Surveillance colonoscopy in patients with inflammatory bowel disease: comparison of random biopsy vs. targeted biopsy protocols

Endoscopic surveillance in patients with long-standing inflammatory bowel disease (IBD) improves early detection of intraepithelial neoplasia (IEN). We aimed to compare three different endoscopic surveillance strategies in the detection of IEN. One hundred fifty surveillance colonoscopies (ulcerative colitis, UC n = 141; Crohn's disease, CD n = 9) were carried out. Random quadrant biopsies were taken (group I, n = 50). Chromoendoscopy with indigo carmine was performed and subsequently quadrant biopsies were collected (group II, n = 50). Patients in group III (n = 50) underwent confocal endomicroscopy (CEM), and CEM-guided as well as random quadrant biopsies were taken (group III, n = 50). The findings of CEM were correlated to conventional histology. Patients with high-grade IEN underwent surgery or strict follow-up by patients' request. In group I (1531 biopsies), no IEN was detected by histology. In group II (1,811 biopsies), chromoendoscopy-guided biopsies revealed high-grade IEN in two patients (4% detection rate). In four patients of group III (1477 biopsies), areas with high-grade IEN were clearly visible by CEM and confirmed by histology (8% detection rate, p < 0.05). Of six patients with high-grade IEN, five patients underwent proctocolectomy. Colorectal cancer was detected in one out of five patients. Targeted biopsy protocols guided by either chromoendoscopy or CEM led to higher detection rates of IEN and are thus mandatory for surveillance colonoscopies in patients with long-standing UC. Random biopsy protocols should be replaced by chromoendoscopy-guided protocols.