Models of "artificial intelligence" applied to pancreatic endoscopic ultrasound

The integration of artificial intelligence (AI) into endoscopic ultrasound (EUS) has transformed the diagnosis and management of pancreatic diseases. AI, particularly deep learning models and neural networks, significantly enhances diagnostic accuracy and procedural efficiency, aiding both trainee and experienced endo-sonographers. Models like BP MASTER and CH-EUS MASTER have demonstrated high accuracy and sensitivity in both training and diagnostic processes. AI enables precise differential diagnosis of pancreatic diseases, including solid and cystic lesions, as shown by studies improving diagnostic performance with clinical variables integration. Despite promising results, limitations such as small sample sizes and the need for further external validation persist. Continuous monitoring is essential to ensure AI’s reliability in clinical practice. In conclusion, AI is a valuable asset for pancreatic EUS, improving diagnosis and training. With ongoing research, AI holds the potential to revolutionise diagnostic and therapeutic approaches, contributing to more precise and personalised medicine.