Exploring the Potential of Artificial Intelligence CAD Software Application for Initial Screening of COVID-19 Patients

Chest radiograph is a frequent imaging of diagnosis for pulmonary diseases such as pneumonia. To assist radiologists in interpreting chest radiographs for pulmonary abnormalities, a deep-learning based lung nodule detection CAD(Computer Aided Detection) software was utilized. To validate its efficacy in screening or isolation of abnormal patients, a retrospective, multi-center study was performed. The results have demonstrated that the performance of deep-learning CAD software is comparable to that of possible methods recommended for COVID-19(coronavirus disease 2019) and the potential utility of the pre-trained commercial deep-learning CAD software for initial screening of infectious pulmonary disease as an assistant tool.