We aimed to retrospectively analyze the imaging changes detected in the follow-up of coronavirus disease 2019 (COVID-19) patients on thin-section computed tomography (CT). METHODSWe included 54 patients diagnosed with COVID-19. The mean interval between the initial and follow-up CT scans was 7.82±3.74 days. Patients were divided into progression and recovery groups according to their outcomes. We evaluated CT images in terms of distribution of lesions and imaging manifestations. The manifestations included ground-glass opacity (GGO), crazy-paving pattern, consolidation, irregular line, and air bronchogram sign. RESULTSCOVID-19 lesions showed mainly subpleural distribution, which was accompanied by bronchovascular bundle distribution in nearly 30% of the patients. The lower lobes of both lungs were the most commonly involved. In the follow-up, the progression group showed more involvement of the upper lobe of the left lung than the recovery group. GGO was the most common sign. As the disease progressed, round GGO decreased and patchy GGO increased. On follow-up CT, consolidation increased in the progression group while decreasing in the recovery group. Air bronchogram sign was more commonly observed at the initial examination (90.9%) than at follow-up (30%) in the recovery group, but there was no significant change in the progression group. Pleural effusion and lymphadenopathy were absent in the initial examination, but pleural effusion was observed in three cases after follow-up. CONCLUSIONAs COVID-19 progressed, round GGOs tended to evolve into patchy GGOs, consolidation increased, and pleural effusion could be occasionally observed. As COVID-19 resolved, the crazy-paving pattern and air bronchogram significantly decreased.

Highlights The incidence of venous thromboembolism is high among mild/moderate COVID-19 cases. Despite of thrombosis, tests for inflammatory, coagulation and biochemistry parameters were all in normal scope. The conventional method by Doppler ultrasound tend to underestimate the rate of thrombosis in comparisons to CTPA + CTV method.

These results suggest that CTPA is a practical and accurate means for evaluating RV dysfunction of pulmonary embolism in patients without any underlying cardiopulmonary disease and can discriminate between severe and non-severe APE.