Endoscopy 2025; 57(S 02): S207-S208
DOI: 10.1055/s-0045-1805511
Abstracts | ESGE Days 2025
Moderated poster
Diagnosis and performance in endoscopy: an update 03/04/2025, 16:00 – 17:00 Poster Dome 1 (P0)

Experienced Endoscopist and Computer-Aided Detection of Colorectal Polyps : just an interesting gadget or „must have“ in every endoscopy unit?

Authors

  • K Grubelic Ravic

    1   University Hospital Centre Zagreb, Zagreb, Croatia

  • I Skrlec

    1   University Hospital Centre Zagreb, Zagreb, Croatia

  • L Zlopasa

    1   University Hospital Centre Zagreb, Zagreb, Croatia

  • L Franolic

    1   University Hospital Centre Zagreb, Zagreb, Croatia

  • Ž Krznarić

    1   University Hospital Centre Zagreb, Zagreb, Croatia
Further Information
Also available at
 

Aims One of the first applications of Artificial Intelligence (AI) in gastroenterology was the detection of colorectal polyps. According to the literature, there are concerns regarding the usefulness of endoscopic technological innovations in colorectal cancer (CRC) screening programs, since they do not always increase the ADR (adenoma detection rate) when patients undergo high-quality colonoscopy by experienced endoscopist with a high ADR. According to existing studies performed, an ADR increase of 1% was associated with a CRC incidence decrease of 3%.

Methods In our single-center retrospective study, we analyzed ADR of senior experienced endoscopists (each with more than 600 colonoscopies performed per year) during a period of 1 year without AI assistance (control group) and compared the results with real-time use of AI CADe system during colonoscopy (CADe group) through a 1-year period also. Both endoscopists together performed 3150 colonoscopies (n=1777 colonoscopies performed in control group, n=1373 in CADe group). All resected polyps underwent patohistological analysis, and all hyperplastic polyps were excluded from the analysis, which corresponds to 10% of the total number of resected polyps [1] [2].

Results In control group, n=1777, there were 693 polyps detected and resected which results in ADR of 38.99%. In CADe group, n=1373, there were 727 polyps detected which results in ADR of 52.94%.

Conclusions We can conclude that even at experienced endoscopists AI assisted colonoscopy increased ADR by 13.95% compared to conventional colonoscopy. This study confirms that novel CADe systems improve ADR while used by experienced endoscopists and lead to improved diagnostics for precancerous lesions and consequently CRC.



Publication History

Article published online:
27 March 2025

© 2025. European Society of Gastrointestinal Endoscopy. All rights reserved.

Georg Thieme Verlag KG
Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany