Non-contrast CT scans were performed using 128-slice CT scanners (Siemens, Erlangen, Germany) in a craniocaudal direction in a single breath-hold with helical scans obtained in
A supine position with 100-120 kVp (automatic kV setting based on patient size – “Care kV”), automatic tube current modulation, pitch 1.2, collimation 0.6 mm, and matrix 512 × 512. Images were reconstructed with a slice thickness of 1.5 mm using a Br59 kernel with ADMIRE iterative algorithm (level strength 1).
One radiologist (Reader 1, with 13-year experience) and two 4th-year radiology residents (Readers 2 and 3), with experience of at least 500 COVID-19-positive chest CT readings, independently reassigned CO-RADS score (1 to 5 points-scale) [23 (link)] and ACR COVID classification (negative, non-typical, indeterminate, typical) [24 (link),25 (link)] to all examinations. The readers also independently evaluated the CT scans according to the following: visual quantification of pulmonary involvement expressed as the percentage of total lung volume and corresponding CT severity score (low involvement < 25%, high involvement ≥ 25%, as proposed by Au-Yong et al. [26 (link)] and used by Lee et al. [21 (link)]); CT patterns (presence of ground glass opacities, consolidations, crazy paving areas, mono- or bi-lateral involvement, mono- or multi-focal involvement); and findings distribution (mainly central, mainly peri-pheral, or mixed central and peripheral). The main CT pattern when more than one was present (ground glass opacities, consolidations, crazy paving areas) was assigned by the most experienced radiologist (Reader 1). All readers were blinded to the vaccination status of the patients.
A supine position with 100-120 kVp (automatic kV setting based on patient size – “Care kV”), automatic tube current modulation, pitch 1.2, collimation 0.6 mm, and matrix 512 × 512. Images were reconstructed with a slice thickness of 1.5 mm using a Br59 kernel with ADMIRE iterative algorithm (level strength 1).
One radiologist (Reader 1, with 13-year experience) and two 4th-year radiology residents (Readers 2 and 3), with experience of at least 500 COVID-19-positive chest CT readings, independently reassigned CO-RADS score (1 to 5 points-scale) [23 (link)] and ACR COVID classification (negative, non-typical, indeterminate, typical) [24 (link),25 (link)] to all examinations. The readers also independently evaluated the CT scans according to the following: visual quantification of pulmonary involvement expressed as the percentage of total lung volume and corresponding CT severity score (low involvement < 25%, high involvement ≥ 25%, as proposed by Au-Yong et al. [26 (link)] and used by Lee et al. [21 (link)]); CT patterns (presence of ground glass opacities, consolidations, crazy paving areas, mono- or bi-lateral involvement, mono- or multi-focal involvement); and findings distribution (mainly central, mainly peri-pheral, or mixed central and peripheral). The main CT pattern when more than one was present (ground glass opacities, consolidations, crazy paving areas) was assigned by the most experienced radiologist (Reader 1). All readers were blinded to the vaccination status of the patients.
