Impact of Availability of Computer-Aided Detection Devices on Adenoma Detection During Colonoscopy: A Cluster Randomized Study.

Abstract

Endoscopists' colonoscopy adenoma detection rates (ADRs) are inversely associated with their patients' risk of postcolonoscopy colorectal cancer death. Although randomized controlled trials have found that computer-aided detection (CADe) devices significantly improve ADRs, real-world studies have generally failed to demonstrate a benefit. In this pragmatic quality improvement study, randomly selected Veterans Health Administration facilities were equipped with CADe devices between October 2022 and February 2023. Control facilities were assigned a random deployment date. Adenoma detection was determined for colonoscopies, and mixed effects logistic regression was used to assess the independent effect of CADe availability on colonoscopy outcomes. Colonoscopy data were available from 269 endoscopists at 42 facilities with CADe performing 71,594 pre- and 35,399 postdeployment colonoscopies; and from 547 endoscopists at 97 control facilities performing 151,792 pre- and 75,415 postdeployment colonoscopies. Comparing post- with predeployment colonoscopies, the absolute increase in ADR was 4.2 percentage points (95% CI, 3.48 to 4.74 percentage points), from 50.7% to 54.9%, at CADe sites and ADR decreased by 0.7 percentage points (95% CI, -1.16 to -0.28 percentage points), from 51.8% to 51.1%, at control sites. Using multivariable regression, CADe availability was associated with a significantly increased odds of adenoma detection (odds ratio, 1.22; 95% CI, 1.15 to 1.28), but was not associated with detection of other lesions or changes in withdrawal time. Baseline endoscopist ADR quintiles were not associated with the impact of CADe availability on ADR. Availability of CADe devices was associated with a significant increase in adenoma detection compared with usual care. The impact on cancer outcomes and potential endoscopist deskilling remains undetermined. gov Number: NCT05888623.