Computer Aided Detection (CADE) In Colonoscopy: An End-User Experience Using Two Systems

Aims CADe is a novel technology developed to increase adenoma detection rate (ADR). Recent studies have shown its effectiveness, however data on its acceptance in the endoscopy suite is lacking. This study investigates the perception of endoscopists towards CADe.

Methods We performed a prospective, multicenter study including endoscopists at different levels of experience. No endoscopists had prior experience with CADe. 2different systems were used: Medtronic GI Genius and Fujifilm CAD EYE. For 10weeks, all colonoscopies performed with CADe were prospectively registered and assessed via a questionnaire, using LimeSurvey. In these, using slider bars (0-100), we explored the subjective experience and perceived performance of CADe, as well as the number of detected relevant lesions by the endoscopist and/or CADe. Relevant lesions were defined as adenomas and sessile serrated lesions.

The data were analyzed using SPSSv26, normality was tested with a Shapiro-Wilk test and continuous variables were compared with an independent sample T-test.

Results 791 colonoscopies were performed by 19 endoscopists of whom 3 trainees and 4 young (<5 years experience) consultants.

In total 720 lesions were detected, 54 thanks to CADe, which entails an increase of 8.1 %lesions detected and an increase in lesion detection rate from 41.2 % to 43.0 %.

Endoscopists scored CADe as user-friendly (63.9), not distracting (25.1), nor timewasting (26.6). In contrast, the auditory signal wasn’t positively perceived (23.1).

While endoscopists state they’re triggered to better characterize the polyps (70.6), they don’t feel their ADR increases (53.3), nor do they feel more confident (57.8). Overall CADe was experienced as having some added value (61.9), without significant differences between the 2systems.

Trainees feel more positive about CADe, giving higher scores for perceived ADR increase (64.9; p = 0.002), confidence (65.1; p = 0.037) and overall added value (74.5; p < 0.001).

Conclusions The added value of CADe is perceived lower than the increase in lesion detection probably warrants. This might hinder the implementation of CADe in daily practice. More effort will have to be invested in convincing endoscopists to start using this technology.

Citation Van Langendonck S, Corens P, Stragier E et al. eP476 COMPUTER AIDED DETECTION (CADE) IN COLONOSCOPY: AN END-USER EXPERIENCE USING TWO SYSTEMS. Endoscopy 2021; 53: S254.

Publication History

Article published online:
19 March 2021