Letter to the Editor: Management of incomplete resection of small rectal neuroendocrine tumours: An open question

Abstract

Management of incomplete resection of rectal neuroendocrine tumours remains an open clinical question. In the issue of the World Journal of Gastroenterology, Kim et al highlighted the heterogeneous incomplete resection rates across endoscopic techniques, while analysing impact of salvage treatment in this subset of patients. We further searched for evidence favouring a specific endoscopic technique for resection and the diagnostic and therapeutic role of salvage therapy in the long term. We performed a sensitivity meta-analysis restricted to studies with ≥ 36 months follow-up (5 studies; 413 patients) to compare recurrence rates between salvage therapy and observation. Our analysis showed no significant difference in recurrence (odds ratio = 0.89; 95% confidence interval: 0.37-2.18; P = 0.80), confirming the original findings even with extended follow-up. Subgroup analysis for local and distant recurrence did not show any significant difference.

Key Words: Rectum; Neuroendocrine tumours; Salvage therapy; Observation; Residual neoplasms

Core Tip: This sensitivity meta-analysis of studies with ≥ 36 months follow-up confirms that salvage therapy does not significantly reduce recurrence rates compared with observation after incomplete resection of small rectal neuroendocrine tumours (odds ratio = 0.89, 95% confidence interval: 0.37-2.18). Despite a 25% residual tumour rate at salvage, long-term outcomes do not favour routine re-intervention. These findings challenge current clinical practice and support individualised management based on risk stratification rather than automatic salvage. However, given the low event rates and limited follow-up, current evidence supports shared decision-making with patients.

TO THE EDITOR

We read with great interest the meta-analysis published in the issue of the World Journal of Gastroenterology by Kim et al[1] on salvage therapy after incomplete endoscopic resection of rectal neuroendocrine tumours (NETs). The authors address two clinically relevant questions, focusing on minimisation of incomplete resection and the potential benefit of salvage therapy for incomplete resections.

Incomplete resection of rectal NETs, in terms of both positive or indeterminate margins and lymphovascular invasion (LVI), remains a significant clinical challenge. In this context, Kim et al[1] report heterogeneous incomplete resection rates across endoscopic techniques: 73.1% for cold snare polypectomy, 29.8% for conventional endoscopic mucosal resection (EMR), 28.4% for modified EMR (mEMR), and 14.7% for endoscopic submucosal dissection (ESD).

In a published meta-analysis on 1168 patients with rectal NETs ≤ 10 mm in size[2], no differences were found in terms of incomplete resection rates between EMR and ESD. However, the authors concluded that, owing to the shorter operative time and lower costs, EMR outperformed ESD. In the 2016 meta-analysis published by Zhang et al[3] on patients with rectal NETs < 16 mm in size, EMR was further subdivided into mEMR and conventional EMR (cEMR). The authors found that ESD and mEMR were associated with higher rates of complete resection when compared to cEMR, with no differences between ESD and mEMR.

Thus, we agree with the results reported by Kim et al[1] that ESD or mEMR should be preferred for upfront resection of small rectal NETs (< 2 cm), with technique selection tailored to lesion characteristics, endoscopist expertise, and resource availability.

The second question - whether salvage therapy is necessary after incomplete resection - is more contentious. Kim et al[1] report a residual tumor rate of 25% in salvage procedures. While crude recurrence rates favoured salvage therapy (0.96% vs 2.96%, P = 0.003), adjusted comparative analysis showed no significant benefit [odds ratio (OR) = 0.89, 95% confidence interval (CI): 0.40-2.02]. The authors appropriately acknowledge limitations including the observational study designs and limited follow-up duration.

In light of the 25% residual tumour rate, the lack of benefit from salvage therapy may be explained by several factors. First, not all residual tumours progress to clinically significant recurrence during the available follow-up, especially for small, low-grade lesions. Second, the morbidity of salvage procedures may offset the benefit of removing residual disease in some patients. Third, the 25% residual rate may include lesions that would have been adequately managed by surveillance alone, underscoring the need for better selection criteria to identify which patients would benefit from re-intervention.

Among the included studies, the multicentric retrospective study by Duan et al[4] reported worse disease free survival in patients who underwent observation, with no impact on overall survival, with a median follow-up of 72 months. Small rectal NETs (< 2 cm in diameter) are often characterised by a more indolent course due to low early metastatic risk, with recurrences presenting beyond 3-5 years[5]. Thus, we decided to investigate whether including only studies with a median follow-up period of ≥ 36 months would lead to different conclusions regarding the incidence of recurrence.

Of the selected studies, only 5 had a median follow-up of ≥ 36 months and reported outcomes separately for salvage therapy vs observation[4,6-9]. “Review Manager Version 5.4. The Cochrane Collaboration, 2020” was used to perform our meta-analysis. As our variables were presented as dichotomous outcomes, odds ratios and corresponding 95%CIs were computed according to the Mantel-Haenszel method. Statistical significance of the overall effect was set at P < 0.05. I² statistics were used to assess statistical heterogeneity. Less than 25%, 25%-50%, and > 50% I² values were classified as follows: Low, moderate, and high. The choice between a random-effects model and a fixed effects model was dictated by statistical heterogeneity, favouring the random-effects model in case of high heterogeneity (I² > 50%)[10].

The five included studies comprised 413 patients, with median follow-up periods ranging from 36 months to 72 months and sample sizes ranging from 19 patients to 181 patients. Recurrence patterns were reported separately in the studies by Duan et al[4], Lee et al[8] and Sung et al[9], while the remaining studies did not specify the type of recurrence.

The pooled analysis demonstrated no statistically significant difference in recurrence rates between salvage therapy and observation (OR = 0.89, 95%CI: 0.37-2.18, P = 0.80; I² = 0%), reaffirming the findings of Kim et al[1] (Figure 1). Even with extended follow-up, salvage therapy did not confer a measurable benefit.

Figure 1 Forest plot of recurrence rates: Salvage therapy vs observation. Meta-analysis of studies with median follow-up of ≥ 36 months comparing recurrence rates after salvage therapy vs observation for incomplete resection of small rectal neuroendocrine tumours. Pooled analysis demonstrates no significant difference between strategies (odds ratio = 0.89, 95% confidence interval: 0.37-2.18, P = 0.80; I² = 0%). M-H: Mantel-Haenszel; CI: Confidence interval.

We then performed subgroup analyses for local and distant recurrence separately in long-term follow-up studies (Figure 2). For local recurrence, the pooled analysis showed no significant difference between salvage therapy and observation (OR = 1.28, 95%CI: 0.06-28.09, P = 0.88). For distant recurrence, no significant difference was observed (OR = 0.86, 95%CI: 0.23-3.22, P = 0.82). These findings must be interpreted cautiously given the small number of events (1 local and 3 distant in the salvage group; 5 local and 8 distant in the observation group) and the limited number of studies reporting recurrence patterns.

Figure 2 Forest plot of local and distant recurrence rates: Salvage therapy vs observation. A: Subgroup analysis for local recurrence. Pooled analysis demonstrates no significant difference between strategies [odds ratio (OR) = 1.28, 95% confidence intervals (CI): 0.06-28.09, P = 0.88]; B: Subgroup analysis for distant recurrence. Pooled analysis demonstrates no significant difference between strategies (OR = 0.86, 95%CI: 0.23-3.22, P = 0.82). M-H: Mantel-Haenszel; CI: Confidence interval.

Regarding the type of salvage therapy, only Duan et al[4] reported recurrences stratified by salvage modality; the remaining studies either did not specify the salvage technique used or pooled different modalities in their analyses. Consequently, the available data did not permit a pooled subgroup analysis by salvage treatment type.

Several methodological considerations warrant attention. First, pooling positive and indeterminate margins may not be appropriate, as cautery artefacts may limit the histopathological interpretation of the completeness of the initial resection. Second, combining margin positivity with LVI may impact long-term outcomes, as the two features may be more accurate in predicting local and nodal risk of recurrence, respectively. Finally, the inclusion of different salvage treatments (repeat ESD, transanal surgery, radical surgery) may represent a potential confounder in the interpretation of results.

With expected event rates of 1%-3%, these comparative analyses remain underpowered to detect clinically meaningful differences, leaving some unresolved questions about the best management option for these patients. Moreover, long-term oncological outcomes were separately reported for both groups only in the study by Duan et al[4], which does not adequately address the question regarding the potential long-term treatment benefit in these patients.

In our opinion, while Kim et al[1] provided interesting results, future research should explore the impact of multiple disease features on optimal treatment, including positive vs indeterminate margins, positive margins vs LVI, Ki-67 grade, and tumour size. Further stratification by treatment strategy (i.e., repeat ESD, transanal excision, radical surgery) may be appropriate in the context of long-term follow-ups.

In clinical practice, for patients with indeterminate margins (e.g., due to cautery artefact) the European Neuroendocrine Tumor Society 2023 guidance for colorectal NETs[11] recommends further resection in order to achieve R0 status. It is also recommended that, after R1 resection without second endoscopic resection, patients should undergo endoscopy and EUS (or magnetic resonance imaging) at 12 months for at least 5 years. However, considering the reported findings, a structured surveillance programme could also represent a viable alternative to immediate salvage therapy. EUS could become a key component in this context, with recent evidence demonstrating a sensitivity of 94% and an accuracy of 87% for detecting residual disease, allowing a more precise assessment compared with visual evaluation[7].

Patients with Ki-67 indices approaching the upper limit of G2 (≥ 5%) may represent a distinct subgroup that could reflect a higher-risk condition and may derive greater benefit from salvage therapy. A 2025 study on 121 patients with rectal NETs found that, among stage I tumours, patients who developed disease recurrence showed higher Ki-67 values compared with those without recurrence, although this difference did not reach statistical significance (P = 0.054), likely due to the small number of events (only 10 recurrences)[12]. Future studies should stratify outcomes by tumour grade to test this hypothesis.

Thus, in order to mitigate these confounding factors, we recommend a structured approach: (1) Indeterminate margins should not be automatically equated with positive margins; these patients may be offered structured surveillance rather than immediate salvage therapy; (2) When salvage is considered, the choice of modality should be guided by tumour characteristics, including size, depth of invasion and nodal status. Repeat ESD may suffice for small, non-invasive residuals, while transanal excision or radical surgery may be reserved for lesions with suspected deep invasion, higher Ki-67 values or nodal involvement; and (3) Multidisciplinary discussion and, where feasible, central pathology review can help standardise the interpretation of resection margins and LVI, reducing variability in clinical decision-making.

In conclusion, the meta-analysis by Kim et al[1] highlights that there is no significant support for salvage therapy over careful observation after incomplete endoscopic resection of small rectal NETs. We believe that future decisions should be individualised based on patient characteristics, histopathological findings and endoscopist expertise, while awaiting data from larger, more robust studies.

ACKNOWLEDGEMENTS

This study was partially supported by Italian Ministry of Health - Ricerca Corrente Annual Program.