Study device not available

Computer aided detection (CADe) algorithms have been developed to overcome human errors and assist endoscopists in detecting more polyps during colonoscopy. The aim of this study was to investigate the accuracy of the novel Pentax Discovery CADe system (Discovery-AI) against pre-recorded videos of colon polyps of various size, shape and pathology while using videos of normal colon segments as controls from two different institutes.

Details regarding polyp size and location, morphology including Paris classification, optical assessment, and bowel preparation were prospectively collected and recorded in the online Redcap software. Final results of polyp histology were also collected. The video library was then independently reviewed for quality assessment by 3 experienced gastroenterologists. Videos that passed the initial quality assessment were then independently evaluated using the Discovery-AI system.

Aim of the study was to evaluate overall performance of Discovery-AI system for polyps of various size, morphology and pathology utilizing a prospectively developed video library.

Inclusion Criteria:

• Patients scheduled for a regular or screening colonoscopy at the institution as standard of care.
• Over 18 years of age

Exclusion Criteria:

• None if they qualify based on inclusion criteria
• Pregnant patients
• Inmate or prisoners
• Unable to provide informed consent

Currently, the most commonly used quality marker in conventional colonoscopy (CC) is adenoma detection rate (ADR), which has been shown to inversely correlate with the incidence of CRC.3 Although colonoscopy is generally considered to be the most accurate screening modality, a substantial number of polyps are still missed. Hence population scheduled for a colonoscopy at either of the institution were part of the project.