Comprehensive Diagnostic Performance of Real-Time Characterization of Colorectal Lesions Using an Artificial Intelligence–Assisted System: A Prospective Study

The "resect-and-discard" and "leave in situ" strategies for colorectal polyps, which were proposed to reduce polypectomy-related costs, are attracting great attention.1 The American Society for Gastrointestinal Endoscopy proposed diagnostic performance thresholds (negative predictive value or surveillance interval agreement ≥90%) for these optical biopsy strategies.2 Artificial intelligence (AI)–assisted computer-aided characterization (CADx) has been investigated as a possible solution for implementing optical biopsy strategies.