Key points

Video of Endoscopic Submucosal Dissection (ESD) of a non-polypoid dysplastic lesion in ulcerative colitis accompanies this article at http://www.giendo.theclinics.com/

Introduction

The risk of developing IBD-colitis-related colorectal cancer has been highlighted for many years. Early data suggested that the risk increased year on year with an 18% risk at 30 years and the initial British guidelines advocating shortening of surveillance intervals with each decade of disease. Subsequent data suggested the stronger influence of patient factors, including disease extent and activity, family history of colorectal cancer, endoscopic features (strictures or postinflammatory polyps) and previous dysplasia, rather than duration of disease alone, with the current generation of European guidelines advocating risk-based stratification. More recently, some population-based studies have suggested that previous results overestimate the risk of IBD dysplasia and cancer because of case selection from academic and tertiary centers.

Alongside risk-based stratification, a new concept emerged for the management of polypoid dysplasia in IBD, in that polypoid circumscribed lesions (adenoma like masses) even within the colitic segment, might be safely managed by endoscopic resection and close follow-up rather than by panproctocolectomy. A recent meta-analysis of 10 studies with more than 370 patients and 1700 years of patient follow-up supports this concept: 5 (95% confidence interval, 3–10) cancers developed per 1000 years of patient follow-up. The rate of dysplasia detected at subsequent colonoscopy was 65 cases per 1000 years of patient follow-up, emphasizing that close colonoscopic surveillance is mandatory. However, all the studies in this meta-analysis predate the use of chromoendoscopy. The need for proctocolectomy when dysplasia is detected in IBD is based on older data, which suggested a 19% cancer rate at immediate proctocolectomy when low-grade dysplasia was detected and as much as 42% when high-grade dysplasia was found. These data almost certainly related to a previous generation of endoscopes and endoscopists, the latter being less familiar than present-day endoscopists are with the appearances of nonpolypoid colorectal neoplasms, dysplasia, and cancer in IBD and hampered by a lack of high-quality endoscopic imaging. Furthermore, these endoscopists did not enjoy the advantages of high-definition, wide-angle endoscopes and dye-spray or image-enhanced endoscopy including structure enhancement, narrow-spectrum endoscopy (narrow band imaging [NBI, Olympus, Tokyo, Japan], Fujinon intelligent chromoendoscopy [FICE, Fujinon, Tokyo, Japan], i-Scan, image-enhanced endoscopy [Pentax, Tokyo, Japan]), autofluorescence, or confocal endomicroscopy (see the article on advanced imaging elsewhere in this issue). Therefore, dysplasia detected in the current era of endoscopes and endoscopists is likely to be at an early stage and can be safely managed by endoscopic resection if polypoid and circumscribed.

However, not all dysplasia detected at endoscopy in IBD is polypoid. The concept of flat dysplasia or endoscopically invisible dysplasia, detectable only by random biopsies has been commonly accepted, particularly in the prechromoendoscopy era, leading to previous generations of guidelines advocating the use of quadratic biopsies every 10 cm of colonoscopic withdrawal to detect this invisible dysplasia. This recommendation is poor for detection of early dysplasia, with one simulation paper based on colonic surface areas and dysplasia size suggesting that the standard 32 nontargeted biopsies would only detect an area of dysplasia encompassing 5% or more of the colonic surface with 80% certainty. The use of the word flat for biopsy-only-detected dysplasia is unfortunate because this word has also been used to describe nonpolypoid dysplasia in the endoscopic literature as part of the Paris classification. Flat or nonpolypoid in the endoscopic literature corresponds to Paris 0-IIa, flat elevated lesion; Paris 0-IIb, completely flat lesions; and Paris 0-IIc, depressed lesions. Many instances of patients diagnosed with flat biopsy-only dysplasia can be converted to circumscribed areas of dysplasia described as Paris 0-IIa, IIb, or IIc by reexamination with meticulous bowel preparation, with the patient in full remission, with an experienced endoscopist familiar with dysplasia in IBD, and with the use of high-definition endoscopes with dye-spray and image enhancement. If one accepts that circumscribed areas of flat dysplasia may be safely endoscopically resected with close endoscopic surveillance afterward, a concept that is by no means proven, then one needs to consider the special circumstances of how to safely and comprehensively resect such lesions. The technique for endoscopic resection is the focus of this review.