Impact of Computer-Aided Detection on Endoscopists' Learning Curve for Adenoma Detection: A Prospective Observational Study

Aims Computer-aided detection (CADe) systems have been introduced in clinical practice to enhance adenoma detection rates (ADR). Despite their use, uncertainty remains about the influence of CADe on endoscopists' skill development and potential deskilling. This study aims to evaluate the impact of CADe on the learning curve of endoscopists for ADR.

Methods This was a prospective, single-center, observational study conducted at a tertiary care hospital from January 2021 to December 2023. In this study, CADe systems were installed in half of the endoscopy units, and patients were equally distributed between the rooms. Patients in CADe-equipped rooms were examined using EndoBRAIN-EYE (Olympus Corp., Tokyo, Japan), while the control group underwent standard colonoscopy. The high-definition endoscope was used for all procedures.

Inclusion criteria included patients aged≥20 years. Exclusion criteria were polyposis, inflammatory bowel disease, incomplete colonoscopy, those examined by novices or known polyps. To address selection bias, we applied propensity score matching considering known ADR-related factors (indication, age, sex, and bowel preparation quality). The primary outcome was to evaluate the learning effect of CADe using cumulative summation (CUSUM) analysis. The CUSUM score was tested by dividing the sequence of assessment into blocks of consecutive 50 colonoscopies. Endoscopists were classified as high-detectors (ADR≥25%) or low-detectors (ADR<25%) based on the ADR recorded before CADe implementation. The CUSUM score was calculated based on data from standard colonoscopies conducted during the study period, as it was intended to evaluate how CADe affects the learning effect on standard colonoscopy performance. The secondary outcomes included ADR, advanced adenoma detection rate (AADR), sessile serrated lesion detection rate (SSLDR), and adenomas per colonoscopy (APC). The study received ethics approval (no. 19H072) and complied with the Declaration of Helsinki.

Results Out of 5,717 patients initially enrolled, 4,712 patients (2,726 male; mean age 65.3±12.6 years; 65.8% for FIT/screening) were analyzed after propensity score matching. CUSUM analysis showed that CADe led to an accelerated learning curve for high-detectors, as indicated by plots below the acceptable performance line, suggesting enhanced skill acquisition. In contrast, for low-detectors, CADe resulted in no significant change in learning trajectory, as plots remained between acceptable and unacceptable limits. The CADe group had a significantly higher ADR (40.5% [95% CI: 38.5-42.5%]) compared to the standard colonoscopy (32.4% [30.5-34.4%]; p<0.001). SSLDR was also significantly higher in the CADe group (6.9% [5.9-8.0%] vs. 4.5% [3.7-5.4%]; p=0.001), while AADR showed no significant difference (4.1% [3.4-5.0%] vs. 4.7% [3.9-5.6%]; p=0.39).

Conclusions The implementation of the CADe system did not result in deskilling regarding adenoma detection.

Publication History

Article published online:
27 March 2025