Artificial Intelligence in Pancreatobiliary Endoscopy: Current Advances, Opportunities, and Challenges

Pancreaticobiliary endoscopy, encompassing endoscopic ultrasound (EUS), endoscopic retrograde cholangiopancreatography (ERCP), and digital single-operator cholangioscopy (DSOC), is essential for diagnosing and managing pancreatic and biliary diseases. However, these procedures are limited by operator dependency, variable diagnostic accuracy, and technical complexity. Artificial intelligence (AI), particularly through machine learning (ML) and deep learning (DL), has emerged as a promising tool to address these challenges. Early studies show that AI can enhance lesion detection, improve differentiation of pancreatic masses, classify cystic lesions, and aid in diagnosing malignant biliary strictures. AI has also been used to predict post-ERCP pancreatitis risk and reduce radiation exposure during ERCP. Despite this promise, current AI models are largely experimental-limited by small, single-center datasets, lack of external validation, and no FDA-approved systems for these indications. Major barriers include inconsistent data acquisition, limited interoperability across hardware platforms, and integration into real-time workflows. Future progress depends on multicenter data sharing, standardized imaging protocols, interpretable AI design, and regulatory pathways for model deployment and updates. AI can be developed as a valuable partner to endoscopists, enhancing diagnostic accuracy, reducing complications, and supporting more efficient, personalized care in pancreaticobiliary endoscopy.