Endoscopic retrograde cholangiopancreatography (ERCP) is a highly specialized and technically demanding endoscopic procedure requiring advanced cognitive and manual skills. Due to its complexity, limited procedural exposure and concentration of expertise within tertiary centers, structured and safe training opportunities are essential. Simulation-based education has been increasingly recognized as an effective strategy to bridge the training gap, allowing skill acquisition without compromising safety. While several mechanical, animal and computer-based simulators have been validated for luminal endoscopy, ERCP-specific simulators remain few. Recently, the Mikoto simulator series, a line of high-fidelity silicone-based models featuring integrated automated performance assessment, has been introduced. This exploratory study aimed to evaluate the construct validity of the novel Mikoto ERCP Simulator by comparing the performance of expert and non-experts.

The study was conducted during the 8th Educational Meeting of the Lombardy Section of the Italian Society of Digestive Endoscopy (SIED). Twenty-two endoscopists (11 experts and 11 non-experts) performed simulated ERCP procedures using the Mikoto ERCP Simulator (R Zero Inc., Japan, distributed by Fujifilm, Tokyo, Japan). The model provides realistic anatomical representation of the upper gastrointestinal tract, including interchangeable papilla of Vater modules and automatically records multiple performance metrics such pharyngeal pressure, the time to reach the subsequent anatomical segment, biliary cannulation time and procedural scores. Each endoscopist had 10 minutes to achieve deep biliary cannulation. Experts were defined as operators with >250 ERCPs performed, whereas non-experts were trainees competent in reaching the second duodenal portion (D2) and attempting cannulation. Primary outcomes were the simulator’s ability to differentiate experts from non-experts in biliary cannulation rate, cannulation time, total score (0–100) and total procedure time. Secondary outcomes included subgroup comparisons of pharyngeal passage time and pressure, time and score to reach D2 and changes in repeated performance attempts. Perceived realism was rated on a 5-point Likert scale. 

Experts (mean age 46±6.7 yrs, 11 M) and non-experts (median age 32±3.8 years, 6 M, 4 residents) demonstrated distinct performance profiles. Biliary cannulation was achieved by 91% of experts versus 64% of non-experts (p=0.15). Among successful cannulations, the median cannulation time was shorter for experts compared with non-experts (76s, IQR 53–172 vs 89s, IQR 79–231,p=0.07), even if not statistically significant. Experts completed the overall procedure faster (142s, IQR 102–257 vs 492s, IQR 158–600; p=0.04) and achieved higher total scores (96/100, IQR 88.5–98.5 vs 82/100, IQR 61.5–91.5; p=0.02). During the pharyngeal phase, no significant differences were noted in time or pressure; however, experts achieved significantly higher pharyngeal pressure control scores (median 10/10 vs 7/10; p=0.02). Time to reach D2 and corresponding score both favoured experts (median 35 vs 85s,p=0.014; mean D2 score 19.9±7.07 vs 16.5±10.6, p=0.01). Table 1 synthetise differences between experts and non-experts. Among the five participants (four experts) who repeated simulations, sequential improvements in total time and total score were observed and confirmed by Wilcoxon signed-rank tests (p=0.002;p=0.003,respectively),suggesting a measurable learning effect. Overall, participants rated the simulator’s realism highly, with a mean fidelity score of 4/5 (“very good”), particularly commending the accuracy of the pharyngeal passage simulation.

This study shows that the Mikoto ERCP Simulator reliably distinguishes between expert and non-expert operators, reinforcing its construct validity as a training modality. Its automated scoring and detailed performance analytics offer clear advantages by supporting objective self-assessment and promoting individualized, self-directed learning. The simulator also highlights procedural steps with the largest performance gaps, helping educators target training more effectively. These results justify future, larger studies to confirm validity across additional performance domains and to clarify the simulator’s place within structured ERCP curricula. Overall, the Mikoto platform appears to be a high-fidelity, data-driven supplement to traditional ERCP education, providing a safe, reproducible and efficient setting for training