The virus SARS-CoV2, which causes coronavirus disease (COVID-19) has become a pandemic and has spread to every inhabited continent. Given the increasing caseload, there is an urgent need to augment clinical skills in order to identify from among the many mild cases the few that will progress to critical illness. We present a first step towards building an artificial intelligence (AI) framework, with predictive analytics (PA) capabilities applied to real patient data, to provide rapid clinical decision-making support. COVID-19 has presented a pressing need as a) clinicians are still developing clinical acumen to this novel disease and b) resource limitations in a surging pandemic require difficult resource allocation decisions. The objectives of this research are: (1) to algorithmically identify the combinations of clinical characteristics of COVID-19 that predict outcomes, and (2) to develop a tool with AI capabilities that will predict patients at risk for more severe illness on initial presentation. The predictive models learn from historical data to help predict who will develop acute respiratory distress syndrome (ARDS), a severe outcome in COVID-19. Our results, based on data from two hospitals in Wenzhou, Zhejiang, China, identified features on initial presentation with COVID-19 that were most predictive of later development of ARDS. A mildly elevated alanine aminotransferase (ALT) (a liver enzyme), the presence of myalgias (body aches), and an elevated hemoglobin (red blood cells), in this order, are the clinical features, on presentation, that are the most predictive. The predictive models that learned from historical data of patients from these two hospitals achieved 70% to 80% accuracy in predicting severe cases.

AbstractsThe clinical features and treatment of pulmonary tuberculosis patients with COVID-19 is unclear and understudied. Here, three pulmonary tuberculosis patients with COVID-19 infection were prospectively followed from hospital admission to discharge. We provide information and experience with treatment of pulmonary tuberculosis cases with confirmed COVID-19 infection.

Common lung diseases are first diagnosed via chest X-rays. Here, we show that a fully automated deep-learning pipeline for chest-X-ray-image standardization, lesion visualization and disease diagnosis can identify viral pneumonia caused by Coronavirus disease 2019 (COVID-19), assess its severity, and discriminate it from other types of pneumonia. The deep-learning system was developed by using a heterogeneous multicentre dataset of 145,202 images, and tested retrospectively and prospectively with thousands of additional images across four patient cohorts and multiple countries. The system generalized across settings, discriminating between viral pneumonia, other types of pneumonia and absence of disease with areas under the receiver operating characteristic curve (AUCs) of 0.88–0.99, between severe and non-severe COVID-19 with an AUC of 0.87, and between severe or non-severe COVID-19 pneumonia and other viral and non-viral pneumonia with AUCs of 0.82–0.98. In an independent set of 440 chest X-rays, the system performed comparably to senior radiologists, and improved the performance of junior radiologists. Automated deep-learning systems for the assessment of pneumonia could facilitate early intervention and provide clinical-decision support.

Background: Coronavirus disease 2019 (COVID-19) has infected more than 4 million people within 4 months. There is an urgent need to properly identify high-risk cases that are more likely to deteriorate even if they present mild diseases on admission. Methods: A multicenter nested case-control study was conducted in four designated hospitals in China enrolling confirmed COVID-19 patients who were mild on admission. Baseline clinical characteristics were compared between patients with stable mild illness (stable mild group) and those who deteriorated from mild to severe illness (progression group). Results: From Jan 17, 2020, to Feb 1, 2020, 85 confirmed COVID-19 patients were enrolled, including 16 in the progression group and 69 in the stable mild group. Compared to stable mild group (n = 69), patients in the progression group (n = 16) were more likely to be older, male, presented with dyspnea, with hypertension, and with higher levels of lactase dehydrogenase and c-reactive protein. In multivariate logistic regression analysis, advanced age (odds ratio [OR], 1.012; 95% confidence interval [CI], 1.020-1.166; P = 0.011) and the higher level of lactase dehydrogenase (OR, 1.012; 95% CI, 1.001-1.024; P = 0.038) were independently associated with exacerbation in mild COVID-19 patients. Conclusion: Advanced age and high LDH level are independent risk factors for exacerbation in mild COVID-19 patients. Among the mild patients, clinicians should pay more attention to the elderly patients or those with high LDH levels.

medRxiv preprint WHAT IS ALREADY KNOWN ON THIS TOPICThere are several reports about the serum antibodies against SARS-CoV-2. However, most of them evaluate diagnostic accuracy. Only two articles report dynamics of SARS-CoV-2 viral RNA and antibodies with serial samples, but the observation periods are within 30 days. None of the studies investigate the profiles of SARS-CoV-2 viral load and antibodies in a long period. Three reports investigate profiles in respiratory samples, but there are no reports on the dynamics of the viral load in stool samples. WHAT THIS STUDY ADDSIn both sputum and stool, SARS-CoV-2 RNA persists for a long time. The anti-RBD antibodies may involve in the clearance of SARS-CoV-2 infection. After eight weeks from symptom onset, IgM were negative in many of the previously positive patients, and IgG levels remained less than 50% of the peak levels in more than 20% of the patients. In about 40% of the patients, anti-RBD IgG levels increased 4-time higher in convalescence than in acute phase. Long persistence of SARS-CoV-2 viral RNA in sputum and stool presents challenges for management of the infection. The IgM/IgG comb test is better than single IgM test as a supplement diagnostic tool. Anti-RBD may be a protective antibody, and is valuable for development of vaccines. ABSTRACT OBJECTIVETo investigate the dynamics of viral RNA, IgM, and IgG and their relationships in patients with SARS-CoV-2 pneumonia over an 8-week period. DESIGNRetrospective, observational case series. SETTING Wenzhou Sixth People's HospitalPARTICIPANTS Thirty-three patients with laboratory confirmed SARS-CoV-2 pneumonia admitted to hospital. Data were collected from MAIN OUTCOME MEASURES Throat swabs, sputum, stool, and blood samples were collected, and viral load was measured by reverse transcription PCR (RT-PCR). Specific IgM and IgG against spike protein (S), spike protein receptor binding domain (RBD), and nucleocapsid (N) were analyzed. RESULTSAt the early stages of symptom onset, SARS-CoV-2 viral load is higher in throat swabs and sputum, but lower in stool. The median (IQR) time of undetectable viral RNA in throat swab, sputum, and stool was 18.5 (13.25-22) days, 22 (18.5-27.5) days, and 17 (11.5-32) days, respectively. In sputum, 17 patients (51.5%) had undetectable viral RNA within 22 days (short persistence), and 16 (48.5%) had persistent viral RNA more than 22 days (long persistence). Three patients (9.1%) had a detectable relapse of viral RNA in sputum within two weeks of their discharge from the hospital. One patient had persistent viral RNA for 59 days or longer. The median (IQR) seroconversion time of anti-S IgM, anti-RBD IgM, and anti-N IgM was 10.5 (7.75-15.5) days, 14 (9-24) days, and 10 (7-14) days, respectively. The median (IQR) seroconversion time of anti-S IgG, anti-RBD IgG, and anti-N IgG was 10 (7.25-16.5) days, 13 (9-17) days, and 10 (7-14) days, respectively. By week 8 after symptom onset, IgM were negative in many of the previously positive patients, and IgG levels remained less than 50% of the p...