Prediction of Complication Risk in Computed Tomography-guided Thoracic Biopsy: A Prescription for Improving Procedure Safety

Hwan Lee,Christina Murphy,Eduardo J Mortani Barbosa Jr
DOI: https://doi.org/10.1097/RTI.0000000000000689
2023-03-01
Abstract:Purpose: Computed tomography-guided transthoracic biopsy (CTTB) is a minimally invasive procedure with a high diagnostic yield for a variety of thoracic diseases. We comprehensively assessed a large CTTB cohort to predict procedural and patient factors associated with the risk of complications. Materials and methods: The medical record and computed tomography images of 1430 patients who underwent CTTB were reviewed individually to obtain clinical information and technical procedure factors. Statistical analyses included descriptive and summary statistics, univariate analysis with the Fisher test, and multivariate logistic regression. Results: The most common type of complication was pneumothorax (17.4%), followed by bleeding (5.9%). Only 26 patients (1.8%) developed a major complication. Lung lesions carried a higher risk of complications than nonlung lesions. For lung lesions, the nondependent position of the lesion, vertical needle approach, trespassing aerated lung, and involvement of a trainee increased the risk of complication, whereas the use of the coaxial technique was a protective factor. The time with the needle in the lung, the number of biopsy samples, and the distance crossing the aerated lung were identified as additional risk factors in multivariate analysis. For nonlung lesions, trespassing the pleural space was the single best predictor of complications. A logistic regression-based model achieved an area under the receiver operating characteristic curve of 0.975, 0.699, and 0.722 for the prediction of major, minor, and no complications, respectively. Conclusions: Technical procedural factors that can be modified by the operator are highly predictive of the risk of complications in CTTB.
What problem does this paper attempt to address?