Comparison of Targeted vs Random Biopsies for Surveillance of Ulcerative Colitis-Associated Colorectal Cancer

摘要

BACKGROUND & AIMS: A random biopsy is recommended for surveillance of ulcerative colitis (UC)-associated colorectal cancer. However, a targeted biopsy might be more effective. We conducted a randomized controlled trial to compare rates of neoplasia detection by targeted vs random biopsies in patients with UC. METHODS: We performed a study of 246 patients with UC for 7 years or more, seen at 52 institutions in Japan from October 1, 2008 through December 31, 2010. Patients were randomly assigned to the random group (4 random biopsies collected every 10 cm in addition to targeted biopsies, n = 122) or the target group (biopsies collected from locations of suspected neoplasia, n = 124). The primary end point was the number of neoplastic lesions detected in a single surveillance colonoscopy. We estimated the ratio and difference in the mean number of neoplastic lesions between the groups. We also evaluated the non-inferiority between the groups as an exploratory study. A non-inferiority margin of 0.65 (0.13 of 0.20) was considered for the ratio of the mean number of neoplastic lesions between groups. RESULTS: The mean number of biopsies found to contain neoplastic tissue per colonoscopy was 0.211 (24 of 114) in the target group and 0.168 (18 of 107) in the random group (ratio of 1.251; 95% confidence interval, 0.679-2.306). The lower limit was above the non-inferiority margin of 0.65. Neoplasias were detected in 11.4% of patients in the target group and 9.3% of patients in the random group (P = .617). Larger numbers of biopsy samples per colonoscopy were collected in the random group (34.8 vs 3.1 in the target group; P < .001), and the total examination time was longer (41.7 vs 26.6 minutes in the target group; P < .001). In the random group, all neoplastic tissues found in random biopsies were collected from areas of the mucosa with a history or presence of inflammation. CONCLUSIONS: In a randomized controlled trial, we found that targeted and random biopsies detect similar proportions of neoplasias. However, a targeted biopsy appears to be a more cost-effective method. Random biopsies from areas without any signs of present or past inflammation were not found to contain neoplastic tissues.