P0393 - Artificial Intelligence Improves Adenoma Detection Amongst Gastroenterology Fellows

Nikita Chadha, DO¹, Marcus Healey, DO¹, Puneet Puri, MD², HoChong Gilles, NP², Michael Fuchs, MD², Joseph Spataro, MD², Alvin Zfass, MD^2
1Virginia Commonwealth University Health System, Richmond, VA; ²Hunter Holmes McGuire VA Medical Center, Richmond, VA

Introduction: Increased adenoma detection rates (ADR) are associated with a reduction of interval colorectal cancer (CRC). To improve ADR, the real-time computer-aided polyp detection (CADe) systems were validated. This study aims to evaluate the impact of applying a FDA approved CADe to evaluate the ADR in colonoscopy amongst supervised first through third year gastroenterology (GI) fellow trainees.

Methods: This single-center, IRB approved, prospective study was performed at a Veterans Affair Medical Center from 2020 to 2022. Twelve US board-certified faculty supervised GI fellow trainees during colonoscopy a year with and without implementation of a CADe. Inclusion criteria included Veterans undergoing high-definition white light colonoscopy for screening, post-polypectomy surveillance, or follow up screening after a positive fecal immunohistochemical test. Those with a history of CRC, inflammatory bowel disease, prior colon resection, on antithrombotic therapy which precludes polyp resection, inadequate bowel preparation, incomplete colonoscopy, or a withdrawal time of less than six minutes were excluded. Descriptive statistics were performed and a p value < 0.05 was significant.

Results: 1,020 patients were included (with CADe n=435, without CADe n=585). Compared to colonoscopy without, colonoscopy with CADe had higher polyp detection rate (74.3% vs. 62.6%; p< 0.0001) and adenoma detection rate (63.9% vs. 52.8%; p=0.0003). Withdrawal time was comparable in non-interventional colonoscopies (with CADe 12.31 minutes vs. without CADe 12.34 minutes). Additionally, non-neoplastic resection rates were similar (with CADe 10.3% vs. without CADe 9.7%). When stratified by fellowship year, under faculty supervision, there was a 20% increase in ADR amongst first year fellows while using CADe (with 68.6% vs. without 48.5%; p=0.0003). A majority of colonoscopies performed by first year fellows were supervised by faculty with more than 10 years of experience.

Discussion: Colonoscopy with CADe markedly improved ADR without compromising safety. These data demonstrate the value of incorporating technology in both fellowship training and long-term practice.


Disclosures:


Nikita Chadha, DO¹, Marcus Healey, DO¹, Puneet Puri, MD², HoChong Gilles, NP², Michael Fuchs, MD², Joseph Spataro, MD², Alvin Zfass, MD². P0393 - Artificial Intelligence Improves Adenoma Detection Amongst Gastroenterology Fellows, ACG 2024 Annual Scientific Meeting Abstracts. Philadelphia, PA: American College of Gastroenterology.