Clinical Research on Navigation and Quality Control System of Pancreatic Ultrasound Endoscopy Based on Deep Learning

What this Trial Is About

Researchers are evaluating a deep learning-based artificial intelligence (AI) system designed to assist in pancreatic endoscopic ultrasound (EUS) scanning. The study aims to compare the image recognition accuracy of this AI system with that of ultrasound endoscopists and to determine whether the AI system can improve the efficiency of pancreatic scanning. The investigation focuses on the navigation and quality control roles of AI during pancreatic ultrasound endoscopy. Participants will undergo pancreatic EUS examinations either with or without the support of the AI system. The study will collect videos and images from these procedures to establish the AI system's ability to identify pancreatic ultrasound sites and anatomical structures. The AI system's recognition results will be compared to those of experienced endoscopists. Additionally, the study will compare the completeness of standard scanning sites and the scanning duration between AI-assisted and non-assisted procedures. During the trial, participants will have pancreatic EUS performed, and researchers will assess the accuracy of image recognition and the completeness of standard station scanning over a two-year period. Secondary measures include agreement between AI and endoscopist assessments, completeness of anatomical landmarks, and operation time. The study involves randomized allocation into AI-assisted or non-assisted groups, with double masking to reduce bias. Participant involvement includes signing informed consent and undergoing the scheduled procedures and assessments until the study concludes.

Clinical Research on a Novel Deep-learning Based System in Pancreatic Endoscopic Ultrasound Scanning