Assessing risk factors for complications in computer tomography-guided lung biopsy: quantitative analysis for predicting pneumothorax

Abstract

BACKGROUND: Computer tomography (CT)-guided lung biopsy carries the risk of pneumothorax. A variety of other risk factors affect the occurrence of pneumothorax. OBJECTIVE: Assess the incidence and risk factors associated with pneumothorax complications in CT-guided lung biopsy, and to conduct a quantitative analysis of the variables among the significant risk factors to identify more effective indicators for predicting pneumothorax complications. DESIGN: Retrospective logistic SETTING: Single center in China PATIENTS AND METHODS: From June 2017 to May 2021, consecutive patients who underwent CT-guided lung biopsy were identified from the medical record system. Binary logistic regression analysis was used to identify potential risk factors for pneumothorax. Receiver operating characteristic (ROC) curves were constructed for continuous variables to determine cutoff values that optimized sensitivity and specificity. MAIN OUTCOME MEASURES: The incidence and risk factors of pneumothorax in CT-guided lung biopsy. SAMPLE SIZE: 132 patients RESULTS: The incidence of pneumothorax was 28.9% (38/132), with 6.8% (9/132) of patients requiring chest tube insertion. Results indicated that smaller lesion size (OR 0.724; 95% CI 0.619-0.848; P=.0001), longer needle tract length (OR 1.320; 95% CI 1.145-1.521; P=.001), multiple passes through the pleura (OR 4.618; 95% CI 1.378-15.467; P=.013), and needle tract length/lesion diameter (L/D) ratio (OR 0.028; 95% CI 0.002-0.732; P=.007) were independent risk factors for pneumothorax. ROC curve analysis determined a cut-off value of 0.81 for the L/D ratio (sensitivity=89.5%, specificity=71.3%). The area under the ROC curve (AUC) values of maximum diameter, needle tract length, and L/D ratio for pneumothorax were 0.749, 0.812, and 0.850, respectively. CONCLUSIONS: The L/D ratio, multiple passes through the pleura, longer needle tract length, and smaller lesions were independent risk factors for pneumothorax. A L/D ratio of less than 0.81 may indicate a pneumothorax. It may be necessary to use the proper sealing procedure for this patient group. LIMITATIONS: Due to its retrospective nature, there may be inherent selection bias.