Endoscopic eradication therapy (EET) is the standard of care for treatment of Barrett’s esophagus (BE)-related neoplasia. It consists of a combination of endoscopic resection with endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) followed by ablation [1]. Resection enables removal of all visible lesions that may reflect advanced neoplastic lesions. Ablation enables eradication of any remaining flat BE mucosa to achieve complete remission of intestinal metaplasia (CRIM). EMR was initially preferred given its efficiency and simplicity. ESD was later developed to enable en bloc resection regardless of lesion size, but it is technically more challenging and time consuming.

In esophageal (squamous), gastric, and colonic neoplasia, ESD has been shown not only to achieve higher en bloc and R0 resection rates, but also to lower recurrence compared with EMR. However, in Barrett’s neoplasia, despite en bloc resection of visible neoplastic BE lesions, the remaining flat nondysplastic BE mucosa has been shown to harbor genomic and epigenetic alterations similar to those found in resected dysplastic lesions, which could lead to progressive neoplastic transformation into high grade dysplasia (HGD) or esophageal adenocarcinoma (EAC) if left untreated [2]. Recurrence after successful EET is well described and rates are higher (compared with those after CRIM) if only complete remission of dysplasia is achieved [3]. Hence, eradication of all intestinal metaplasia is recommended as the optimal endpoint for EET [1]. Predictors of recurrence after CRIM include older age, male sex, longer BE segments, and baseline HGD/EAC.

In this issue of Endoscopy, Fujiyoshi et al. [4] report recurrence data for patients with Barrett’s neoplasia (HGD or EAC) who underwent either EMR (87 patients) or ESD (70 patients) at a single tertiary referral center in Canada. Recurrence was defined as detection of HGD/EAC on follow-up after achieving complete remission of neoplasia (CRN), defined as the absence of HGD or EAC on preceding surveillance histology. In patients who achieved CRN, the rate of neoplastic recurrence (defined as HGD/EAC on surveillance biopsies) was significantly higher in the EMR group compared with the ESD group (13.0% vs. 1.9%). The mean follow-up period was less than 1 year (362 [SD 304] days) in the ESD group and only slightly longer (437 [SD 301] days) in the EMR group. Rates of “neoplastic recurrence” were higher at 1 and 2 years in the EMR group. The authors concluded that ESD reduces “recurrence” rates compared with piecemeal EMR.

This study raises the important question on how recurrence should be defined in order to be clinically relevant in real-life clinical practice. It is critical to note that the primary endpoint of EET is achieving CRIM, which is defined as the absence of intestinal metaplasia on biopsies from the tubular esophagus and gastroesophageal junction in at least one surveillance endoscopy. The main rationale for achieving CRIM is to reduce the rate of (dysplastic) recurrence after successful EET, which has been shown to be substantially higher in patients who only achieve complete remission of dysplasia [3]. In almost all currently available literature comparing the outcomes of EMR and ESD for Barrett’s neoplasia, including a recently published systematic review and meta-analysis [5], the standard definition of recurrence (after CRIM) has not been used. Similarly, the definition of recurrence used in the current study was the presence of HGD/EAC on surveillance biopsies after CRN (not after CRIM). As per current guidelines, true recurrence should only be determined once CRIM, the primary endpoint of EET, has been achieved, which typically requires both endoscopic resection and ablation. This standard definition of recurrence is likely more clinically relevant, allowing us to accurately distinguish true recurrence from residual disease after incomplete endoscopic treatment. Additionally, in the study by Fujiyoshi et al., CRIM was achieved in only 33.8% of the EMR group and 39.6% of the ESD group, which is much lower than the rates reported in other studies. Additionally, multivariable analysis adjusting for known confounders of recurrence, such as age, sex, and BE segment length, was not conducted.

A large multicenter retrospective cohort study comparing rates of recurrence following EMR vs. ESD and ablation, after CRIM was achieved, was recently published [6]. In this study, rates of any or dysplastic BE recurrence were comparable between the two groups on both multivariable regression and inverse proportional treatment weighted analyses (enabling adjustment for known confounders/predictors of recurrence to mimic a randomized trial). Subgroup analysis to assess the rate of dysplastic recurrence in only patients with baseline HGD and EAC (as in the study by Fujiyoshi et al.), also did not reveal any difference in dysplastic recurrence between the groups. Importantly, this study used the guideline definition of recurrence (BE detected after CRIM).

The optimal endoscopic resection technique for Barrett’s neoplasia needs further study. While ESD unequivocally allows a higher rate of en bloc R0 resection, the clinical relevance of achieving en bloc resection with ESD in BE neoplasia is likely somewhat mitigated by subsequent endoscopic ablative therapy. In high-volume BE centers with ESD experts, both endoscopic resection techniques appear to be a viable option. At this time, ESD may be preferred in some situations where the lesion is difficult to resect with EMR, such as those with submucosal fibrosis/scarring from prior treatment, higher likelihood of submucosal invasion based on macroscopic appearance, and size >15–20 mm (to allow lateral margin assessment and confirmation of margin-free neoplasia) [7] [8]. To further define the landscape of endoscopic resection in Barrett’s neoplasia, specifically the criteria for selecting between EMR and ESD followed by ablation in clinic practice, adequately powered randomized trials incorporating ablation to achieve CRIM, and with long-term follow-up, are needed.

Article published online:
16 October 2024

© 2024. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany