Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 647
Filter
Add filters

Document Type
Year range
1.
Comput Math Methods Med ; 2021: 9269173, 2021.
Article in English | MEDLINE | ID: covidwho-1511543

ABSTRACT

Early diagnosis of the harmful severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), along with clinical expertise, allows governments to break the transition chain and flatten the epidemic curve. Although reverse transcription-polymerase chain reaction (RT-PCR) offers quick results, chest X-ray (CXR) imaging is a more reliable method for disease classification and assessment. The rapid spread of the coronavirus disease 2019 (COVID-19) has triggered extensive research towards developing a COVID-19 detection toolkit. Recent studies have confirmed that the deep learning-based approach, such as convolutional neural networks (CNNs), provides an optimized solution for COVID-19 classification; however, they require substantial training data for learning features. Gathering this training data in a short period has been challenging during the pandemic. Therefore, this study proposes a new model of CNN and deep convolutional generative adversarial networks (DCGANs) that classify CXR images into normal, pneumonia, and COVID-19. The proposed model contains eight convolutional layers, four max-pooling layers, and two fully connected layers, which provide better results than the existing pretrained methods (AlexNet and GoogLeNet). DCGAN performs two tasks: (1) generating synthetic/fake images to overcome the challenges of an imbalanced dataset and (2) extracting deep features of all images in the dataset. In addition, it enlarges the dataset and represents the characteristics of diversity to provide a good generalization effect. In the experimental analysis, we used four distinct publicly accessible datasets of chest X-ray images (COVID-19 X-ray, COVID Chest X-ray, COVID-19 Radiography, and CoronaHack-Chest X-Ray) to train and test the proposed CNN and the existing pretrained methods. Thereafter, the proposed CNN method was trained with the four datasets based on the DCGAN synthetic images, resulting in higher accuracy (94.8%, 96.6%, 98.5%, and 98.6%) than the existing pretrained models. The overall results suggest that the proposed DCGAN-CNN approach is a promising solution for efficient COVID-19 diagnosis.


Subject(s)
Algorithms , COVID-19 Testing/methods , COVID-19/classification , COVID-19/diagnostic imaging , Deep Learning , SARS-CoV-2 , COVID-19 Testing/statistics & numerical data , Databases, Factual , Early Diagnosis , False Positive Reactions , Humans , Neural Networks, Computer , Pandemics , ROC Curve , Radiography, Thoracic/statistics & numerical data , Software Design , Tomography, X-Ray Computed/statistics & numerical data
2.
Sci Rep ; 11(1): 21564, 2021 11 03.
Article in English | MEDLINE | ID: covidwho-1500504

ABSTRACT

The Coronavirus has spread across the world and infected millions of people, causing devastating damage to the public health and global economies. To mitigate the impact of the coronavirus a reliable, fast, and accurate diagnostic system should be promptly implemented. In this study, we propose EpistoNet, a decision tree-based ensemble model using two mixtures of discriminative experts to classify COVID-19 lung infection from chest X-ray images. To optimize the architecture and hyper-parameters of the designed neural networks, we employed Epistocracy algorithm, a recently proposed hyper-heuristic evolutionary method. Using 2500 chest X-ray images consisting of 1250 COVID-19 and 1250 non-COVID-19 cases, we left out 500 images for testing and partitioned the remaining 2000 images into 5 different clusters using K-means clustering algorithm. We trained multiple deep convolutional neural networks on each cluster to help build a mixture of strong discriminative experts from the top-performing models supervised by a gating network. The final ensemble model obtained 95% accuracy on COVID-19 images and 93% accuracy on non-COVID-19. The experimental results show that EpistoNet can accurately, and reliably be used to detect COVID-19 infection in the chest X-ray images, and Epistocracy algorithm can be effectively used to optimize the hyper-parameters of the proposed models.


Subject(s)
COVID-19 , Deep Learning , Image Processing, Computer-Assisted , Radiography, Thoracic
3.
BMC Cardiovasc Disord ; 21(1): 522, 2021 10 29.
Article in English | MEDLINE | ID: covidwho-1486551

ABSTRACT

BACKGROUND: With the high prevalence of COVID-19 infections worldwide, the multisystem inflammatory syndrome in adults (MIS-A) is becoming an increasingly recognized entity. This syndrome presents in patients several weeks after infection with COVID-19 and is associated with thrombosis, elevated inflammatory markers, hemodynamic compromise and cardiac dysfunction. Treatment is often with steroids and intravenous immunoglobulin (IVIg). The pathologic basis of myocardial injury in MIS-A, however, is not well characterized. In our case report, we obtained endomyocardial biopsy that revealed a pattern of myocardial injury similar to that found in COVID-19 cardiac specimens. CASE PRESENTATION: A 26-year-old male presented with fevers, chills, headache, nausea, vomiting, and diarrhea 5 weeks after his COVID-19 infection. His SARS-CoV-2 PCR was negative and IgG was positive, consistent with prior infection. He was found to be in cardiogenic shock with biventricular failure, requiring inotropes and diuretics. Given concern for acute fulminant myocarditis, an endomyocardial biopsy (EMB) was performed, showing an inflammatory infiltrate consisting predominantly of interstitial macrophages with scant T lymphocytes. The histologic pattern was similar to that of cardiac specimens from COVID-19 patients, helping rule out myocarditis as the prevailing diagnosis. His case was complicated by persistent hypoxemia, and a computed tomography scan revealed pulmonary emboli. He received IVIg, steroids, and anticoagulation with rapid recovery of biventricular function. CONCLUSIONS: MIS-A should be considered as the diagnosis in patients presenting several weeks after COVID-19 infection with severe inflammation and multi-organ involvement. In our case, EMB facilitated identification of MIS-A and guided therapy. The patient's biventricular function recovered with IVIg and steroids.


Subject(s)
Anticoagulants/administration & dosage , COVID-19 , Myocarditis/diagnosis , Shock, Cardiogenic , Systemic Inflammatory Response Syndrome , Adult , Biopsy/methods , COVID-19/complications , COVID-19/diagnosis , COVID-19/drug therapy , COVID-19/immunology , COVID-19/physiopathology , Cardiotonic Agents/administration & dosage , Diagnosis, Differential , Diuretics/administration & dosage , Electrocardiography/methods , Humans , Immunoglobulins, Intravenous/administration & dosage , Male , Myocardium/pathology , Radiography, Thoracic/methods , SARS-CoV-2 , Shock, Cardiogenic/diagnosis , Shock, Cardiogenic/drug therapy , Shock, Cardiogenic/etiology , Shock, Cardiogenic/physiopathology , Systemic Inflammatory Response Syndrome/diagnosis , Systemic Inflammatory Response Syndrome/drug therapy , Systemic Inflammatory Response Syndrome/physiopathology , Treatment Outcome
4.
AJR Am J Roentgenol ; 217(5): 1093-1102, 2021 Nov.
Article in English | MEDLINE | ID: covidwho-1484970

ABSTRACT

BACKGROUND. Previous studies compared CT findings of COVID-19 pneumonia with those of other infections; however, to our knowledge, no studies to date have included noninfectious organizing pneumonia (OP) for comparison. OBJECTIVE. The objectives of this study were to compare chest CT features of COVID-19, influenza, and OP using a multireader design and to assess the performance of radiologists in distinguishing between these conditions. METHODS. This retrospective study included 150 chest CT examinations in 150 patients (mean [± SD] age, 58 ± 16 years) with a diagnosis of COVID-19, influenza, or non-infectious OP (50 randomly selected abnormal CT examinations per diagnosis). Six thoracic radiologists independently assessed CT examinations for 14 individual CT findings and for Radiological Society of North America (RSNA) COVID-19 category and recorded a favored diagnosis. The CT characteristics of the three diagnoses were compared using random-effects models; the diagnostic performance of the readers was assessed. RESULTS. COVID-19 pneumonia was significantly different (p < .05) from influenza pneumonia for seven of 14 chest CT findings, although it was different (p < .05) from OP for four of 14 findings (central or diffuse distribution was seen in 10% and 7% of COVID-19 cases, respectively, vs 20% and 21% of OP cases, respectively; unilateral distribution was seen in 1% of COVID-19 cases vs 7% of OP cases; non-tree-in-bud nodules was seen in 32% of COVID-19 cases vs 53% of OP cases; tree-in-bud nodules were seen in 6% of COVID-19 cases vs 14% of OP cases). A total of 70% of cases of COVID-19, 33% of influenza cases, and 47% of OP cases had typical findings according to RSNA COVID-19 category assessment (p < .001). The mean percentage of correct favored diagnoses compared with actual diagnoses was 44% for COVID-19, 29% for influenza, and 39% for OP. The mean diagnostic accuracy of favored diagnoses was 70% for COVID-19 pneumonia and 68% for both influenza and OP. CONCLUSION. CT findings of COVID-19 substantially overlap with those of influenza and, to a greater extent, those of OP. The diagnostic accuracy of the radiologists was low in a study sample that contained equal proportions of these three types of pneumonia. CLINICAL IMPACT. Recognized challenges in diagnosing COVID-19 by CT are furthered by the strong overlap observed between the appearances of COVID-19 and OP on CT. This challenge may be particularly evident in clinical settings in which there are substantial proportions of patients with potential causes of OP such as ongoing cancer therapy or autoimmune conditions.


Subject(s)
COVID-19/diagnostic imaging , Cryptogenic Organizing Pneumonia/diagnostic imaging , Influenza, Human/diagnostic imaging , Pneumonia, Viral/diagnostic imaging , Tomography, X-Ray Computed , Diagnosis, Differential , Female , Humans , Influenza, Human/virology , Male , Massachusetts , Middle Aged , Observer Variation , Pneumonia, Viral/virology , Radiography, Thoracic , Retrospective Studies , SARS-CoV-2
5.
Sci Rep ; 11(1): 20384, 2021 10 14.
Article in English | MEDLINE | ID: covidwho-1469995

ABSTRACT

Chest X-rays (CXRs) are the first-line investigation in patients presenting to emergency departments (EDs) with dyspnoea and are a valuable adjunct to clinical management of COVID-19 associated lung disease. Artificial intelligence (AI) has the potential to facilitate rapid triage of CXRs for further patient testing and/or isolation. In this work we develop an AI algorithm, CovIx, to differentiate normal, abnormal, non-COVID-19 pneumonia, and COVID-19 CXRs using a multicentre cohort of 293,143 CXRs. The algorithm is prospectively validated in 3289 CXRs acquired from patients presenting to ED with symptoms of COVID-19 across four sites in NHS Greater Glasgow and Clyde. CovIx achieves area under receiver operating characteristic curve for COVID-19 of 0.86, with sensitivity and F1-score up to 0.83 and 0.71 respectively, and performs on-par with four board-certified radiologists. AI-based algorithms can identify CXRs with COVID-19 associated pneumonia, as well as distinguish non-COVID pneumonias in symptomatic patients presenting to ED. Pre-trained models and inference scripts are freely available at https://github.com/beringresearch/bravecx-covid .


Subject(s)
COVID-19/diagnostic imaging , Lung/diagnostic imaging , Radiography, Thoracic/methods , Algorithms , Artificial Intelligence , COVID-19 Testing/methods , Emergency Service, Hospital , Humans , Neural Networks, Computer , Prospective Studies , SARS-CoV-2/isolation & purification , Sensitivity and Specificity
6.
PLoS One ; 16(10): e0257884, 2021.
Article in English | MEDLINE | ID: covidwho-1468160

ABSTRACT

Recent studies show the potential of artificial intelligence (AI) as a screening tool to detect COVID-19 pneumonia based on chest x-ray (CXR) images. However, issues on the datasets and study designs from medical and technical perspectives, as well as questions on the vulnerability and robustness of AI algorithms have emerged. In this study, we address these issues with a more realistic development of AI-driven COVID-19 pneumonia detection models by generating our own data through a retrospective clinical study to augment the dataset aggregated from external sources. We optimized five deep learning architectures, implemented development strategies by manipulating data distribution to quantitatively compare study designs, and introduced several detection scenarios to evaluate the robustness and diagnostic performance of the models. At the current level of data availability, the performance of the detection model depends on the hyperparameter tuning and has less dependency on the quantity of data. InceptionV3 attained the highest performance in distinguishing pneumonia from normal CXR in two-class detection scenario with sensitivity (Sn), specificity (Sp), and positive predictive value (PPV) of 96%. The models attained higher general performance of 91-96% Sn, 94-98% Sp, and 90-96% PPV in three-class compared to four-class detection scenario. InceptionV3 has the highest general performance with accuracy, F1-score, and g-mean of 96% in the three-class detection scenario. For COVID-19 pneumonia detection, InceptionV3 attained the highest performance with 86% Sn, 99% Sp, and 91% PPV with an AUC of 0.99 in distinguishing pneumonia from normal CXR. Its capability of differentiating COVID-19 pneumonia from normal and non-COVID-19 pneumonia attained 0.98 AUC and a micro-average of 0.99 for other classes.


Subject(s)
Artificial Intelligence , COVID-19/diagnostic imaging , Pneumonia/diagnostic imaging , Thorax/diagnostic imaging , Humans , Predictive Value of Tests , Radiography, Thoracic , Sensitivity and Specificity
9.
BMC Pregnancy Childbirth ; 21(1): 658, 2021 Sep 28.
Article in English | MEDLINE | ID: covidwho-1440917

ABSTRACT

BACKGROUND: Whilst the impact of Covid-19 infection in pregnant women has been examined, there is a scarcity of data on pregnant women in the Middle East. Thus, the aim of this study was to examine the impact of Covid-19 infection on pregnant women in the United Arab Emirates population. METHODS: A case-control study was carried out to compare the clinical course and outcome of pregnancy in 79 pregnant women with Covid-19 and 85 non-pregnant women with Covid-19 admitted to Latifa Hospital in Dubai between March and June 2020. RESULTS: Although Pregnant women presented with fewer symptoms such as fever, cough, sore throat, and shortness of breath compared to non-pregnant women; yet they ran a much more severe course of illness. On admission, 12/79 (15.2%) Vs 2/85 (2.4%) had a chest radiograph score [on a scale 1-6] of ≥3 (p-value = 0.0039). On discharge, 6/79 (7.6%) Vs 1/85 (1.2%) had a score ≥3 (p-value = 0.0438). They also had much higher levels of laboratory indicators of severity with values above reference ranges for C-Reactive Protein [(28 (38.3%) Vs 13 (17.6%)] with p < 0.004; and for D-dimer [32 (50.8%) Vs 3(6%)]; with p < 0.001. They required more ICU admissions: 10/79 (12.6%) Vs 1/85 (1.2%) with p=0.0036; and suffered more complications: 9/79 (11.4%) Vs 1/85 (1.2%) with p=0.0066; of Covid-19 infection, particularly in late pregnancy. CONCLUSIONS: Pregnant women presented with fewer Covid-19 symptoms but ran a much more severe course of illness compared to non-pregnant women with the disease. They had worse chest radiograph scores and much higher levels of laboratory indicators of disease severity. They had more ICU admissions and suffered more complications of Covid-19 infection, such as risk for miscarriage and preterm deliveries. Pregnancy with Covid-19 infection, could, therefore, be categorised as high-risk pregnancy and requires management by an obstetric and medical multidisciplinary team.


Subject(s)
COVID-19 , Intensive Care Units/statistics & numerical data , Pregnancy Complications, Infectious , Premature Birth , Radiography, Thoracic , Symptom Assessment , Abortion, Spontaneous/epidemiology , Abortion, Spontaneous/etiology , C-Reactive Protein/analysis , COVID-19/blood , COVID-19/epidemiology , COVID-19/therapy , COVID-19/transmission , Case-Control Studies , Female , Fibrin Fibrinogen Degradation Products/analysis , Humans , Infant, Newborn , Infectious Disease Transmission, Vertical/prevention & control , Male , Pregnancy , Pregnancy Complications, Infectious/epidemiology , Pregnancy Complications, Infectious/physiopathology , Pregnancy Complications, Infectious/therapy , Pregnancy Complications, Infectious/virology , Pregnancy Outcome/epidemiology , Pregnancy, High-Risk , Premature Birth/epidemiology , Premature Birth/etiology , Radiography, Thoracic/methods , Radiography, Thoracic/statistics & numerical data , SARS-CoV-2/isolation & purification , Severity of Illness Index , Symptom Assessment/methods , Symptom Assessment/statistics & numerical data , United Arab Emirates/epidemiology
10.
Ghana Med J ; 54(4 Suppl): 46-51, 2020 Dec.
Article in English | MEDLINE | ID: covidwho-1436194

ABSTRACT

Introduction: The novel corona virus disease 2019 (COVID-19) was diagnosed in Wuhan, China in December 2019 and, in Ghana, in March 2020. As of 30th July 2020, Ghana had recorded 35,142 cases. COVID-19 which can be transmitted by both symptomatic and asymptomatic individuals usually manifest as pneumonia with symptoms like fever, cough, dyspnoea and fatigue. The current non-availability of a vaccine or drug for COVID-19 management calls for early detection and isolation of affected individuals. Chest imaging has become an integral part of patient management with chest radiography serving as a primary imaging modality in many centres. Methods: The study was a retrospective study conducted at Ga East Municipal Hospital (GEMH). Chest radiographs of patients with mild to moderate disease managed at GEMH were evaluated. The age, gender, symptom status, comorbidities and chest x-ray findings of the patients were documented. Results: 11.4 % of the patients had some form of respiratory abnormality on chest radiography with 88.9% showing COVID-19 pneumonia features. 93.8% showed ground glass opacities (GGO), with 3.1% each showing consolidation (CN) only and CN with GGO. There was a significant association between COVID-19 radiographic features and patient's age, symptom status and comorbidities but not with gender. Conclusion: Most radiographs were normal with only 11% showing COVID-19-like abnormality. There was a significant association between age, symptom status and comorbidities with the presence of COVID-19 like features but not for gender. There was no association between the extent of the lung changes and patient characteristics. Funding: None declared.


Subject(s)
COVID-19/diagnostic imaging , Radiography, Thoracic/statistics & numerical data , SARS-CoV-2 , Adolescent , Adult , Age Factors , Aged , COVID-19/epidemiology , Comorbidity , Female , Ghana/epidemiology , Hospitals, Urban , Humans , Lung/diagnostic imaging , Lung/virology , Male , Middle Aged , Retrospective Studies , Severity of Illness Index , Symptom Assessment/methods , Young Adult
11.
Intern Med ; 60(18): 2911-2917, 2021 Sep 15.
Article in English | MEDLINE | ID: covidwho-1413645

ABSTRACT

Objective Severe acute respiratory syndrome coronavirus 2 has spread globally, and it is important to utilize medical resources properly, especially in critically ill patients. We investigated the validity of chest radiography as a tool for predicting aggravation in coronavirus disease (COVID-19) cases. Methods A total of 104 laboratory-confirmed COVID-19 cases were referred from the cruise ship "Diamond Princess" to the Self-Defense Forces Central Hospital in Japan from February 11 to 25, 2020. Fifty-nine symptomatic patients were selected. Chest radiography was performed upon hospitalization; subsequently, patients were categorized into the positive radiograph (Group A) and negative radiograph (Group B) groups. Radiographic findings were analyzed with a six-point semiquantitative score. Group A was further classified into two additional subgroups: patients who required oxygen therapy during their clinical courses (Group C) and patients who did not (Group D). Clinical records, laboratory data, and radiological findings were collected for an analysis. Results Among 59 patients, 34 were men with a median age of 60 years old. Groups A, B, C, and D consisted of 33, 26, 12, and 21 patients, respectively. The number of patients requiring oxygen administration was significantly larger in Group A than in Group B. The consolidation score on chest radiographs was significantly higher in Group C than in Group D. When chest radiographs showed consolidation in more than two lung fields, the positive likelihood ratio of deterioration was 10.6. Conclusions Chest radiography is a simple and easy-to-use clinic-level triage tool for predicting the severity of COVID-19 and may contribute to the allocation of medical resources.


Subject(s)
COVID-19 , Triage , Humans , Male , Middle Aged , Primary Health Care , Radiography , Radiography, Thoracic , Retrospective Studies , SARS-CoV-2
12.
BMC Pulm Med ; 21(1): 293, 2021 Sep 16.
Article in English | MEDLINE | ID: covidwho-1412819

ABSTRACT

BACKGROUND: Re-expansion pulmonary edema is an uncommon complication following drainage of a pneumothorax or pleural effusion. While pneumothorax is noted to complicate COVID-19 patients, no case of COVID-19 developing re-expansion pulmonary edema has been reported. CASE REPRESENTATION: A man in his early 40 s without a smoking history and underlying pulmonary diseases suddenly complained of left chest pain with dyspnea 1 day after being diagnosed with COVID-19. Chest X-ray revealed pneumothorax in the left lung field, and a chest tube was inserted into the intrathoracic space without negative pressure 9 h after the onset of chest pain, resulting in the disappearance of respiratory symptoms; however, 2 h thereafter, dyspnea recurred with lower oxygenation status. Chest X-ray revealed improvement of collapse but extensive infiltration in the expanded lung. Therefore, the patient was diagnosed with re-expansion pulmonary edema, and his dyspnea and oxygenation status gradually improved without any intervention, such as steroid administration. Abnormal lung images also gradually improved within several days. CONCLUSIONS: This case highlights the rare presentation of re-expansion pulmonary edema following pneumothorax drainage in a patient with COVID-19, which recovered without requiring treatment for viral pneumonia. Differentiating re-expansion pulmonary edema from viral pneumonia is crucial to prevent unnecessary medication for COVID-19 pneumonia and pneumothorax.


Subject(s)
COVID-19/complications , Chest Tubes , Pneumothorax/therapy , Pulmonary Edema/etiology , Adult , COVID-19/diagnosis , Humans , Male , Radiography, Thoracic , SARS-CoV-2/isolation & purification , Tomography, X-Ray Computed
13.
Int J Med Sci ; 18(15): 3395-3402, 2021.
Article in English | MEDLINE | ID: covidwho-1409696

ABSTRACT

Computed tomography (CT) of the chest is one of the main diagnositic tools for coronavirus disease 2019 (COVID-19) infection. To document the chest CT findings in patients with confirmed COVID-19 and their association with the clinical severity, we searched related literatures through PubMed, MEDLINE, Embase, Web of Science (inception to May 4, 2020) and reviewed reference lists of previous systematic reviews. A total of 31 case reports (3768 patients) on CT findings of COVID-19 were included. The most common comorbid conditions were hypertension (18.4%) and diabetes mellitus (8.3%). The most common symptom was fever (78.7%), followed by cough (60.2%). It took an average of 5.6 days from symptom onset to admission. The most common chest CT finding was vascular enlargement (84.8%), followed by ground-glass opacity (GGO) (60.1%), air-bronchogram (47.8%), and consolidation (41.4%). Most lung lesions were located in the lung periphery (72.2%) and involved bilateral lung (76%). Most patients showed normal range of laboratory findings such as white blood cell count (96.4%) and lymphocyte (87.2%). Compared to previous published meta-analyses, our study is the first to summarize the different radiologic characteristics of chest CT in a total of 3768 COVID-19 patients by compiling case series studies. A comprehensive diagnostic approach should be adopted for patients with known COVID-19, suspected cases, and for exposed individuals.


Subject(s)
COVID-19/diagnostic imaging , Radiography, Thoracic/methods , Tomography, X-Ray Computed/methods , COVID-19/blood , Humans , Lung/diagnostic imaging , Lung Diseases/diagnostic imaging , Lymphocyte Count , Oxygen/therapeutic use , Prognosis
15.
Br J Radiol ; 94(1126): 20210221, 2021 Oct 01.
Article in English | MEDLINE | ID: covidwho-1406740

ABSTRACT

OBJECTIVES: For optimal utilization of healthcare resources, there is a critical need for early identification of COVID-19 patients at risk of poor prognosis as defined by the need for intensive unit care and mechanical ventilation. We tested the feasibility of chest X-ray (CXR)-based radiomics metrics to develop machine-learning algorithms for predicting patients with poor outcomes. METHODS: In this Institutional Review Board (IRB) approved, Health Insurance Portability and Accountability Act (HIPAA) compliant, retrospective study, we evaluated CXRs performed around the time of admission from 167 COVID-19 patients. Of the 167 patients, 68 (40.72%) required intensive care during their stay, 45 (26.95%) required intubation, and 25 (14.97%) died. Lung opacities were manually segmented using ITK-SNAP (open-source software). CaPTk (open-source software) was used to perform 2D radiomics analysis. RESULTS: Of all the algorithms considered, the AdaBoost classifier performed the best with AUC = 0.72 to predict the need for intubation, AUC = 0.71 to predict death, and AUC = 0.61 to predict the need for admission to the intensive care unit (ICU). AdaBoost had similar performance with ElasticNet in predicting the need for admission to ICU. Analysis of the key radiomic metrics that drive model prediction and performance showed the importance of first-order texture metrics compared to other radiomics panel metrics. Using a Venn-diagram analysis, two first-order texture metrics and one second-order texture metric that consistently played an important role in driving model performance in all three outcome predictions were identified. CONCLUSIONS: Considering the quantitative nature and reliability of radiomic metrics, they can be used prospectively as prognostic markers to individualize treatment plans for COVID-19 patients and also assist with healthcare resource management. ADVANCES IN KNOWLEDGE: We report on the performance of CXR-based imaging metrics extracted from RT-PCR positive COVID-19 patients at admission to develop machine-learning algorithms for predicting the need for ICU, the need for intubation, and mortality, respectively.


Subject(s)
COVID-19/diagnostic imaging , Machine Learning , Pneumonia, Viral/diagnostic imaging , Radiography, Thoracic , Adult , Aged , COVID-19/therapy , Critical Care/statistics & numerical data , Early Diagnosis , Female , Health Services Needs and Demand , Humans , Male , Middle Aged , Pneumonia, Viral/therapy , Pneumonia, Viral/virology , Predictive Value of Tests , Prognosis , Respiration, Artificial/statistics & numerical data , Retrospective Studies , SARS-CoV-2
16.
Eur Respir J ; 58(3)2021 09.
Article in English | MEDLINE | ID: covidwho-1403207

ABSTRACT

INTRODUCTION: For the management of patients referred to respiratory triage during the early stages of the severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) pandemic, either chest radiography or computed tomography (CT) were used as first-line diagnostic tools. The aim of this study was to compare the impact on the triage, diagnosis and prognosis of patients with suspected COVID-19 when clinical decisions are derived from reconstructed chest radiography or from CT. METHODS: We reconstructed chest radiographs from high-resolution CT (HRCT) scans. Five clinical observers independently reviewed clinical charts of 300 subjects with suspected COVID-19 pneumonia, integrated with either a reconstructed chest radiography or HRCT report in two consecutive blinded and randomised sessions: clinical decisions were recorded for each session. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and prognostic value were compared between reconstructed chest radiography and HRCT. The best radiological integration was also examined to develop an optimised respiratory triage algorithm. RESULTS: Interobserver agreement was fair (Kendall's W=0.365, p<0.001) by the reconstructed chest radiography-based protocol and good (Kendall's W=0.654, p<0.001) by the CT-based protocol. NPV assisted by reconstructed chest radiography (31.4%) was lower than that of HRCT (77.9%). In case of indeterminate or typical radiological appearance for COVID-19 pneumonia, extent of disease on reconstructed chest radiography or HRCT were the only two imaging variables that were similarly linked to mortality by adjusted multivariable models CONCLUSIONS: The present findings suggest that clinical triage is safely assisted by chest radiography. An integrated algorithm using first-line chest radiography and contingent use of HRCT can help optimise management and prognostication of COVID-19.


Subject(s)
COVID-19 , Triage , Humans , Lung/diagnostic imaging , Radiography , Radiography, Thoracic , SARS-CoV-2 , Tomography, X-Ray Computed
18.
Sci Rep ; 11(1): 15523, 2021 09 01.
Article in English | MEDLINE | ID: covidwho-1392879

ABSTRACT

Chest radiography (CXR) is the most widely-used thoracic clinical imaging modality and is crucial for guiding the management of cardiothoracic conditions. The detection of specific CXR findings has been the main focus of several artificial intelligence (AI) systems. However, the wide range of possible CXR abnormalities makes it impractical to detect every possible condition by building multiple separate systems, each of which detects one or more pre-specified conditions. In this work, we developed and evaluated an AI system to classify CXRs as normal or abnormal. For training and tuning the system, we used a de-identified dataset of 248,445 patients from a multi-city hospital network in India. To assess generalizability, we evaluated our system using 6 international datasets from India, China, and the United States. Of these datasets, 4 focused on diseases that the AI was not trained to detect: 2 datasets with tuberculosis and 2 datasets with coronavirus disease 2019. Our results suggest that the AI system trained using a large dataset containing a diverse array of CXR abnormalities generalizes to new patient populations and unseen diseases. In a simulated workflow where the AI system prioritized abnormal cases, the turnaround time for abnormal cases reduced by 7-28%. These results represent an important step towards evaluating whether AI can be safely used to flag cases in a general setting where previously unseen abnormalities exist. Lastly, to facilitate the continued development of AI models for CXR, we release our collected labels for the publicly available dataset.


Subject(s)
COVID-19/diagnostic imaging , Radiographic Image Interpretation, Computer-Assisted/methods , Tuberculosis/diagnostic imaging , Adult , Aged , Algorithms , Case-Control Studies , China , Deep Learning , Female , Humans , India , Male , Middle Aged , Radiography, Thoracic , United States
19.
Int J Med Sci ; 18(2): 520-527, 2021.
Article in English | MEDLINE | ID: covidwho-1389720

ABSTRACT

Background: Multiple societies including the Fleischner Society do not recommend that CT is routinely used in asymptomatic SARS-CoV-2 infections; however, this advice is based on the limited evidence. In this study, we aim to confirm whether it is necessary to do CT scans in SARS-CoV-2 asymptomatic infections by summarizing the longitudinal chest CT and clinical features of asymptomatic SARS-CoV-2 infections. Methods: A total of 33 individuals (14 men and 19 women) with asymptomatic SARS-CoV-2 infections were retrospectively enrolled. Clinical data of CT positive and negative groups were compared. Longitudinal chest CT scans were reviewed for CT features and analyzed for temporal change. Results: Thirty-two (97%) individuals had positive results for first RT-PCR testing. For clinical data, only monocyte count showed a significant difference between CT positive and negative groups. For first chest CT, only eighteen (54.5%) individuals had abnormal manifestations, common CT features were GGO (88.9%) and consolidation (33.3%), the median number of segments involved was 3.0 (1.0-7.5). No case in CT negative group was abnormal on the follow-up CT. Three patterns of evolution throughout series of CT were observed in CT positive group, including gradual improvement (12, 66.7%), mismatch to improvement (3, 16.7%) and mild progression to improvement (3, 16.7%). On last CT scans, most cases had radiographic improvement but residual abnormalities. Significant differences were exhibited in density, long diameter, number of lung segments involved, and percentage of consolidation between the first and last CT scans. All cases had stable conditions and finally confirmed negative for SARS-CoV-2 RT-PCR tests without developing into severe pneumonia. Conclusion: Considering poor performance of CT in screening, stable conditions during followup, and good outcomes in asymptomatic SARS-CoV-2 infections, we confirm that it is unnecessary to do CT scans in asymptomatic SARS-CoV-2 infections.


Subject(s)
Asymptomatic Infections , COVID-19/diagnostic imaging , Radiography, Thoracic , Tomography, X-Ray Computed , Adult , Female , Humans , Longitudinal Studies , Male , Middle Aged , Retrospective Studies , Unnecessary Procedures
SELECTION OF CITATIONS
SEARCH DETAIL
...