The role of chest CT quantitative pulmonary inflammatory index in the evaluation of the course and treatment outcome of COVID-19 pneumonia

Abstract

To explore the clinical application value of chest CT quantitative pulmonary inflammation index (PII) in the evaluation of the course and treatment outcome of COVID-19 pneumonia. One hundred and eighteen patients with COVID-19 pneumonia diagnosed by RT-PCR were analyzed retrospectively. The correlation between chest CT PII, clinical symptoms and laboratory examinations during the entire hospitalization period was compared. The average age of the patients was 46.0 ± 15 (range: 1-74) years. Of the 118 patients, 62 are male (52.5%) and 56 are female (47.5%). Among them, 116 patients recovered and were discharged, 2 patients died, and the median length of hospital stay was 22 (range: 9-41) days. On admission, 76.3% of the patients presented with fever, and the laboratory studies showed a decrease in lymphocyte (LYM) count and an increase in lactate dehydrogenase (LDH) levels, C-reactive protein (CRP) levels, and erythrocyte sedimentation rate (ESR). Within the studies' chest CTs, the median number of involved lung lobes was 4 (range: 0-5) and the median number of involved lung segments was 9 (range 0-20). The left lower lobe and the right lower lobe were the most likely areas to be involved (89.0% and 83.9%), and 84.7% of the patients had inflammatory changes in both lungs. The main manifestations on chest CT were ground glass opacities (31.4%), ground glass opacities and consolidation (20.3%), ground glass opacities and reticular patterns (32.2%), mixed type (13.6%), and white lungs (1.7%); common accompanying signs included linear opacities (55.9%), air bronchograms (46.6%), thick small vessel shadows (36.4%), and pleural hypertrophy (13.6%). The chest CT at discharge showed complete absorption of lesions in 19 cases (16.1%), but not in the remaining 99 cases. Lesions remained in a median of 3 lung lobes (range: 0-5). Residual lesions remained in a median of 5 lung segments (range: 0-20). The residual lesions mainly presented as ground glass opacities (61.0%), and the main accompanying sign was linear opacities (59.3%). Based on chest CT, the median maximum PII of lungs was 30.0% (range: 0-97.5%), and the median PII after discharge in the patients excluding the two deaths was 12.5% (range: 0-53.0%). PII was significantly negatively correlated with the LYM count and significantly positively correlated with body temperature, LDH, CRP, and ESR. There was no significant correlation between the PII and the white blood cell count, but the grade of PII correlated well with the clinical classification. PII can be used to monitor the severity and the treatment outcome of COVID-19 pneumonia, provide help for clinical classification, assist in treatment plan adjustments and aid assessments for discharge.

Figure 1

Main image manifestations and accompanying signs. (A) 34-year-old man; the main sign is bilateral subpleural ground glass opacities with left lower lung dorsal linear opacities (arrow); (B) 68-year-old man; the main signs are bilateral ground glass opacities and large consolidation; (C) 74-year-old woman; the main signs are bilateral ground glass opacities and reticular patterns, “crazy-paving” pattern, thickened small vessel shadows can be seen in the right lung lesions; (D) 59-year-old man; extensive lung glass shadow and partial consolidation shadow; lungs are white; air bronchial signs (arrows) can be seen in the background of bilateral exudation, and the dorsal pleura is slightly thicker. Imaging were acquired on GE AW VolumeShare software (Version number: 4.6; http://www.gehealthcare.com/products/advanced-visualization/platforms/aw-server).

Figure 2

Case of complete absorption of lung lesions. A 44-year-old female patient was hospitalized for 23 days and her lung lesions were gradually absorbed to disappearance. (A) The first CT on admission (2020–1–27): flaky ground glass on the outside of the right lung; (B) The second CT (2020–1–31): progression of the right lung lesion, expanded scope; (C) Third CT (2020–2–05): obvious absorption of right lung lesions, visible small plate-shaped ground glass opacities and stripe shadows; (D) Fourth CT (2020–2–12): the right lung lesions are further absorbed and reduced; (E) Fifth CT (2020–2–19): Complete absorption of the right lung lesion. Imaging were acquired on GE AW VolumeShare software (Version number: 4.6; http://www.gehealthcare.com/products/advanced-visualization/platforms/aw-server).

Figure 3

The reliability of quantitative methods for PII between Radiologist 1 and Radiologist 2. The left hand panel depicts the Bland–Altman plot of the two radiologists assessed the maximum PII. The right hand panel depicts the Bland–Altman plot of the two radiologists assessed the PII upon discharge. The upper and lower dashed lines represented the 95% confidence intervals.

Figure 4

Grade and clinical classification of inflammatory factors. (A–C) 41-year-old female, ground glass opacities in both lungs, involving 5 lung segments, PII was 12.5%, inflammation grade I, clinically classified as moderate type; (D–F) 56-year-old female, scattered lung glass shadows, involving 13 lung segments, PII was 37.5%, degree of inflammation was grade II, clinically classified as moderate type; (G–I) 41-year-old female, double lung ground glass opacities with consolidation and thickened lobular septa, the image showed mixed type, involving 16 lung segments, PII is 62.5%, the degree of inflammation is grade III, clinically classified as severe type; (J–L) 57 years old, bilateral ground glass opacities combined with consolidation and stripe shadow, bilateral pleural effusion, involving 16 lung segments, PII is 77.5%, the degree of inflammation is grade IV, clinically classified as severe type. Imaging were acquired on GE AW VolumeShare software (Version number: 4.6; http://www.gehealthcare.com/products/advanced-visualization/platforms/aw-server).