This study developed and evaluated a deep learning model, COVNet, to detect COVID-19 from chest CT images. The model was trained on a dataset of 4,356 chest CT exams from 3,322 patients and tested on an independent set. COVNet achieved high sensitivity (90%) and specificity (96%) in detecting COVID-19, with an AUC of 0.96. It also performed well in distinguishing community-acquired pneumonia (CAP) from other non-pneumonic lung diseases, with an AUC of 0.95. The study highlights the potential of deep learning in accurately detecting and differentiating COVID-19 from other lung diseases using chest CT, which could aid in early diagnosis and treatment. However, the model's performance may be influenced by overlapping imaging features between different types of pneumonias, emphasizing the need for a multidisciplinary approach to final diagnosis.This study developed and evaluated a deep learning model, COVNet, to detect COVID-19 from chest CT images. The model was trained on a dataset of 4,356 chest CT exams from 3,322 patients and tested on an independent set. COVNet achieved high sensitivity (90%) and specificity (96%) in detecting COVID-19, with an AUC of 0.96. It also performed well in distinguishing community-acquired pneumonia (CAP) from other non-pneumonic lung diseases, with an AUC of 0.95. The study highlights the potential of deep learning in accurately detecting and differentiating COVID-19 from other lung diseases using chest CT, which could aid in early diagnosis and treatment. However, the model's performance may be influenced by overlapping imaging features between different types of pneumonias, emphasizing the need for a multidisciplinary approach to final diagnosis.