Endoscopy 2021; 53(S 01): S51
DOI: 10.1055/s-0041-1724380
Abstracts | ESGE Days
ESGE Days 2021 Oral presentations
Friday, 26 March 2021 14:00 – 14:45 AI in the colon: Better detection and characterisation of polyps? Room 6

Prospective Evaluation of a New Artificial Intelligence System For Detection of Colon Polyps

C Zippelius
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
J Schedel
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
D Brookman-Amissah
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
K Muehlenberg
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
W Schorr
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
A Salzberger
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
C Federle
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
,
F Zeman
2   University Medical Centre Regensburg, Centre for Clinical Studies, Regensburg, Germany
,
O Pech
1   Krankenhaus Barmherzige Brueder Regensburg, Department of Gastroenterology and interventional Endoscopy, Regensburg, Germany
› Author Affiliations
› Further Information
Also available at
 
 

    Aims The adenoma detection rate (ADR) varies significantly between different endoscopists leading to an adenoma miss rate (AMR) of up to 26 %. To improve endoscopic quality and to reduce the rate of interval cancer artificial intelligence (AI) systems can be valuable. We evaluated the efficacy of an AI system in real time colonoscopy and its influence on the AMR and the ADR in a clinical setting.

    Methods In this prospective study we analyzed 150 patients (age 65±14, 69 women, 81 men) undergoing diagnostic colonoscopy at a single endoscopy center in Germany from June to October 2020. The AI system GI Genius (Medtronic) detects polyps during real time colonoscopy by highlighting lesions with a frame. Every patient was examined at the same time by the endoscopist and the AI using two different opposing screens. The AI, which was overseen by a second observer, was not visible to the endoscopist. Primary outcome was the AMR. Absolute and relative frequencies are presented with 95 %-Confidence Intervals.

    Results There was no significant difference (p = 0.754) concerning the AMR between the AI system (6/197, 3.0 % [1.1-6.5]) and routine colonoscopy (4/197, 2.0 % [0.6-5.1]). The polyp miss rate of the AI system (14/311, 4.5 % [2.5-7.4]) was not significantly different (p = 0.720) from routine colonoscopy (17/311, 5.5 % [3.2-8.6]). There was no significant difference (p = 0.500) between the ADR with routine colonoscopy (78/150, 52.0 % [43.7-60.2]) and the AI system (76/150, 50.7 % [42.4-58.9]). Routine colonoscopy detected adenomas in two patients that were missed by the AI system.

    Conclusions We found that the AI system proves as a valuable second observer during real time colonoscopy and can keep up even with experienced endoscopists with an ADR >50 %. Its application in routine colonoscopy could decrease the performance variability between endoscopists and increase the overall ADR in less experienced endoscopists. DRKS00022279.

    Citation: Zippelius C, Schedel J, Brookman-Amissah D et al. OP121 PROSPECTIVE EVALUATION OF A NEW ARTIFICIAL INTELLIGENCE SYSTEM FOR DETECTION OF COLON POLYPS. Endoscopy 2021; 53: S51.


    #

    Publication History

    Article published online:
    19 March 2021

    © 2021. European Society of Gastrointestinal Endoscopy. All rights reserved.

    Georg Thieme Verlag KG
    Rüdigerstraße 14, 70469 Stuttgart, Germany