REVIEW   

Minerva Surgery 2023 February;78(1):81-5

DOI: 10.23736/S2724-5691.22.09810-0

Copyright © 2022 EDIZIONI MINERVA MEDICA

lingua: Inglese

Artificial intelligence and upper gastrointestinal endoscopy: what is the optimal study design?

Natalie HALVORSEN ¹, Yuichi MORI ^{1, 2, 3} ✉

¹ Clinical Effectiveness Research Group, University Hospital of Oslo, University of Oslo, Oslo, Norway; ² Department of Transplantation Medicine, University Hospital of Oslo, Oslo, Norway; ³ Digestive Disease Center, Showa University Northern Yokohama Hospital, Yokohama, Japan

Upper gastrointestinal cancers (i.e., esophageal and gastric cancers) are common cancers worldwide with high mortality and morbidity. Although there is no randomized controlled trial-based evidence, early detection with endoscopy is expected to positively affect prognosis and morbidity. However, endoscopic procedures are always accompanied by human-induced errors such as overlooking of neoplasia and cancers. Recently, the use of artificial intelligence (AI) during upper gastrointestinal endoscopy is catching attention because it is expected to reduce human-induced variability of the examination. This review article introduces the overview of the expectation and current status of the AI tools for upper gastrointestinal endoscopy and shares possible challenges and corresponding solutions with readers.

KEY WORDS: Artificial intelligence; Endoscopy, digestive system; Research design