Comparison between endoscopic resection and transanal surgery for treatment of rectal tumors: a systematic review and meta‑analysis
Article information
Abstract
Background/Aims
Both endoscopic resection (ER) and transanal surgery (TAS) are minimally invasive treatment options that allow organ preservation in early rectal tumors. We conducted a meta-analysis to compare treatment outcomes between the 2 treatments.
Methods
We searched all relevant studies published until January 2024 that examined the comparative outcomes between ER and TAS for rectal tumors, including adenoma, adenocarcinoma, and neuroendocrine tumor (NET). TAS included transanal excision, transanal endoscopic microsurgery, and transanal minimally invasive surgery.
Results
Seventeen studies with a total of 1,569 patients were included in this meta-analysis. For adenoma/adenocarcinoma, the R0 resection rate did not differ between ER and TAS (risk ratio [RR], 0.99; 95% confidence interval [CI], 0.94–1.03). For NET, the R0 resection rate was lower in the ER group than in the TAS group (RR, 0.76; 95% CI, 0.68–0.84) and the procedure time for ER was shorter than that for TAS. For both adenoma/adenocarcinoma and NET, ER and TAS did not differ in terms of complication rates, additional surgery, and recurrence.
Conclusions
ER and TAS showed similar treatment outcomes for adenoma/adenocarcinoma. Considering that TAS typically incurs higher costs than ER, ER may be favored in the treatment of rectal adenoma/adenocarcinoma. For rectal NET, TAS showed a superior R0 resection rate than ER. However, given that TAS requires a long procedure time, expensive equipment, and complex manipulations, TAS may be considered selectively for large NETs with suspected deep tumorous infiltration.
INTRODUCTION
Despite detailed guidelines for screening and surveillance, colorectal cancer (CRC) still has a high incidence and remains the leading cause of cancer-related death [1-5]. According to 2020 global cancer statistics, CRC is the third most frequently diagnosed cancer and the second most common cause of cancer death [1,6]. In particular, the proportion of rectal cancer in the incidence and mortality rates of CRC is high at 38.9% and 37.0%, respectively [1]. Early rectal cancer or rectal premalignant lesions can be removed using minimally invasive techniques, such as transanal surgery (TAS), instead of performing major surgeries, such as low anterior resection or abdominoperineal resection, which can cause significant morbidity and greatly impact the quality of life.
TAS includes transanal excision (TAE), transanal endoscopic microsurgery (TEM), and transanal minimally invasive surgery (TAMIS). TAE is a method initially developed to remove lesions by direct visualization using standard retractors [7]. Therefore, this method does not allow sufficient visualization of the mid and upper rectum, allowing resection of only distal rectal lesions. TEM, introduced in the 1980s, uses special instruments, including proctoscopes, laparoscopic camera, and laparoscopic instruments, to allow the operator to reach lesions in the mid and upper rectum [].7 However, this method presents challenges when the lesion is located where it cannot be properly removed in standard patient positioning. TAMIS, developed in 2009 as an alternative technique, uses a single multichannel port inserted into the anus, enabling more flexibility in patient positioning [8].
Another minimally invasive technique to remove early rectal cancer or rectal premalignant lesions is endoscopic resection (ER), such as endoscopic submucosal dissection (ESD) and endoscopic mucosal resection (EMR) [9]. ER is less invasive and has lower morbidity; however, sometimes resection margins, especially deep margins, cannot be accurately determined, requiring additional surgical treatment. In these cases, TAS, which allows for full-thickness resection of the rectal wall can be a more curative method. However, TEM and TAMIS require special equipment and general anesthesia or epidural anesthesia, and can place economic and psychological burdens on patients.
Clinicians often encounter lesions that are difficult to determine which method is more appropriate, ER or TAS. To address this concern, several studies have compared ER and TAS, and recently, a meta-analysis was performed on this topic [10]. A meta-analysis of 11 studies searched until January 2020 demonstrated that TAS has a higher R0 resection rate for early rectal neoplasms than ER [10]. However, this meta-analysis presented the results of combining these lesions without distinguishing between adenoma/adenocarcinoma and neuroendocrine tumor (NET). Most rectal NETs are <1 cm in size and appear as small yellowish submucosal lesions with intact overlying mucosa [11]. As with rectal adenoma/adenocarcinoma, treatment options for rectal NET include ER and TAS; however, since the 2 tumors have different characteristics, analyzing them separately is reasonable. Moreover, several relevant studies have been published and accumulated since then. Therefore, complementing and updating the information on this topic is necessary.
In this study, we conducted a systematic review and meta-analysis to analyze the comparative outcomes between ER and TAS for rectal tumors, including adenoma/adenocarcinoma and NET. In contrast to previous meta-analysis, our study compared adenoma/adenocarcinoma and NET separately.
METHODS
1. Search Strategy
We searched for all relevant studies published between January 1990 and January 2024 that compared clinical outcomes between ER and TAS in patients with rectal adenoma, adenocarcinoma, or NET through MEDLINE, EMBASE, and Cochrane Library databases. The following search string was used: ([endoscopic submucosal dissection] OR [endoscopic submucosal dissections] OR [ESD] OR [endoscopic mucosal resection] OR [endoscopic mucosal resections] OR [EMR] OR [endoscopic resection] OR [endoscopic resections]) AND ([transanal surgery] OR [transanal surgeries] OR [TAS] OR [transanal excision] OR [transanal excisions] OR [TAE] OR [local excision] OR [local excisions] OR [transanal resection] OR [transanal resections] OR [transanal endoscopic microsurgery] OR [transanal endoscopic microsurgeries] OR [TEM] OR [transanal endoscopic surgery] OR [transanal endoscopic surgeries] OR [transanal minimally invasive surgery] OR [transanal minimally invasive surgeries] OR [TAMIS]) AND ([rectal] OR [rectum] OR [colorectal] OR [colorectum]). The detailed search strategies used for each database are shown in Appendix 1.
2. Inclusion/Exclusion Criteria
The inclusion criteria were as follows: (1) patients—patients who underwent ER or TAS for rectal adenoma, adenocarcinoma, or NET; (2) intervention—ER, including EMR and ESD; (3) comparator—TAS, including TAE, TEM, and TAMIS; and (4) outcome—procedure time, en bloc resection, R0 resection, complications (including bleeding and perforation or postoperative leakage), additional surgery, and recurrence. Nonoriginal studies, nonhuman studies, abstract-only publications, and non-English publications were excluded. In addition, studies in which more than 25% of the study participants had rectal tumors other than adenoma, adenocarcinoma, or NET were excluded from the analysis.
3. Study Selection
We conducted a comprehensive review of the identified studies through our keyword search methodology. Initially, we excluded duplicates obtained from various search engines. Subsequently, we applied our predetermined inclusion and exclusion criteria to eliminate irrelevant studies based on a thorough assessment of their titles and abstracts. Following this, we meticulously examined the full texts of the remaining studies. Eligibility evaluation was independently carried out by 2 investigators (C.H.P. and Y.S.J.), and any disagreements were resolved through discussion and consensus. In cases where consensus could not be reached, a third investigator (B.W.J.) made the final determination. Furthermore, we conducted a manual search of potentially relevant literature by scrutinizing the references of the included studies.
4. Quality Assessment
Two investigators (C.H.P. and Y.S.J.) independently performed a formal quality assessment of observational studies using the Newcastle–Ottawa Scale [12]. The scoring encompassed 3 categories: selection (4 points), comparability of study groups (2 points), and ascertainment of exposure or outcome (3 points). Studies with a cumulative score of ≥ 7 points were classified as high-quality studies. We used the Cochrane risk of bias assessment tool for randomized controlled trials (RCTs), to evaluate the risk of bias in individual studies [13].
5. Data Extraction
Data extraction was executed using a predeveloped form. Two investigators (C.H.P. and Y.S.J.) independently extracted information, including the first author, year of publication, study design, country, study period, publication language, tumor size and location, and clinical outcomes (including procedure time, en bloc resection, R0 resection, complications, additional surgery, and recurrence).
6. Study Endpoints
The primary endpoint of our meta-analysis was R0 resection. R0 resection was defined as complete resection with a pathologically negative margin. The secondary endpoint included procedure time, en bloc resection, complications (bleeding and perforation or postoperative leakage), additional surgery, and recurrence. Baseline lesion characteristics, including tumor size and location (distance from anal verge), were compared between ER and TAS. Additional surgery was defined as surgery due to noncurative resection; however, surgery due to recurrence was not considered.
7. Statistical Analyses
We conducted meta-analyses to calculate pooled odds ratios or mean differences (MDs) with 95% confidence intervals (CIs) using a random-effects model. Meta-analyses were performed according to the predominant histology ( >75%) of rectal tumors in each individual study (adenoma/adenocarcinoma vs. NET). Individual studies that included rectal tumors with various histological types but with no predominant histology were excluded from the meta-analyses. The primary comparison groups were ER and TAS. However, short-term outcomes, including procedure time, en bloc resection, and R0 resection, were further compared between ESD and TAS as sensitivity analyses.
Heterogeneity was assessed through Cochran’s Q test, with P-values <0.1 indicating significant heterogeneity and I2 statistics, with values >50% suggesting significant heterogeneity [14]. We qualitatively examined funnel plots to assess any publication bias. Additionally, we conducted a quantitative assessment of publication bias using Egger’s test, considering P-values <0.1 as statistically significant [15]. Based on the recommendations of the Cochrane group, the funnel plot asymmetry test was not conducted when <10 studies were included [16]. All P-values were two-tailed, and significance was set at P<0.05 for all tests, except heterogeneity and publication bias tests. The analysis and reporting adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [17]. All statistical analyses were conducted using Review Manager 5.3 (version 5.3.5; Cochrane Collaboration, Copenhagen, Denmark).
RESULTS
1. Study Selection and Characteristics
Seventeen studies, involving a total of 1,569 participants, met the inclusion criteria for this meta-analysis (Fig. 1). Table 1 provides a summary of the baseline characteristics of these included studies [18-34]. Their publication dates span between 2009 and 2023, with enrollment periods ranging from 1998 to 2021 [18-34]. Among the 17 studies, 1 was a RCT [27], while the remaining 16 were retrospective observational studies [18-26,28-34]. Eleven studies predominantly included patients with adenoma/adenocarcinoma [18-22,26,27,29-31,34], whereas 5 studies exclusively enrolled patients with NET [23-25,32,33]. The remaining 1 study reported results for both epithelial tumors (adenoma/adenocarcinoma) and subepithelial tumors [28]. However, the subepithelial tumor groups exhibited heterogeneity due to the inclusion of various histology types: 42.9% NET, 28.6% gastrointestinal stromal tumor, 14.3% leiomyoma, and 14.3% mucinous cystadenoma. Consequently, only results related to adenoma/adenocarcinoma were incorporated into the meta-analysis.
Study flow diagram.
Baseline Characteristics of Included Studies
The Newcastle–Ottawa Scale quality scores for the included studies are shown in Table 1. All observational studies were rated as high quality. The single RCT was deemed to have a low risk of bias across all domains, encompassing selection, performance, detection, attrition, reporting, and other biases [27].
2. Baseline Lesion Characteristics between ER and TAS
Supplementary Fig. 1A shows the difference in tumor size between ER and TAS groups. For adenoma/adenocarcinoma, the tumor size did not differ between the 2 groups (ER vs. TAS: MD, 2.55 mm; 95% CI, –1.42 to 6.53 mm), with significant heterogeneity (df =8, P<0.01, I2 =66%). However, for NET, the tumor size was smaller in the ER group than in the TAS group (ER vs. TAS: MD, –2.25 mm; 95% CI, –3.56 to –0.93 mm).
The difference in the distance from the anal verge is shown in Supplementary Fig. 1B. In the adenoma/adenocarcinoma subgroup, the lesions were located further away from the anal verge in the ER group, compared to that in the TAS group (ER vs. TAS: MD, 2.14 cm; 95% CI, 1.15 to 3.13 cm). Significant heterogeneity was observed in this comparison (df =7, P<0.01, I2 =78%). In contrast, for NET, the distance from the anal verge did not differ between ER and TAS (ER vs. TAS: MD, 0.28 cm; 95% CI, –0.59 to 1.14 cm), without significant heterogeneity (df=1, P=0.41, I2 =0%).
3. Short-term Clinical Outcomes
Short-term clinical outcomes, including procedure time, en bloc resection, and R0 resection, are demonstrated in Fig. 2. For adenoma/adenocarcinoma, the procedure time did not differ between ER and TAS (ER vs. TAS: MD, 1.55 minutes; 95% CI, –20.62 to 23.72 minutes) (Fig. 2A). However, the result should be interpreted cautiously because a significant heterogeneity was observed (df =7, P<0.001, I2 =94%). For NET, on the contrary, the procedure time was shorter in the ER group than in the TAS group, without heterogeneity (ER vs. TAS: MD, –31.32 mm; 95% CI, –37.22 to –25.42 mm; df=1, P=0.35, I2 =0%).
Forest plots for short-term clinical outcomes between endoscopic resection and transanal surgery. (A) Procedure time, (B) en bloc resection, and (C) R0 resection. SD, standard deviation; IV, inverse variance; CI, confidence interval; NET, neuroendocrine tumor; M-H, Mantel<Haenszel.
For adenoma/adenocarcinoma, the en bloc resection rate was lower in the ER group than in the TAS group (ER vs. TAS: risk ratio [RR], 0.92; 95% CI, 0.86–0.99), with significant heterogeneity (df=7, P=0.01, I2 =62%) (Fig. 2B). For NET, the en bloc resection rate tended to be lower in the ER group than in the TAS group (ER vs. TAS: RR, 0.93; 95% CI, 0.80–1.09), with significant heterogeneity (df=2, P=0.02, I2 =74%).
The R0 resection rate, the primary endpoint of the current study, is shown in Fig. 2C. For adenoma/adenocarcinoma, the R0 resection rate of ER was comparable with that of TAS (ER vs. TAS: RR, 0.99; 95% CI, 0.94–1.03), without significant heterogeneity (df =9, P=0.22, I2 =24%). However, for NET, the R0 resection rate was lower in the ER group than in the TAS group (ER vs. TAS: RR, 0.76; 95% CI, 0.68–0.84), without significant heterogeneity (df=3, P=0.95, I2 =0%).
We further performed meta-analyses for short-term clinical outcomes between ESD and TAS (Supplementary Fig. 2). Overall, the results of ESD and TAS were similar to those of ER and TAS. However, no significant differences in en bloc resection rate between ER and TAS for adenoma/adenocarcinoma were observed in the comparison of ESD and TAS (ESD vs. TAS: RR, 0.95; 95% CI, 0.89–1.01).
4. Adverse Events
Adverse events, including bleeding and perforation or postoperative leakage, are shown in Fig. 3. For adenoma/adenocarcinoma, the risk of bleeding did not differ between the 2 groups, without significant heterogeneity (ER vs. TAS: RR, 1.17; 95% CI, 0.78–1.77). For NET, the risk of bleeding also did not differ between ER and TAS, without significant heterogeneity (ER vs. TAS: RR, 1.49; 95% CI, 0.27–8.31). The risk of perforation or postoperative leakage did not differ between ER and TAS for adenoma/adenocarcinoma, without significant heterogeneity (ER vs. TAS: RR, 1.05; 95% CI, 0.40–2.77). For NET, only 1 study was included in the meta-analysis for perforation or postoperative leakage, and it did not show a significant difference in the risk (ER vs. TAS: RR, 3.21; 95% CI, 0.13–77.28).
Forest plots for adverse events between endoscopic resection and transanal surgery. (A) Bleeding and (B) perforation or postoperative leakage. M-H, Mantel–Haenszel; CI, confidence interval; NET, neuroendocrine tumor.
5. Additional Surgery and Recurrence
Additional surgery due to noncurative resection from the initial ER or TAS is shown in Supplementary Fig. 3A. The additional surgery did not differ between ER and TAS for both adenoma/adenocarcinoma and NET (ER vs. TAS: RR, 1.15; 95% CI, 0.39–3.36 for adenoma/adenocarcinoma; RR, 1.46; 95% CI, 0.64–3.34 for NET). Supplementary Fig. 3B shows the difference in recurrence between ER and TAS. For adenoma/adenocarcinoma, the risk of recurrence did not differ between the groups (ER vs. TAS: RR, 1.15; 95% CI, 0.17–7.61). For NET, only 1 study was included in the analysis of the recurrence, and it did not show any difference between ER and TAS (ER vs. TAS: RR, 0.27; 95% CI, 0.03–2.33).
6. Publication Bias
Publication bias was assessed for R0 resection and bleeding in the individual studies on adenoma/adenocarcinoma (Supplementary Fig. 4). Asymmetry of funnel plots was not identified. Additionally, no significant publication bias was identified using Egger test (P=0.449 for R0 resection, P=0.303 for bleeding).
DISCUSSION
In this study, we performed meta-analyses for rectal adenoma/adenocarcinoma and NET. In the meta-analysis for adenoma/adenocarcinoma, the TAS group had a superior en bloc resection rate than the ER group; however, no difference was observed between the 2 groups in terms of R0 resection rate and procedure time. For NET, the R0 resection rate was higher in the TAS group than in the ER group, and the en bloc resection rate also tended to be higher in the TAS group than in the ER group. However, TAS showed a longer procedure time than that for ER. For both adenoma/adenocarcinoma and NET, the rates of complications (such as bleeding and perforation), additional surgery, and recurrence, did not differ between TAS and ER.
For adenoma/adenocarcinoma, the en bloc resection rate was lower in the ER group than in the TAS group, with 88.8% in the ER group and 97.2% in the TAS group. A reason for this result may be that ER included EMR. This is supported by the finding of no difference in the en bloc resection rate when comparing the ESD (90.7%) and TAS (96.8%) groups (Supplementary Fig. 2). However, the R0 resection rates were similar between the ER and TAS groups, with 88.1% and 88.5%, respectively. Clinically, when deciding whether to perform additional surgery for endoscopically resected adenoma/adenocarcinoma lesions, R0 resection, especially the negative vertical resection margin, is more important than en bloc resection. Similar R0 resection rates between the 2 groups likely resulted in no differences in additional surgery and recurrence rates.
A previous meta-analysis study comparing ESD and TEM/ TAMIS for the treatment of early rectal tumor is consistent with our results. The meta-analysis study, which included 4 original articles and 2 abstracts searched until November 2018, demonstrated that there were no significant differences between the 2 groups for all outcomes, including R0 resection, en bloc resection, local recurrence, procedure duration, and complications [35]. Our meta-analysis additionally included studies published after 2018 and compared ER and TAS more comprehensively than the aforementioned meta-analysis study. Our results strengthen the evidence that there is no difference in outcomes and safety between ER and TAS for the treatment of adenoma/adenocarcinoma. Although our study did not evaluate cost-effectiveness, several studies have reported that ER is more cost-effective than TAS [36,37]. A previous study using the Markov model reported that ER was less expensive than TEM in the management of benign rectal polyps [36]. Another study revealed that median total hospital costs for the treatment of rectal tumors were significantly lower for ESD than for TEM (1,214 United States dollars [USD] vs. 1,686 USD) [37]. Given the similar treatment outcomes between ER and TAS and the high cost of TAS, considering ER over TAS as a primary treatment for rectal adenoma/adenocarcinoma may be reasonable. Guidelines also recommend that ER should be considered as the standard treatment for colorectal neoplasms with dysplasia confined to the mucosa, while ESD should be considered for en bloc resection of colorectal (but particularly rectal) lesions with suspicion of superficial submucosal invasion [38-40]. However, although ESD is a commonly used technique for the treatment of early-stage CRC in Asian countries, its use is still quite limited in Western countries. The management for early rectal adenocarcinoma should be determined by comprehensively considering local expertise, the endoscopist’s skill, equipment availability, and cost. For lesions where determining the direction of treatment is difficult, it may be wise to make a decision on an individual basis through a multidisciplinary team.
For NET, the R0 resection rate was higher in the TAS group than in the ER group, with 91.9% and 64.6% in the TAS and ER groups, respectively. Given that adenoma/adenocarcinoma originates from the mucosal layer, whereas rectal NET originates from the submucosal layer, the higher R0 resection rate of TAS is due to the full-thickness resection of TAS. The R0 resection rate was also significantly higher in the TAS group than in the ESD group, although the difference narrowed (ESD 72.9%) (Supplementary Fig. 2). These results suggest that ESD may be more appropriate than EMR for complete resection of NETs. However, the procedure time in the TAS group was longer than that in the ER group by a mean of 31 minutes. When compared to the ESD group, the procedure time in the TAS group was also longer, with a mean of 25 minutes. The longer procedure time for TAS is because TAS requires general or spinal anesthesia in the operating room. Specifically, TEM requires learning curves, relatively higher expertise, and expensive specialized instruments [41,42]. Although the R0 resection rate for rectal NET was higher in the TAS group, ER may be more advantageous considering its cost-effectiveness. Furthermore, unlike in adenoma/adenocarcinoma, R0 resection may be less important in the prognosis of rectal NETs smaller than 10 mm without lymphovascular invasion and atypical features. When the margins are positive after ER of rectal NETs <1 cm in size, the positive vertical margin rate is much higher than the positive lateral margin rate [43]. Rectal NET exists in the submucosal layer and ER has limitations in resecting the deep submucosal layer; therefore, it is not uncommon for vertical resection margins to be positive even after complete ER. However, this may not necessarily mean that there are residual lesions. A multicenter study examining long-term clinical outcomes of endoscopically resected rectal NETs supports this [44]. The study showed that among 137 patients with positive or indeterminate resection margins, local recurrence occurred in only 2 patients (1.5%) [44]. Similarly, our study also revealed that despite the lower R0 resection rate in the ER group, the recurrence rate was not different from that in the TAS group. In particular, among 338 patients who underwent ER, recurrence occurred in only 1 patient (Supplementary Fig. 3B). Of note, majority of the rectal NETs analyzed in the aforementioned study and in our study were <1 cm. In summary, ER should be chosen as the primary treatment for small rectal NETs. However, for large NETs with suspected deep tumorous infiltration where performing ER is challenging, TAS may be preferred.
Our meta-analysis provides a better understanding of the outcomes of ER and TAS for the treatment of rectal tumors. Nevertheless, our study has several limitations. First, although TAE, TEM, and TAMIS are different surgical techniques, they were analyzed together as TAS. Only 2 studies used TAMIS; therefore, the number of studies was too small to analyze each. As more relevant papers are accumulated in the future, ranking them using network meta-analysis seems necessary. Second, although no significant difference in recurrence was observed between the ER and TAS groups, there were limitations in accurately comparing recurrence because the follow-up period differed between studies. Third, differences in procedural or surgical skills among endoscopists and surgeons were not considered, even though these could greatly affect the outcomes. Fourth, the size of the resected lesions between the studies was heterogeneous. Lastly, all studies included in the meta-analysis, except one, were retrospective observational studies; therefore, there are concerns about selection bias, important variables could not be controlled, and the level of evidence was lowered. In the future, more RCTs on this topic should be conducted.
In conclusion, TAS had a superior en bloc resection rate in rectal adenoma/adenocarcinoma and a superior R0 resection rate and longer procedure time in rectal NET; other treatment outcomes and safety were comparable between ER and TAS. Given the cost-effectiveness, prioritizing ER as a treatment for early rectal tumors and considering TAS only for lesions that are large and suspected to have deep tumorous infiltration seems desirable. However, the medical environment may be different in each country and hospital; therefore, the choice of treatment modality should be made carefully by comprehensively considering local expertise, the endoscopist’s skill, and equipment availability.
Notes
Funding Source
The authors received no financial support for the research, authorship, and/or publication of this article.
Conflict of Interest
Jung YS is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.
Data Availability Statement
Data sharing is not applicable as no new data were created or analyzed in this study.
Author Contributions
Conceptualization; Data curation: Park CH, Jung YS. Data interpretation: Jung BW. Formal analysis: Park CH. Investigation: all authors. Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization: Park CH, Jung YS. Writing–original draft: Park CH, Jung YS. Writing–review & editing: all authors. Approval of final manuscript: all authors.
Supplementary Material
Supplementary materials are available at the Intestinal Research website (https://www.irjournal.org).
Supplementary Fig. 1.
Forest plots for baseline lesion characteristics between endoscopic resection and transanal surgery. (A) Tumor size and (B) distance from anal verge. SD, standard deviation; IV, inverse variance; CI, confidence interval; NET, neuroendocrine tumor.
Supplementary Fig. 2.
Forest plots for short-term clinical outcomes between endoscopic submucosal dissection and transanal surgery. (A) Procedure time, (B) en bloc resection, and (C) R0 resection. ESD, endoscopic submucosal dissection; TAS, transanal surgery; SD, standard deviation; IV, inverse variance; CI, confidence interval; NET, neuroendocrine tumor; M-H, Mantel–Haenszel.
Supplementary Fig. 3.
Forest plots for other outcomes between endoscopic resection and transanal surgery. (A) Additional surgery and (B) recurrence. M-H, Mantel–Haenszel; IV, inverse variance; CI, confidence interval; NET, neuroendocrine tumor.
Supplementary Fig. 4.
Funnel plots for analysis of publication bias. (A) R0 resection and (B) bleeding. Only comparisons that included ≥10 individual studies were included in the analysis for publication bias.