Tomographic characteristics of lung lesions in hospitalized patients with COVID-19 and its prognostic value

The severity of lung involvement on chest tomography (CT) images in COVID-19 patients may have a prognostic value. This study assesses the type, severity and frequency of the different images of lung CT in hospitalized patients with COVID-19, and the differences in clinical characteristics and in-hospital outcomes, according to the CT severity score. This represents an observational study (retrospective cohort) of hospitalized patients with COVID-19. The ISARIC-WHO form was used to collect data. The type of lung lesions, affected lobes, and total CT severity score were determined at hospital admission. The first, second and third quartiles of the total CT score were calculated to divide the sample into four equal parts (Q1, Q2, Q3 and Q4). A total of 556 patients were included, 336 men (60.4%) and 220 women (39.6%), with a mean age of 61.9 +/- 15.8 years; and 532 of them had CT scan at admission. Patients in the more severe quartiles had more days of symptoms evolution (Q1 6.4 +/- 3.5, Q2 7.9 +/- 4.1, Q3 8.2 +/- 4.1, Q4 8.1 +/- 4.4), desaturation (Q1 95.3 +/- 3.7, Q2 94.4 +/- 3.1%, Q3 91.7 +/- 4.8%, Q4 86.5 +/- 9.1%), alterations of inflammatory markers, hospital stay (Q1 6.4 +/- 2.9, Q2 7.4 +/- 4.1, Q3 9.6 +/- 5.8, Q4 13.1 +/- 10.4 days), admission to ICU (Q1-2.5%, Q2-5.8%, Q3-12.5%, Q4-49.1%), mortality (Q1-3.8%, Q2-4.5%, Q3-9.4%, Q4-33.3%), mixed CT lesions (ground glass opacity-consolidation), linear opacities, crazy-paving pattern, reverse halo sign, and bronchiectasis. The total CT score significantly correlated with leukocyte, neutrophil and lymphocyte counts, and with other inflammatory markers. Semi-quantitative evaluation of pulmonary involvement in the initial chest CT can help to establish the severity of the case and predict relevant clinical outcomes in COVID-19 patients.