The Barrett-esophagus (BE) guideline of 2018 gives recommendations regarding the endoscopic evaluation, description and biopsy sampling of BE. This chapter reviews whether recent evidence allows adapting these recommendations.

High-definition endoscopy is widely available in Netherlands and is current standard of care in endoscopy.

The evidence for supporting (virtual) chromoendoscopy is weak, however using virtual chromoendoscopy introduce next to its intrinsic properties of imporved vizualisation of the mucosa additional factors (e.g. extra cleaning, additional inspection and increased inspection)  that might improve quality in upper endoscopy in BE.

The evidence for adopting AI in Barrett’s surveillance is at current moment insufficient.

Several studies show an increase in dysplasia detection with longer BE inspection times while low adherence to the Seattle Biopsy protocol was associated with decreased of dysplasia detection.

The text below is adapted from the international guideline (Weusten, 2023). Details are descripted in Table 6, 7, 8 and 13 from the evidence tables.

High definition endoscopy systems (endoscope, processor, and screen) provide superior image resolution and are widely available; however, the role of high definition endoscopy systems in BE surveillance is based on limited low quality studies. In a retrospective study, a high definition endoscopy system was superior to a standard definition system in detecting dysplastic lesions, and in detecting high grade dysplasia (HGD) or cancer on random and targeted biopsies (Sami, 2024).

Given its inferior imaging quality and limited biopsy sampling capability, transnasal endoscopy should not be used for BE surveillance.

With regard to narrow-band imaging (NBI), there are studies demonstrating a significantly higher rate of dysplasia detection with fewer biopsies when the use of NBI is compared with standard resolution white-light endoscopy (WLE) (Wolfsen, 2008), but no differences in detection rate when NBI is compared with high definition WLE with four-quadrant biopsies (Sharma, 2013).

Acetic acid chromoendoscopy (AAC) has been studied for its usefulness to detect Barrett neoplasia in several studies (Coletta, 2016). Longcroft-Wheaton et al. reported on a feasibility study in which patients scheduled for BE surveillance underwent two gastroscopies 6–8 weeks apart: one regular endoscopy with random biopsies according to the Seattle protocol, and one endoscopy using AAC with AAC-targeted biopsies only. The authors found a similar dysplasia/EAC detection rate between the two protocols, with a significant reduction in the number of biopsies in the AAC arm (Longcroft-Wheaton, 2019); however, larger studies are needed.

Although the evidence to support the use of (virtual) chromoendoscopy is weak, the ESGE working group favors its use. Currently, all of the high definition endoscopy systems are equipped with virtual chromoendoscopy, and the use of AAC is associated with very limited additional costs. An important additional advantage is that the use of virtual chromoendoscopy and AAC requires decent cleaning of the esophagus. In addition, the use of (virtual) chromoendoscopy is generally associated with an extra pull through, which translates to increased inspection time. These factors, apart from the possible intrinsic properties of (virtual) chromoendoscopy, might improve quality in BE surveillance.

Evidence that these techniques can replace the Seattle protocol in a standard surveillance setting is still lacking. Therefore, it is recommended that these techniques be used prior to and in addition to Seattle protocol biopsy sampling.

Additional considerations on the role of virtual chromoendoscopy with NBI, blue-light imaging, and i-SCAN in surveillance of patients with BE are provided in Appendix 4s of the international guideline.

Role of artificial intelligence (AI)

ESGE has recently published a Position Statement on the expected value of AI in gastrointestinal (GI) endoscopy (Messmann, 2022). It is anticipated that AI will improve the quality of routine endoscopy. In view of the fact that many lesions are missed on referral or in daily practice (Nieuwenhuis, 2022; Visrodia, 2016), expectations after the first pilot study to improve this outcome are high. The value of AI in BE surveillance lies not in exceeding expert performance but in raising routine practice to expert level performance. Different research groups have demonstrated high sensitivities of AI systems for detecting dysplasia/EAC during real-time endoscopy, ranging from 83.7 % to 95.4 % (Ebigbo, 2020; Hashimoto, 2020; de Groof, 2020). Two systematic reviews and meta-analyses have indicated high detection performances, ranging between 88 % and 96 % (Arribas, 2020; Lui, 2020). Nonetheless, most studies are pilot feasibility studies in enriched populations, and therefore more evidence is needed before AI can be generally accepted as an adjunct to – or a replacement of – the Seattle biopsy protocol during surveillance. At the advent of the launch of commercial devices, emphasis must remain on basic endoscopy quality standards with regard to technical performance and cleaning of the esophagus: if the lesion is not adequately shown to the system because of hurried pull through, insufficient insufflation, or insufficient cleaning, AI will not be of any help.

Quality standards of gastroscopy

There are no randomized controlled data supporting a minimum inspection time for BE surveillance; however, a few retrospective studies suggest an increase in dysplasia detection with longer BE inspection times (Gupta, 2012; Vithayathil 2023). Even more important than inspection time is what endoscopists do during the time they are inspecting. Proper inspection includes cleaning of the esophagus, multiple pull throughs with both high definition WLE and (virtual) chromoendoscopy, as well as accurate photo documentation of landmarks and the BE segment. If this is applied consistently, BE inspection time easily exceeds 1 minute per cm of BE length prior to biopsy taking.

Seattle biopsy

Biopsy samples should be taken of all visible mucosal abnormalities. One to two biopsies, targeted on the most suspicious part of the lesion, are considered enough for lesions (Paris type 0-I, 0-II) that are potentially amenable to endoscopic resection (ER) in order to confirm the diagnosis and not compromise subsequent ER (Pouw, 2021). In addition, random four-quadrant biopsies should be collected every 2 cm within the Barrett segment, starting from the upper end of the gastric folds. Biopsies from each level are preferably collected in separate, marked containers. Several studies have indicated low compliance with guidelines with regard to obtaining a sufficient number of random biopsies (Wani, 2019; Westerveld, 2018; Antony, 2018), with lower dysplasia detection rates if the Seattle biopsy protocol was not adhered to (Abela, 2008). These studies highlight points of improvement in current practice and the importance of high quality endoscopy. These standards are key to minimizing the risk of undetected lesions and avoiding redundant repeat endoscopies that are scheduled because of nonadherence to quality guidelines. This will help to reduce the carbon footprint and environmental burden of elective endoscopy (Rodriquez de Santiago, 2022).

In patients with reflux esophagitis grade C or D, no random biopsies should be taken, and BE surveillance endoscopy should be repeated at least 6 weeks after optimization of antireflux therapy. However, even in the presence of severe reflux esophagitis, a careful inspection of the BE segment is warranted and targeted biopsies of suspected lesions are still recommended (Figuur 1 Stroomschema Diagnostische gastroscopie).

The ESGE working group feels that it is essential to allocate at least 30 minutes to surveillance procedures, preferably in well-sedated patients, in order to allow for adequate inspection time, photo documentation, and biopsy sampling, increasing to 40 minutes for ultralong BE segments (Vithayathil, 2023).

To answer the clinical question the guideline ‘Diagnosis and management of Barrett esophagus: European Society of Gastrointestinal Endoscopy (ESGE) Guideline’ was adapted (Weusten, 2023). In this module we refer to this guideline.

Research questions:

• What is the role of chromoendoscopy and virtual chromoendoscopy in the detection and characterization of BE lesions?
• What is the role of artificial intelligence in detection and characterization of neoplastic lesions in BE?
• What is the impact of inspection time on lesion detection, and what is the role of dedicated endoscopy lists in Barrett's surveillance?
• What is the validity of the Seattle biopsy protocol in BE surveillance?

Relevant outcome measures

The task force did not consider critical nor important outcome measures.

Search and select (Methods)

The results of data extraction are available in the evidence tables viewable at the ESGE website: https://www.esge.com/assets/downloads/pdfs/guidelines/a_2176_2440_Evidence_tables.pdf. See Table 6, 7, 8 and 13 for the current question.