Enhancing safety in CT-guided lung biopsies: correlation of MinIP imaging with pneumothorax risk prediction

Objectives This study aimed to evaluate whether minimum-intensity projection (MinIP) images could predict complications in CT-guided lung biopsies. Methods We retrospectively analyzed 72 procedures from January 2019 to December 2023, categorizing patients by pneumothorax and the severity of hemorrhage (grade 2 or higher). Radiodensity measurements were performed using lung window (LW) and MinIP (10-mm slab) images. Regions of interest (ROIs) were placed at sites of the lowest density along the biopsy pathway. Absolute values were recorded, categorized by a radiodensity level of -850 HU, and assessed using our bridged radiological observations with measurement-optimized model (BROM-OLB) model with validation from three additional ROIs. Emphysema was visually scored. Statistical analysis included univariate analysis (Fisher's exact and Mann-Whitney U-tests) and binomial logistic regression to identify confounders. Results Lower radiodensity values in MinIP images in the access route, particularly with the BROM-OLB MinIP method, were significantly associated with a higher risk of pneumothorax (5/39, 13% vs 27/33, 82%, p < 0.01; Sensitivity 81.8% and Specificity 87.2%). Pneumothorax was more common with longer procedures (p < 0.05). Lower LW density values correlated with higher pulmonary hemorrhage rates (p < 0.01). Binomial logistic regression identified positive BROM-OLB MinIP results (OR 28.244, 95% CI: 7.675-103.9, p < 0.01) and lower LW density (OR 0.992, 95% CI: 0.985-0.999, p = 0.025) as independent risk factors. The optimal threshold values to predict pneumothorax were -868 HU in MinIP images and -769 HU in LW. Conclusion The assessment of MinIP images is superior, and in combination with relative quantitative measurement of radiodensity for access route planning, it can reduce the risk of pneumothorax in CT-guided lung biopsies.