Review of the Impact of a Computer-aided Real-time Polyp Detection System on Adult Colonoscopy

Colonic Polyp

Colonic polyp, colonoscopy, artificial intelligence, implementation science

Background: Removal of adenomatous polyps during colonoscopy is associated with long-term prevention of colorectal cancer-related deaths. Recently, there have been much interest in the use of artificial intelligence (AI) platforms to augment the routine endoscopic assessment of the colon to enhance adenoma detection rate (ADR). To date, computer assisted detection of polyps (CADe) have been shown to be safe, with a significant increase in ADR, without any concomitant increase in post-procedural complications. Aims: The investigators aim to evaluate the use of GI GeniusTM Intelligent Endoscopy Module in a multi-ethnic Asian population (Singapore) to increase in ADR and adenoma detected per colonoscopy (ADPC)to justify its effectiveness as an adjunct in polyp detection and training for colonoscopy. Methods: This study will be a single-institution cohort study, conducted over a 2-year period. Sengkang General Hospital (SKH) does an estimated 12,500 colonoscopies per year, with an average of 1,040 colonoscopies performed every month. Thus, given the case volume, the investigators expect to detect differences in ADR amongst endoscopists if any during this study period. As part of the subgroup analysis, the investigators also aim to compare the ADR rates of trainee endoscopists with and without the GI GeniusTM Intelligent Endoscopy Module to ascertain its utility as an education tool/training adjunct

This study will be a single-institution cohort study, conducted over a 2-year period. The investigators will recruit patients prospectively who would have their colonoscopy performed with AI guidance from the GI Genius, which is a standard of care feature in the endoscopy rooms fitted with the GI Genius, and compare it against historical data when the AI technology was not available (Jan 2018-Jan 2021). This historical data has already been collected as part of a published audit performed for SKH's endoscopy. The data collected for this historical cohort are from an anonymised database collected from Operating Theatre Management Unit (OTMU) (who is not involved in the study). For the prospective cohort undergoing AI aided endoscopy (1/4/2023 - 31/3/2025): This will include all adult patients going for colonoscopy in the our institution. Patients with incomplete or failed colonoscopy, flexible sigmoidoscopy, colonoscopy done after previous colorectal cancers or previous colonic resections, patients with poor bowel preparation, when deemed by the endoscopist to have an incomplete assessment of the colon, will be excluded from the analysis. patient demographical data, procedural related data as well as histology from histological reports from their colonoscopy would be collected by research coordinators in order to ascertain ADR, Adenoma Detection Per Colonoscopy (ADPC) and Polyp Detection Rate (PDR). These coordinators are independent of the study team and will not be part of the analysis of the data. In order to have an effective barrier (for the study team not to be able to identify the patients from the data collected), therefore the coordinators act as a trusted independent party who will extract/de-identify the data before it is handed to the study team for analysis of the ADR ADPC and PDR. The study team will not attempt to re- identify the patients back. Analysis of data: Subsequent analysis will be conducted at every 3 months by reviewing for the absolute ADR and ADPC amongst endoscopists by the study team. By comparing these prospectively collected data with the historical cohort data, the investigators would be able to evaluate the effectiveness of the GI Genius in improving the ADR, PDR and ADPC. Cost analysis can also be done to analyse the cost effectiveness of the added AI feature.

No related papers found yet.

Singapore