Chest CT Severity Score: An Imaging Tool for Assessing Severe COVID-19

Assessing the severity of respiratory diseases remains a challenge in clinical practice. While laboratory tests and clinical symptoms provide some guidance, they often lack precision in capturing the full extent of pulmonary involvement. Prior research has shown that imaging modalities like chest CT can reveal detailed lung involvement patterns. However, no prior work had resolved how to translate these visual findings into a standardized numerical score for severity. This gap motivated the development of a chest CT severity score (CT-SS) to better differentiate mild and severe cases of respiratory illness. Existing methods rely on subjective interpretation or limited anatomical regions. That uncertainty drove the need for a reproducible, anatomically comprehensive scoring system. It was already known that lower lobes are frequently affected in respiratory infections. No prior work had resolved how to quantify this involvement systematically. This study aimed to address that limitation by introducing a 20-region scoring framework.

The goal of this research was to determine whether a chest CT severity score could reliably distinguish between mild and severe forms of COVID-19. The specific problem addressed was the lack of an objective imaging-based tool for severity assessment. The motivation stemmed from the need for rapid and accurate triage of patients during a pandemic. The authors sought to validate a CT-SS that could be applied consistently across different clinical settings. By assigning scores to 20 lung regions, the study aimed to capture the spatial distribution of lung involvement. This approach was intended to improve diagnostic accuracy and guide clinical decisions. The study focused on patients with confirmed SARS-CoV-2 infection to ensure relevance to the disease being studied. The ultimate aim was to provide a tool that could be used in real-time clinical practice.

The study included 102 patients with laboratory-confirmed COVID-19 and available chest CT scans. The CT-SS was calculated by summing scores from 20 predefined lung regions. Each region was assigned a score of 0, 1, or 2 based on the extent of parenchymal opacification. The scoring system was designed to reflect the proportion of affected tissue in each region. Clinical data and laboratory results were collected for correlation with CT findings. Patients were classified as mild or severe according to national guidelines. The distribution of lung involvement was analyzed across lobes and segments. Statistical comparisons were made between mild and severe groups to determine score thresholds.

The CT-SS was significantly higher in severe cases compared to mild ones (p < 0.05). The maximum possible score was 40, but the highest recorded was 34. The most frequently affected regions were posterior segments of upper lobes and posterior basal segments of lower lobes. Lower lobe involvement was more common than middle or upper lobe involvement. No significant differences were found in disease distribution between left and right lungs. The optimal threshold for identifying severe disease was 19.5. At this threshold, the score achieved 83.3% sensitivity and 94% specificity. The area under the curve was 0.892, indicating strong diagnostic performance.

The authors concluded that the CT-SS is a valid tool for assessing the severity of pulmonary involvement in COVID-19. The score provides a rapid and objective way to differentiate between mild and severe cases. The findings suggest that the CT-SS could be used in clinical settings to guide triage and treatment decisions. The study supports the use of a 20-region scoring system for consistent evaluation. The threshold of 19.5 was proposed as a practical cutoff for identifying severe disease. The results align with the hypothesis that higher scores correlate with more extensive lung involvement. The authors emphasize the importance of using standardized regions for accurate comparisons. These conclusions are based on the observed statistical differences and diagnostic accuracy metrics.