Abstract Background Small bowel capsule endoscopy (SBCE) is central to Crohn’s Disease (CD) care but limited by lengthy review and reader variability. We evaluated whether an AI-triaged, human-reviewed workflow (SmartScan, SS) can replace routine full-video (FV) in the assessment of inflammatory burden, while reducing reading time and maintaining diagnostic safety under real-world conditions. Methods Single-center, prospective, randomized, assessor-blinded, dual-reader, 2 × 2 crossover non-inferiority study on a retrospective cohort of OMOM® HD studies. Adults with established CD underwent paired FV and SS reads - Figure 1. The primary endpoint was non-inferiority of the CAD-free proportion (agreement on Lewis Score, LS, category and stricture status). Secondary endpoints included categorical agreement, continuous equivalence for LS and CECDAI, within-tolerance non-inferiority, stricture sensitivity, stress testing under rapid transit and low Brotz QI, and reading-time superiority. Results We analyzed 153 paired studies from 141 patients. SS and FV agreed on both LS category and stricture in 96.1% (147/153), surpassing the non-inferiority margin (0.917, p = 0.005). LS category exact-match was 96.7%(148/153). Continuous scores were equivalent: LS-inflammation Δ = −16.4 (90% CI − 30.5 to − 2.3; margin ±120); CECDAI Δ = −0.24 (−0.39 to − 0.10; margin ±3); total LS Δ = −44.9 (−77.4 to − 12.4) - Figure 2. Within-tolerance agreement was 91.5% (140/153) and stricture sensitivity 100%. Stress-testing under short small-bowel transit time and low Brotz QI yielded similar results. SS reduced mean reading time fell from 41.5 to 12.8 minutes (p 0.001). Conclusion SS achieved decisional non-inferiority, preserved continuous score equivalence, and markedly reduced reading time, while maintaining safety. These data support an AI-first, human-verified SBCE workflow with selective escalation to FV and call for multicenter, multi-ecosystem validation. References: 1. Maaser C, Sturm A, Vavricka SR, et al. ECCO-ESGAR Guideline for Diagnostic Assessment in IBD Part 1: Initial diagnosis, monitoring of known IBD, detection of complications. J Crohns Colitis. 2019;13:144–164K. 2. Xie X, Xiao YF, Zhao XY, et al. Development and validation of an artificial intelligence model for small bowel capsule endoscopy video review. JAMA Netw Open. 2022;5:e2221992. 3. Giordano A, Romero-Mascarell C, González-Suárez B, et al. Integration of artificial intelligence-enhanced capsule endoscopy in clinical practice: a review of market-available tools for clinical practice. Dig Dis Sci. 2025; doi:10.1007/s10620-025-09099-4. 4. O’Hara FJ, Mc Namara D. Capsule endoscopy with artificial intelligence-assisted technology: real-world usage of a validated AI model for capsule image review. Endosc Int Open. 2023;11:E970–E975. Conflict of interest: Dr. Mesquita, Pedro: No conflict of interest Pinho, Rolando: No conflict of interest costa, catarina: No conflict of interest Teixeira, Pedro: No conflict of interest Ferreira, Rita: No conflict of interest Estevinho, Maria Manuela: No conflict of interest Ponte, Ana: No conflict of interest
Mesquita et al. (Thu,) studied this question.