Exposed endoscopic full-thickness resection vs thoracoscopic surgery for complex esophageal subepithelial lesions arising from the muscularis propria

Abstract

BACKGROUND

Tumor enucleation via thoracoscopic surgery (TS) and submucosal tunneling endoscopic resection are both effective and widely used therapies for resecting most esophageal subepithelial lesions (SELs) originating from the muscularis propria (MP). However, when a lesion exceeds 35 mm in diameter and/or is located in the cervical esophagus, it is classified as a complex esophageal tumor. In such cases, submucosal tunneling endoscopic resection is not feasible due to insufficient tunnel space and the inability to successfully extract the lesion from the tunnel. Exposed endoscopic full-thickness resection (EFTR), which does not require tunnel creation, is a potential alternative treatment for complex esophageal SELs-MP. However, data on esophageally exposed EFTR remain limited.

AIM

To compare the feasibility and safety of esophageal exposed EFTR with TS for the resection of complex esophageal SELs.

METHODS

Between November 2016 and October 2023, the clinical records of patients with esophageal SELs-MP who underwent resection at the First Affiliated Hospital of Zhengzhou University were retrospectively reviewed. Patients with lesions larger than 35 mm in diameter and/or located in the cervical esophagus were included in the study. Clinicopathological characteristics, perioperative outcomes, complications, and follow-up data were collected and compared between the EFTR and TS groups.

RESULTS

A total of 60 patients with complex esophageal SELs-MP were included, with 15 patients in the EFTR group and 45 in the TS group. The EFTR and TS groups demonstrated comparable technical success rates (100% vs 97.8%, P = 0.574) and en bloc resection rates (86.7% vs 75.6%, P = 0.423). Compared to the TS group, the EFTR group had a significantly longer median procedure time (240.0 minutes vs 120.0 minutes, P < 0.001) but a shorter postoperative nasogastric decompression period (5.6 ± 4.9 days vs 10.7 ± 13.2 days, P = 0.016). In the EFTR group, complete defect closure was achieved in 3 patients (20.0%), incomplete closure in 8 patients (53.3%), and non-closure in 4 patients (26.7%). Postoperative adverse events occurred in 3 patients in the EFTR group and 4 patients in the TS group. Both groups experienced cases of esophageal stricture and fistula. Notably, chylothorax was observed exclusively in the TS group.

CONCLUSION

Esophageal-exposed EFTR demonstrated clinical outcomes comparable to those of TS, suggesting that it is a safe and feasible option for the treatment of complex esophageal SEL-MP. Further studies are warranted to validate these findings.

Key Words: Esophageal subepithelial lesion; Exposed endoscopic full-thickness resection; Thoracoscopic surgery; Defect closure; Cervical esophagus

Core Tip: Esophageal exposed endoscopic full-thickness resection (EFTR) proved to be a safe procedure for complex esophageal subepithelial lesions-muscularis propria. Postoperative adverse events were conservatively managed and occurred at an acceptable rate. Exposed EFTR achieved 100% technical success and curative resection. Even for large tumors requiring piecemeal resection for retrieval, patient outcomes were not compromised. Patients undergoing esophageal exposed EFTR demonstrated faster recovery, evidenced by earlier initiation of oral intake and shorter nasogastric intubation durations.

INTRODUCTION

Esophageal subepithelial lesions (SELs) originate from the muscularis mucosa, submucosa, or muscularis propria (MP) and are predominantly benign. Among them, leiomyomas, mesenchymal tumors, and lipomas are the most common pathologic types[1]. The detection rate of esophageal SELs has risen markedly in recent years, primarily due to the widespread adoption of gastrointestinal endoscopy and advances in endoscopic ultrasound. Esophageal SELs smaller than 2 cm are generally considered low risk and require only close surveillance. However, depending on the size of the lesion and the extent of its extension into the esophageal lumen, symptoms such as dysphagia and thoracic pain may occur, and resection may be recommended[2]. Treatment options include endoscopic and surgical procedures, including esophagectomy, and the choice of treatment should depend on lesion size, location, and local expertise[3].

Surgical enucleation is an effective treatment for most SELs, and Mutrie et al[4] reported that it is technically feasible in up to 97 % of cases. In the past, open transthoracic enucleation of esophageal SELs was associated with significant adverse events. However, with the advent of minimally invasive procedures, such as video-assisted thoracoscopic surgery (TS), postoperative complications and mortality have decreased significantly. In recent years, with the rapid development of endoscopic techniques, endoscopic resection has become a standard approach for removing SELs, including transmural lesions, offering a less invasive and safer alternative. Accepted endoscopic techniques include endoscopic submucosal resection, endoscopic submucosal dissection, submucosal tunneling endoscopic resection (STER), and endoscopic full-thickness resection (EFTR). Among these, STER is a highly feasible and safe treatment for lesions arising from the MP because it maintains mucosal integrity over the SEL, thereby reducing the risk of perforation[5]. However, STER is technically challenging for large SELs-MP or for lesions located in the cervical esophagus because the excised tumor may not be retrievable from the tunnel, the submucosal tunneling space is limited, and endoscopic maneuverability is restricted[6]. In such cases, tumor enucleation by TS is often the preferred approach. Nevertheless, EFTR offers an alternative to STER by enabling en bloc, curative resection without the need for a submucosal tunnel. Although EFTR is widely performed for gastric and colorectal tumors[7-11], its application to esophageal SELs-MP remains limited. Therefore, in this study, we aimed to compare the feasibility and safety of esophageal exposed EFTR with TS for the resection of complex esophageal SELs.

MATERIALS AND METHODS

Study design and enrollment

We retrospectively analyzed 328 patients with SEL-MP who underwent resection at the First Affiliated Hospital of Zhengzhou University between November 2016 and October 2023. The inclusion criteria were as follows: (1) Preoperative endoscopic ultrasound confirmed that the esophageal SEL originated from the MP; and (2) The lesion was > 35 mm in diameter or located in the cervical esophagus, regardless of size. The exclusion criteria: (1) Age < 18 years; and (2) Endoscopic resection not performed as exposed EFTR or surgical resection not performed as TS.

Of the 328 patients with esophageal SELs-MP, 159 had lesions located in the cervical esophagus or measuring > 35 mm; of these, 82 underwent surgical resection and 77 underwent endoscopic resection. We then screened patients who underwent TS or EFTR. Ultimately, 60 patients met the inclusion criteria and were allocated to the EFTR (n = 15) and TS (n = 45) groups. Clinicopathological characteristics, perioperative outcomes, complications, and follow-up data were collected. The flowchart of enrollment is shown in Figure 1.

Figure 1 Flowchart of patient inclusion and grouping. SEL-MP: Subepithelial lesion originating from muscularis propria; EFTR: Endoscopic full-thickness resection; TS: Thoracoscopic surgery; ESD: Endoscopic submucosal dissection; STER: Submucosal tunneling endoscopic resection.

Exposed EFTR

Exposed EFTR procedures were performed by four experienced endoscopists at our endoscopy center under general anesthesia with CO₂ insufflation. The technique included lesion marking, submucosal dissection, full-thickness resection, and defect closure using metallic clips, nylon loops, or stents as needed. Details of the procedure are outlined in Figure 2 and in prior publications[12]. Figure 3 shows the different endoscopic defect-closure techniques and outcomes.

Figure 2 Procedures of endoscopic full-thickness resection for a case with esophageal subepithelial lesion originating from muscularis propria. A: Endoscopic view of esophageal subepithelial lesion originating from muscularis propria; B: Marking dots placed on the surface of the lesion; C: Dissection of the lesion from the muscularis propria; D: Esophageal wall defect following lesion resection; E: Incomplete closure of the esophageal wall defect using metallic clips; F: Resected specimen.

Figure 3 Endoscopic techniques for defect closure. A: Incomplete closure of esophageal wall defect using metallic clips; B: Kissing suture technique for the esophageal wall defect using a nylon loop and metallic clips; C: Esophageal wall defect managed withe the placement of a fully-covered stent.

TS

Procedures were performed with patients in the left lateral decubitus position under combined inhalational and intravenous anesthesia, and a double-lumen endotracheal tube was used for single-lung ventilation. The procedure was as follows: (1) A viewing port was established in the seventh intercostal space (ICS) along the right midaxillary line, and three working ports were established in the fourth ICS along the right anterior axillary line and the sixth and ninth ICSs along the right posterior axillary line, respectively; (2) Thoracoport trocars were inserted into these ports, and an artificial pneumothorax was established; (3) The mediastinal pleura above the azygos vein was opened, and the esophagus was mobilized; (4) The esophageal muscular layer over the tumor was incised longitudinally using an electric hook, and the tumor was dissected thoroughly; and (5) Mucosal integrity was assessed, the muscular layer was intermittently sutured with 3-0 Vicryl, and a mediastinal drainage tube was placed through the ninth ICS along the posterior axillary line.

Postoperative management

Patients in both groups received standard postoperative care, including fasting, nasogastric decompression, and total parenteral nutrition. In the TS group, both a thoracic catheter and a nasogastric tube were placed during surgery. The intrathoracic tube was typically removed within 7 days, and oral intake was initiated after removal of the nasogastric decompression tube, provided there were no signs of thoracic infection.

Follow-up

The follow-up protocol was determined based on histopathological findings. All patients were advised to undergo upper endoscopy and computed tomography 3 to 6 months after treatment to assess for residual tumor. Follow-up consultations were scheduled every 6 to 12 months. When in-person visits were not feasible, outcomes were monitored via telephone.

Outcomes

The primary outcomes were compared between esophageal exposed EFTR and TS using the following criteria: (1) Curative resection, defined as a histologically negative margin with no residual tumor on postoperative pathology; (2) En bloc resection, defined as complete tumor removal without capsular disruption; (3) Technical success, defined as successful, complete resection of the lesion; and (4) Recurrence-free survival, defined as the time from the procedure to the last follow-up visit without tumor recurrence. Secondary outcomes included: (1) Procedure-related parameters, including procedure time, length of hospital stay, antibiotic prescription, intraesophageal decompression time, fasting period, and hospitalization expenses; and (2) Intraoperative and postoperative adverse events, defined as Clavien-Dindo grade III-IV complications.

Statistical analysis

Continuous variables are presented as the mean ± SD or as the median (range). Categorical variables were compared using the Pearson χ² test or Fisher’s exact test, whereas continuous variables were analyzed using the Student’s t test or the Mann-Whitney U test. Recurrence-free survival was assessed using the log-rank test. Statistical analyses were performed using SPSS version 27.0 (SPSS, Inc., Chicago, IL, United States). Statistical significance was defined as a P value < 0.05.

RESULTS

Baseline clinicopathological characteristics

A total of 60 patients met the inclusion criteria for complex esophageal SELs-MP, including 10 (16.7%) with cervical lesions, 35 with large lesions (> 35 mm in diameter), and 15 (25.0%) with both characteristics. Based on treatment modality, 15 patients underwent esophageal exposure to EFTR, and 45 underwent TS. As shown in Table 1, there were no significant differences between the EFTR and TS groups in terms of age, sex, or American Society of Anesthesiologists classification. Although lesion size did not differ significantly between groups, the mean maximum lesion diameter was slightly larger in the EFTR group than in the TS group (59.3 ± 21.7 mm vs 48.7 ± 19.3 mm, P = 0.087).

Table 1 Baseline clinicopathologic characteristics of enrolled patients, mean ± SD/n (%).

Variables	EFTR (n = 15)	TS (n = 45)	P value
Age, years	47.2 ± 15.6	41.6 ± 11.3	0.148
Male sex	8 (57.1)	28 (62.2)	0.734
Location			0.775
Cervical	7 (46.6)	18 (40.0)
Thoracic	4 (26.7)	17 (37.8)
Abdominal	4 (26.7)	10 (22.2)
Lesion size, mm	59.3 ± 21.7	48.7 ± 19.3	0.087
Pathology			0.041
Leiomyoma	10 (66.7)	40 (88.9)
GIST	3 (20.0)	1 (2.2)
Schwannomas	0	4 (8.9)
Others1	2 (13.3)	0 (0.0)
ASA class			0.325
I	12 (80.0)	40 (88.9)
II	3 (20.0)	5 (11.1)
III	0	0

¹Hemangioma, granuloblastoma.

EFTR: Endoscopic full-thickness resection; TS: Thoracoscopic surgery; ASA: American Society of Anesthesiologists; GIST: Gastrointestinal stromal tumor.

In the EFTR group, defect closure methods were tailored to lesion location and size. Complete closure was achieved in 3 patients (20.0%) using a nylon loop combined with metallic clips (n = 1), a fully covered retrievable metallic stent (n = 1), or metallic clips alone (n = 1). Incomplete closure was achieved in 8 patients (53.3%) using metallic clips, whereas 4 patients (26.7%) had unclosed defects. Of the four patients with unclosed defects, three had cervical lesions, and one had a 100-mm-diameter lesion. Intraesophageal decompression was mandatory in all cases with unclosed defects but was not used in some cases with complete or incomplete closure.

Primary outcomes

The technical success rate was comparable between the EFTR and TS groups (100% vs 97.8%, respectively; P = 0.574). Both groups achieved a 100% curative resection rate (Table 2). Although the overall en bloc resection rate was numerically higher in the EFTR group (86.7% vs 75.6%), the difference was not statistically significant (P = 0.423). Subgroup analysis by lesion location provided additional clinical insights. Exposed EFTR may offer a technical advantage in achieving en bloc resection for lesions in the abdominal esophagus (100% vs 50%), although this difference did not reach statistical significance (P = 0.118), likely due to the small sample size. Compared with the TS group, the EFTR group showed similarly high en bloc resection rates in the cervical and thoracic segments. For lesions < 75 mm in diameter, en bloc resection was achieved in 90.0% of the EFTR group and 73.9% of the TS group. However, in 2 EFTR cases (13.3%), lesions measuring 100 mm and 50 mm in diameter that were resected en bloc were divided into two fragments to facilitate retrieval using a snare. During follow-up, no recurrence was observed in the EFTR group, whereas 2 patients in the TS group experienced recurrence at 14 months and 64 months, respectively. Both recurrences were successfully managed with endoscopic resection. Detailed perioperative data for both groups are provided in Table 2.

Table 2 Perioperative data of the endoscopic full-thickness resection and thoracoscopic surgery groups, median (range)/n (%)/mean ± SD.

Variables	EFTR (n = 15)	TS (n = 45)	P value
Primary outcomes
Curative resection	15 (100)	45 (100)	> 0.999
En bloc resection	13 (86.7)	34 (75.6)	0.423
Cervical	71.4 (5/7)	77.8 (14/18)	0.739
Thoracic	100 (4/4)	88.2 (15/17)	0.471
Abdominal	100 (4/4)	50 (5/10)	0.118
Recurrence-free survival	15 (100)	43 (95.6)	0.422
Technical success	15 (100)	44 (97.8)	0.574
Clinical success	15 (100)	45 (100)	> 0.999
Secondary outcomes
Procedure time, minutes	240.0 (172.3, 262.5)	120.0 (87.50, 145.0)	< 0.001
Hospital stay, days	18.1 ± 7.4	18.78 ± 7.4	0.757
Postoperative hospital stay, days	12.2 ± 4.9	11.1 ± 5.2	0.474
Duration of antibiotic use after surgery, days	8.9 ± 5.3	8.9 ± 5.0	0.967
Nasogastric decompression tube placement	12 (80.0)	43 (95.6)	0.059
Nasogastric decompression duration, days	5.6 ± 4.9	10.7 ± 13.2	0.016
Postoperative fasting time, days	7.36 ± 3.8	10.2 ± 13.2	0.430
Hospitalization expenses, yuan	55608.1 (43969.7, 65099.5)	48680.6 (43429.0, 56037.4)	0.291
Intraoperative adverse events			0.591
Bleeding	0	2 (4.4)
Perforation	0	1 (2.2)
Postoperative adverse events			0.234
Stricture	2 (12.5)	1 (2.2)
Fistula	1 (6.3)	1 (2.2)
Chylothorax	0	2 (4.4)
Follow-up, months	36.9 ± 19.9	58.9 ± 25.1	0.004

EFTR: Endoscopic full-thickness resection; TS: Thoracoscopic surgery.

Secondary outcomes

The median procedure time was significantly longer for exposed EFTR than for TS (240 minutes vs 120 minutes, P < 0.001; Table 2). Nasogastric decompression tubes were placed in 12 patients (80.0%) in the EFTR group and 43 patients (95.6%) in the TS group (P = 0.059). However, the duration of postoperative nasogastric decompression was significantly shorter in the EFTR group than in the TS group (5.6 ± 4.9 days vs 10.7 ± 13.2 days, P = 0.016; Table 2). Postoperative length of hospital stay, antibiotic administration, fasting duration, and hospitalization expenses were similar between groups.

During the procedure, severe adverse events occurred in 3 patients (6.7%) in the TS group (Table 2). One patient experienced an esophageal perforation and required open surgery as definitive treatment. Two patients (4.4%) had intraoperative bleeding, with blood losses of 400 mL and 300 mL, requiring intraoperative transfusion. One of these patients ultimately underwent radical esophagectomy with digestive tract reconstruction. In the EFTR group, no severe intraoperative adverse events were reported.

Postoperative adverse events occurred in 4 patients in the TS group (Table 3). These included esophageal stricture in 1 patient (2.2%), which was successfully managed with a retrievable metallic stent for 9 days, and esophageal fistula in 1 patient (2.2%), which was treated conservatively with intraesophageal decompression and nasojejunal feeding for 35 days. Two patients (4.4%) developed chylothorax and required thoracoscopic duct ligation (Table 3).

Table 3 Clinical features and outcomes of patients with postoperative adverse events.

Case No.	Sex/age (years)	Adverse events	Method	Location	Size (mm)	Procedure time (minutes)	Hospital stay (days)	Use of antibiotic	Treatment
1	Female/51	Stricture	EFTR	Thoracic	85	333	21	Yes	Balloon dilation
2	Male/55	Stricture	EFTR	Abdominal	50	250	18	Yes	Stent placement
3	Male/44	Fistula	EFTR	Abdominal	57	140	40	Yes	Conservative
4	Male/22	Fistula	TS	Abdominal	60	278	19	Yes	Conservative
5	Male/67	Stricture	TS	Abdominal	75	170	20	Yes	Stent placement
6	Male/16	Chylothorax	TS	Cervical	25	156	23	Yes	Thoracic duct ligation
7	Female/72	Chylothorax	TS	Thoracic	130	226	40	Yes	Thoracic duct ligation

EFTR: Endoscopic full-thickness resection; TS: Thoracoscopic surgery.

In the EFTR group, postoperative adverse events occurred in 3 patients (Table 3). These included esophageal stricture in 2 cases, managed with balloon dilation and a retrievable metallic stent, respectively, and esophageal fistula in 1 case, which was successfully managed conservatively with percutaneous drainage and intraesophageal decompression. The percutaneous drainage tube was removed, and oral intake was resumed within 30 days after the EFTR procedure.

DISCUSSION

STER is widely considered the optimal method for removing SELs-MP because of its feasibility and safety[13]. However, STER is unsuitable for complex esophageal lesions that are large and/or located in the cervical esophagus, because there is insufficient space to create a submucosal tunnel and to allow effective endoscopic manipulation. EFTR offers a non-tunneling approach while still achieving en bloc and curative resection. Zhang et al[14] described six cases of esophageal SELs resected using EFTR with satisfactory outcomes. Our study provides additional evidence, with the following key findings: (1) Safety: Exposed esophageal EFTR is a safe procedure for complex esophageal SELs-MP. To our knowledge, this is the largest cohort of complex esophageal EFTR procedures reported to date. Although no severe intraoperative adverse events occurred, the rate of postoperative adverse events was acceptable and was managed conservatively; (2) Feasibility: Exposed esophageal EFTR achieved 100% technical success and curative resection rates. Although large tumors required piecemeal resection, this did not compromise outcomes; and (3) Recovery: Patients undergoing exposed EFTR experienced faster recovery, as evidenced by earlier postoperative oral intake and shorter nasogastric intubation duration. However, in our study, exposed EFTR required longer procedure times than TS.

Given the limited space within the esophageal tunnel, STER is typically indicated for lesions < 35 mm. Even with this limitation, en bloc resection is not always achievable. Chai et al[15] compared STER with video-assisted thoracoscopic enucleation for esophageal SELs-MP and reported a higher en bloc resection rate with thoracoscopic enucleation. Theoretically, EFTR can achieve en bloc resection regardless of tumor size; however, the difficulty increases with larger tumors. In our study, en bloc resection rates were comparable between the EFTR and TS groups, even for large cervical SELs (71.4% vs 77.8%, P = 0.739). In addition, piecemeal resection is often necessary for tumors > 80 mm in diameter. Some SELs, such as gastrointestinal stromal tumor, have malignant potential, making it critical to avoid fragmentation during surgical intervention. However, endoscopic piecemeal resection is considered safe, with a negligible risk of tumor seeding because of the protective role of the gastrointestinal wall[16,17]. Our findings support this, as no recurrences or metastases were observed during follow-up among tumors resected piecemeal.

Secure defect closure is the most challenging and critical step in exposed esophageal EFTR, particularly for transmural post-EFTR defects[18]. Various methods are widely used, including clip closure, endoloop-assisted closure, over-the-scope clip closure, and endosuturing[12,19]. Zhou et al[20] first reported successful post-EFTR defect closure using multiple metallic clips in 26 cases. Shi et al[21] introduced a nylon loop combined with metallic clips for closure of large gastric defects, demonstrating its reliability. Over-the-scope clip has also been shown to be effective for gastric EFTR, as demonstrated in a cohort of 23 patients[22]. Zhou et al[20] further proposed the use of fully covered stents to maintain esophageal integrity after EFTR and reduce complications. In our study, complete defect closure was achieved in 3 patients (20%) using a nylon loop with metallic clips, a fully covered stent, or metallic clips alone. However, 8 patients (53.3%) experienced incomplete closure, and 4 (26.7%) had unclosed defects due to the cervical esophageal location and absence of leakage. Importantly, we demonstrated that intra-esophageal decompression is a viable alternative to promote wound healing in cases of incomplete or unclosed defects.

Adverse events and outcomes

Because EFTR is still in the early stages of application for esophageal lesions, operator experience remains limited, particularly for large or cervical lesions. Rigorous intraoperative and postoperative surveillance is crucial to mitigate adverse events. No severe intraoperative adverse events occurred in the EFTR group. Major adverse events associated with EFTR include bleeding, perforation, secondary infection, pneumoperitoneum, and hydrothorax[23,24].

Proper hemostasis and defect closure are critical for preventing bleeding and perforation. Esophageal stricture was the most common adverse event in the EFTR group and was managed conservatively. In contrast to previous reports, no cases of pneumoperitoneum requiring intervention occurred in the EFTR group, likely because carbon dioxide insufflation was used during the procedure[25]. Endoscopists minimized gas insufflation intraoperatively and aspirated residual gas through the full-thickness defect before withdrawing the endoscope. Importantly, no surgical intervention or intensive care unit transfer was required in the EFTR group.

In contrast, the TS group had one case requiring conversion to open surgery because of active intraoperative bleeding, as well as two cases of postoperative chylothorax requiring thoracoscopic catheter ligation. These findings support the use of exposed EFTR as a reasonable treatment option for complex esophageal lesions, with manageable adverse events and satisfactory outcomes.

In this study, exposed EFTR required a longer procedure time but offered no advantages in hospital length of stay, antibiotic duration, fasting time, or total hospitalization expenses. These findings may be attributable to the larger and more challenging cervical lesions in the EFTR group, in which the median tumor size was 59 mm vs 49 mm in the TS group. Increased lesion complexity likely contributed to the prolonged procedure time. We anticipate that the advantages of EFTR will become more apparent as technical proficiency improves with experience along the learning curve.

Several limitations should be acknowledged. First, the retrospective design introduces potential selection bias. Second, because exposed esophageal EFTR is a novel procedure, the cohort size and number of events were relatively small, which restricted our comprehensive evaluation of its advantages and limited our ability to detect differences with adequate statistical power. Finally, longer-term follow-up is needed to evaluate the impact of exposed EFTR on quality of life.

CONCLUSION

Esophageal EFTR demonstrated clinical outcomes comparable to those of TS, suggesting that it is a safe and feasible option for complex esophageal SEL-MP. Further studies are needed to validate these findings.