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.