Development of a Nomogram Model to Predict the Risk of Pneumothorax after CT-guided Coaxial Core Needle Lung Biopsy: A Cross-sectional Study

Introduction: A sharp increase in the incidence of chest tumours has led to an increased demand for Computed Tomography (CT)-guided lung biopsy. Pneumothorax remains the most common complication of CT-guided coaxial core needle lung biopsy. Pneumothorax triggers shortness of breath and chest pain, which leads to prolonged inpatient stay. To reduce the incidence of pneumothorax, a predictive model in the form of nomogram was developed using the potential risk factors to identify patients at high-risk for pneumothorax. Aim: To investigate the risk factors of pneumothorax and develop a nomogram model using the independent risk factors to predict pneumothorax after CT-guided coaxial core needle lung biopsy and to validate the prediction model in a test population. Materials and Methods: A cross-sectional time-bound study was done at Barnard Institute of Radiology, Madras Medical College, and Rajiv Gandhi Government General Hospital in Chennai, Tamil Nadu, India from October 2023 to October 2024. The study included 335 patients who were referred to the Department of interventional radiology for CT-guided coaxial core needle lung biopsy to diagnose lung nodules or masses suspected of malignancy in chest CT. A total of 28 variables were assessed, which included baseline patient characteristics, primary pulmonary diseases, target lesion image characteristics, and biopsy-related variables. Of the total sample size of 335 patients, 233 (70%) were part of the development group, and 102 patients were included in the validation group (30%). Multivariate and univariate logistic regression analysis methods were used to identify the independent risk factors of pneumothorax, which were used to develop the nomogram risk prediction model. The prediction model was validated in a test population of 102 patients. Results: A total of 51 (21.88%) patients developed pneumothorax post CT-guided lung biopsy among the development group, and 20 (19.60%) patients among the validation group also developed pneumothorax. Seven independent risk factors were determined among the assessed variables using multivariate logistic regression analysis, which included age, emphysema, pleural thickening, fissure traversed, patient position, size grade, and lesion depth. The seven factors were used to construct a nomogram risk prediction model. Receiver Operating Characteristic (ROC) curves were used for validate the nomogram model. In the development group, Area Under Curve (AUC) was 0.921, and in the validation group, AUC was 0.897. Conclusion: A nomogram risk prediction model for pneumothorax was developed using the seven independent risk factors and has proven to be accurate and statistically valid. The nomogram model can thus help predict pneumothorax in clinical practice, assisting interventional radiologists in avoiding risk factors.