High-definition chromoendoscopy significantly outperformed high-definition white-light endoscopy for detecting dysplastic lesions in patients with inflammatory bowel disease, according to the findings of a single-center prospective randomized trial.
In the intention-to-diagnose analysis, rates of dysplasia detection were 11% for high-definition chromoendoscopy and 5% for high-definition white-light endoscopy (P = .032). The per-protocol analysis produced a similar result (12% vs. 5%, respectively; P = .027). High-definition chromoendoscopy also detected significantly more dysplastic lesions per 10 minutes of colonoscope withdrawal time in the per-protocol analysis, although the difference did not reach statistical significance in the intention-to-diagnose analysis.
Overall, the findings “support the use of chromoendoscopy for surveillance of patients with inflammatory bowel diseases,” Bjarki Alexandersson, a PhD student at Karolinska University Hospital in Solna, Stockholm, and his associates wrote in Clinical Gastroenterology and Hepatology. Patients with inflammatory bowel disease are at increased risk for colorectal cancer. Most guidelines support chromoendoscopy for the surveillance of these patients, as do the results of two recent meta-analyses in which chromoendoscopy detected significantly more dysplasias among patients with inflammatory bowel disease than did white light endoscopy. However, in subgroup analyses of these studies, the difference only emerged when comparing chromoendoscopy with standard (not high-definition) white-light endoscopy. “Thus, the evidence in support of chromoendoscopy using high-definition endoscopes is weak,” the researchers wrote.
For the study, they prospectively enrolled 305 patients with ulcerative colitis or Crohn’s disease who were referred for surveillance colonoscopy at an academic hospital in Sweden from March 2011 through April 2016. Participants were randomly assigned to receive either high-definition chromoendoscopy with indigo carmine (152 patients) or high-definition white-light endoscopy (153 patients).
In the intention-to-diagnose analysis, dysplasias were detected in 17 (11%) patients evaluated by high-definition chromoendoscopy, compared with 7 (5%) patients evaluated by high-definition white-light endoscopy (P = .032). After excluding 20 patients for inadequate bowel preparation, 18 patients for protocol violations, and 4 patients for incomplete colonoscopies, the per-protocol population consisted of 263 patients. Dysplasias were detected in 12% of patients evaluated by high-definition chromoendoscopy and in 5% of those evaluated by high-definition white-light endoscopy (P = .027).
All patients also had 32 samples collected by random biopsy, which used to be standard for detecting dysplasia in inflammatory bowel disease but has become more controversial in the era of video endoscopy, the researchers noted. In all, random biopsy evaluation identified dysplasias in nine patients, including six in the high-definition chromoendoscopy group and three in the high-definition white-light endoscopy group. Random biopsies were low-yield, identifying dysplasias in 0.092% of all specimens and 3% of colonoscopies. However, 20% of patients with dysplasias were identified only through random biopsy. This finding resembles that of another recent randomized trial in which 13% of patients with inflammatory bowel disease had dysplasias detected only through random biopsy (Gut. 2018;67:616-24), the researchers noted.
They also evaluated the number of macroscopic dysplastic lesions identified for every 10 minutes of colonoscope withdrawal time. These numbers were not significantly different in the intention-to-diagnose analysis (0.066 lesions in the high-definition chromoendoscopy group vs. 0.027 lesions in the high-definition white-light endoscopy group; P = .056). However, the per-protocol analysis revealed a significant difference (0.073 vs. 0.029 dysplastic lesions, respectively; P = .031).
“Based on our findings, we recommend the use of high-definition chromoendoscopy in inflammatory bowel disease surveillance,” the researchers concluded. They acknowledged several limitations: the study included patients from only one center, most dyplastic lesions were small and, thus, had an unclear natural history, and the endoscopists included both experts and nonexperts.
The Stockholm City Council provided funding. The researchers reported having no conflicts of interest.
SOURCE: Alexandersson B et al. Clin Gastroenterol Hepatol. 2020 Apr 27. doi: 10.1016/j.cgh.2020.04.049.