Logistic regression analysis and a risk prediction model of pneumothorax after CT-guided needle biopsy

摘要

Background: Pneumothorax is the most common complication of computed tomography (CT)-guided needle biopsy. The purpose of this study was to investigate independent risk factors of pneumothorax, other than emphysema, after CT-guided needle biopsy and to establish a risk prediction model. Methods: A total of 864 cases of CT-guided needle biopsy with an 18-gauge cutting needle were enrolled in this study. The relevant risk factors associated with pneumothorax included age, sex, emphysema, shortaxis size of the lesion, depth of the lesion, body position, and the number of pleural punctures. Several independent risk factors of pneumothorax were found, and a predictive model for pneumothorax was established using univariate and multivariate logistic regression analyses. Results: Pneumothorax occurred in 31.4% (271/864) of cases. Univariate analysis showed that significant risk factors of pneumothorax included age, emphysema, small lesion size, no contact between the lesion and the pleura, prone or lateral body position, and multiple punctures. Independent risk factors of pneumothorax in the multivariate logistic regression analysis included emphysema (P=0.000), no contact between the lesion and the pleura (P=0.000), prone or lateral body position (P=0.002), and the number of pleural punctures (P=0.000). The sensitivity, specificity, and accuracy of the predictive model for pneumothorax were 56.8%, 79.6%, and 72.5%, respectively. Conclusions: Pneumothorax is a common complication of CT-guided lung biopsy. Independent risk factors of pneumothorax include emphysema, no contact between the lesion and the pleura, and prone or lateral body position. The predictive model developed in this study was highly accurate in predicting the incidence of pneumothorax.