Evaluating an AI System to Detect and Diagnose Solid Pancreatic Lesions Using Endoscopic Ultrasound

What this Trial Is About

Researchers are evaluating an artificial intelligence system called iEUS-SPL intelligent endoscopic ultrasound system-solid pancreatic lesion designed to detect and classify solid pancreatic lesions during endoscopic ultrasound EUS examinations. This observational, prospective cohort study aims to validate how well iEUS-SPL performs in diagnosing various types of solid pancreatic lesions by analyzing EUS images, clinical data, and imaging features collected from patients. Participants are patients aged 18 years or older who are scheduled for EUS due to suspected solid pancreatic lesions based on symptoms, medical history, lab tests, or imaging. The study uses the iEUS-SPL device to automatically detect lesions in real-time EUS scanning videos and categorize them into five types pancreatic cancer, pancreatic neuroendocrine tumor, solid pseudopapillary tumor, autoimmune pancreatitis, and chronic pancreatitis. During the study, participants undergo EUS while iEUS-SPL analyzes the images and data to detect and classify lesions. Researchers will measure the systems accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and lesion detection rate during the procedure. They will also compare these results with those of expert endosonographers. Participant involvement includes consenting to the procedure and allowing their data to be used for evaluation, with the study continuing until June 2028.

An Artificial Intelligence System for Multimodal, Multi-class Diagnosing Solid Pancreatic Lesions Based on Endoscopic Ultrasound