Low local recurrence after colorectal endoscopic submucosal dissection: a case for better endoscopic mucosal resection, not universal adoption

10.1055/a-2870-7294 10.1055/a-2875-1254 As technical expertise in colorectal endoscopic submucosal dissection (ESD) in Western countries continues to improve, support for its adoption as the mainstay of treatment for large (≥20 mm) nonpedunculated colorectal polyps (LNPCPs) is rapidly increasing. Although dependent on institutional practice, ESD generally demands substantially greater time and resource use compared with endoscopic mucosal resection (EMR), and carries a higher risk for immediate and delayed adverse events [ 1 ] [ 2 ]. In the RESECT-COLON randomized trial, ESD took three times longer than EMR, and in many Western settings, ESD is associated with a greater requirement for deep sedation or general anesthesia and a longer hospital stay [ 3 ]. While the role of ESD in the management of T1 colorectal cancer is established, superiority of ESD in the treatment of benign lesions, which constitute the vast majority of LNPCPs, remains unproven and in increasingly resource constrained health care environments requires clear justification. In this issue of Endoscopy , Saino et al. report the outcomes for 1838 colorectal lesions (median size 25 mm, interquartile range [IQR] 20–35 mm) removed by ESD in 11 centers in Hiroshima, Japan [ 4 ]. A mixture of low- (<100 ESDs/year) and high-volume centers were included. Using Western histopathologic criteria, 91% of lesions were benign, 9% were T1 cancers, and 6% were superficial (<1000 mm), although the proportion of additional high-risk histopathologic features within this subgroup is not clear. En bloc and R0 rates were 97% and 89%, respectively. Adverse events included delayed bleeding (2%), intraprocedural perforation (2.7%), and delayed perforation (0.1%), while a further 0.4% of the original cohort of 2478 lesions required emergency surgery for an ESD-related complication. “…among predominantly benign, mucosal lesions removed by ESD, principally in Asian centers with high R0 rates, recurrence rates remain <2%...” The authors report cumulative recurrence rates of 0.06%. 0.64%, and 0.75% at 1, 3, and 5 years respectively (median follow-up 48 months, IQR 24–62 months). The final cohort excluded 17% of cases owing to a lack of surveillance beyond 3 months; the authors conducted a comparison of included vs. excluded cases, which revealed broadly similar features. Notably, attrition increased over time, with only 34% of the cohort undergoing surveillance at 5 years. Given this attrition, the true long-term recurrence rate may be underestimated and these figures should be interpreted with caution. Multivariable analysis revealed the following independent predictors for post ESD recurrence: rectal location (hazard ratio [HR] 3.93, 95%CI 1.09–14.1; P = 0.04), high-risk pathologic features (HR 14.3, 95%CI 3.60–56.8; P < 0.001), R1 resection (HR 7.59, 95%CI 2.03–28.4; P = 0.003), and delayed bleeding (HR 7.72, 95%CI 1.66–35.8; P = 0.009). There was no recurrence among curative resections. Interestingly, horizontal margin involvement (HM1) did not emerge as a key factor for the development of recurrence; a comparison between technical (HM1 only) vs. oncologic noncurative resection revealed significantly lower rates of recurrence in the former group (1.5% vs. 11.7%; P = 0.003). Concurrently, Ramai et al. present pooled recurrence and metachronous disease rates from 5306 colorectal lesions, across 10 studies, 9 of which originated from high-volume Asian centers [ 5 ]. The pooled R0 resection rate was 82% (95%CI 79%–85%). Histology was only available for 4051/5306 lesions; of these, 73% were benign. Adverse events were not reported. Recurrence rates were 1.2% (95%CI 0.4%–2.6%), 1.4% (95%CI 1.0%–1.9%), and 1.9% (95%CI 0.9%–3.1%) at 1, 3, and 5 years, respectively. R1 resection and advanced histopathologic features were again associated with recurrence, while margin-positive T1 cancers conferred a 25% recurrence rate in one series. Once again, the clinical significance of HM1 among benign polyps is called into question, with recurrence rates of around 2% observed in this group. Significant heterogeneity, retrospective design in the majority of included studies, nonstandardized surveillance protocols, substantial missing long-term follow-up data, and inclusion of data from predominantly Asian high-volume centers all limit interpretation and applicability to a Western health care setting. Both studies examined metachronous disease. Overall, Saino et al. observed metachronous advanced neoplasia and invasive cancer in 12% and 0.92% of cases, respectively [ 4 ]. In the meta-analysis, four studies reported on metachronous neoplasia with a pooled rate of 10% (95%CI 2%–24%). The cumulative incidence was around 1% at 1 year, rising to 8% at 5 years. The common message from these two papers is that among predominantly benign, mucosal lesions removed by ESD, principally in Asian centers with high R0 rates, recurrence rates remain <2% and are stable over time. In the context of low metachronous neoplastic disease rates, surveillance intervals following R0 ESD for low-risk lesions may reasonably be extended compared with current post-EMR protocols, and on this basis, ESD may appear effective in terms of cost and resource utilization over the long term. Beyond the methodologic limitations outlined above, two broader conceptual points caution against concluding that ESD should supplant EMR for benign LNPCPs. First, both papers cite recurrence rates after piecemeal EMR that are no longer representative of contemporary practice. Rates of <2%–3%, regardless of lesion size, have been demonstrated consistently across multiple high-volume, multicenter studies where EMR is combined with margin thermal ablation (MTA) [ 6 ] [ 7 ]. Critically, O’Sullivan et al. demonstrated that among more than 1000 LNPCPs treated with EMR and MTA, the rate of late recurrence (beyond the first surveillance colonoscopy) was just 0.2%, suggesting that even the 18-month surveillance colonoscopy may become unnecessary following high-quality EMR [ 8 ]. Second, the burden of metachronous disease requires careful interpretation. This is hard to quantify accurately as it will be determined by patient and lesion factors (e.g. higher risk with nongranular LNPCP), the quality and completeness of the index colonoscopy, and on whether follow-up examinations represent full colonoscopy or scar-only site checks. Therefore, data reporting metachronous disease yield variable results. In fact, a North American analysis of 728 patients with LNPCPs revealed far higher rates than reported in these two studies, with 40.4% found to have an advanced adenoma within 1 year of follow-up [ 9 ]. These results should prompt high-quality synchronous lesion assessment at the index colonoscopy and surveillance tailored to macroscopic resection completeness, final histology, and overall polyp burden. Notably, recurrence after HM1-only resection was low in both papers, and similar to recurrence rates after high-quality EMR with MTA, suggesting lateral margin status carries limited clinical consequence in benign disease. The correct response to historically high recurrence rates after EMR is therefore not to abandon EMR, with its attendant safety and low sedation/ resource requirements, in favor of universal ESD but to strive for better EMR. In centers where high-quality EMR with MTA is achievable, this remains the appropriate first-line treatment for the majority of LNPCPs, with ESD reserved for lesions with overt or high-risk covert submucosal invasion. Publication History Article published online: 06 August 2026 © 2026. Thieme. All rights reserved. Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany Refers to: Long-term local recurrence and survival after colorectal endoscopic submucosal dissection: a prospective multicenter real-world cohort study Endoscopy eFirst DOI: 10.1055/a-2870-7294 Refers to: Neoplastic recurrence after colorectal endoscopic submucosal dissection at scheduled endoscopic surveillance: a systematic review and meta-analysis Endoscopy eFirst DOI: 10.1055/a-2875-1254