Causal Inference Using Multimodal Propensity Score Adjustment for Assessing Gelatin Sponge Efficacy in Preventing Lung Biopsy Related Hemorrhage

Abstract

Purpose

Pulmonary hemorrhage is a potentially serious complication of CT-guided percutaneous lung biopsy. While gelatin sponge embolization of the needle tract is widely used for pneumothorax prevention, its effectiveness against hemorrhage remains uncertain. We aimed to assess whether prophylactic gelatin sponge tract embolization reduces the incidence of radiologically significant pulmonary hemorrhage, using advanced propensity score–based techniques.

Materials and methods

We retrospectively analyzed 1812 patients who underwent CT-guided lung biopsy between 2018 and 2022. Clinically significant hemorrhage was defined as grade ≥ 2 according to a standardized radiologic scale. Logistic regression and three propensity score methods—matching, inverse probability of treatment weighting (IPTW), and propensity score weighting (PSW)—were used to adjust for baseline differences between the gelatin and control groups.

Results

Clinically significant hemorrhage occurred in 38.4 % of patients. While unadjusted regression did not show a significant effect (OR 0.88; p = 0.24), IPTW (OR 0.81; p = 0.019) and PSW (OR 0.78; p = 0.013) analyses revealed statistically significant reductions in hemorrhage risk with gelatin sponge use. Subgroup analyses indicated enhanced protective effects in patients with smaller lesions, multiple needle passes, or non-perpendicular angles.

Conclusion

Gelatin sponge embolization modestly but significantly reduces pulmonary hemorrhage during lung biopsy, especially in high-risk procedures. These findings suggest a role for individualized embolization strategies in procedural planning.