Surveillance in ulcerative colitis: is chromoendoscopy-guided endomicroscopy always better than conventional colonoscopy? A randomized trial.

Ulcerative colitis (UC) is associated with an increased risk of colorectal cancer. Chromoendoscopy showed superiority to conventional colonoscopy (CC) in surveillance studies including high-risk patients. We aimed to compare chromoendoscopy-guided endomicroscopy (CGE) with CC for intraepithelial neoplasia (IN) detection in patients with longstanding UC without primary sclerosing cholangitis and/or history of IN. One hundred sixty-two patients with longstanding (≥ 8 yr) distal/extensive UC and without primary sclerosing cholangitis and/or history of IN were prospectively randomized to undergo CGE (group A) or CC (group B). Seventeen patients were excluded. In group A (n = 72), circumscribed lesions highlighted by pan-chromoendoscopy were evaluated by endomicroscopy, and targeted biopsy/polypectomy was performed. In group B (n = 73), 4 random biopsies every 10 cm and targeted biopsy/polypectomy of detected lesions were performed. Thirteen IN, all low grade, were detected: 7 IN in group A and 6 in group B (P > 0.05), distributed, respectively, by 6 and 4 patients (P > 0.05). Significantly, more biopsies were performed in group B (4.7 ± 4.9 versus 36.0 ± 6.2, P < 0.001), and the per-biopsy yield of IN was higher in group A (1/48 versus 1/438, P < 0.001). Examination time was 61.5 ± 15.6 minutes in group A and 40.7 ± 8.7 minutes in group B (P < 0.001). The IN detection by endomicroscopy revealed: sensitivity = 85.7%, specificity = 97.9%, positive predictive value = 75.0%, and negative predictive value = 98.9%. CGE does not improve the detection of IN in the endoscopic screening of patients with longstanding UC without primary sclerosing cholangitis and/or history of IN. CGE takes longer than CC, but it decreases the number of biopsies performed and significantly increases the per-biopsy yield of IN. Endomicroscopy is an accurate tool for IN detection.