Colometer: A real-time quality feedback system for screening colonoscopy

Abstract

The results from a novel automated real-time colonoscopy quality feedback system strongly agreed with the endoscopists' quality assessments. Further study is required to validate this approach.

Artificial intelligence and computer-aided diagnosis in colonoscopy: current evidence and future directions

“…Filip et al developed software to provide automated outputs in three key areas: 1) Real-time visual feedback of image quality taking into account blurriness and changing velocity 2) Post procedure statistics including % time of adequate visualisation and withdrawal times 3) automated bowel preparation documentation. 64 In this study, 14 screening colonoscopy videos were analysed by the software and compared to three gastroenterologists who rated the videos using a scale for overall quality, withdrawal velocity, bowel preparation and image quality. The automated overall quality rating was strongly correlated with the reviewers' overall quality rating, although there was no correlation with mean endoscopists image quality rating.…”

Using artificial intelligence to improve adequacy of inspection in gastrointestinal endoscopy

“…Two groups have done studies to improve adequacy of inspection in the colon. In 2012 a Canadian group described a software program called colometer; the software reported on blurriness of individual frames and on the speed of withdrawal [59]. Velocity was measured as frame difference: rapid withdrawal leads to large differences between successive frames where as slow withdrawal does not.…”

Simulation-Based Training for Colonoscopy

“…Plooy et al 31 introduced an indirect measure of peak force using a force plate interposed between the table and the colon model and showed construct validity of their method. In a clinical setting, a recent study by Filip 32 tested a software (colometer) for real-time assessment of colonoscopy, and further studies is needed to validate this approach. Use of a scope guide has shown better completion rates in trainees, 33 and future studies are needed to see if a 3-D imager in combination with the scope guide could provide automatic nonbiased assessment of performance.…”