P0549 - Endoscopist and Procedure-Level Factors Associated With Increased Adenoma Detection With the Use of a Computer-Aided Detection Device

Award: Outstanding Research Award in the General Endoscopy Category

Award: Presidential Poster Award

Aasma Shaukat, MD, MPH¹, David Lichtenstein, MD², Daniel Chung, MD³, Sahin Coban, MD⁴, Yeli Wang, PhD⁵, Emma Navajas, BS⁵, Daniel Colucci, BS⁵, Shrujal Baxi, MD⁵, William Brugge, MD^6
1New York University Grossman School of Medicine, New York, NY; ²Boston Medical Center, Boston University School of Medicine, Boston, MA; ³Massachusetts General Hospital, Harvard Medical School, Boston, MA; ⁴University of North Dakota, School of Medicine & Health Sciences, Grand Forks, ND; ⁵Iterative Health Inc., Cambridge, MA; ⁶Mount Auburn Hospital/Harvard Medical School, Boston, MA

Introduction: Colonoscopy quality variability due to endoscopist characteristics reduces its effectiveness in detecting potential precursors to colorectal cancer (CRC). Recent studies using computer-aided detection (CADe) devices have shown increases in adenoma detection rate (ADR) and adenomas per colonoscopy (APC). However, the effect of baseline endoscopist and colonoscopy factors on CADe remains unexplored. This study analyzes the association of procedure and provider-level factors associated with APC in a randomized study of CADe vs. standard colonoscopy.

Methods: We randomized patients from the SKOUT registration study aged 40+ undergoing average risk screening or surveillance into CADe and standard colonoscopy arms. Colonoscopy factors included practice type, procedure time of day, withdrawal time, and bowel preparation. Endoscopist factors included baseline ADR and years of experience (YOE). The outcome variable was endoscopist APC, compared between standard and CADe arms in each colonoscopy, and operator factor subgroup by Z-statistic.

Results: A total of 1,423 participants (standard = 709; CADe = 714) were randomized. Among 22 participating endoscopists, 50% were from community practices, averaged 21 YOE, and had a mean baseline ADR of 45.5%. For procedures, 72.1% were performed before 12 PM, 98.3% had adequate preparation, and mean procedure and withdrawal times were 18.0 and 11.3 minutes, respectively. APC increased with CADe in both community and academic centers, and in procedures before/after 12 PM. Patients with a withdrawal time of ≥8 minutes and adequate bowel preparation experienced an increase in APC with CADe. Endoscopists with >20 YOE and baseline ADR < 45% had the highest increase in APC. There were no significant differences in the effect size for increased APC with CADe based on physician baseline ADR and YOE.

Discussion: Our study evaluated endoscopist and colonoscopy procedure-level factors to determine the most effective settings for CADe devices. We found that CADe was beneficial in both community and academic centers at all times of day. In procedures with adequate bowel preparation and withdrawal times of ≥8 minutes, CADe increased APC suggesting that existing guidelines for high quality colonoscopy must be maintained with the use of CADe. Additionally, our results suggest that experienced endoscopists may have a larger increase in their APC with CADe. With real world application of CADe, settings where CADe may be most beneficial will need to be carefully considered.




Disclosures:


Aasma Shaukat, MD, MPH¹, David Lichtenstein, MD², Daniel Chung, MD³, Sahin Coban, MD⁴, Yeli Wang, PhD⁵, Emma Navajas, BS⁵, Daniel Colucci, BS⁵, Shrujal Baxi, MD⁵, William Brugge, MD⁶. P0549 - Endoscopist and Procedure-Level Factors Associated With Increased Adenoma Detection With the Use of a Computer-Aided Detection Device, ACG 2023 Annual Scientific Meeting Abstracts. Vancouver, BC, Canada: American College of Gastroenterology.