TOPLINE

Medical centers with the provision of AI-based computer-aided detection (CADe) devices during colonoscopy had higher adenoma detection rates (ADRs) than centers without these devices, with benefits seen even among high-performing endoscopists.

METHODOLOGY

• Although randomized controlled trials have demonstrated that the use of CADe with colonoscopy significantly improves ADR, real-world studies on the implementation of CADe devices have yielded mixed results.
• Researchers conducted a pragmatic, quality improvement trial across Veterans Health Administration facilities using colonoscopy data between October 2021 and June 2023; 42 facilities were selected to receive CADe devices, and 97 facilities served as control sites; installation of devices in the intervention sites occurred between October 2022 and February 2023.
• They analyzed 334,200 colonoscopies performed by 816 endoscopists: 269 endoscopists at 42 CADe facilities (71,594 colonoscopies before and 35,399 after device deployment) and 547 endoscopists at 97 control facilities (151,792 before and 75,415 after a random deployment date).
• Endoscopists who performed at least 100 colonoscopies both before and after device deployment were grouped by baseline ADRs, and the analysis tested whether the device benefit varied across these performance levels.
• The primary outcome was the change in ADR between before and after device deployment. ADR was defined as the proportion of colonoscopies, regardless of indication, with at least one adenoma detected.

TAKEAWAY

• Before device deployment, ADR was lower at CADe facilities than at control facilities (50.7% vs 51.8%), whereas after deployment, ADR was higher at CADe sites (54.9% vs 51.1%; P < .0001 for both).
• After adjustment, the availability of CADe devices was associated with 22% increased odds of adenoma detection (adjusted odds ratio, 1.22; P < .05 ).
• No significant associations were observed between the availability of CADe devices and the detection of adenocarcinoma, the detection of sessile serrated lesions, the likelihood of obtaining pathology specimens without adenoma or adenocarcinoma, or withdrawal time.
• ADR increased for endoscopists across all baseline performance levels after the introduction of CADe devices, with similar gains regardless of prior endoscopist performance.

IN PRACTICE

“Considering these results and generally positive endoscopist feedback, the VA [Veterans Health Administration] has subsequently provided CADe devices to all facilities to help ensure high-quality colonoscopy,” the authors of the study wrote.

SOURCE

The study was led by Jason A. Dominitz, MD, MHS, National Gastroenterology and Hepatology Program, Veterans Health Administration in Washington, DC. It was published online in Gastroenterology.

LIMITATIONS

Data inaccuracies are uncommon in the Veterans Health Administration Quality Assurance Program and are unlikely to have differed between CADe and control sites. Although information about the staff physician who performed the colonoscopies was available, the involvement of trainees could not be determined. The lack of data on polyp size prevented the assessment of the association between the availability of CADe devices and the detection of advanced neoplasia.

DISCLOSURES

This study received support from the Veterans Health Administration Quality Enhancement Research Initiative. One author disclosed serving as a consultant for a medical technology company.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.