Prediction model using clinical factors for radiation exposure during endoscopic retrograde cholangiopancreatography.

Background And Aim: Endoscopic retrograde cholangiopancreatography (ERCP) requires radiation. This study aimed to assess the clinical factors influencing radiation exposure and devise a scoring model for predicting high-dose radiation exposure.

Methods: Endoscopic retrograde cholangiopancreatography cases recorded between 2016 and 2019 in a single tertiary teaching hospital were retrospectively reviewed. A scoring model was created by bootstrap method in a derivation cohort (2016-2018) and was assessed in a validation cohort (2019).

Results: Out of 4223 ERCPs, 2983 and 1240 cases were included in the derivation and validation cohorts, respectively. In the derivation cohort, 746 cases (top 25%) comprised the high-dose exposure group, and 2237 cases (bottom 75%) comprised the low-dose exposure group. Nine clinical parameters associated with high-dose exposure were male, pancreatic sphincterotomy, balloon dilatation, biliary or pancreatic drainage, procedures with contrast dye, endoscopist, in-hospital ERCP, and spot image. Stone removal was included by bootstrap analysis. As presented in a nomogram, the weight score of each variable was as follows: male, 1; pancreatic sphincterotomy, 3; balloon dilatation, 7; stone removal, 3; biliary or pancreatic drainage, 5; procedures with contrast dye, 1; endoscopist B, 4; endoscopist C, 5; in-hospital procedure, 3; and spot image, 3. A total score ≥ 15 suggested a high-dose radiation exposure. The sensitivity and specificity of the model for high-dose exposure were 0.562 and 0.813, respectively. In the validation cohort, the model showed reasonable predictability.

Conclusions: Various factors were associated with radiation exposure. The simple scoring system in this study could guide endoscopists in predicting the risk of high-dose radiation exposure.