As the use of endoscopy in medical examination has become more common, the frequency of gastric subepithelial lesions that are accidentally discovered is also increasing. Most gastric subepithelial lesions result in a good clinical course and final outcome. However, some lesions can progress to malignant tumors, especially those originating from the muscularis propria layer, represented by gastrointestinal stromal tumors (GISTs) [1]. To date, surgical treatment has been the first choice for complete resection of subepithelial lesions originating from the muscularis propria layer [2]. However, surgical resection is invasive, often involves the loss of a large part of the organ, and can lead to postoperative functional disorders. In addition, evaluation of the extent of resection necessary is limited for tumors that grow toward the gastric lumen because the inside of the stomach cannot be observed during surgery. Moreover, lesions adjacent to the gastroesophageal junction and posterior wall of the gastric body are often difficult to access surgically and endoscopic assistance may be required.

Recently, the safety of endoscopic treatment for gastrointestinal perforation that occurs after the procedure has increased to a great extent [3]. Therefore, to increase the en bloc resection rate of subepithelial lesions originating from the muscularis propria, endoscopic full-thickness resection (EFTR) is actively performed, and its efficiency and safety have been reported [4].

However, data in South Korea are insufficient regarding EFTR without laparoscopic assistance for subepithelial lesions originating from the muscularis propria. Therefore, we report our experiences with EFTR for gastric subepithelial lesions.

Laparoscopic wedge resection is a common treatment for subepithelial lesions originating from the muscularis propria. However, an excessive amount of normal stomach tissue can be removed in this procedure, which results in structural deformation, and depending on the location of the lesion, laparoscopic access may be difficult. In addition, when the lesion protrudes toward the lumen of the gastrointestinal tract, the extent of the lesion on the peritoneal side may be difficult to determine. As a less invasive alternative, endoscopic resection can be considered. However, if the lesion originates from the muscularis propria layer, complete resection using the conventional submucosal resection method is challenging and perforation is difficult to avoid when attempting a sufficient resection. With the development of endoscopic techniques, the gastrointestinal perforation that occurs during the procedure can be managed without any significant problems by performing a proper endoscopic closure. To completely remove the lesion without remnant tissue, complete removal of the lesion including all layers of the gastric wall and anastomosis of the perforation site may be more effective after removal of the lesion including that from the gastrointestinal layer.

Zhou et al. were the first to report 26 cases of subepithelial lesions arising from the muscularis propria that were treated using EFTR [5]. They sutured using metal clips and an endoloop without laparoscopic assistance in 2011 [5]. All the lesions were completely resected, with no serious complications and recurrence during the follow-up period. Feng et al. reported 48 cases treated with resection and suturing using the same method [6]. Huang et al. reported 35 cases and Ye et al. reported 51 cases in 2014 [7,8]. In all these reports, no serious complications or recurrence were observed. However, in the study by Ye et al., the excised specimen in 1 of the 51 cases was dropped into the abdominal cavity, and the procedure was completed with laparoscopic assistance [8]. Similarly, we had difficulty avoiding dropping of the specimen into the gastric wall defect during the procedure. However, a dual channel gastroscope can be used for the procedure, and gripping the excised tissue with forceps during the last excision step can prevent sample loss.

In our patients, all GISTs were endoscopically resected en bloc, but 4 of the 8 patients were considered for RX resection for histological examination. This is because the gastric mucosa and submucosa were removed at the first stage of the procedure and sent to the pathology department with the capsule exposed. Moreover, in the process of removing the lesion, the submucosal tissue near the capsule was damaged. As a method that can be considered to avoid RX resection, endoscopic submucosal dissection may be considered. However, in the case of GIST originating in the proper muscle layer, complete excision of the lesion while leaving the serosal layer is difficult and time consuming. This is especially true if the lesion protrudes out of the stomach [5,9]. In the case of surgical resection of GISTs, tumor rupture or intraperitoneal bleeding seems to be a major risk factor of recurrence. In addition, the 3-year relapse-free survival was reported to be similar between R0 resection (80%) and rupture-free R1 resection (79%) [10].

After the procedure, postoperative adjuvant therapy should be considered for patients with high risk of malignancy (tumor size >5 cm and mitotic count >5/50 HPF; tumor size >10 cm and mitotic count >10/50 HPF). At this time, imatinib mesylate therapy may be considered as the primary treatment [11].

If the size of the gastric wall defect after the procedure is large, it is not easy to suture with only a metal clip. This can be supplemented with an additional endoloop. In addition, devices such as an over-the-scope clip and endoscopic suture device seem useful for easier closure [12,13], although we did not use any of these devices in our study. In conclusion, development of various devices for resection and closure and accumulation of further experience through additional studies are necessary to prevent complications and specimen loss.