Background: Procedure length and agreement in detection of abnormalities may limit implementation of chromoendoscopy (CE) for dysplasia surveillance in ulcerative colitis (UC). We investigated these factors among endoscopists inexperienced in this technique.
Methods: Six investigators performed surveillance colonoscopy with white light endoscopy (WLE) followed by CE on 75 patients with long-standing UC. Interobserver agreement for WLE and CE images of polyps and nonpolypoid mucosa was determined. Withdrawal times from the cecum were compared based on number of colonoscopies performed. Dysplasia detection rate with WLE was compared with CE.
Results: The analysis of 586 images (266 WLE and 320 CE) from 57 patients included images of 160 polyps (64 flat) with 29 dysplastic lesions. All investigators identified 10/11 WLE images of dysplasia and 4 identified all 18 CE dysplasia images, 1 missed 1 and 1 missed 3. Four dysplastic lesions were not identified by 1 or more investigators and all measured <5 mm. Interobserver agreement for lesions was high with kappa scores of 0.91 and 0.86 for WLE and CE, respectively. Among the 75 patients enrolled, dysplasia was found in 9.3% with WLE compared with 21.3% with WLE and CE (P = 0.007). Median colonoscopy withdrawal time improved from 31 minutes for endoscopists performing fewer than 5 procedures to 18 minutes for 5 to 14 and 19 minutes for more than 15 procedures.
Conclusions: Indigo carmine CE for UC surveillance resulted in high rates of interobserver agreement for polyp detection, acceptable withdrawal times, and enhanced dysplasia detection. These results are encouraging for the implementation of CE programs for chronic UC.