205 - Radiation Risk in Routine Interventional Radiology Procedures: Outlier Profiles, Predictive Factors, and Implications for Bedside Fluoroscopy

Purpose: Radiation safety is essential in interventional radiology (IR). Current monitoring emphasizes threshold-based reporting at 5 Gy and 10 Gy. Routine procedures usually fall below thresholds and are not regularly monitored for outliers. Our objective was to determine which IR procedures are most appropriate for bedside mobile fluoroscopy by identifying those with predictable radiation profiles and low outlier rates and to characterize patient factors contributing to outliers.

Materials and Methods: We retrospectively analyzed 53,230 IR procedures. Four routine procedures were included: central venous catheter (CVC) placement, nephrostomy tube exchange, biliary drain exchange, and gastrostomy tube exchange. Outliers were defined using the Tukey method when air kerma rate or fluoroscopy time exceeded the upper bound. Predictive modeling assessed associations with age, sex, BMI, race/ethnicity, and ICU status. Logistic regression (LR) identified predictors; random forest (RF) compared predictor importance. Model discrimination was evaluated using area under the receiver operating curve (AUC).

Results: A total of 18,516 routine cases were included. CVC placement (n=12,286) had 6.0% outliers, with significant predictors being BMI ≥50 (OR 4.6, 95% CI 2.1–10.1, p< 0.001), BMI ≥60 (OR 3.9, 95% CI 1.3–11.8, p=0.02), and ICU status (OR 2.2, 95% CI 1.5–3.1, p< 0.001). G-tube exchange (n=2,341) had 10.6% outliers, with BMI ≥40 (OR 16.3, 95% CI 2.9–91.1, p=0.001) as the only significant predictor. Nephrostomy tube exchange (n=1,927) had 8.9% outliers, with male sex (OR 2.5, 95% CI 1.5–3.9, p< 0.001) being a significant predictor. Biliary tube exchange (n=1,962) had 5.7% outliers with no significant predictors. Model performance was weak overall with AUCs 0.58–0.70 (LR) and 0.55–0.68 (RF), with nephrostomy performing best at 0.70 (LR) and 0.68 (RF).

Conclusion: Predictive models were weak overall but highlighted consistent risk factors relevant to bedside appropriateness. Severe obesity was the strongest predictor; ICU status, and sex contributed in select procedures. Procedure type and patient factors may guide bedside fluoroscopy use and minimize radiation risk.



The authors used GPT-5 for language editing and formatting. Authors reviewed and edited the work and take full responsibility for the content of the abstract.