Computer-assisted optical diagnosis of colorectal polyps up to 10mm

Background and study aims: Computer aided (CADx) diagnosis of colorectal polyps during colonoscopy could replace pathology for certain polyps. This study aimed to evaluate CADx-assisted diagnosis for polyps up to 10mm in size in the context of established quality benchmarks. Patients and methods: We performed a post-hoc analysis of a randomized controlled trial evaluating assistive versus autonomous CADx diagnosis (NCT05236790). Our primary outcome was achievement of the ASGE PIVI1 threshold for resect-and-discard implementation when CADx-assisted diagnosis was used for polyps ≤3mm. Secondary outcomes included PIVI1 threshold achievement when using CADx-assisted diagnosis for polyps ≤5mm, and ≤10mm, as well as diagnostic performance and advanced histology prevalence across polyp size groups. Results We included 313 patients with a total of 463 polyps ≤10mm undergoing CADx-assisted diagnosis. Surveillance interval agreement was 94.6% (95% CI 91.3-96.7), 89.5% (95% CI 85.4-92.5), and 85.9% (95% CI 81.5-89.5) when CADx-assisted diagnosis was used compared to pathology for polyps up to 3mm, 5mm, and 10mm, respectively. Diagnostic accuracy of CADx-assisted diagnosis was 76.2%, 76.6%, and 72.5% for ≤3mm, >3 to ≤5mm, and >5 to ≤10mm polyps. Rectosigmoid adenoma NPV was above 90% for all size groups (PIVI2). Prevalence of advanced or serrated pathology increased in polyps >3mm resulting in an increasing number of incorrectly assigned surveillances intervals. Conclusions: In our study, CADx-assisted diagnosis met the resect-and-discard PIVI1 threshold only for diagnosis of polyps up to 3mm and the diagnose-and-leave PIVI2 threshold for polyps up to 10mm.

Publication History

Received: 27 January 2025

Accepted after revision: 22 July 2025

Accepted Manuscript online:
23 July 2025