Narrow-band imaging versus I-Scan for the real-time histological prediction of diminutive colonic polyps: a prospective comparative study by using the simple unified endoscopic classification

Abstract

Background: Digital chromoendoscopy (DCE) has the potential for the in vivo optical diagnosis of colon polyps. However, no comparison of different DCE technologies currently exists.

Objective: To compare the diagnostic efficacies of narrow-band imaging (NBI) with those of I-Scan for the real-time histological prediction of diminutive colonic polyps (DCPs) (≤5 mm) by using the simple unified endoscopic classification.

Design: Prospective cohort study.

Setting: Academic hospital.

Patients: In total, 296 DCPs from 142 patients undergoing screening or surveillance colonoscopy were assessed.

Interventions: All DCPs detected during withdrawal were evaluated for the surface details by using high-definition white-light colonoscopy, and thereafter by using DCE (NBI or I-Scan) without optical magnification. The histology of all polyps was predicted in real-time and confirmed through the evaluation of biopsy or polypectomy specimens.

Main outcome measurements: Diagnostic efficacies of NBI and I-Scan.

Results: NBI and I-Scan displayed a significantly higher sensitivity and improved accuracy compared with high-definition white-light colonoscopy for the prediction of adenomas (P < .05). No significant differences were evident between the NBI and I-Scan (sensitivity, 88.8% vs 94.6%; specificity, 86.8% vs 86.4%; accuracy, 87.8% vs 90.7%, respectively; P > .05). Additionally, substantial levels of intra- and interobserver agreement between the NBI and I-Scan were measured (κ values >0.7).

Limitations: No randomized or crossover design.

Conclusions: NBI and I-Scan displayed a similar efficacy for the real-time histological prediction of DCPs. The simple unified endoscopic classification can be used for the interpretation of DCE, regardless of the type of technology. (

Clinical trial registration number: NCT1133041.).

Trial registration: ClinicalTrials.gov NCT01133041.