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

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence

Background:

• Pancreaticobiliary endoscopy (EUS, ERCP, DSOC) is crucial for diagnosing and managing pancreatic and biliary diseases.
• Current endoscopic procedures face limitations in operator dependency, diagnostic accuracy, and technical complexity.

Purpose of the Study:

• To explore the potential of artificial intelligence (AI), including machine learning (ML) and deep learning (DL), in addressing the challenges of pancreaticobiliary endoscopy.
• To review early studies on AI applications in enhancing lesion detection, mass differentiation, lesion classification, and diagnosis of malignant biliary strictures.

Main Methods:

• Review of early studies and applications of AI, ML, and DL in pancreaticobiliary endoscopy.
• Analysis of AI's role in predicting post-ERCP pancreatitis risk and reducing radiation exposure.

Main Results:

• AI demonstrates potential in improving lesion detection, differentiating pancreatic masses, classifying cystic lesions, and diagnosing malignant biliary strictures.
• AI applications have shown promise in predicting post-ERCP pancreatitis risk and reducing radiation exposure during ERCP.

Conclusions:

• Current AI models are experimental, facing limitations such as small datasets, lack of validation, and absence of FDA approval.
• Overcoming barriers like data inconsistency, interoperability, and workflow integration is crucial for future AI development in this field.
• AI has the potential to become a valuable partner for endoscopists, enhancing accuracy, reducing complications, and enabling personalized care in pancreaticobiliary endoscopy.