Abstract

Coronavirus disease 2019 (Covid-19) is highly contagious with limitedtreatment options. Early and accurate diagnosis of Covid-19 is crucial inreducing the spread of the disease and its accompanied mortality. Currently,detection by reverse transcriptase polymerase chain reaction (RT-PCR) is thegold standard of outpatient and inpatient detection of Covid-19. RT-PCR is arapid method, however, its accuracy in detection is only ~70-75%. Anotherapproved strategy is computed tomography (CT) imaging. CT imaging has a muchhigher sensitivity of ~80-98%, but similar accuracy of 70%. To enhance theaccuracy of CT imaging detection, we developed an open-source set of algorithmscalled CovidCTNet that successfully differentiates Covid-19 fromcommunity-acquired pneumonia (CAP) and other lung diseases. CovidCTNetincreases the accuracy of CT imaging detection to 90% compared to radiologists(70%). The model is designed to work with heterogeneous and small sample sizesindependent of the CT imaging hardware. In order to facilitate the detection ofCovid-19 globally and assist radiologists and physicians in the screeningprocess, we are releasing all algorithms and parametric details in anopen-source format. Open-source sharing of our CovidCTNet enables developers torapidly improve and optimize services, while preserving user privacy and dataownership.