Chest computed tomography findings in different phases of SARS-CoV-2 infection. / Hallazgos en la tomografía computarizada de tórax en las fases evolutivas de la infección por SARS-CoV-2.

OBJECTIVE: To compare the findings on chest computed tomography (CT) in patients with COVID-19 during different phases of the disease and to evaluate the reproducibility of a visual radiologic score for estimating the extent of lung involvement. METHODS: We retrospectively reviewed chest CT studies from 182 patients with RT-PCR findings positive for SARS-CoV-2. Patients were classified according to the time elapsed from the onset of symptoms, as follows: early (0-4 days), intermediate/progressive (5-9 days), or advanced (≥10 days). We analyzed the frequency of each radiologic finding, as well as the pattern, appearance, and predominant distribution of lung involvement. A visual tomographic score (range, 0-25) was used to estimate the extent of involvement in each lobe and in the total lung volume. RESULTS: The predominant CT finding was the ground-glass pattern (n=110; 60.4%), the most common distribution was peripheral (n = 116; 66.7%), and the most prevalent appearance was typical (n=112; 61.5%). The halo sign was seen most frequently in the early phase (25%), whereas ground-glass opacities were more common in the intermediate/progressive and advanced phases. The median severity score was 10 (IQR: 5-13), and the scores increased as the disease progressed. The interobserver agreement (kappa) was 0.92 for the appearance, 0.84 for the distribution, 0.70 for the predominant pattern, and 0.89 for the visual score. CONCLUSION: The CT findings in patients with COVID-19 vary with the course of the infection. The proposed visual radiologic score is a simple, reproducible, and reliable tool for assessing lung involvement in COVID-19 pneumonia.

[Chest computed tomography findings in different phases of SARS-CoV-2 infection]. / Hallazgos en la tomografía computarizada de tórax en las fases evolutivas de la infección por SARS-CoV-2.

Objective: To compare the findings on chest computed tomography (CT) in patients with COVID-19 during different phases of the disease and to evaluate the reproducibility of a visual radiologic score for estimating the extent of lung involvement. Methods: We retrospectively reviewed chest CT studies from 182 patients with RT-PCR findings positive for SARS-CoV-2. Patients were classified according to the time elapsed from the onset of symptoms, as follows: early (0-4 days), intermediate/progressive (5-9 days), or advanced (≥10 days). We analyzed the frequency of each radiologic finding, as well as the pattern, appearance, and predominant distribution of lung involvement. A visual tomographic score (range, 0-25) was used to estimate the extent of involvement in each lobe and in the total lung volume. Results: The predominant CT finding was the ground-glass pattern (n=110; 60.4%), the most common distribution was peripheral (n = 116; 66.7%), and the most prevalent appearance was typical (n=112; 61.5%). The halo sign was seen most frequently in the early phase (25%), whereas ground-glass opacities were more common in the intermediate/progressive and advanced phases. The median severity score was 10 (IQR: 5-13), and the scores increased as the disease progressed. The interobserver agreement (kappa) was 0.92 for the appearance, 0.84 for the distribution, 0.70 for the predominant pattern, and 0.89 for the visual score. Conclusion: The CT findings in patients with COVID-19 vary with the course of the infection. The proposed visual radiologic score is a simple, reproducible, and reliable tool for assessing lung involvement in COVID-19 pneumonia.