Re: ‘Dedicated service for Barrett’s oesophagus surveillance endoscopy yields higher dysplasia detection and guideline adherence in a non-tertiary setting in the UK: a 5-year comparative cohort study’ by Ratcliffeet al

I read this paper with great interest and would like to congratulate Ratcliffe and coauthors for their welldesigned comparative cohort study carried out in a nontertiary centre. This paper emphasises the need for adequate time for assessment of Barrett’s oesophagus during surveillance, and confirms already published evidence recommending a Barrett’s inspection time of 1 min/cm length of the Barrett’s segment using highdefinition white light endoscopy. 3 There are a few important issues that need addressing when planning services for Barrett’s oesophagus surveillance in the UK. First, while it is ideal to have a full dedicated Barrett’s list, with adequate time and a trained Barrett’s endoscopist, to achieve the best outcomes for dysplasia detection, most secondary centres in the NHS will struggle to implement this practice in the present environment of endoscopy backlog and increased clinical pressures. Second, it is still unclear from the literature as to whether having a full list of Barrett’s surveillance cases or simply adequate time (30–45 min per case) in an upper GI endoscopy list carried out by a trained endoscopist will result in a statistically significant difference in dysplasia detection rate (DDR). In this paper, the authors have not clarified whether all lesions detected in the dedicated endoscopy group were using highdefinition white light endoscopy only and/or with the help of image enhanced endoscopy (such as narrow band imaging, near focus magnification or acetic acid chromoendoscopy) as compared with the surveillance carried out in the nondedicated list, which was predominantly carried out using white light endoscopy. This is an important aspect in the detection of early Barrett’s neoplasia, since the authors mention that narrow band imaging and acetic acid chromoendoscopy was used much more frequently in the dedicated surveillance group compared with the nondedicated group. Finally, the advent of new technologies such as sponge based cell collection techniques (Cytosponge, Medtronic Inc) and artificial intelligence (AI) and computer aided detection (CADe) and characterisation of Barrett’s related neoplasia may make DDRs in nontertiary centres as good as in expert tertiary centres as long as the Barrett’s mucosa is inspected with adequate time to allow the computer to analyse the mucosal surface and vascular characteristics to make a diagnosis. The use of sponge based cell analysis has been shown to improve the selection of patients for Barrett’s surveillance. Although adequate inspection time will still be needed for an accurate diagnosis, confirming the need for 30–40 min time for each patient on a Barrett’s surveillance programme, the use of AI tools in a Cytosponge positive patient may remove the expertise needed of an endoscopist to make an optical diagnosis. This is yet to be formally validated in a comparative study of the use of AIassisted CADe between tertiary and nontertiary centres. Overall, the authors of this study deserve to be complimented for raising the awareness of the need to change our current practice of Barrett’s surveillance for better dysplasia/ neoplasia detection.