Rubin DT et al. (2007) Are dysplasia and colorectal cancer endoscopically visible in patients with ulcerative colitis? Gastrointest Endosc 65: 998–1004

Interpretation of systematic random biopsies of colonic mucosa is the leading approach for identifying dysplasia and colorectal cancer (CRC) in patients with IBD; however, the accuracy of this approach is limited by the small mucosal surface area sampled and by variation in the interpretation of pathology.

To determine whether endoscopy can visualize CRC and dysplasia in ulcerative colitis patients and, therefore, allow targeted biopsies to be taken, Rubin et al. retrospectively analyzed reports from 1,339 endoscopies performed in 622 patients diagnosed with ulcerative colitis and compared them with the corresponding pathology reports. The endoscopic procedures involved standard technology and both random and targeted biopsies of lesions were taken.

A total of 75 neoplastic lesions were identified on the basis of pathological examination of biopsies but only 46 were visible during endoscopy, the remainder were 'invisible' (i.e. were not described as suspicious by the endoscopist but were identified as neoplastic by the pathologist). The overall endoscopic sensitivities for neoplasia per-patient and per-lesion were 76.1% and 61.3% respectively. Per-patient endoscopic visibilities for dysplasia and CRC were 71.8% and 100% respectively.

The authors acknowledge that there was no quality assurance for biopsy sampling, endoscopic procedural variation or variation in endoscopist's experience. They suggest that their findings, which challenge the traditional dogma of 'invisible dysplasia' in ulcerative colitis, might result from the increasing optical resolution of colonoscopes. They conclude that endoscopically targeted biopsies complement random biopsies and should be included in future practice guidelines.