Developing and Validating UroAgent, an AI Assistant for Diagnosing and Treating Urinary Stones, Prostate Cancer, and Bladder Cancer

What this Trial Is About

Urological diseases such as urinary stones, prostate cancer, and bladder cancer are common and often require specialized diagnosis and treatment. This research aims to develop and validate an artificial intelligence AI assistant called UroAgent, designed specifically for urology. UroAgent uses a large language model combined with tools for retrieving medical knowledge and analyzing images to support doctors in making faster, more consistent decisions. The study is observational, meaning patients medical treatments will not be changed. The study will use a total of 2,000 urological disease cases, including 1,500 retrospective cases recorded from 2010 to 2025 for developing the AI model and 500 new patient cases collected prospectively for independent validation. UroAgents diagnostic and treatment recommendations will be compared to those of senior urology specialists to evaluate accuracy, appropriateness, completeness, and safety across different disease subtypes. All patient records will be de-identified to protect privacy. Participants will have their clinical information, imaging data, and surgical videos reviewed without any changes to their care. The researchers will measure how well UroAgent agrees with expert diagnoses and treatment plans, focusing on diagnostic accuracy as the primary outcome. Secondary measures include expert ratings of accuracy, treatment recommendation appropriateness, clinical safety, and concordance with clinician diagnoses. The study will follow strict ethical standards and statistical analysis plans from 2025 to 2027.

Development of an AI-Agent for Urological Disease Diagnosis and Treatment