-
1
This retrospective multicenter study compared secondary-review outcomes before and after AI-assisted endoscopy was introduced in a mobile gastric cancer screening program in Gifu, Japan.
-
2
The cohort included 5013 screening examinations: 2021 in 2021–2022 before AI implementation and 2992 in 2023–2025 during AI use.
-
3
During the AI period, primary endoscopists used real-time CADe alerts, but they retained responsibility for interpretation and referral decisions.
-
4
The structured secondary review process remained unchanged across study periods and involved independent review of recorded images and reports before referral was finalized.
-
5
Secondary-review modifications were descriptive process outcomes, not measures of diagnostic accuracy or clinical benefit, because no external reference standard was available.
