CT-Based Quantitative Assessment of Coronavirus Disease 2019 Using a Deep Learning-Based Segmentation System: A Longitudinal Study

Abstract BackgroundCoronavirus disease 2019 (COVID-19) is a global catastrophic disease that has severely affected more than 185 countries. The key steps in fighting against COVID-19 involve early detection and tracking of the treatment effects. A large number of studies highlighted computed tomography (CT) as a reliable method for early diagnosis and follow-up monitoring of the disease. However, there are limited data on quantitative analysis of the follow-up images. In this study, we used a deep learning model using a neural network with high accuracy in automatic segmentation and quantification to analyze the infected lesions on chest CT images.MethodsWe used a deep learning model using a neural network with high accuracy in automatic segmentation and quantification to analyze the infected lesions on chest CT images. A total of 14 patients (mean age, 53±14 years; age range, 23–74 years; 42.9% men and 57.1% women) with confirmed mild-type COVID-19 from January 1 to May 7, 2020, were retrospectively reviewed. Initial and follow-up original CT images were collected, and CT quantitative parameters, including percentage of infection (POI) and density variation of pneumonia, were determined.ResultsThe median initial POI was 3.4% (interquartile range, IQR 0.5%–8.4%) for the whole lung, 0.8% (IQR 0.2%–6.7%) for the left lung, and 5.8% (IQR 0.5%–9.7%) for the right lung. The infection was more serious in the right than in the left lung. The infected region mainly involved bilateral lower lobes, more pronounced on the right side. Quantitative CT showed that POI significantly decreased throughout the follow-up period in all 14 patients (p < 0.001). Among them, 50% of the patients had a more significant decrease in POI (51.3%) after a negative nucleic acid test. Moreover, there was a significant decrease in the CT number range of ground-glass opacities (GGO) and consolidation (p < 0.001).ConclusionsThis study demonstrated the quantitative analysis of follow-up CT scans plays an important role in the monitoring of COVID-19 treatment, which could help in treatment planning and standardizing the assessment for discharge.