Endoscopic submucosal dissection and submucosal tunneling endoscopic resection of gastric subepithelial lesions originating from the muscle layer: a multicenter retrospective study

Introduction

Subepithelial lesions (SELs) in the gastrointestinal tract are usually incidental findings during upper gastrointestinal (GI) endoscopy. The incidence is about 1 in every 300 endoscopies, and 2/3 are located in the stomach. SELs include a large spectrum of lesions with varied potential for malignancy, ranging from completely benign pathologies, such as ectopic pancreatic rest, to potentially malignant lesions, such as neuroendocrine tumors and GI stromal tumors (GISTs) [1-4].

Among gastric SELs, GISTs are the most commonly encountered, followed by pancreatic rests, lipomas, leiomyomas and neuroendocrine tumors. These lesions require further investigation to determine their nature and malignant potential [5].

The diagnostic workup for gastric SELs includes computed tomography (CT), magnetic resonance imaging and endoscopic ultrasound (EUS), often complemented by fine-needle aspiration (EUS-FNA) or biopsy [5-7]. However, despite these modalities, obtaining a definitive diagnosis without tissue resection remains challenging. Even with EUS-FNA or biopsy, histopathological confirmation may be inconclusive, particularly for lesions originating from the muscularis propria layer [8].

Endoscopic surveillance at regular intervals for SELs smaller than 2 cm is suggested by the existing guidelines, while surgical resection is recommended for lesions exceeding this size [1,5,6]. However, emerging evidence suggests that small SELs, more specifically GISTs, could be potentially malignant despite their size. Furthermore, prolonged surveillance has its own drawbacks, including patient anxiety, loss to follow up, and higher healthcare costs due to repeated examinations [9]. The most recent recommendation from the European Society for Medical Oncology (ESMO) suggests the resection of histologically confirmed GISTs, even when they are less than 2 cm [10].

Until now, surgical excision has been considered the gold standard for the treatment of gastric GISTs [4,6]. However, advancements in endoscopic techniques and accessories have significantly expanded the feasibility of en bloc resection with effective defect closure [11]. Endoscopic resection offers distinct advantages, particularly for lesions in anatomically challenging locations, such as the esophagogastric junction or cardia, where laparoscopic access may be complex [12]. In addition, the endoscopic resection preserves the anatomy of the stomach, thus sparing the risk of postoperative anatomical deformity, gastroparesis or dyspepsia often related to surgical wedge resections. In particular, knife-assisted resection (KAR) enucleates the tumor and, if necessary, intentional perforation and dissection beyond the serosa are applied. In most publications the term ESD is used for KARs (although dissection may expand beyond the submucosal layer to the muscle layer or even the serosa), while the term STER is reserved for endoscopic resections of SELs through a submucosal tunnel that is closed at the end of the procedure. The choice between these techniques depends on the location, the size and the need for dissection beyond the serosa. Most frequent complications related to either ESD or STER include difficulty to close the wall defect, intraprocedural or late bleeding, peritonitis, and gas-related effects [13]. However, most of these adverse events may be managed conservatively. In addition, in cases with preoperative uncertainty about the need for surgical closure of the defect, the endoscopic procedure is undertaken inside the surgical theater, so that cooperative laparoscopic endoscopic surgery can be applied if needed [14]. The aforementioned techniques demand advanced endoscopic skills, high-end endoscopic equipment, surgical backup, and copious training in reference centers that perform peroral endoscopic myotomy (POEM) and ESD. Moreover, they are mainly applied in Asia, while only limited experience has been published from other parts of the world [15-17].

In this retrospective study, we aimed to evaluate the feasibility, efficacy and safety of endoscopic resection for gastric SELs arising from the muscularis propria layer, mainly in countries outside of East Asia.

Patients and methods

Technical details

All procedures were performed under general anesthesia using a gastroscope fitted with a transparent hood under CO₂ insufflation. Depending on the risk of perforation and the availability of laparoscopic endoscopic cooperative surgery (LECS), the operation took place in the endoscopy unit or inside the operating theater. A tip-cutting knife was used in the majority of cases; the type varied among centers (Flush Knife, FujiFilm Tokyo; Dual Knife Olympus, Tokyo; Hybrid Knife I type, ERBE, Tubingen). In case of bleeding, visible vessels were coagulated with the tip of the knife, or with a hemostatic forceps. Various techniques were used for defect closure, such as clip-and-loop closure, conventional clip closure, over-the-scope clip closure, or endoscopic suturing. In cases of STER, the mucosal window was closed in all cases, while in cases of ESD it was left at the discretion of the endoscopist, based on the risk of late perforation or of bleeding. Representative examples of endoscopic resection of gastric SELs originating from the muscle layer are shown in Fig. 1. In our study, we defined the resection technique as KAR when intentional serosal dissection was performed, and as ESD when the serosa remained intact. On the other hand, the term endoscopic full-thickness resection is broad and encompasses various endoscopic techniques—such as KAR, full-thickness resection device and STER—which may lead to inconsistencies in communication among endoscopists. Fig. 2 shows our endoscopic approach based on lesion size, depth of invasion, and site of origin.

Figure 1 (1A) Knife assisted resection of a gastric GIST of the posterior wall of the gastric body. (1B) The tumor invades the muscularis propria. (1C) Exposure of the peritoneal fat. (1D, E) Clip and band endoloop closure. (1F) Final specimen. 2 Demonstration of STER. (2A) A leiomyoma of the posterior wall of the gastric body. (2B) Tunneling. (2C) Enucleation of the tumor. (2D) Resection of the tumor through the tunnel. (2E, F) Mucosal window before and after closure. (3A) Knife assisted subserosal resection of a GIST of the gastric angulus. (3B) Exposure of the tumor. (3C) Subserosal dissection. (3D, E) Closure of the defect with a suturing device. (3F) Final specimen GIST, gastrointestinal stromal tumors; STER, submucosal tunneling endoscopic resection

Figure 2 Endoscopic options for resection of epithelial lesions and SEL depending on size (<1.5 cm vs. >1.5 cm) and depth of invasion SEL, subepithelial lesions; FTRD, full-thickness resection device; EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; KAR, knife-assisted resection; LECS, laparoscopic endoscopic cooperative surgery

Results

A total of 82 patients (38% men) were included in the study, recruited from 18 centers. Clinical and demographic characteristics are summarized in Table 1. The patients’ median age was 58.5 years, and the median lesion size was 20 mm (range 6-80 mm). The lesions were located in the cardia, fundus, body and antrum in 18%, 18%, 40% and 23% of the cases, respectively. Among these 4 gastric regions, the median tumor size did not differ significantly and was 21 mm in the cardia, 18 mm in the fundus, 20 mm in the body, and 21 mm in the antrum. Most tumors were endoluminal, 16 cases presented extraluminal growth, and 6 extended beyond the serosa.

Table 1 Clinical and demographic characteristics of 82 patients undergoing endoscopic resection of gastric subepithelial lesions across 18 centers

Indications for endoscopic resection included a confirmed diagnosis of GIST (38%), increasing lesion size with uncertain histology (25%), symptomatic presentation (3%), patient preference for removal over surveillance (14%), and a combination of these factors (20%).

ESD was performed in 54 cases (66%) and STER in 28 (34%), with a median procedure time of 120 min (range 24-305). In 26 cases a deliberate perforation was performed, 24 of which were managed endoscopically, while 2 cases were converted to surgical intervention (LECS) after failure to close the defect. Histopathological examination diagnosed GISTs in 78% of cases, while leiomyomas and schwannomas were identified in 20% and 2%, respectively. Seven GISTs of 64 (11%) had a high mitotic rate, whereas 57 (89%) showed a low mitotic rate. Table 2 presents an overview of the procedure details by technique (ESD or STER).

Table 2 Procedure-related details by endoscopic technique (ESD vs. STER)

En bloc resection was successfully achieved in 79 of 82 cases (96%). Moderate to severe complications, classified as AGREE Grade II, III and IV—including perforation, bleeding, conversion to surgery and peritonitis—occurred in 6 cases (7%). The majority of adverse events were mild, consisting of pneumoperitoneum (9 cases) and abdominal pain, which was managed conservatively. Two cases were converted into LECS, as mentioned above. One patient experienced perforation leading to peritonitis and prolonged hospitalization; however, no mortality was reported. Local recurrence occurred in 2 cases after a mean follow up of 12 months (range 3-60 months). Primary and secondary outcomes by technique are shown in Table 3.

Table 3 Primary and secondary outcomes by endoscopic technique (ESD vs. STER)

Discussion

This is the first multicenter retrospective study outside northeast Asia that evaluates the feasibility and safety of endoscopic resection of gastric SELs originating from the muscularis propria. Our study included cases exclusively involving the deep wall layers, including lesions with extraluminal extension, given the paucity of data concerning their endoscopic management. In our patients, 26.8% of the lesions presented with extramural growth. Intentional full-thickness resection with controlled serosal perforation was performed in 31.7% of cases. En bloc resection was achieved in 96%, and endoscopic closure was feasible in the vast majority of cases. The recurrence rate was low (3.7%) during a median follow up of 12 months, consistent with previous publications for endoscopically resected SELs (2.2%), and comparable to the surgical series (approximately 5%) [20]. More specifically, the recurrence rate in GISTs <4 cm ranges from 0-6.7% [21].

No recurrence was observed among patients with leiomyomas or pancreatic rests, aligning with their benign behavior. Overall, the complication rate in our study was 7%, and the majority of the adverse events were successfully managed endoscopically. There were no fatalities. The efficacy and safety were similar in the STER and ESD groups, although STER was used more frequently to treat leiomyomas, which could possibly reflect anatomical accessibility or even tumor morphology.

According to the current guidelines, endoscopic surveillance is recommended for gastric SELs smaller than 2 cm, while surgical resection is suggested in lesions greater than that size [1,5]. However, recent histopathological studies suggest that even small SELs could have a malignant potential, which can range from approximately 15% in early studies [22] to about 40% in more recent ones [23]. Moreover, almost half of SELs <2 cm are found to be GISTs [24]. Despite these indications, there is no consensus on the optimal management of these small lesions.

Surveillance strategies are imperfect and could come with some downsides. Through repeated endoscopies the cumulative costs rise significantly, diagnosis of a malignancy could be delayed, and the uncertainty could lead patients to become non-compliant or to experience psychological distress [25]. In view of these factors, a therapeutic endoscopic resection may offer an alternative option in selected cases, especially in experienced centers.

In contrast to the 4.4% macroperforation rate requiring surgical closure reported by Joo et al (25), in our study a deliberate macroperforation was performed in 26 cases (31.5% in ESD and 32.5% in STER), 24 of which were managed successfully using different endoscopic closure modalities. These modalities included through-the-scope clip placement, loop-and-clip combination, cap-assisted clip closure and endoscopic suturing. In most patients, a CT scan with oral contrast was performed the next day to confirm adequate closure. The majority of the patients were discharged within 3 days, highlighting the safety of the procedure when performed by experienced endoscopists.

The AGREE classification system, an update to the ASGE lexicon, was used to classify the complications [19,26]. Mostly minor events were documented. Two patients required surgical closure of wall defects (Grade III), and 1 patient developed peritonitis post-ESD, which was managed conservatively with through-the-scope and over-the-scope endoscopic clip placement, followed by radiologically guided drainage. This patient’s hospital stay was 20 days, which was the longest hospital stay overall, in both groups, and most likely influenced the overall higher mean stay in the ESD group.

Lesions treated with ESD were significantly larger than those in the STER group (26.7 vs. 19.0 mm, P=0.0026). Additionally, tumors in anatomically complex locations, such as the cardia or esophagogastric junction, were resected more frequently via ESD. This fact may have contributed to the longer procedure times and greater risk of adverse events in the ESD group, although the difference in procedure duration was not statistically significant (126 vs. 105 min, P=0.07).

Endoscopic resection can potentially have some advantages over surgery in anatomically challenging areas. Near the esophagogastric junction or pylorus, endoscopic resection could be organ-preserving, whereas conventional surgery could require partial gastrectomy [12]. In our study, 18.3% of lesions were located in the cardia, and through the endoscopic resection a more extensive surgical intervention was most likely avoided. In Fig. 3, we propose an algorithm for the endoscopic resection of SELs arising from the muscle layer.

Figure 3 Endoscopic resection strategy for SEL originating from the muscle layer based on the feasibility for tunneling and for endoscopic closure SEL, subepithelial lesions; KAR, knife assisted resection; ESD, endoscopic submucosal dissection; STER, submucosal tunneling endoscopic resection; LECS, laparoscopic endoscopic cooperative surgery

Interestingly, about one third of the procedures in this study were performed by experienced endoscopists in non-university tertiary care centers, suggesting that advanced endoscopic techniques can be successfully performed outside academic institutions or university hospitals.

The main strengths of this study are the large sample size and the inclusion of multiple centers from different regions, which reflects the real-world heterogeneity in practice and experience of the endoscopist. On the other hand, there are some important limitations, such as the retrospective design of the study and the lack of standardized procedural protocols. Further limitations are the potential selection bias and the variability in follow-up intensity and reporting.

In conclusion, endoscopic resection of gastric subepithelial tumors arising from the muscularis propria is a feasible, safe and effective approach, including in selected extraluminal cases, when performed by skilled endoscopists in specialized tertiary centers. Surgical backup should be readily available, particularly in cases involving full-thickness resections. Future prospective studies with longer follow up and risk stratification are needed to define the long-term oncologic outcomes of this minimally invasive approach.