPP SARS-CoV-2 Antigen test in the context of an Omicron-dominant real-life setting. We evaluated 347 unselected individuals (all-comers) from a public testing centre in Brazil, performing the rapid antigen test diagnosis at point-of-care with fresh samples. The combinatory result from two distinct real-time quantitative PCR (RT-qPCR) methods was employed as a reference and 13 samples with discordant PCR results were excluded.

RESULTS

The assessment of the rapid test in 67 PCR-positive and 265 negative samples revealed an overall sensitivity of 80.5% (CI 95% = 69.1%-89.2%), specificity of 99.2% (CI 95% = 97.3%-99.1%) and positive/negative predictive values higher than 95%. However, we observed that the sensitivity was dependent on the viral load (sensitivity in Ct < 31 = 93.7%, CI = 82.8%-98.7%; Ct > 31 = 47.4%, CI = 24.4%-71.1%). The positive samples evaluated in the study were Omicron (BA.1/BA.1.1) by whole-genome sequencing (n = 40) and multiplex RT-qPCR (n = 17).

CONCLUSIONS

Altogether, the data obtained from a real-life prospective cohort supports that the rapid antigen test sensitivity for Omicron remains high and underscores the reliability of the test for COVID-19 diagnosis in settings with high disease prevalence and limited PCR testing capability.

Improved deep convolutional neural networks using chimp optimization algorithm for Covid19 diagnosis from the X-ray images

Applying Deep Learning (DL) in radiological images (i.e., chest X-rays) is emerging because of the necessity of having accurate and fast COVID-19 detectors. Deep Convolutional Neural Networks (DCNN) have been typically used as robust COVID-19 positive case detectors in these approaches. Such DCCNs tend to utilize Gradient Descent-Based (GDB) algorithms as the last fully-connected layers' trainers. Although GDB training algorithms have simple structures and fast convergence rates for cases with large training samples, they suffer from the manual tuning of numerous parameters, getting stuck in local minima, large training samples set requirements, and inherently sequential procedures. It is exceedingly challenging to parallelize them with Graphics Processing Units (GPU). Consequently, the Chimp Optimization Algorithm (ChOA) is presented for training the DCNN's fully connected layers in light of the scarcity of a big COVID-19 training dataset and for the purpose of developing a fast COVID-19 detector with the capability of parallel implementation. In addition, two publicly accessible datasets termed COVID-Xray-5 k and COVIDetectioNet are used to benchmark the proposed detector known as DCCN-Chimp. In order to make a fair comparison, two structures are proposed: i-6c-2 s-12c-2 s and i-8c-2 s-16c-2 s, all of which have had their hyperparameters fine-tuned. The outcomes are evaluated in comparison to standard DCNN, Hybrid DCNN plus Genetic Algorithm (DCNN-GA), and Matched Subspace classifier with Adaptive Dictionaries (MSAD). Due to the large variation in results, we employ a weighted average of the ensemble of ten trained DCNN-ChOA, with the validation accuracy of the weights being used to determine the final weights. The validation accuracy for the mixed ensemble DCNN-ChOA is 99.11%. LeNet-5 DCNN's ensemble detection accuracy on COVID-19 is 84.58%. Comparatively, the suggested DCNN-ChOA yields over 99.11% accurate detection with a false alarm rate of less than 0.89%. The outcomes show that the DCCN-Chimp can deliver noticeably superior results than the comparable detectors. The Class Activation Map (CAM) is another tool used in this study to identify probable COVID-19-infected areas. Results show that highlighted regions are completely connected with clinical outcomes, which has been verified by experts.

Application of ultrasensitive assay for SARS-CoV-2 antigen in nasopharynx in the management of COVID-19 patients with comorbidities during the peak of 2022 Shanghai epidemics in a tertiary hospital

OBJECTIVES

Various comorbidities associated with COVID-19 add up in severity of the disease and obviously prolonged the time for viral clearance. This study investigated a novel ultrasensitive MAGLUMI^® SARS-CoV-2 Ag chemiluminescent immunoassay assay (MAG-CLIA) for diagnosis and monitoring the infectivity of COVID-19 patients with comorbid conditions during the pandemic of 2022 Shanghai.

METHODS

Analytical performances of the MAG-CLIA were evaluated, including precision, limit of quantitation, linearity and specificity. Nasopharyngeal specimens from 232 hospitalized patients who were SARS-CoV-2 RT-qPCR positive and from 477 healthy donors were included. The longitudinal studies were performed by monitoring antigen concentrations alongside with RT-qPCR results in 14 COVID-19 comorbid participants for up to 22 days. The critical antigen concentration in determining virus infectivity was evaluated at the reference cycle threshold (Ct) of 35.

RESULTS

COVID-19 patients were well-identified using an optimal threshold of 0.64 ng/L antigen concentration, with sensitivity and specificity of 95.7% (95% CI: 92.2-97.9%) and 98.3% (95% CI: 96.7-99.3%), respectively, while the Wondfo LFT exhibited those of 34.9% (95% CI: 28.8-41.4%) and 100% (95% CI: 99.23-100%), respectively. The sensitivity of MAG-CLIA remained 91.46% (95% CI: 83.14-95.8%) for the samples with Ct values between 35 and 40. Close dynamic consistence was observed between MAG-CLIA and viral load time series in the longitudinal studies. The critical value of 8.82 ng/L antigen showed adequate sensitivity and specificity in evaluating the infectivity of hospitalized convalescent patients with comorbidities.

CONCLUSIONS

The MAG-CLIA SARS-CoV-2 Ag detection is an effective and alternative approach for rapid diagnosis and enables us to evaluate the infectivity of hospitalized convalescent patients with comorbidities.

Is Dementia Associated with COVID-19 Mortality? A Multicenter Retrospective Cohort Study Conducted in 50 Hospitals in Germany

BACKGROUND

Dementia has been identified as a major predictor of mortality associated with COVID-19.

OBJECTIVE

The objective of this study was to investigate the association between dementia and mortality in COVID-19 inpatients in Germany across a longer interval during the pandemic.

METHODS

This retrospective study was based on anonymized data from 50 hospitals in Germany and included patients with a confirmed COVID-19 diagnosis hospitalized between March 11, 2020 and July, 20, 2022. The main outcome of the study was the association of mortality during inpatient stays with dementia diagnosis, which was studied using multivariable logistic regression adjusted for age, sex, and comorbidities as well as univariate logistic regression for matched pairs.

RESULTS

Of 28,311 patients diagnosed with COVID-19, 11.3% had a diagnosis of dementia. Prior to matching, 26.5% of dementia patients and 11.5% of non-dementia patients died; the difference decreased to 26.5% of dementia versus 21.7% of non-dementia patients within the matched pairs (n = 3,317). This corresponded to an increase in the risk of death associated with dementia (OR = 1.33; 95% CI: 1.16-1.46) in the univariate regression conducted for matched pairs.

CONCLUSION

Although dementia was associated with COVID-19 mortality, the association was weaker than in previously published studies. Further studies are needed to better understand whether and how pre-existing neuropsychiatric conditions such as dementia may impact the course and outcome of COVID-19.

A novel COVID diagnosis and feature extraction based on discrete wavelet model and classification using X-ray and CT images

Recently, the Covid-19 pandemic has affected several lives of people globally, and there is a need for a massive number of screening tests to diagnose the existence of coronavirus. For the medical specialist, detecting COVID-19 cases is a difficult task. There is a need for fast, cheap and accurate diagnostic tools. The chest X-ray and the computerized tomography (CT) play a significant role in the COVID-19 diagnosis. The advancement of deep learning (DL) approaches helps to introduce a COVID diagnosis system to achieve maximum detection rate with minimum time complexity. This research proposed a discrete wavelet optimized network model for COVID-19 diagnosis and feature extraction to overcome these problems. It consists of three stages pre-processing, feature extraction and classification. The raw images are filtered in the pre-processing phase to eliminate unnecessary noises and improve the image quality using the MMG hybrid filtering technique. The next phase is feature extraction, in this stage, the features are extracted, and the dimensionality of the features is diminished with the aid of a modified discrete wavelet based Mobile Net model. The third stage is the classification here, the convolutional Aquila COVID detection network model is developed to classify normal and COVID-19 positive cases from the collected images of the COVID-CT and chest X-ray dataset. Finally, the performance of the proposed model is compared with some of the existing models in terms of accuracy, specificity, sensitivity, precision, f-score, negative predictive value (NPV) and positive predictive value (PPV), respectively. The proposed model achieves the performance of 99%, 100%, 98.5%, and 99.5% for the CT dataset, and the accomplished accuracy, specificity, sensitivity, and precision values of the proposed model for the X-ray dataset are 98%, 99%, 98% and 97% respectively. In addition, the statistical and cross validation analysis is conducted to validate the effectiveness of the proposed model.

AMTLDC: a new adversarial multi-source transfer learning framework to diagnosis of COVID-19

In recent years, deep learning techniques have been widely used to diagnose diseases. However, in some tasks, such as the diagnosis of COVID-19 disease, due to insufficient data, the model is not properly trained and as a result, the generalizability of the model decreases. For example, if the model is trained on a CT scan dataset and tested on another CT scan dataset, it predicts near-random results. To address this, data from several different sources can be combined using transfer learning, taking into account the intrinsic and natural differences in existing datasets obtained with different medical imaging tools and approaches. In this paper, to improve the transfer learning technique and better generalizability between multiple data sources, we propose a multi-source adversarial transfer learning model, namely AMTLDC. In AMTLDC, representations are learned that are similar among the sources. In other words, extracted representations are general and not dependent on the particular dataset domain. We apply the AMTLDC to predict Covid-19 from medical images using a convolutional neural network. We show that accuracy can be improved using the AMTLDC framework, and surpass the results of current successful transfer learning approaches. In particular, we show that the AMTLDC works well when using different dataset domains, or when there is insufficient data.

Machine learning to analyse omic-data for COVID-19 diagnosis and prognosis

BACKGROUND

With the global spread of COVID-19, the world has seen many patients, including many severe cases. The rapid development of machine learning (ML) has made significant disease diagnosis and prediction achievements. Current studies have confirmed that omics data at the host level can reflect the development process and prognosis of the disease. Since early diagnosis and effective treatment of severe COVID-19 patients remains challenging, this research aims to use omics data in different ML models for COVID-19 diagnosis and prognosis. We used several ML models on omics data of a large number of individuals to first predict whether patients are COVID-19 positive or negative, followed by the severity of the disease.

RESULTS

On the COVID-19 diagnosis task, we got the best AUC of 0.99 with our multilayer perceptron model and the highest F1-score of 0.95 with our logistic regression (LR) model. For the severity prediction task, we achieved the highest accuracy of 0.76 with an LR model. Beyond classification and predictive modeling, our study founds ML models performed better on integrated multi-omics data, rather than single omics. By comparing top features from different omics dataset, we also found the robustness of our model, with a wider range of applicability in diverse dataset related to COVID-19. Additionally, we have found that omics-based models performed better than image or physiological feature-based models, proving the importance of the omics-based dataset for future model development.

CONCLUSIONS

This study diagnoses COVID-19 positive cases and predicts accurate severity levels. It lowers the dependence on clinical data and professional judgment, by leveraging the utilization of state-of-the-art models. our model showed wider applicability across different omics dataset, which is highly transferable in other respiratory or similar diseases. Hospital and public health care mechanisms can optimize the distribution of medical resources and improve the robustness of the medical system.

SARS-CoV-2 Diagnosis Using Transcriptome Data: A Machine Learning Approach

SARS-CoV-2 pandemic is the big issue of the whole world right now. The health community is struggling to rescue the public and countries from this spread, which revives time to time with different waves. Even the vaccination seems to be not prevents this spread. Accurate identification of infected people on time is essential these days to control the spread. So far, Polymerase chain reaction (PCR) and rapid antigen tests are widely used in this identification, accepting their own drawbacks. False negative cases are the menaces in this scenario. To avoid these problems, this study uses machine learning techniques to build a classification model with higher accuracy to filter the COVID-19 cases from the non-COVID individuals. Transcriptome data of the SARS-CoV-2 patients along with the control are used in this stratification using three different feature selection algorithms and seven classification models. Differently expressed genes also studied between these two groups of people and used in this classification. Results shows that mutual information (or DEGs) along with naïve Bayes (or SVM) gives the best accuracy (0.98 ± 0.04) among these methods.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42979-023-01703-6.

Capacitive immunosensor for COVID-19 diagnosis

COVID-19 has spread worldwide and early detection has been the key to controlling its propagation and preventing severe cases. However, diagnostic devices must be developed using different strategies to avoid a shortage of supplies needed for tests' fabrication caused by their large demand in pandemic situations. Furthermore, some tropical and subtropical countries are also facing epidemics of Dengue and Zika, viruses with similar symptoms in early stages and cross-reactivity in serological tests. Herein, we reported a qualitative immunosensor based on capacitive detection of spike proteins of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of COVID-19. The sensor device exhibited a good signal-to-noise ratio (SNR) at 1 kHz frequency, with an absolute value of capacitance variation significantly smaller for Dengue and Zika NS1 proteins (|ΔC| = 1.5 ± 1.0 nF and 1.8 ± 1.0 nF, respectively) than for the spike protein (|ΔC| = 7.0 ± 1.8 nF). Under the optimized conditions, the established biosensor is able to indicate that the sample contains target proteins when |ΔC| > 3.8 nF, as determined by the cut-off value (CO). This immunosensor was developed using interdigitated electrodes which require a measurement system with a simple electrical circuit that can be miniaturized to enable point-of-care detection, offering an alternative for COVID-19 diagnosis, especially in areas where there is also a co-incidence of Zika and Dengue.

CovidExpert: A Triplet Siamese Neural Network framework for the detection of COVID-19

Patients with the COVID-19 infection may have pneumonia-like symptoms as well as respiratory problems which may harm the lungs. From medical images, coronavirus illness may be accurately identified and predicted using a variety of machine learning methods. Most of the published machine learning methods may need extensive hyperparameter adjustment and are unsuitable for small datasets. By leveraging the data in a comparatively small dataset, few-shot learning algorithms aim to reduce the requirement of large datasets. This inspired us to develop a few-shot learning model for early detection of COVID-19 to reduce the post-effect of this dangerous disease. The proposed architecture combines few-shot learning with an ensemble of pre-trained convolutional neural networks to extract feature vectors from CT scan images for similarity learning. The proposed Triplet Siamese Network as the few-shot learning model classified CT scan images into Normal, COVID-19, and Community-Acquired Pneumonia. The suggested model achieved an overall accuracy of 98.719%, a specificity of 99.36%, a sensitivity of 98.72%, and a ROC score of 99.9% with only 200 CT scans per category for training data.

Learning from pseudo-lesion: a self-supervised framework for COVID-19 diagnosis

The Coronavirus disease 2019 (COVID-19) has rapidly spread all over the world since its first report in December 2019, and thoracic computed tomography (CT) has become one of the main tools for its diagnosis. In recent years, deep learning-based approaches have shown impressive performance in myriad image recognition tasks. However, they usually require a large number of annotated data for training. Inspired by ground glass opacity, a common finding in COIVD-19 patient's CT scans, we proposed in this paper a novel self-supervised pretraining method based on pseudo-lesion generation and restoration for COVID-19 diagnosis. We used Perlin noise, a gradient noise based mathematical model, to generate lesion-like patterns, which were then randomly pasted to the lung regions of normal CT images to generate pseudo-COVID-19 images. The pairs of normal and pseudo-COVID-19 images were then used to train an encoder-decoder architecture-based U-Net for image restoration, which does not require any labeled data. The pretrained encoder was then fine-tuned using labeled data for COVID-19 diagnosis task. Two public COVID-19 diagnosis datasets made up of CT images were employed for evaluation. Comprehensive experimental results demonstrated that the proposed self-supervised learning approach could extract better feature representation for COVID-19 diagnosis, and the accuracy of the proposed method outperformed the supervised model pretrained on large-scale images by 6.57% and 3.03% on SARS-CoV-2 dataset and Jinan COVID-19 dataset, respectively.

SARS-CoV-2 Detection and COVID-19 Diagnosis: A Bird's Eye View

The battle against severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) associated coronavirus disease 2019 (COVID-19) is continued worldwide by administering firsttime emergency authorized novel mRNA-based and conventional vector-antigen-based anti- COVID-19 vaccines to prevent further transmission of the virus as well as to reduce the severe respiratory complications of the infection in infected individuals. However; the emergence of numerous SARS-CoV-2 variants is of concern, and the identification of certain breakthrough and reinfection cases in vaccinated individuals as well as new cases soaring in some low-to-middle income countries (LMICs) and even in some resource-replete nations have raised concerns that only vaccine jabs would not be sufficient to control and vanquishing the pandemic. Lack of screening for asymptomatic COVID-19-infected subjects and inefficient management of diagnosed COVID-19 infections also pose some concerns and the need to fill the gaps among policies and strategies to reduce the pandemic in hospitals, healthcare services, and the general community. For this purpose, the development and deployment of rapid screening and diagnostic procedures are prerequisites in premises with high infection rates as well as to screen mass unaffected COVID-19 populations. Novel methods of variant identification and genome surveillance studies would be an asset to minimize virus transmission and infection severity. The proposition of this pragmatic review explores current paradigms for the screening of SARS-CoV-2 variants, identification, and diagnosis of COVID-19 infection, and insights into the late-stage development of new methods to better understand virus super spread variants and genome surveillance studies to predict pandemic trajectories.

The Capacity of Artificial Intelligence in COVID-19 Response: A Review in Context of COVID-19 Screening and Diagnosis

Artificial intelligence (AI) has been shown to solve several issues affecting COVID-19 diagnosis. This systematic review research explores the impact of AI in early COVID-19 screening, detection, and diagnosis. A comprehensive survey of AI in the COVID-19 literature, mainly in the context of screening and diagnosis, was observed by applying the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. Data sources for the years 2020, 2021, and 2022 were retrieved from google scholar, web of science, Scopus, and PubMed, with target keywords relating to AI in COVID-19 screening and diagnosis. After a comprehensive review of these studies, the results found that AI contributed immensely to improving COVID-19 screening and diagnosis. Some proposed AI models were shown to have comparable (sometimes even better) clinical decision outcomes, compared to experienced radiologists in the screening/diagnosing of COVID-19. Additionally, AI has the capacity to reduce physician work burdens and fatigue and reduce the problems of several false positives, associated with the RT-PCR test (with lower sensitivity of 60-70%) and medical imaging analysis. Even though AI was found to be timesaving and cost-effective, with less clinical errors, it works optimally under the supervision of a physician or other specialists.

Hydrogenated Boron Phosphide THz-Metamaterial-Based Biosensor for Diagnosing COVID-19: A DFT Coupled FEM Study

Recent reports focus on the hydrogenation engineering of monolayer boron phosphide and simultaneously explore its promising applications in nanoelectronics. Coupling density functional theory and finite element method, we investigate the bowtie triangle ring microstructure composed of boron phosphide with hydrogenation based on structural and performance analysis. We determine the carrier mobility of hydrogenated boron phosphide, reveal the effect of structural and material parameters on resonance frequencies, and discuss the variation of the electric field at the two tips. The results suggest that the mobilities of electrons for hydrogenated BP monolayer in the armchair and zigzag directions are 0.51 and 94.4 cm²·V-1·s-1, whereas for holes, the values are 136.8 and 175.15 cm²·V-1·s-1. Meanwhile, the transmission spectra of the bowtie triangle ring microstructure can be controlled by adjusting the length of the bowtie triangle ring microstructure and carrier density of hydrogenated BP. With the increasing length, the transmission spectrum has a red-shift and the electric field at the tips of equilateral triangle rings is significantly weakened. Furthermore, the theoretical sensitivity of the BTR structure reaches 100 GHz/RIU, which is sufficient to determine healthy and COVID-19-infected individuals. Our findings may open up new avenues for promising applications in the rapid diagnosis of COVID-19.

Advances in Biosensing Technologies for Diagnosis of COVID-19

The COVID-19 pandemic has severely impacted normal human life worldwide. Due to its rapid community spread and high mortality statistics, the development of prompt diagnostic tests for a massive number of samples is essential. Currently used traditional methods are often expensive, time-consuming, laboratory-based, and unable to handle a large number of specimens in resource-limited settings. Because of its high contagiousness, efficient identification of SARS-CoV-2 carriers is crucial. As the advantages of adopting biosensors for efficient diagnosis of COVID-19 increase, this narrative review summarizes the recent advances and the respective reasons to consider applying biosensors. Biosensors are the most sensitive, specific, rapid, user-friendly tools having the potential to deliver point-of-care diagnostics beyond traditional standards. This review provides a brief introduction to conventional methods used for COVID-19 diagnosis and summarizes their advantages and disadvantages. It also discusses the pathogenesis of COVID-19, potential diagnostic biomarkers, and rapid diagnosis using biosensor technology. The current advancements in biosensing technologies, from academic research to commercial achievements, have been emphasized in recent publications. We covered a wide range of topics, including biomarker detection, viral genomes, viral proteins, immune responses to infection, and other potential proinflammatory biomolecules. Major challenges and prospects for future application in point-of-care settings are also highlighted.

RESCOVIDTCNnet: A residual neural network-based framework for COVID-19 detection using TCN and EWT with chest X-ray images

Since the advent of COVID-19, the number of deaths has increased exponentially, boosting the requirement for various research studies that may correctly diagnose the illness at an early stage. Using chest X-rays, this study presents deep learning-based algorithms for classifying patients with COVID illness, healthy controls, and pneumonia classes. Data gathering, pre-processing, feature extraction, and classification are the four primary aspects of the approach. The pictures of chest X-rays utilized in this investigation came from various publicly available databases. The pictures were filtered to increase image quality in the pre-processing stage, and the chest X-ray images were de-noised using the empirical wavelet transform (EWT). Following that, four deep learning models were used to extract features. The first two models, Inception-V3 and Resnet-50, are based on transfer learning models. The Resnet-50 is combined with a temporal convolutional neural network (TCN) to create the third model. The fourth model is our suggested RESCOVIDTCNNet model, which integrates EWT, Resnet-50, and TCN. Finally, an artificial neural network (ANN) and a support vector machine were used to classify the data (SVM). Using five-fold cross-validation for 3-class classification, our suggested RESCOVIDTCNNet achieved a 99.5 percent accuracy. Our prototype can be utilized in developing nations where radiologists are in low supply to acquire a diagnosis quickly.

Monitoring of SARS-CoV-2 circulation using saliva testing in school children in Rome, Italy

Objectives

To describe the trend of SARS-CoV-2 RNA in saliva samples from children attending nine schools in Rome in the local surveillance unit RM3 during the period of September 2021-March 2022, in parallel with the trend of SARS-CoV-2 RNA observed in nasopharyngeal swabs (NPSs) from the population in the same catchment area that was routinely tested at our laboratory in the same period.

Methods

Saliva samples were collected using the Copan LolliSponge^TM device and analyzed by Aptima® SARS-CoV-2 Assay on the Panther® System. NPSs were tested using either Aptima® SARS-CoV-2 Assay or Alinity m SARS-CoV-2 Assay.

Results

The percentage of positivity in the two populations was different; of the 2222 saliva samples from students, 0.99% had positive results, whereas the percentage was higher (33.43%) in the 8994 NPSs representing the population from local surveillance unit RM3. Interestingly, the trend of SARS-CoV-2 RNA in saliva samples from students was consistent with that observed in NPSs from the population in same catchment area, although with peaks slightly anticipated.

Conclusion

Overall, screening of saliva in the schools represents a good system to monitor SARS-CoV-2 circulation in the population, allowing early detection and quick isolation of students who are asymptomatic with positive test results and thus prevention of virus transmission.

Evaluation of the relationship between the disease severity and the level of physical activity in patients followed up with COVID-19 diagnosis

OBJECTIVES

To investigate the relationship between physical activity level and disease severity, anxiety level, sleep quality, and fatigue in patients followed up with COVID-19 diagnosis.

METHODS

This was a cross-sectional study of 111 volunteer patients who were receiving treatment with COVID-19 diagnosis at the Chest Diseases Polyclinic, Sanko University, Sani Konukoğlu Practice and Research Hospital, Gaziantep, Turkey between May 2021 and July 2021 were included in the study and classified clinically and radiologically. They were evaluated on the basis of demographic characteristics, International Physical Activity Questionnaire, Beck Anxiety Inventory, Pittsburgh Sleep Quality Index, and Fatigue Severity Scale.

RESULTS

Approximately 63% of the patients did not have a habit of exercise, while 52.3% of our patients were clinically mild cases, and 33.3% had normal lung tomography. While clinical disease severity was not associated with exercise habits, sleep quality was impaired in clinically severe patients.

CONCLUSION

The results of our study suggested that physical inactivity is common. Anxiety is a frequent symptom in COVID-19 cases and also COVID-19 negatively affects sleep quality.

Comparison between Suspected and Confirmed COVID-19 Respiratory Patients: What Is beyond the PCR Test

COVID-19 modified the healthcare system. Nasal-pharyngeal swab (NPS), with real-time reverse transcriptase-polymerase (PCR), is the gold standard for the diagnosis; however, there are difficulties related to the procedure that may postpone it. The study aims to evaluate whether other elements than the PCR-NPS are reliable and confirm the diagnosis of COVID-19. This is a cross-sectional study on data from the Lung Unit of Pavia (confirmed) and at the Emergency Unit of Palermo (suspected). COVID-19 was confirmed by positive NPS, suspected tested negative. We compared clinical, laboratory and radiological variables and performed Logistic regression to estimate which variables increased the risk of COVID-19. The derived ROC-AUCcurve, assessed the accuracy of the model to distinguish between COVID-19 suspected and confirmed. We selected 50 confirmed and 103 suspected cases. High Reactive C-Protein (OR: 1.02; CI95%: 0.11–1.02), suggestive CT-images (OR: 11.43; CI95%: 3.01–43.3), dyspnea (OR: 10.48; CI95%: 2.08–52.7) and respiratory failure (OR: 5.84; CI95%: 1.73–19.75) increased the risk of COVID-19, whereas pleural effusion decreased the risk (OR: 0.15; CI95%: 0.04–0.63). ROC confirmed the discriminative role of these variables between suspected and confirmed COVID-19 (AUC 0.91). Clinical, laboratory and imaging features predict the diagnosis of COVID-19, independently from the NPS result.

Machine Learning-Based Research for COVID-19 Detection, Diagnosis, and Prediction: A Survey

The year 2020 experienced an unprecedented pandemic called COVID-19, which impacted the whole world. The absence of treatment has motivated research in all fields to deal with it. In Computer Science, contributions mainly include the development of methods for the diagnosis, detection, and prediction of COVID-19 cases. Data science and Machine Learning (ML) are the most widely used techniques in this area. This paper presents an overview of more than 160 ML-based approaches developed to combat COVID-19. They come from various sources like Elsevier, Springer, ArXiv, MedRxiv, and IEEE Xplore. They are analyzed and classified into two categories: Supervised Learning-based approaches and Deep Learning-based ones. In each category, the employed ML algorithm is specified and a number of used parameters is given. The parameters set for each of the algorithms are gathered in different tables. They include the type of the addressed problem (detection, diagnosis, or detection), the type of the analyzed data (Text data, X-ray images, CT images, Time series, Clinical data,...) and the evaluated metrics (accuracy, precision, sensitivity, specificity, F1-Score, and AUC). The study discusses the collected information and provides a number of statistics drawing a picture about the state of the art. Results show that Deep Learning is used in 79% of cases where 65% of them are based on the Convolutional Neural Network (CNN) and 17% use Specialized CNN. On his side, supervised learning is found in only 16% of the reviewed approaches and only Random Forest, Support Vector Machine (SVM) and Regression algorithms are employed.

Diagnostic performance of RT-PCR-based sample pooling strategy for the detection of SARS-CoV-2

BACKGROUND

The rapid spread of SARS-CoV-2 has created a shortage of supplies of reagents for its detection throughout the world, especially in Latin America. The pooling of samples consists of combining individual patient samples in a block and analyzing the group as a particular sample. This strategy has been shown to reduce the burden of laboratory material and logistical resources by up to 80%. Therefore, we aimed to evaluate the diagnostic performance of the pool of samples analyzed by RT-PCR to detect SARS-CoV-2.

METHODS

A cross-sectional study of diagnostic tests was carried out. We individually evaluated 420 samples, and 42 clusters were formed, each one with ten samples. These clusters could contain 0, 1 or 2 positive samples to simulate a positivity of 0, 10 and 20%, respectively. RT-PCR analyzed the groups for the detection of SARS-CoV-2. The area under the ROC curve (AUC), the Youden index, the global and subgroup sensitivity and specificity were calculated according to their Ct values that were classified as high (H: ≤ 25), moderate (M: 26-30) and low (L: 31-35) concentration of viral RNA.

RESULTS

From a total of 42 pools, 41 (97.6%) obtained the same result as the samples they contained (positive or negative). The AUC for pooling, Youden index, sensitivity, and specificity were 0.98 (95% CI, 0.95-1); 0.97 (95% CI, 0.90-1.03); 96.67% (95% CI; 88.58-100%) and 100% (95% CI; 95.83-100%) respectively. In the stratified analysis of the pools containing samples with Ct ≤ 25, the sensitivity was 100% (95% CI; 90-100%), while with the pools containing samples with Ct ≥ 31, the sensitivity was 80% (95% CI, 34.94-100%). Finally, a higher median was observed in the Ct of the clusters, with respect to the individual samples (p < 0.001).

CONCLUSIONS

The strategy of pooling nasopharyngeal swab samples for analysis by SARS-CoV-2 RT-PCR showed high diagnostic performance.

No significant association between COVID-19 diagnosis and the incidence of depression and anxiety disorder? A retrospective cohort study conducted in Germany

Little is known about the effects of coronavirus disease 2019 (COVID-19) on mental health compared with other respiratory infections. Thus, the aim of this retrospective cohort study was to investigate whether COVID-19 diagnosis is associated with a significant increase in the incidence of depression and anxiety disorder in patients followed in general practices in Germany compared with acute upper respiratory infection diagnosis. This study included all patients diagnosed with symptomatic or asymptomatic COVID-19 for the first time in 1198 general practices in Germany between March 2020 and May 2021. Patients diagnosed with acute upper respiratory infection were matched to those with COVID-19 using propensity scores based on sex, age, index month, and Charlson Comorbidity Index. The index date corresponded to the date on which either COVID-19 or acute upper respiratory infection was diagnosed. Differences in the incidence of depression and anxiety disorder between the COVID-19 and the acute upper respiratory infection group were studied using conditional Poisson regression models. This study included 56,350 patients diagnosed with COVID-19 and 56,350 patients diagnosed with acute upper respiratory infection (52.3% women; mean [SD] age 43.6 [19.2] years). The incidence of depression (IRR = 1.02, 95% CI = 0.95-1.10) and anxiety disorder (IRR = 0.94, 95% CI = 0.83-1.07) was not significantly higher in the COVID-19 group than in the upper respiratory infection group. Compared with acute upper respiratory infection diagnosis, COVID-19 diagnosis was not associated with a significant increase in the incidence of depression and anxiety disorder in patients treated in general practices in Germany.

WEENet: An Intelligent System for Diagnosing COVID-19 and Lung Cancer in IoMT Environments

The coronavirus disease 2019 (COVID-19) pandemic has caused a major outbreak around the world with severe impact on health, human lives, and economy globally. One of the crucial steps in fighting COVID-19 is the ability to detect infected patients at early stages and put them under special care. Detecting COVID-19 from radiography images using computational medical imaging method is one of the fastest ways to diagnose the patients. However, early detection with significant results is a major challenge, given the limited available medical imaging data and conflicting performance metrics. Therefore, this work aims to develop a novel deep learning-based computationally efficient medical imaging framework for effective modeling and early diagnosis of COVID-19 from chest x-ray and computed tomography images. The proposed work presents “WEENet” by exploiting efficient convolutional neural network to extract high-level features, followed by classification mechanisms for COVID-19 diagnosis in medical image data. The performance of our method is evaluated on three benchmark medical chest x-ray and computed tomography image datasets using eight evaluation metrics including a novel strategy of cross-corpse evaluation as well as robustness evaluation, and the results are surpassing state-of-the-art methods. The outcome of this work can assist the epidemiologists and healthcare authorities in analyzing the infected medical chest x-ray and computed tomography images, management of the COVID-19 pandemic, bridging the early diagnosis, and treatment gap for Internet of Medical Things environments.

Review and classification of AI-enabled COVID-19 CT imaging models based on computer vision tasks

This article presents a systematic overview of artificial intelligence (AI) and computer vision strategies for diagnosing the coronavirus disease of 2019 (COVID-19) using computerized tomography (CT) medical images. We analyzed the previous review works and found that all of them ignored classifying and categorizing COVID-19 literature based on computer vision tasks, such as classification, segmentation, and detection. Most of the COVID-19 CT diagnosis methods comprehensively use segmentation and classification tasks. Moreover, most of the review articles are diverse and cover CT as well as X-ray images. Therefore, we focused on the COVID-19 diagnostic methods based on CT images. Well-known search engines and databases such as Google, Google Scholar, Kaggle, Baidu, IEEE Xplore, Web of Science, PubMed, ScienceDirect, and Scopus were utilized to collect relevant studies. After deep analysis, we collected 114 studies and reported highly enriched information for each selected research. According to our analysis, AI and computer vision have substantial potential for rapid COVID-19 diagnosis as they could significantly assist in automating the diagnosis process. Accurate and efficient models will have real-time clinical implications, though further research is still required. Categorization of literature based on computer vision tasks could be helpful for future research; therefore, this review article will provide a good foundation for conducting such research.

The Rapid Antigen Detection Test for SARS-CoV-2 Underestimates the Identification of COVID-19 Positive Cases and Compromises the Diagnosis of the SARS-CoV-2 (K417N/T, E484K, and N501Y) Variants

Timely detection of severe acute respiratory syndrome due to coronavirus 2 (SARS-CoV-2) by reverse transcription quantitative polymerase chain reaction (RT-qPCR) has been the gold- strategy for identifying positive cases during the current pandemic. However, faster and less expensive methodologies are also applied for the massive diagnosis of COVID-19. In this way, the rapid antigen test (RAT) is widely used. However, it is necessary to evaluate its detection efficiency considering the current pandemic context with the circulation of new viral variants. In this study, we evaluated the sensitivity and specificity of RAT (SD BIOSENSOR, South Korea), widely used for testing and SARS-CoV-2 diagnosis in Santiago of Chile. The RAT showed a 90% (amplification range of 20 ≤ Cq <25) and 10% (amplification range of 25 ≤ Cq <30) of positive SARS-CoV-2 cases identified previously by RT-qPCR. Importantly, a 0% detection was obtained for samples within a Cq value>30. In SARS-CoV-2 variant detection, RAT had a 42.8% detection sensitivity in samples with RT-qPCR amplification range 20 ≤ Cq <25 containing the single nucleotide polymorphisms (SNP) K417N/T, N501Y and E484K, associated with beta or gamma SARS-CoV-2 variants. This study alerts for the special attention that must be paid for the use of RAT at a massive diagnosis level, especially in the current scenario of appearance of several new SARS-CoV-2 variants which could generate false negatives and the compromise of possible viral outbreaks.

New conductive filament ready-to-use for 3D-printing electrochemical (bio)sensors: Towards the detection of SARS-CoV-2

The 3D printing technology has gained ground due to its wide range of applicability. The development of new conductive filaments contributes significantly to the production of improved electrochemical devices. In this context, we report a simple method to producing an efficient conductive filament, containing graphite within the polymer matrix of PLA, and applied in conjunction with 3D printing technology to generate (bio)sensors without the need for surface activation. The proposed method for producing the conductive filament consists of four steps: (i) mixing graphite and PLA in a heated reflux system; (ii) recrystallization of the composite; (iii) drying and; (iv) extrusion. The produced filament was used for the manufacture of electrochemical 3D printed sensors. The filament and sensor were characterized by physicochemical techniques, such as SEM, TGA, Raman, FTIR as well as electrochemical techniques (EIS and CV). Finally, as a proof-of-concept, the fabricated 3D-printed sensor was applied for the determination of uric acid and dopamine in synthetic urine and used as a platform for the development of a biosensor for the detection of SARS-CoV-2. The developed sensors, without pre-treatment, provided linear ranges of 0.5-150.0 and 5.0-50.0 μmol L^-1, with low LOD values (0.07 and 0.11 μmol L^-1), for uric acid and dopamine, respectively. The developed biosensor successfully detected SARS-CoV-2 S protein, with a linear range from 5.0 to 75.0 nmol L^-1 (0.38 μg mL^-1 to 5.74 μg mL^-1) and LOD of 1.36 nmol L^-1 (0.10 μg mL^-1) and sensitivity of 0.17 μA nmol^-1 L (0.01 μA μg^-1 mL). Therefore, the lab-made produced and the ready-to-use conductive filament is promising and can become an alternative route for the production of different 3D electrochemical (bio)sensors and other types of conductive devices by 3D printing.

COVID-19 disease diagnosis from paper-based ECG trace image data using a novel convolutional neural network model

Clinical reports show that COVID-19 disease has impacts on the cardiovascular system in addition to the respiratory system. Available COVID-19 diagnostic methods have been shown to have limitations. In addition to current diagnostic methods such as low-sensitivity standard RT-PCR tests and expensive medical imaging devices, the development of alternative methods for the diagnosis of COVID-19 disease would be beneficial for control of the COVID-19 pandemic. Further, it is important to quickly and accurately detect abnormalities caused by COVID-19 on the cardiovascular system via ECG. In this study, the diagnosis of COVID-19 disease is proposed using a novel deep Convolutional Neural Network model by using only ECG trace images created from ECG signals of COVID-19 infected patients based on the abnormalities caused by the COVID-19 virus on the cardiovascular system. An overall classification accuracy of 98.57%, 93.20%, 96.74% and AUC value of 0.9966, 0.9771, 0.9905 is achieved for COVID-19 vs. Normal, COVID-19 vs. Abnormal Heartbeats, COVID-19 vs. Myocardial Infarction binary classification tasks, respectively. In addition, an overall classification accuracy of 86.55% and 83.05% is achieved for COVID-19 vs. Abnormal Heartbeats vs. Myocardial Infarction and Normal vs. COVID-19 vs. Abnormal Heartbeats vs. Myocardial Infarction multi-classification tasks. This study is believed to have great potential to speed up the diagnosis and treatment of COVID-19 patients, saving clinicians time and facilitating the control of the pandemic.

Factors influencing the performance of rapid SARS-CoV-2 antigen tests under field condition

Background

Globally, real‐time reverse transcription–polymerase chain reaction (rRT‐PCR) is the reference detection technique for SARS‐CoV‐2, which is expensive, time consuming, and requires trained laboratory personnel. Thus, a cost‐effective, rapid antigen test is urgently needed. This study evaluated the performance of the rapid antigen tests (RATs) for SARS‐CoV‐2 compared with rRT‐PCR, considering different influencing factors.

Methods

We enrolled a total of 214 symptomatic individuals with known COVID‐19 status using rRT‐PCR. We collected and tested paired nasopharyngeal (NP) and nasal swab (NS) specimens (collected from same individual) using rRT‐PCR and RATs (InTec and SD Biosensor). We assessed the performance of RATs considering specimen types, viral load, the onset of symptoms, and presenting symptoms.

Results

We included 214 paired specimens (112 NP and 100 NS SARS‐CoV‐2 rRT‐PCR positive) to the analysis. For NP specimens, the average sensitivity, specificity, and accuracy of the RATs were 87.5%, 98.6%, and 92.8%, respectively, when compared with rRT‐PCR. While for NS, the overall kit performance was slightly lower than that of NP (sensitivity 79.0%, specificity 96.1%, and accuracy 88.3%). We observed a progressive decline in the performance of RATs with increased Ct values (decreased viral load). Moreover, the RAT sensitivity using NP specimens decreased over the time of the onset of symptoms.

Conclusion

The RATs showed strong performance under field conditions and fulfilled the minimum performance limit for rapid antigen detection kits recommended by World Health Organization. The best performance of the RATs can be achieved within the first week of the onset of symptoms with high viral load.

Coronavirus Disease 2019 (COVID-19) Diagnostic Clinical Decision Support: A Pre-Post Implementation Study of CORAL (COvid Risk cALculator)

BACKGROUND

Isolation of hospitalized persons under investigation (PUIs) for coronavirus disease 2019 (COVID-19) reduces nosocomial transmission risk. Efficient evaluation of PUIs is needed to preserve scarce healthcare resources. We describe the development, implementation, and outcomes of an inpatient diagnostic algorithm and clinical decision support system (CDSS) to evaluate PUIs.

METHODS

We conducted a pre-post study of CORAL (COvid Risk cALculator), a CDSS that guides frontline clinicians through a risk-stratified COVID-19 diagnostic workup, removes transmission-based precautions when workup is complete and negative, and triages complex cases to infectious diseases (ID) physician review. Before CORAL, ID physicians reviewed all PUI records to guide workup and precautions. After CORAL, frontline clinicians evaluated PUIs directly using CORAL. We compared pre- and post-CORAL frequency of repeated severe acute respiratory syndrome coronavirus 2 nucleic acid amplification tests (NAATs), time from NAAT result to PUI status discontinuation, total duration of PUI status, and ID physician work hours, using linear and logistic regression, adjusted for COVID-19 incidence.

RESULTS

Fewer PUIs underwent repeated testing after an initial negative NAAT after CORAL than before CORAL (54% vs 67%, respectively; adjusted odd ratio, 0.53 [95% confidence interval, .44-.63]; P < .01). CORAL significantly reduced average time to PUI status discontinuation (adjusted difference [standard error], -7.4 [0.8] hours per patient), total duration of PUI status (-19.5 [1.9] hours per patient), and average ID physician work-hours (-57.4 [2.0] hours per day) (all P < .01). No patients had a positive NAAT result within 7 days after discontinuation of precautions via CORAL.

CONCLUSIONS

CORAL is an efficient and effective CDSS to guide frontline clinicians through the diagnostic evaluation of PUIs and safe discontinuation of precautions.

Diagnostic Approach for Accurate Diagnosis of COVID-19 Employing Deep Learning and Transfer Learning Techniques through Chest X-ray Images Clinical Data in E-Healthcare

COVID-19 is a transferable disease that is also a leading cause of death for a large number of people worldwide. This disease, caused by SARS-CoV-2, spreads very rapidly and quickly affects the respiratory system of the human being. Therefore, it is necessary to diagnosis this disease at the early stage for proper treatment, recovery, and controlling the spread. The automatic diagnosis system is significantly necessary for COVID-19 detection. To diagnose COVID-19 from chest X-ray images, employing artificial intelligence techniques based methods are more effective and could correctly diagnosis it. The existing diagnosis methods of COVID-19 have the problem of lack of accuracy to diagnosis. To handle this problem we have proposed an efficient and accurate diagnosis model for COVID-19. In the proposed method, a two-dimensional Convolutional Neural Network (2DCNN) is designed for COVID-19 recognition employing chest X-ray images. Transfer learning (TL) pre-trained ResNet-50 model weight is transferred to the 2DCNN model to enhanced the training process of the 2DCNN model and fine-tuning with chest X-ray images data for final multi-classification to diagnose COVID-19. In addition, the data augmentation technique transformation (rotation) is used to increase the data set size for effective training of the R2DCNNMC model. The experimental results demonstrated that the proposed (R2DCNNMC) model obtained high accuracy and obtained 98.12% classification accuracy on CRD data set, and 99.45% classification accuracy on CXI data set as compared to baseline methods. This approach has a high performance and could be used for COVID-19 diagnosis in E-Healthcare systems.

COVID-19 anomaly detection and classification method based on supervised machine learning of chest X-ray images

The term COVID-19 is an abbreviation of Coronavirus 2019, which is considered a global pandemic that threatens the lives of millions of people. Early detection of the disease offers ample opportunity of recovery and prevention of spreading. This paper proposes a method for classification and early detection of COVID-19 through image processing using X-ray images. A set of procedures are applied, including preprocessing (image noise removal, image thresholding, and morphological operation), Region of Interest (ROI) detection and segmentation, feature extraction, (Local binary pattern (LBP), Histogram of Gradient (HOG), and Haralick texture features) and classification (K-Nearest Neighbor (KNN) and Support Vector Machine (SVM)). The combinations of the feature extraction operators and classifiers results in six models, namely LBP-KNN, HOG-KNN, Haralick-KNN, LBP-SVM, HOG-SVM, and Haralick-SVM. The six models are tested based on test samples of 5,000 images with the percentage of training of 5-folds cross-validation. The evaluation results show high diagnosis accuracy from 89.2% up to 98.66%. The LBP-KNN model outperforms the other models in which it achieves an average accuracy of 98.66%, a sensitivity of 97.76%, specificity of 100%, and precision of 100%. The proposed method for early detection and classification of COVID-19 through image processing using X-ray images is proven to be usable in which it provides an end-to-end structure without the need for manual feature extraction and manual selection methods.

A novel deep neuroevolution-based image classification method to diagnose coronavirus disease (COVID-19)

COVID-19 has had a detrimental impact on normal activities, public safety, and the global financial system. To identify the presence of this disease within communities and to commence the management of infected patients early, positive cases should be diagnosed as quickly as possible. New results from X-ray imaging indicate that images provide key information about COVID-19. Advanced deep-learning (DL) models can be applied to X-ray radiological images to accurately diagnose this disease and to mitigate the effects of a shortage of skilled medical personnel in rural areas. However, the performance of DL models strongly depends on the methodology used to design their architectures. Therefore, deep neuroevolution (DNE) techniques are introduced to automatically design DL architectures accurately. In this paper, a new paradigm is proposed for the automated diagnosis of COVID-19 from chest X-ray images using a novel two-stage improved DNE Algorithm. The proposed DNE framework is evaluated on a real-world dataset and the results demonstrate that it provides the highest classification performance in terms of different evaluation metrics.