Monitoring pneumonia onset and severe respiratory infectious diseases using quantitative computed tomography, validation amidst the coronavirus disease 2019 outbreak at the end of 2022 in the Chinese mainland.

Authors

Ting Cheng, Tiantian Zhang, Jingci Xu, Yong Li, Yaqing Jin, Shengyu Wang, Li Luo

Published in

Journal of thoracic disease. Volume 18. Issue 5. Pages 523. May 31, 2026. Epub May 27, 2026.

Abstract

Effective surveillance of severe respiratory infectious diseases is crucial. This study aimed to explore an automated quantitative analysis method based on chest computed tomography (CT) for distinguishing individuals with pneumonia and to evaluate its effectiveness for monitoring pneumonia incidence.
Hospitalized pneumonia patients and healthy controls with normal chest CTs were enrolled. CT images were analyzed using "uAI Discover Pneumonia" software (United Imaging Intelligence) to derive CT-derived quantitative metrics related to pneumonia, including total lung infection volume and proportion of infected lung volume relative to total lung volume. Differences in these quantitative CT (QCT) infection metrics between the two groups were compared. The cutoff value with the highest Youden index was selected and its sensitivity and specificity were evaluated. CT scans requested from pneumonia-related emergency departments and outpatient departments (PR-ED/OPD) between Week 27, 2022, and Week 26, 2023 were extracted. Chest CT request volume was counted. The proportion of pneumonia-positive CT scans and the estimated pneumonia patients volume within these requests were calculated weekly to reflect temporal changes in pneumonia incidence. These indicators were compared with other surveillance metrics using cross-correlation analysis to evaluate temporal relationships. The proportion of pneumonia-positive CT scans and estimated pneumonia patient volume were compared across typical periods following policy changes.
A total of 209 pneumonia patients and 221 control subjects were included. The median [interquartile range (IQR)] total lung infection volume was 545.1 (255.7-966.7) mL in the pneumonia group and 0.0 (0.0-13.1) mL in the non-pneumonia control group (P<0.001). Median infected volume proportion was 19.90% (7.15-33.85%) vs. 0.00% (0.00-0.30%) (P<0.001). Area under the curve (AUC) values were 0.978±0.005 and 0.980±0.005, respectively. Optimal cutoffs were ≥65 mL (sensitivity 92.34%, specificity 91.86%) and ≥1.75% (sensitivity 93.30%, specificity 93.21%). CT scans from PR-ED/OPD comprised 99,723 encounters. Quantitative analysis was performed on 39,366 eligible encounters, of which 8,955 were classified as pneumonia-positive. Cross-correlation analysis showed that the CT-estimated pneumonia patient volume led regional pneumonia-associated mortality by 1 week (r=0.954, P<0.001), lagged the proportion of influenza-like illness (ILI) cases testing positive for coronavirus disease 2019 (COVID-19) in Chinese sentinel hospitals by 1 week (r=0.824, P<0.001), and was highly synchronized with fluctuations in the number of adult pneumonia inpatients during the same period (r=0.968, P<0.001).
CT-derived quantitative metrics exhibit excellent diagnostic accuracy for pneumonia, with recommended cutoffs of ≥65 mL total infected volume and ≥1.75% infected proportion. This automated method effectively monitors pneumonia incidence, reflecting the epidemic status of severe respiratory infections.

PMID:
42306677
Bibliographic data and abstract were imported from PubMed on 17 Jun 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.