Targeted Biopsies Identify Larger Proportions of Patients With Colonic Neoplasia Undergoing High-Definition Colonoscopy, Dye Chromoendoscopy, or Electronic Virtual Chromoendoscopy.

Abstract

It is unclear what are the best and most appropriate endoscopic procedures for detecting colonic neoplasia in patients with long-term colonic inflammatory bowel disease (IBD). Dye chromoendoscopy (DCE) is the standard used in IBD surveillance colonoscopies. However, studies are needed to determine the optimal endoscopic technique for detecting dysplastic lesions. We investigated current practices used in surveillance colonoscopies by IBD gastroenterologists at a single tertiary center. We also determined the rate of neoplasia detection among different surveillance endoscopic techniques in an analysis of random or targeted biopsies. We collected data on 454 patients with IBD (54.5% male; mean age, 50 y; mean disease duration, 14.5 y; 55.9% with ulcerative colitis, 42.7% with Crohn's disease, and 1.3% with indeterminate colitis) who underwent surveillance colonoscopy from April 2011 through March 2014 at the University of Calgary in Canada. Subjects were examined using white-light standard-definition endoscopy (WLE), high-definition (HD) colonoscopy, virtual electronic chromoendoscopy (VCE), or DCE; random or targeted biopsy specimens were collected. Endoscopic and histologic descriptions with suspected neoplasia were recorded. Rates of neoplasia detection by the different endoscopic procedures were compared using chi-square analysis. Of the patients analyzed, 27.7% had WLE endoscopy with random collection of biopsy specimens, 27.3% had HD colonoscopy with random collection of biopsy specimens, 14.1% had VCE with random collection of biopsy specimens, 0.9% had DCE with random collection of biopsy specimens, 12.8% had HD colonoscopy with collection of targeted biopsy specimens, 11.9% had VCE with collection of targeted biopsy specimens, and 5.3% had DCE with collection of targeted biopsy specimens. Neoplastic lesions were detected in 8.2% of the procedures performed in the random biopsy group (95% confidence interval, 5.6-11.7) and 19.1% of procedures in the targeted biopsy group (95% confidence interval, 13.4-26.5) (P < .001). Neoplasias were detected in similar proportions of patients by HD colonoscopy, VCE, or DCE, with targeted biopsy collection. In a large cohort of IBD patients undergoing surveillance colonoscopy, targeted biopsies identified greater proportions of subjects with neoplasia than random biopsies. Targeted collection of biopsy specimens appears to be sufficient for detecting colonic neoplasia in patients undergoing HD colonoscopy, DCE, or VCE, but not WLE.