Imaging in IBD: Capsule, NBI, and chromoendoscopy.

Abstract

Crohn's disease (CD) affects the small bowel in most patients and between 20% and 40% of CD patients have isolated small bowel CD. Given that mucosal healing is widely accepted as a target of therapy in CD, assessing the mucosa of the small bowel is important. In clinical practice, however, assessment of small bowel mucosa is frequently sub-optimal. This may be because it is difficult to inspect endoscopically; cross–sectional imaging does not give sufficient detail of mucosal disease or because video capsule endoscopy (VCE) is not available or is felt to be associated with unacceptable risks. Several options are available for small bowel mucosal inspection; VCE, single or double balloon enteroscopy via anal or oral approaches, or push enteroscopy are widely available in the Asia-Pacific region. Multiple studies have demonstrated efficacy of VCE in identifying CD-associated changes including ulcers, erosions, erythema, aphthae, and strictures and in evaluating response to therapy.1, 2 There are however, some impediments to the universal use of VCE in evaluating small bowel mucosa in CD. First amongst these is the cost of the procedure, which needs to be considered in the specific health economic setting of each country. The caecum is reached less often than with VCE for other indications.3 Small bowel preparation and therefore image quality are slightly poorer than seen for other indications.4 An additional procedure to obtain biopsies for histology (typically a device-assisted enteroscopy) is often required, which adds to cost and invasiveness for health systems and patients, respectively. Capsule retention is a much feared complication of VCE. For VCE used to investigate suspected CD, rates of capsule retention are low and comparable with VCE for other indications. For VCE in the context of established CD, the risk of capsule retention is higher and testing with a patency capsule prior to VCE is recommended.3, 4 The Lewis score (LS) is a composite score reflecting CD activity identified at VCE5 and is built into some VCE reporting software. Patients with a LS confirming moderate–severe activity have been shown to be more likely to have features associated with poor prognosis in CD (smoking and immunomodulator use) and have greater risk for corticosteroid use and admission in follow-up (relative risk 5.0; P = 0.011, relative risk 13.7; P = 0.028 respectively).6 VCE also has demonstrated efficacy in assessing disease progression or response and can demonstrate substantial response in LS, where less objective markers (Crohn's Disease Activity Index and Inflammatory Bowel Disease Questionnaire) are unpredictable and correlate poorly with endoscopic disease activity. Video capsule endoscopy is effective in diagnosing CD and is safe for diagnosing CD. VCE performs at least as well as magnetic resonance enterography in identifying small bowel CD and colonoscopy for evaluating terminal ileal CD. Scoring systems of disease activity exist that correlate with short-term and long-term outcomes and can demonstrate response. Chromoendoscopy in IBD colonoscopy has found an important role in recent years. It is important to remember that chromoendoscopy can only be of optimal use if other factors, patient attendance, bowel preparation, performing surveillance in remission, endoscopist training and technique, and high-definition endoscopes, are attended to and optimized. The recent SCENIC guidelines have summarized and systematically reviewed the literature on chromoendoscopy in IBD dysplasia surveillance. There is a demonstrated incremental yield of chromoendoscopy over white-light endoscopy (6%, 95% confidence interval 3–9%). Narrow band imaging (NBI) has not demonstrated the same benefit over white-light endoscopy. Curiously, comparative studies of NBI and chromoendoscopy have not shown a significant benefit of chromoendoscopy over NBI (6% incremental yield, 95% confidence interval 1–14%). More data is needed comparing these two modalities. There is good evidence to suggest that targeted biopsies of abnormal appearing areas have a far greater diagnostic yield than random biopsies and that not performing random biopsies does not carry a high risk of missing dysplasia. The uptake of NBI and chromoendoscopy is increasing in Australia although only 3% of gastroenterologists in Australia performing surveillance colonoscopy in IBD do not take random biopsies.7