Editorial: Surveillance Colonoscopy for Detection of Neoplasia in Inflammatory Bowel Disease—Is Dye-Based Chromoendoscopy Always the Final Answer?

Abstract

Surveillance colonoscopy has been recommended in patients with inflammatory bowel disease (IBD) for over 20 years to manage their risk of colorectal neoplasia. However, many aspects of a high-quality examination are still debated [1].

Advances in endoscopic technology and evolving clinical paradigms, including updated nomenclature for serrated lesions and more accurate diagnostic criteria for lesion characterisation, have challenged our technical approach to endoscopic surveillance in IBD [2]. Consequently, purely statistical comparisons between different surveillance strategies may be methodologically weak when they involve technologies and protocols that are too heterogeneous due to the numerous variables influencing their application.

Shehab et al. presented a network meta-analysis (NMA) including 22 randomised controlled trials and three prospective cohort studies covering nine different technologies used for neoplasia detection in IBD [3]. They concluded that dye-based chromoendoscopy (DCE), especially with high-definition (HD), HD-white light endoscopy with segmental reinspection (HD-WLE-SR) and chromoendoscopy-guided endomicroscopy (CEM) outperformed other methods, including virtual chromoendoscopy (VCE), which includes Narrow Band Imaging (NBI), Fuji Intelligent Colour Enhancement (FICE) and i-SCAN, with similar results between HD-WLE-SR and DCE.

The finding in favour of HD-WLE with SR is interesting because it would provide an alternative to DCE that would simplify surveillance. However, this result stems from the inclusion of just one study [4]. Equally interesting is the good performance of CEM, although supported by older proof-of-concept studies that used endoscope-based confocal laser endomicroscopy (eCLE) currently no longer commercially available [2].

What is the clinical message? Does DCE remain the only appropriate, yet unloved, surveillance tool? Should we ultimately abandon the hope of using VCE, a tool both easy to activate and widely available in many endoscopy units, despite evidence from some head-to-head studies demonstrating its non-inferiority to DCE? Notably, NBI appeared inferior to DCE based on studies limited by the use of non-magnified instruments and by diagnostic criteria that later proved to be inaccurate (i.e. Kudo's) in IBD. Conversely, FICE was not inferior to DCE, noting that the only included study stressed the need to use specific diagnostic criteria to better enhance the accuracy of this technology [5].

Could the recurrent supremacy of DCE in many NMAs be attributed to operational factors that a purely statistical approach fails to capture? In this regard, Toruner et al. analysed technical and clinical aspects influencing the quality of surveillance colonoscopy: among all, the “time factor” dominated (i.e. the attention devoted), not only in chronometric terms [6]. That same “time factor” is implicit in DCE protocols and might be what allows the endoscopist to analyse the details of the mucosal and vascular patterns with VCE. Even the potential success of HD-WLE with SR or CEM may partially derive from the increased attention stimulated by segmental re-inspection and endomicroscopic analysis. Are we then to suspect that the lower accuracy sometimes described with recent technologies is linked to an unconscious need to delegate to technology itself the awareness and dedication that should characterise any surveillance endoscopy? We shall see what artificial intelligence comes up with.

Andrea Cassinotti: conceptualization, writing – original draft, supervision. Lorenzo Canova: writing – review and editing.

The authors declare no conflicts of interest.

This article is linked to Shehab et al papers. To view these articles, visit https://doi.org/10.1111/apt.18500 and https://doi.org/10.1111/apt.70057.