Quantitative chest CT imaging characteristics and outcome of patients with COVID-19 associated pulmonary artery thrombosis: A single-center retrospective cohort study

Abstract

Coronavirus disease 2019 (COVID-19)-associated pulmonary thrombotic events occur frequently and are associated with disease severity and worse clinical outcomes. We aimed to describe the clinical and quantitative chest computed tomography (CT) imaging characteristics based on density ranges (Hounsfield units) and the outcomes of patients with COVID-19 associated pulmonary artery thrombosis. This retrospective cohort study included all patients with COVID-19 hospitalized in a tertiary care hospital between March 2020 and June 2022 who underwent a CT pulmonary angiography. We included 73 patients: 36 (49.3%) with and 37 (50.7%) without pulmonary artery thrombosis. The in-hospital all-cause mortality was 22.2 versus 18.9% (P = .7), and the intensive care unit admission rates were 30.5 versus 8.1% (P = .01) at the time of diagnosis of pulmonary artery thrombosis. Except for D-dimers (median of 3142 vs 533, P = .002), the other clinical, coagulopathy, and inflammatory markers were similar. Logistic regression analysis revealed that only D-dimers were associated with pulmonary artery thrombosis (P = .012). ROC curve analysis of D-dimers showed that a value greater than 1716 ng/mL predicted pulmonary artery thrombosis with an area under the curve of 0.779, 72.2% sensitivity, and 73% specificity (95% CI 0.672–0.885). Peripheral distribution of pulmonary artery thrombosis was recorded in 94.5% of cases. In the lower lobes of the lungs, the incidence of pulmonary artery thrombosis was 6 times higher than that in the upper lobes (58–64%), with a percentage of lung injury of 80% to 90%. Analysis of the distribution of arterial branches with filling defects revealed that 91.6% occurred in lung areas with inflammatory lesions. Quantitative chest CT imaging provides valuable information regarding the extent of COVID-19 associated lung damage and can be used to anticipate the co-location of pulmonary immunothrombotic events. In patients with severe COVID-19, in-hospital all-cause mortality was similar regardless of the presence of associated distal pulmonary thrombosis.