Randomized trial compares efficacy of linked color imaging and narrow-band imaging for adenoma detection in colonoscopy, suggesting implications for clinical practice.
Introduction: Right-sided lesions are frequently missed during colonoscopy, and the diagnostic performance of image-enhanced endoscopy (IEE) modalities remains debated. This multicenter randomized controlled trial compared linked color imaging (LCI) with second-generation narrow-band imaging (2G-NBI) for adenoma detection and miss rates in the right colon. Methods: Patients undergoing colonoscopy at six centers in Japan between November 2020 and November 2023 were randomly assigned (1:1) to primary right-colon observation with LCI or 2G-NBI, followed by secondary white-light imaging (WLI). The primary outcome was adenoma detection rate (ADR) in the right-sided colon during primary IEE observation. Secondary outcomes included polyp detection rate (PDR) and miss rates of adenomas and polyps. Results: A total of 1,528 patients were analyzed (LCI, n = 776; 2G-NBI, n = 752). ADR did not significantly differ between the LCI group and the 2G-NBI group (26.7% vs 23.4%, P = 0.14). PDR was also comparable (32.6% vs 28.9%, P = 0.12). However, secondary WLI identified fewer additional polyps after LCI than after 2G-NBI (12.4% vs 16.4%, P = 0.026). Per-lesion miss rates were significantly lower with LCI for both adenomas (19.9% vs 31.4%) and polyps (21.5% vs 31.9%) (both P < 0.001). Multivariable analyses confirmed the association between LCI and a reduced risk of missed lesions, with more pronounced associations observed in cases of poor bowel preparation, procedures by expert endoscopists, and observation times ≤70 seconds. Conclusion: Although ADR did not differ significantly between modalities, lower miss rates in the right-sided colon were observed with LCI.
No takes yet. Share an insight, caveat, or question.
Ono et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: