Chest CT is emerging as a valuable diagnostic tool for clinical management of COVID-19 associated lung disease. Artificial intelligence (AI) has the potential to aid in rapid evaluation of CT scans for differentiation of COVID-19 findings from other clinical entities. Here we show that a series of deep learning algorithms, trained in a diverse multinational cohort of 1280 patients to localize parietal pleura/lung parenchyma followed by classification of COVID-19 pneumonia, can achieve up to 90.8% accuracy, with 84% sensitivity and 93% specificity, as evaluated in an independent test set (not included in training and validation) of 1337 patients. Normal controls included chest CTs from oncology, emergency, and pneumonia-related indications. The false positive rate in 140 patients with laboratory confirmed other (non COVID-19) pneumonias was 10%. AI-based algorithms can readily identify CT scans with COVID-19 associated pneumonia, as well as distinguish non-COVID related pneumonias with high specificity in diverse patient populations.

Background The ongoing COVID-19 pandemic is a global threat. Identification of markers for symptom onset and disease progression is a pressing issue. We described the clinical features of people infected on board the Diamond Princess cruise ship who were diagnosed with asymptomatic severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection or mild or severe COVID-19, on admission to the Self-Defense Forces Central Hospital (Tokyo, Japan) and at the end of observation.Methods This retrospective, single-centre study included participants with laboratory-detected SARS-CoV-2 infection who were admitted to the Self-Defense Forces Central Hospital from Feb 11 to Feb 25, 2020. Clinical records, laboratory data, and radiological findings were analysed. Clinical outcomes were followed up until discharge or Feb 26, 2020, whichever came first. We defined asymptomatic infection as SARS-CoV-2 infection with no history of clinical signs and symptoms, severe COVID-19 as clinical symptoms of pneumonia (dyspnoea, tachypnoea, peripheral capillary oxygen saturation <93%, and need for oxygen therapy), and mild COVID-19 as all other symptoms. Clinical features on admission were compared among patients with different disease severity, including asymptomatic infection, at the end of observation. We used univariable analysis to identify factors associated with symptomatic illness among asymptomatic people infected with SARS-CoV-2 and disease progression in patients with COVID-19. FindingsAmong the 104 participants included in the final analysis, the median age was 68 years (IQR 47-75) and 54 (52%) were male. On admission, 43 (41%) participants were classified as asymptomatic, 41 (39%) as having mild COVID-10, and 20 (19%) as having severe COVID-19. At the end of observation, 33 (32%) participants were confirmed as being asymptomatic, 43 (41%) as having mild COVID-19, and 28 (27%) as having severe COVID-19. Serum lactate hydrogenase concentrations were significantly higher in the ten participants who were asymptomatic on admission but developed symptomatic COVID-19 compared with the 33 participants who remained asymptomatic throughout the observation period (five [50%] vs four [12%] participants; odds ratio 7•25, 95% CI 1•43-36•70; p=0•020). Compared with patients with mild disease at the end of observation, patients with severe COVID-19 were older (median age 73 years [IQR 55-77] vs 60 years [40-71]; p=0•028) and had more frequent consolidation on chest CT (13 [46%] of 28 vs nine [21%] of 43; p=0•035) and lymphopenia (16 [57%] vs ten [23%]; p=0•0055) on admission. Interpretation Older age, consolidation on chest CT images, and lymphopenia might be risk factors for disease progression of COVID-19 and contribute to improved clinical management. Funding None.

T he scientific, academic, medical and data science communities have come together in the face of the COVID-19 pandemic crisis to rapidly assess novel paradigms in artificial intelligence (AI) that are rapid and secure, and potentially incentivize data sharing and model training and testing without the usual privacy and data ownership hurdles of conventional collaborations 1,2 . Healthcare providers, researchers and industry have pivoted their focus to address unmet and critical clinical needs created by the crisis, with remarkable results [3][4][5][6][7][8][9] . Clinical trial recruitment has been expedited and facilitated by national regulatory bodies and an international cooperative spirit 10-12 . The data analytics and AI disciplines have always fostered open

Graphical abstract