ABSTRACT
Since the beginning of the COVID-19 pandemic, more than 350 million cases and 5 million deaths have occurred. Since day one, multiple methods have been provided to diagnose patients who have been infected. Alongside the gold standard of laboratory analyses, deep learning algorithms on chest X-rays (CXR) have been developed to support the COVID-19 diagnosis. The literature reports that convolutional neural networks (CNNs) have obtained excellent results on image datasets when the tests are performed in cross-validation, but such models fail to generalize to unseen data. To overcome this limitation, we exploit the strength of multiple CNNs by building an ensemble of classifiers via an optimized late fusion approach. To demonstrate the system’s robustness, we present different experiments on open source CXR datasets to simulate a real-world scenario, where scans of patients affected by various lung pathologies and coming from external datasets are tested. Promising performances are obtained both in cross-validation and in external validation, obtaining an average accuracy of 93.02% and 91.02%, respectively. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
ABSTRACT
Since the emergence of the disease in late December 2019, numerous studies have been published to date regarding clinical, laboratory and treatment aspects associated with COVID-19. The present study attempts to compare and unify the clinical, para-clinical and therapeutic aspects that have come to light regarding coronavirus disease-19 (COVID 19), mainly in adults. Between April 2020 and September 2021, a comprehensive systematic literature review was performed, which we added to from our own medical experiences. The search was performed on the PubMed, Scopus and Google Scholar databases, comprising studies with analyzable data that were identified alongside studies and documents containing general scientific data. All published studies were written in English, and were from different countries. A 95% confidence interval (CI95) was also calculated for almost each study using the Wilson formula. When compared with preliminary reports between December 2019 and January 2020, the most frequent symptoms were still identified as being fever (68.6%; CI95: 67.5-69.7) and cough (72.7%; CI95: 71.7-73.8). Nevertheless, asymptomatic cases also increased (by 21.4%; CI95: 16.6-27.1). Severe and critical cases accounted for 10.4% (CI95: 9.6-11.1) of all cases. The mean fatality rate was found to be 4% (CI95: 3.6-4.5). The primary co-morbidity found was hypertension (28.9%; CI95: 27-30.8), followed by other underlying cardiovascular diseases (15.4%; CI95: 13.9-16.9) and diabetes (14.5%; CI95: 13.1-16.1). The majority of studies showed lower white blood cell numbers with neutropenia and lymphopenia, and lower platelet levels. The levels of the biomarkers C-reaction protein and erythrocyte sedimentation rate were positive in all studied cases alongside other lab tests, such as examining the D-dimer levels and those of other hepatic, cardiac and renal injury markers. The procalcitonin level was also found to be elevated in many cases, resulting in high usage of antibiotics (83.7%; CI95: 81.2-85.9). Approximately 31.6% (CI95: 29.1-34.1) of the patients required non-invasive ventilation, whereas 9.9% (CI95: 8.1-12.1) of the patients were intubated or placed on extracorporeal membrane oxygenation. The most used antivirals were ribavirin (67.3%; CI95: 63.4-70.9), oseltamivir (52.5%; CI95: 49.4-55.5) and Arbidol™ (34.5%; CI95: 32-37.1). General admittance to the intensive care unit was ~7.2% (CI95: 6.5-7.9) of patients.
ABSTRACT
OBJECTIVES: Vitamin D deficiency was previously correlated with incidence and severity of coronavirus disease 2019 (COVID-19). We investigated the association between serum 25-hydroxyvitamin D (25(OH)D) level on admission and radiologic stage and outcome of COVID-19 pneumonia. METHODS: A retrospective observational trial was done on 186 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-infected individuals hospitalized from March 1, 2020, to April 7, 2020, with combined chest computed tomography (CT) and 25(OH)D measurement on admission. Multivariate regression analysis was performed to study if vitamin D deficiency (25(OH)D <20 ng/mL) correlates with survival independently of confounding comorbidities. RESULTS: Of the patients with COVID-19, 59% were vitamin D deficient on admission: 47% of females and 67% of males. In particular, male patients with COVID-19 showed progressively lower 25(OH)D with advancing radiologic stage, with deficiency rates increasing from 55% in stage 1 to 74% in stage 3. Vitamin D deficiency on admission was not confounded by age, ethnicity, chronic lung disease, coronary artery disease/hypertension, or diabetes and was associated with mortality (odds ratio [OR], 3.87; 95% confidence interval [CI], 1.30-11.55), independent of age (OR, 1.09; 95% CI, 1.03-1.14), chronic lung disease (OR, 3.61; 95% CI, 1.18-11.09), and extent of lung damage expressed by chest CT severity score (OR, 1.12; 95% CI, 1.01-1.25). CONCLUSIONS: Low 25(OH)D levels on admission are associated with COVID-19 disease stage and mortality.