Endoscopic full-thickness resection: A definitive solution for local complete resection of small rectal neuroendocrine neoplasms

Abstract

BACKGROUND

Recently, several endoscopic techniques have been used to improve the R0 resection rate of rectal neuroendocrine neoplasms (R-NENs). However, none of these methods can achieve 100% complete resection (CR), particularly in the vertical direction. Endoscopic full-thickness resection (EFTR) has proven to be an effective method for the treatment of submucosal tumors but is seldom utilized in the eradication of R-NENs.

AIM

To review cases of R-NENs removed using EFTR and to evaluate the safety and efficacy of this technique.

METHODS

This retrospective cohort study enrolled 160 patients with pathologically confirmed R-NENs, including 132 who underwent endoscopic submucosal dissection (ESD) and 28 who underwent EFTR. Lesions were categorized as < 1 cm, 1-2 cm, and > 2 cm in size. CR rate, en bloc resection rate, operation time, and complications were evaluated. Subgroup analyses and follow-up were also performed.

RESULTS

EFTR achieved 100% CR rates for lesions < 1 cm and 1-2 cm, compared with 67.0% and 50.0%, respectively, in the ESD group. En bloc resection and successful removal of the R-NENs were achieved in all patients. Meanwhile, EFTR showed performance comparable to ESD in terms of operation time, hospitalization cost, and postoperative adverse events, except for a one-day longer hospital stay. We also analyzed the invasion depth of R-NENs based on full-thickness specimens. The data showed that 80% of lesions (< 1 cm) and 85.7% of lesions (1-2 cm) had invaded the SM3 level or deeper at the time of resection. For ESD specimens, 46.6% (< 1 cm) and 89.3% (1-2 cm) of lesions had infiltrated more than 2000 μm beneath the muscularis mucosae.

CONCLUSION

EFTR has shown superior performance in the resection of small R-NENs compared with that of ESD.

Key Words: Endoscopic full-thickness resection; Endoscopic submucosal dissection; Rectal neuroendocrine neoplasms; Complete resection; Endoscopic technique

Core Tip: This retrospective cohort study included 160 patients with rectal neuroendocrine neoplasms who underwent endoscopic full-thickness resection or endoscopic submucosal dissection. Despite similar procedural durations and costs, endoscopic full-thickness resection (EFTR) achieved a 100% complete resection rate, a significant improvement compared to endoscopic submucosal dissection. Additionally, EFTR demonstrated a one-day longer hospital stay with no complications observed. This study indicates EFTR’s potential as a viable alternative for small, deeply infiltrating rectal neuroendocrine neoplasms, highlighting the importance of vertical resection depth for complete tumor eradication.

INTRODUCTION

Neuroendocrine neoplasms (NENs) are rare tumors arising from neuroendocrine cells, with the gastrointestinal (GI) tract and pancreas being the most common sites of involvement. In recent decades, the incidence of NENs has increased because of the popularization of gastroenterological screening via endoscopes. Rectal NENs (R-NENs), appearing as single yellowish subepithelial lesions with intact overlying mucosa, account for approximately 34%-53% of GI NENs[1,2] and 1%-2% of all rectal tumors[3]. R-NENs are generally asymptomatic and exhibit more indolent behavior than that of those in other locations, primarily because they are often diagnosed at an early stage[4] and have relatively high five-year survival rates[5].

Tumor size is regarded as the principal predictor of metastasis in R-NENs[6]. Tumors < 1.0 cm, accounting for approximately 80% of R-NENs, are associated with a very low metastasis rate (< 1%), whereas tumors > 2 cm, accounting for approximately 12%-20%, have a lymph node metastasis rate of approximately 60%-80% at diagnosis[7]. However, the metastatic potential of 1.0 to 2.0 cm tumors remains controversial[7]. Recently, several countries have updated their clinical guidelines for GI NENs. Although many factors, such as tumor location, infiltration depth, and Ki-67 index, have been considered to play key roles in the prognosis of R-NENs[8], a consensus has been reached that complete resection (CR) of grade 1/2 R-NENs smaller than 1 cm is curative, with no further follow-up required[9].

Various endoscopic techniques, such as polypectomy, conventional endoscopic mucosal resection (EMR), and endoscopic submucosal (SM) dissection (ESD), have been applied for R-NEN resection[10,11]. However, the R0 resection rate is generally unsatisfactory; the overall CR rates reportedly range from 39% to 85% for EMR and 84.08% for ESD[12,13]. Therefore, additional treatment and careful surveillance are required for residual lesions. Recently, modified methods, such as EMR with ligation devices[14] and cap-assisted EMR (EMR-C)[15,16], have achieved relatively high R0 resection rates. However, a small proportion of lesions cannot be removed radically, mainly because of insufficient depth of excision.

Endoscopic full-thickness resection (EFTR), a type of natural orifice transluminal endoscopic surgery, is a minimally invasive technique widely applied in the resection of gastric SM tumors arising from the muscularis propria (MP)[17]. Over the past several years, numerous studies have validated the efficacy and safety of EFTR[18], suggesting that it is a promising option for removing SM lesions in the GI tract. However, despite its effectiveness in ensuring adequate excision depth, there have not been any high-quality trials on its usefulness in the treatment of R-NENs. In the present study, we conducted a retrospective cohort study to compare the safety and efficacy of EFTR with ESD for small R-NENs, aiming to provide evidence regarding the optimal endoscopic technique for these patients.

MATERIALS AND METHODS

Study design

We conducted a multicenter, retrospective clinical study to investigate whether EFTR was superior to ESD for R-NENs at Jiangsu Province Hospital of Chinese Medicine and Nanjing Drum Tower Hospital. This study was approved by the Institutional Review Board of both institutions.

Patients

We retrospectively reviewed 164 patients with R-NENs who underwent ESD, EFTR, or surgery at Nanjing Drum Tower Hospital (November 2012 to September 2020) and Jiangsu Province Hospital of Chinese Medicine (March 2022 to August 2023). The inclusion criteria were a semispherical, pale-yellow SM lesion seen on colonoscopy with a smooth mucosal surface, minimal congestion, and no erosion or ulcers, pathologically confirmed as R-NEN after the procedure. The exclusion criteria were lymph node or distant metastases on endoscopic ultrasonography or computed tomography, intolerance to endoscopy, concurrent malignancy, severe comorbidities, and refusal or inability to provide informed consent. All patients provided written informed consent before undergoing the endoscopic procedures.

Interventions

ESD: ESD was performed using a standard adult colonoscope (GIF-Q260J; Olympus, Tokyo, Japan). After marking the periphery of the lesions with an electrosurgical knife (MK-T-1-195; Micro-Tech, Nanjing, China), a mixed solution of 0.01% epinephrine and methylene blue was injected into the submucosa to raise the surface mucosa. A circular incision and subsequent SM dissection were performed using electric cautery (ERBE, Tübingen, Germany) until the lesion was completely resected. Subsequently, the wound was cleaned, and thermal coagulation forceps (HBF-23/1600, Micro-Tech, Nanjing, China) were used to cauterize the wound to prevent delayed bleeding. The wound was carefully sutured using SureClip (ROCC-C-26, Micro-Tech, Nanjing, China). Resected specimens were measured in vitro and sent for pathological and immunohistochemical analyses. The detailed procedure is illustrated in Figure 1.

Figure 1 Endoscopic submucosal dissection. A: A pale yellow mass with a diameter of about 4.8 mm in the rectum; B: Electrocoagulation marking in the peritumoral area by using the front end of the electrosurgical snare; C: Submucosal injection of the mass with an injection needle; D: Wound surface after the removal of the mass; E: Wound clipping with Sureclips after mass resection; F: The resected mass for pathological evaluation.

EFTR: EFTR was performed using a standard adult colonoscope (GIF-Q260J; Olympus, Tokyo, Japan). After marking the periphery of the lesions with an electrosurgical knife (MK-T-1-195; Micro-Tech, Nanjing, China), the mucosal layer was incised along the outer side of the marked edge, and the muscle layer was gradually dissected to the serosal layer. Dissection continued until the lesions were completely resected. After handling with thermal coagulation forceps and confirming the absence of active bleeding, the wound was sutured with a nylon rope or SureClip (ROCC-C-26, Micro-Tech, Nanjing, China) for pouch suturing. Resected specimens were measured in vitro and sent for pathological and immunohistochemical analyses. The detailed procedure is illustrated in Figure 2.

Figure 2 Endoscopic full thickness resection. A: A pale yellow mass with a diameter of about 8 mm in the rectum; B: Electrocoagulation marking in the peritumoral area by using the front end of the electrosurgical knife and submucosal injection of the mass with an injection needle; C: Full thickness resection of the mass with electrosurgical knife; D: Wound surface after the removal of the mass; E: Wound clipping with Sureclips after mass resection; F: The resected mass for pathological evaluation.

Histological evaluation

Efficacy was evaluated by assessing the rates of histological CR (R0), lymphatic vessel invasion (LVI), and perineural invasion (PNI), whereas safety was evaluated by assessing complications. The primary evaluation parameter was the histological R0. Complete histological resection was defined as the en bloc resection of a targeted lesion with horizontal and vertical free margins. A positive margin indicated that, after resection, the pathological evaluation showed tumor involvement at the vertical or horizontal margin or that the tumor was adjacent to the vertical or horizontal margin[19].

The secondary evaluation parameter was the depth of infiltration, which was measured as shown in Figure 3. In the EFTR group and surgical specimens, the resected specimens had a complete MP and the serosal layer; therefore, the SM layer between the mucosal muscle and MP could be divided into three equal parts (SM1, SM2, and SM3) under a microscope. Detailed data on infiltration depth in patients are shown in Table 1 (< 1 cm group) and Supplementary Table 1 (1-2 cm and > 2 cm groups). The ESD group, because of the lack of complete MP, was divided according to the depth of SM infiltration: < 1000 μm, 1000-2000 μm, and > 2000 μm (Supplementary Table 2). The EFTR and surgical pathological specimens are illustrated in Figure 4. All samples were evaluated for histological type, LVI, and PNI via pathological examination and histological diagnosis. Histopathological grades included NEN grade 1 (G1), NEN grade 2 (G2), and neuroendocrine carcinoma (G3), according to the 2019 World Health Organization (WHO) classification[20].

Figure 3 Infiltration depth measurement diagram. EFTR: Endoscopic full-thickness resection; ESD: Endoscopic submucosal dissection.

Figure 4 Pathological images of surgical and endoscopic full-thickness resection specimens. EFTR: Endoscopic full-thickness resection.

Table 1 Infiltration depth in endoscopic full-thickness resection group (< 1 cm), n (%).

	Infiltration depth
SM1	3 (15.0)
SM2	1 (5.0)
SM3	15 (75.0)
PM	1 (5.0)

Follow-up

The patients were monitored via telephone calls and medical records to assess their postoperative follow-up and survival status. Necessary health education and follow-up guidance were provided to the patients. The follow-up endpoint was defined as the date of completion or loss to follow-up, measured in months.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 9.3 (GraphPad, United States), with quantitative data expressed as mean ± SD. Normally distributed variables were compared using the t-test, whereas non-normally distributed data were analyzed using the Mann-Whitney U test. Categorical variables were described as percentages and evaluated using χ² or Fisher’s exact tests. Statistical significance was set at a two-tailed P < 0.05, and the confidence interval was 95%.

RESULTS

Comparison of lesion resection between the two groups

This study found that the baseline characteristics of the patients were similar between the EFTR and ESD groups. Lesions were categorized into three subgroups based on size (< 1 cm, 1-2 cm, and > 2 cm) according to the American Joint Committee on Cancer/Union Internationale Contre le Cancer[21] guidelines. In the EFTR group, 71.4% of the tumors were < 1 cm, 25% were 1-2 cm, and 3.6% were > 2 cm. Similarly, in the ESD group, 78.0% of tumors were < 1 cm, 21.2% were 1-2 cm, and 0.8% were > 2 cm, with no significant differences between the groups. Postoperative hospitalization was slightly longer in the EFTR group compared with that in the ESD group, a difference that was statistically significant (P = 0.007). However, there was no significant difference in the operation time between the two groups (Table 2).

Table 2 Comparison of lesion resection between the two groups, n (%).

Group	Total	ESD	EFTR	P value
Patient number	160 (100)	132 (82.5)	28 (17.5)
Age (year, mean ± SD)	52.24 ± 11.90	52.29 ± 12.20	52.00 ± 10.59	NS
Gender				NS
Male	78 (48.8)	65 (49.2)	13 (46.4)
Female	82 (51.2)	67 (50.8)	15 (53.6)
Tumor location				NS
Lower rectum	98 (61.3)	80 (60.6)	18 (64.3)
Middle rectum	56 (35.0)	47 (35.6)	9 (32.1)
Upper rectum	6 (3.7)	5 (3.8)	1 (3.6)
Tumor size				NS
< 1 cm	123 (76.9)	103 (78.0)	20 (71.4)
1-2 cm	35 (21.9)	28 (21.2)	7 (25.0)
> 2 cm	2 (1.3)	1 (0.8)	1 (3.6)
Post-operative hospitalization (days, mean ± SD)	4.49 ± 1.25	4.35 ± 1.13	5.14 ± 1.60	0.007
procedure time (min, mean ± SD)	31.99 ± 13.39	32.58 ± 12.65	29.21 ± 16.39	NS
Hospitalization cost ($, mean ± SD)	2883.38 ± 229.95	2896.97 ± 231.16	2819.29 ± 216.74	NS

ESD: Endoscopic submucosal dissection; EFTR: Endoscopic full-thickness resection; NS: No significance.

Comparison of postoperative pathological features between the two groups

To compare the postoperative pathological features between the two groups, the resection margin and WHO grading for lesions < 1 cm were carefully assessed. CR was achieved in all patients in the EFTR group (20/20, 100%), whereas ESD achieved CR in 67% of cases (69/103, P = 0.003). Regarding the WHO grading, most lesions were classified as grade G1, but the percentage of grade G2 was statistically different between the two groups (1% vs 10%, P = 0.017), which may be because of the small number of patients in the EFTR group. LVI and PNI were detected in four patients in the ESD group but in none in the EFTR group, although this difference was not statistically significant (Table 3).

Table 3 Comparison of postoperative pathological features between the two groups (< 1 cm), n (%).

Group	Total	ESD	EFTR	P value
Margin				0.003
Positive	34 (27.6)	34 (33.0)	0 (0)
Negative	89 (72.4)	69 (67.0)	20 (100)
WHO grade				0.017
G1	120 (97.6)	102 (99.0)	18 (90.0)
G2	3 (2.4)	1 (1.0)	2 (10.0)
LVI				NS
Positive	4 (3.3)	4 (3.9)	0 (0.0)
Negative	119 (96.7)	99 (96.1)	20 (100)
PNI				NS
Positive	4 (3.3)	4 (3.9)	0 (0.0)
Negative	119 (96.7)	99 (96.1)	20 (100)

ESD: Endoscopic submucosal dissection; EFTR: Endoscopic full-thickness resection; WHO: World Health Organization; LVI: Lymphatic vessel invasion; PNI: Perineural invasion; NS: No significance.

For lesions measuring 1-2 cm, EFTR also exhibited a better R0 resection rate than that of ESD (100% vs 50%, P = 0.016), and for the two lesions > 2 cm, EFTR also ensured R0 resection (Supplementary Table 3). We also evaluated the WHO grading and LVI/PNI status for lesions measuring > 1 cm. Interestingly, tumors measuring 1-2 cm did not show an increasing ratio of G2 or G3 grades but exhibited a considerable increase in the LVI/PNI ratio compared with observations in those measuring < 1 cm. Since there were only two lesions measuring > 2 cm, a comprehensive analysis of the WHO grading and LVI/PNI status was infeasible (Supplementary Table 4). This limitation is related to the fact that guidelines generally do not recommend endoscopic resection for R-NEN lesions > 2 cm[8]. However, since tumor size is considered the only risk factor, the positive rates of LVI and PNI significantly increased with tumor size (Supplementary Table 5).

Pathological features of surgically resected NENs

To analyze the biological behavior of R-NENs, we reviewed four surgically resected lesions. The characteristics of the lesions are listed in Supplementary Table 6. Despite the limited sample size, several noteworthy findings were observed. All lesions exhibited aggressive behavior; the largest tumor invaded the MP layer, and the smallest invaded SM1, whereas the remaining two penetrated SM3 at the time of excision. These findings indicated that deep infiltration can occur even in diminutive lesions. Regarding WHO grading, except for cases < 1 cm, which were of G1 grade, all remaining cases were G2 and G3, suggesting that histological grading correlates with tumor size. In terms of LVI and PNI, tumors > 2 cm in size showed LVI positivity, whereas tumors with a diameter of 3 cm showed both LVI and PNI positivity. These findings suggest that R-NENs, which are usually considered indolent, become more aggressive as their size and depth of invasion increase.

Complications and postoperative follow-up in both groups

No instances of bleeding were observed in the EFTR group, whereas one case of delayed bleeding was reported in the ESD group, which was successfully managed through endoscopic hemostasis. Overall, no statistically significant differences in bleeding rates were found between the two groups. Furthermore, neither group experienced complications, such as delayed perforation or infection, indicating that the effectiveness of EFTR is comparable to that of ESD. Of the 160 patients in the ESD and EFTR groups, 26 were lost to follow-up. The follow-up period ranged from 3 months to 11.25 years, with an average of 39.46 months. Both groups underwent routine colonoscopies and computed tomography for surveillance purposes. Supplementary Table 7 lists the follow-up times and recurrence rates of patients with tumors < 1 cm and 1-2 cm in each subgroup. Notably, no recurrence or metastasis was observed in the EFTR group or in ESD cases that had achieved en bloc resection during the follow-up period, up to the time of submission. The mean follow-up periods were 43.50 ± 42.89 months and 18.86 ± 12.55 months, respectively (Supplementary Table 8).

DISCUSSION

Given that R-NENs < 1 cm demonstrate a very low metastatic risk and favorable prognosis[22], guidelines recommend no routine follow-up for patients with CR in this group of R-NENs[23]. Consequently, attention should be paid to en bloc resection and CR rates of lesions, regardless of which endoscopic technique is selected[24]. Currently, there are no clear recommendations for selecting an endoscopic resection technique. Many studies have compared different approaches for improving the CR rate. Incomplete resection puts patients at risk of metastasis, resulting in repeated endoscopic and follow-up radiologic examinations, as well as the need for salvage therapy[24]. As previous research indicates that 80%-90% of R-NENs are localized to the submucosa[25], it is reasonable to choose ESD or EMR to remove R-NENs. Conventional EMR is an efficient method for treating R-NENs; however, insufficient capture of the SM layer may lead to incomplete histological resection. Various modified EMR techniques have been derived, such as EMR with ligation devices[26,27], EMR-C[28], and underwater EMR[29], to remove R-NENs. However, approximately 91% of lesions fail to achieve R0 resection[30] because of insufficient vertical depth and horizontal margin residuals[27]. Dual-channel EMR is a simple and effective technique that has been proven to achieve a better CR rate[31]; however, most endoscopic centers are not easily equipped with dual-instrument channel endoscopes. Recently, Gao et al[16] developed a modified EMR-C method to achieve a high CR rate that was non-inferior to ESD. However, incorrect aspiration may also lead to incomplete resection of the tumor, especially in the horizontal direction. As for ESD, which was initially developed for early cancers, dissection at a slightly superficial level may easily result in residuals on the vertical margin[32] (the residual rate was approximately 11%[30]). Although many modified techniques have been applied to minimize the occurrence of positive vertical resection margins in R-NENs[33], none of these methods have achieved a 100% success rate. According to our analysis based on EFTR specimens, when the SM layer was divided into three parts (SM1, SM2, and SM3), mimicking the pathological assessment of colorectal cancer, approximately 75% of R-NENs (< 1 cm) and 71.4% (1-2 cm) invaded the SM3 level. In addition, 5% of the R-NENs (< 1 cm) and 14.3% (1-2 cm) had invaded the MP at the time of discovery (Table 1 and Supplementary Table 1). Meanwhile, when the SM layer was divided into two parts (SM1 and SM2 or deeper) based on ESD specimens, 76.7% (< 1 cm) and 96.4% (1-2 cm) of R-NENs invaded the SM2 or deeper layers (Supplementary Table 2). This finding further demonstrates that conventional SM dissection is not sufficient for the local eradication of small R-NENs. Therefore, negative vertical margins rather than horizontal margins are more important for achieving CR in small R-NENs. Based on this fact, some researchers have developed over-the-scope clip-assisted full-thickness resection methods for small R-NENs, which proved safe, efficient, and achieved a high R0 resection rate of 100%[34,35]. These studies indicate a feasible approach that guarantees the complete removal of lesions, especially at the vertical margin, by expanding the resection depth. However, owing to their high costs and irreversible procedures, over-the-scope clip-based methods are unsuitable for most centers.

Transanal endoscopic microsurgery (TEM) is the primary treatment for larger tumors and greater depths of tumor infiltration[36], which can ensure deep removal of the MP. Recently, clinical trials have validated TEM as an effective approach for treating small R-NENs with a relatively high R0 resection rate and similar degree of safety[37]. However, compared with endoscopic approaches, the relatively high cost and complicated manipulation restrict the promotion of TEM. Meanwhile, postoperative fecal incontinence may affect the quality of life of patients[38], especially for lesions close to the anal margin[25]. Notably, our data indicated that 61.3% of the lesions were located in the lower rectum (Table 2), suggesting that TEM should be carefully considered for this population of patients.

In the present study, we proposed EFTR as a good choice for small R-NENs because full-thickness excavation can ensure sufficient depth without increasing the cost or complications. According to our data, the R0 resection rate was 100% at both centers (Table 3), marking a significant advancement over other endoscopic techniques. Meanwhile, EFTR is particularly safe and practical for R-NENs as they are frequently located in the middle to lower rectum (within 10 cm of the anorectal junction)[39] beneath the peritoneal reflection. There is plentiful mesorectum around the rectal wall, which can help reduce the risks of delayed perforation and severe pelvic infection. In fact, in the present study, EFTR of R-NENs did not result in more complications, increased operation time, or higher resource consumption (Table 2). However, we properly extended the hospital stay by approximately 1 day in the EFTR group (P = 0.007) to monitor potential delayed perforations and subtle local infections because of the depth of resection (Table 2). In addition, current studies have concluded that the prognosis, recurrence, and treatment choice of R-NENs < 1 cm are not related to the invasion depth, and the prognosis and recurrence are almost the same regardless of whether the depth of invasion is SM1, SM2, or SM3.

Recently, therapeutic strategies for R-NENs measuring 1-1.9 cm have not reached consensus acceptance[40]. Previous studies indicated a metastatic risk of 10%-15% in this category[41]. Consequently, the ENETS Guidelines suggest a surgical approach for R-NENs > 1 cm in size with features, such as MP invasion and lymphovascular involvement[8]. However, in the present study, R-NENs sized 1-2 cm that underwent endoscopic resection showed favorable clinical outcomes (Supplementary Table 7), which is in accordance with a recent multicenter retrospective study[42]. These data suggest that endoscopic resection might be the first-line treatment for R-NENs 1-2 cm in size. Nonetheless, attention should be paid to preoperative assessments, including endoscopic ultrasonography and/or magnetic resonance imaging, to predict potential locoregional spread or lymph node metastasis. In addition, patients with risk factors, such as female sex, redness, and G2 tumor grade, should be carefully monitored because of their association with regional or distant metastases[43].

This study had some limitations. First, this was a retrospective study with only 28 EFTR cases, which could only partially demonstrate the advantages of this technique. A prospective randomized controlled study is urgently needed in the near future. Second, the postoperative follow-up of EFTR cases is still ongoing, and long-term outcomes are under observation. In addition, the present study merely compared EFTR with ESD, necessitating cohort studies on EFTR and other effective techniques. Future studies will help to further assess the efficacy and safety of EFTR in a broader patient population.

CONCLUSION

This study demonstrated that EFTR is an efficient and safe approach for treating small R-NENs and showed good performance in achieving CR. In addition, the limited data in this study suggested that EFTR was also effective for R-NENs measuring 1-2 cm with low risk. Further high-quality research is required to normalize the indications and applications of EFTR for R-NENs.

ACKNOWLEDGEMENTS

The author is grateful to all members of the research group for their support.