Abstract:

Objective: To investigate the value of chest CT signs combined with peripheral blood eosinophil percentage in identifying pleuropulmonary paragonimiasis and pleural tuberculosis in children. Methods: Pediatric patients with pleuropulmonary paragonimiasis and pleural tuberculosis in children first diagnosed from January 2019 to December 2021 in The Third People’s Hospital in Kunming were retrospectively collected, including 58 pediatric patients with pleuropulmonary paragonimiasis (PG group) and 45 pediatric patients with pleural tuberculosis (TB group). The clinical symptoms, chest CT manifestations and laboratory test results were analyzed; an imaging model of CT signs and a combined model combining CT signs and eosinophils were established using binary logistic regression, the differential diagnostic efficacy of the two diagnostic models was calculated and compared. Results: The imaging model of CT signs showed that, encapsulated effusion (OR=5.11, 95%CI: 1.64-15.97) and pulmonary nodules (OR=12.33, 95%CI: 2.51-60.65) were more likely to be diagnosed as pleural tuberculosis, while pleural calcification (OR=0.17, 95%CI: 0.04-0.68) and tunnel sign (OR=0.01, 95%CI: 0.00-0.15) was less likely to be diagnosed as pleural tuberculosis. According to the ROC curve, when AUC=0.850 (95%CI: 0.777-0.922), the sensitivity was 79.3% (95%CI: 66.6%-88.8%), and the specificity was 75.6% (95%CI: 60.5%-87.1%). The combined model of CT signs and eosinophil percentage showed that, when pulmonary nodules (OR=11.56, 95%CI: 1.76-75.71) was found, pleural tuberculosis was more likely to be diagnosed, while tunnel signs (OR=0.02, 95%CI: 0.00-0.48) and elevated eosinophil percentage (OR=0.02, 95%CI: 0.01-0.08) were found, pleural tuberculosis was less likely to be diagnosed, and the ROC curve was plotted with AUC=0.934 (95%CI: 0.885-0.982), sensitivity of 89.7% (95%CI: 78.8%-96.1%), and specificity of 88.9% (95%CI: 75.9%-96.3%). The combined model had higher differential diagnostic efficacy than the imaging model, with a statistically significant difference in AUC (Z=-2.401, P=0.016). Conclusion: For the differentiation of pleuropulmonary paragonimiasis and pleural tuberculosis in children, the combined model have higher differential diagnostic efficacy than the imaging model; when there are pulmonary nodules and the percentage of eosinophils is not increased, the children are more likely to be diagnosed of tuberculous.

Key words: Tuberculosis, pleural, Paragonimiasis, Diagnosis, differential, Tomography, X-ray computed