Development of a hemoptysis risk prediction model for patients following CT-guided transthoracic lung biopsy

BACKGROUND: Computed tomography-guided transthoracic needle biopsy (CT-TNB) is a widely used method for diagnosis of lung diseases; however, CT-TNB-induced bleeding is usually unexpected and this complication can be life-threatening. The aim of this study was to develop and validate a predictive model for hemoptysis following CT-TNB. METHODS: A total of 436 consecutive patients who underwent CT-TNB from June 2016 to December 2017 at a tertiary hospital in China were divided into derivation (n = 307) and validation (n = 129) cohorts. We used LASSO regression to reduce the data dimension, select variables and determine which predictors were entered into the model. Multivariate logistic regression was used to develop the predictive model. The discrimination capacity of the model was evaluated by the area under the receiver operating characteristic curve (AUROC), the calibration curve was used to test the goodness-of-fit of the model, and decision curve analysis was conducted to assess its clinical utility. RESULTS: Five predictive factors (diagnosis of the lesion, lesion characteristics, lesion diameter, procedure time, and puncture distance) selected by LASSO regression analysis were applied to construct the predictive model. The AUC was 0.850 (95% confidence interval [CI], 0.808–0.893) in the derivation, and 0.767 (95% CI, 0.684–0.851) in the validation. The model showed good calibration consistency (p > 0.05). Moreover, decision curve analysis indicated its clinical usefulness. CONCLUSION: We established a predictive model that incorporates lesion features and puncture parameters, which may facilitate the individualized preoperative prediction of hemoptysis following CT-TNB.