AI-Assisted Pathologist Performance Improvement: A Multicenter, Prospective, Randomized Controlled Trial

The investigators plan to conduct a multicenter, prospective, randomized controlled trial to systematically evaluate the added value of pathology-based AI models in the gastric cancer diagnostic workflow. The study will focus on comparing AI-assisted platform interpretation with conventional independent slide reading in terms of diagnostic accuracy (e.g., AUC), reading efficiency (e.g., comparison of time to diagnosis), quality of diagnostic reports, diagnostic confidence (Likert scale), and pathologists' satisfaction with the AI models. The investigators will also assess superiority for less-experienced (junior) pathologists and noninferiority for more-experienced (senior) pathologists. Successful completion of this project will provide high-level prospective evidence to support the standardized deployment, quality control, and broader application of pathology AI in the gastric cancer care pathway.