Clinical outcomes of underwater endoscopic mucosal resection for esophageal subepithelial tumors

Authors: Soo Bin Synn, Jin Ook Jang, Woo Jin Kim, Cheol Min Lee, Dae Gon Ryu, Cheol Woong Choi, Su Bum Park, Su Jin Kim

Abstract

Esophageal submucosal tumors are increasingly identified through routine upper endoscopy, often performed as part of health screenings. Although many of these lesions are benign, some have malignant potential, necessitating accurate diagnosis and appropriate resection. Underwater endoscopic mucosal resection (UEMR) has gained attention as a novel technique that enables mucosal resection without submucosal injection, offering advantages in terms of safety and efficacy. However, data regarding its use for esophageal submucosal tumors are limited. This retrospective single-center study included 31 patients who underwent UEMRs for esophageal submucosal tumors between 2020 and 2024. The inclusion criteria were lesions <15 mm in diameter, confined to the submucosal layer, as confirmed by endoscopic ultrasound. Clinical variables, including age, sex, lesion characteristics, procedural details, and histopathological findings, were reviewed. Outcomes included en bloc and pathological complete (R0) resection rates, complications, and recurrence during follow-up. Among the 31 patients (mean 57.0 ± 12.5 years; 61% male), the most common lesion location was the lower esophagus (45.2%). Most lesions measured 5 to 10 mm (51.6%). Histopathologic diagnoses included leiomyomas (38.7%), granular cell tumors (32.3%), inflammatory fibroid polyps, and other benign entities. All lesions were resected en bloc using UEMR, with a 100% R0 resection rate. No major complications, including bleeding or perforation, occurred. During a follow-up period of up to 37 months, no recurrence was observed, including in cases of granular cell tumors or adenoid cystic carcinomas. UEMR is a safe and effective technique for managing esophageal submucosal tumors, demonstrating high en bloc and R0 resection rates with no major complications. These findings support UEMR as a favorable option for the resection of small esophageal submucosal tumors, providing a rationale for further prospective studies comparing UEMR with conventional endoscopic mucosal resection, band ligation-assisted endoscopic mucosal resection, and endoscopic submucosal dissection.

1. Introduction

With the increasing use of endoscopy for health screening, asymptomatic esophageal subepithelial tumors have been detected more frequently in South Korea. While most of these small tumors are of no clinical significance, some of them, such as granular cell tumors (GCTs), carcinoid tumors, gastrointestinal stromal tumors, and lymphomas, have malignant potential.^[1,2]^ Therefore, differentiating between benign, malignant, and potentially malignant tumors based on the size and shape of the subepithelial tumors is necessary.^[1]^ According to the 2023 American College of Gastroenterology guidelines, resection is primarily considered for lesions with malignant potential or symptomatic lesions, and endoscopic ultrasound (EUS) could be performed to assess these features when necessary.^[3]^ While EUS can be used to diagnose subepithelial tumors, its diagnostic accuracy can vary depending on the experience and proficiency of the operator, and ultimately, the diagnoses should be confirmed histologically whenever possible.^[1,4,5]^ For smaller tumors, typically <10 mm in diameter, endoscopic mucosal resection (EMR) can be performed. In addition to conventional EMR for deeper resection, special techniques such as band ligation-assisted EMR and cap-assisted EMR can be performed.^[1,6]^

Recently, underwater endoscopic mucosal resection (UEMR) has received increasing attention as an alternative to overcome the limitations of conventional EMR.^[7,8]^ In cases of sessile lesions with large diameters or lesions in complex locations, conventional EMR has some reported issues, such as incomplete resection rates and high recurrence rates.^[7]^ In contrast, UEMR uses water without submucosal injection to float mucosal lesions and isolate them from the muscle layer, thereby increasing the complete resection rate and reducing complications.^[7–10]^

With UEMR, the mucosal and submucosal tissues form wrinkles due to underwater submergence, while the deeper muscle membrane layer retains its original shape without forming wrinkles.^[7,8]^ Upon removal of the air from the tube, the tension of the barrier decreases, and the wall returns to its original thickness.^[7]^ When placed underwater, mucosal lesions lift off of the muscle membrane due to the fat density and buoyancy of the submucosal tissue, making submucosal injections unnecessary, and the natural separation of wall layers reduces the chance of accidental involvement of the muscle membrane.^[7,9]^ In addition, water can protect the surrounding tissue from thermal damage.^[7,9,11]^ UEMR has, in particular, been recommended as an alternative to conventional EMR in the colon due to its high complete resection rate and low complication rate, even for lesions larger than 10 mm in diameter.^[7–9]^ However, research on UEMR in both South Korea and globally is still in its initial stages and has focused mostly on colorectal polyps, with only limited reports of its use for upper gastrointestinal and esophageal lesions.^[9,12–15]^

In this study, we aimed to evaluate the safety and validity of UEMR of esophageal subepithelial tumors, analyze the results of UEMR application, and present technical limitations and future research directions.

2. Material and methods

This is a retrospective single-center study of UEMRs performed at the Department of Gastroenterology, Pusan National University Yangsan Hospital, South Korea, from 2020 to 2024. The study included 31 patients who underwent UEMRs for esophageal submucosal tumors and excluded those who underwent a resection procedure other than UEMR. For patients with tumors <10 mm, conventional EMR or band ligation-assisted EMR was first considered, whereas UEMR was considered for tumors with unclear borders or where there was difficulty separating the lesion from the submucosal layer. If the lesion was between 10 and 15 mm, UEMR was considered first.

A total of 31 patients underwent UEMRs. Their ages ranged from 38 to 87 years, with a predominance of males (61%). Two of the 31 patients were taking antiplatelet medications but discontinued them with sufficient time before the procedure, while the remaining patients were not taking any antiplatelet or anticoagulant medications. No patients reported any specific symptoms before the procedure. Three of the 31 patients had endoscopic biopsies prior to the procedure, which failed to yield meaningful results. All patients underwent EUS (UM-DP20-25R, Olympus, Tokyo, Japan) evaluation before resection to assess lesion characteristics, including echogenicity, homogeneity, layer of origin, and malignant potential. Furthermore, UEMR was not performed for submucosal tumors without malignant potential, such as typical asymptomatic lipomas or leiomyomas, as confidently diagnosed on EUS evaluation. The traditional indications for EMR are lesions <10 mm, hypoechogenic lesions, solid masses, and lesions confined to the submucosal layer as assessed via EUS. To avoid overtreatment, careful patient selection was performed based on EUS findings and clinical judgment, and only tumors with malignant potential or uncertain diagnosis were considered for resection. In this study, UEMR was performed when the lesion size was <15 mm, even if it exceeded 10 mm (Fig. 1).

Two endoscopists (Su-jin Kim and Dae-gon Ryu) with more than 5 years of experience in performing therapeutic endoscopies (including endoscopic submucosal dissections) performed the procedure under conscious sedation (intravenous administration of midazolam and pethidine). For sedation, midazolam (2.5 mg) and pethidine (12.5 mg) were initially administered, with additional midazolam and pethidine administered at the discretion of the endoscopist.

UEMR was performed by filling the lumen with water to float the lesion without conventional submucosal injection. To prevent pulmonary aspiration due to the reflux of water from the esophagus, the head was elevated by 15° via a tilt table before the examination. The procedure was performed as first, the lesion was identified through the endoscope (Fig. 2A). Second, the esophageal lumen was filled with normal saline to float the lesion (Fig. 2B). Third, the lesion was grasped with a snare after assessing its size and location (Fig. 2C). Finally, the resection was completed with a high frequency current (Fig. 2D).^[8,14,15]^

Information was collected on the patient’s age, sex, the use of antiplatelet and anticoagulant drugs, the location, size, shape, and histopathological findings of the lesion, the duration of the procedure, the type of resection (en bloc vs piecemeal), the occurrence of complications, the completeness of the pathological resection (R0 resection), and the duration of follow-up and recurrence after the procedure. The specimens were sliced into 2 mm sections for histopathological examination, during which the size and margin involvement were evaluated. Complete resection was defined as the absence of microscopic evidence of tumor cells in the resection margins.

3. Results

Most of the 31 patients analyzed in this study were male (61%). The mean patient age was 57.0 ± 12.5 (range: 38–87) years. The most common lesion location was the lower esophagus (n = 14, 45.2%), followed by the middle (n = 9, 29.0%) and upper (n = 8, 25.8%) esophagus. The mean tumor size was 8.68 ± 3.91 mm. Most tumors measured 5 to 10 mm (51.6%), followed by tumors >10 mm (35.5%) and those <5 mm (12.9%; Table 1).

The most common histopathological findings were leiomyomas in 12 patients (38.7%) and GCTs in 10 patients (32.3%). Adenoid cystic carcinomas and gastritis cystica profunda lesions were identified in 1 patient (3.2%) each, and other benign lesions were identified in the remaining 7 patients. The average procedure length was 3 ± 2 minute, and all patients underwent en bloc resections. All 31 patients had endoscopically and pathologically complete resections, with no lesions invading the margins. No patients experienced any major complications, such as procedure-related bleeding, delayed bleeding, or perforation (Table 2). No patients diagnosed with GCTs with malignant potential or adenoid cystic carcinomas (a malignant tumor) had tumor recurrence during the follow-up period (up to 37 months; Table 3).

4. Discussion

This study aimed to analyze the clinical pathological characteristics and treatment outcomes of 31 patients who underwent UEMRs for esophageal submucosal tumors. Overall, UEMR was shown to be an effective treatment for esophageal submucosal tumors, exhibiting high technical and pathological complete resection rates and low complication rates. These results demonstrate the safety and efficacy of UEMR and provide important implications for the management of esophageal submucosal tumors in the future. Especially, submucosal tumors with malignant potential, such as GCTs and carcinoid tumors, generally require long-term surveillance. Performing UEMR for these tumors can provide both therapeutic and diagnostic benefits, potentially replacing the need for prolonged surveillance and reducing the burden on patients and physicians.

In this study, UEMR resulted in en bloc resection in all patients and achieved 100% pathologically complete resection (R0 resection) rates. This appears to be an excellent outcome compared to conventional EMR or ESD. In particular, shorter procedure times and lower complication rates would be the main advantages of UEMR. In this study, no major complications related to the procedure, such as bleeding or perforation, were observed. This supports the safety of UEMR, and its complication rate was lower than those reported for conventional EMR or ESD. The procedures in an underwater environment possibly made it easier to secure a field of view and naturally separate the lesion from the muscle layer, thereby reducing the risk of complications.^[7]^

The use of UEMR for esophageal lesions continues to be reported. Doumbe-Mandengue et al reported the successful use of UEMR for well-differentiated intramucosal adenocarcinoma near the esophagogastric junction and the feasibility of safe and rapid resection.^[16]^ In addition, Hwang et al evaluated the usefulness of UEMR for esophageal GCTs >1 cm and concluded that this surgical technique enables rapid and complete resection (en bloc resection).^[17]^ Deng et al reported a case in which early esophageal cancer was resected with UEMR by inducing peristaltic contraction using a high-concentration iodine solution. The same research group subsequently reported a case in which early esophageal cancer was resected via UEMR using precutting and trimming techniques.^[18,19]^

In UEMR, there is a risk that the water used to fill the esophageal lumen can be aspirated into the airway, which may lead to aspiration pneumonia.^[17]^ In particular, there is a high risk of water being swallowed by the patient during the procedure or of water entering the airway due to reflux during endoscopic manipulation.^[17]^ These aspiration risks can be reduced by applying the lateral decubitus position and the head-up tilt position during the UEMR procedure.^[17]^ These postures use the effects of gravity to prevent water from flowing back into the airway and allow for easy drainage of residual fluid if necessary.^[17]^ Tilt tables facilitate the adjustment of the patient’s posture during the procedure. As excessive water injection during the procedure may increase the risk of aspiration, it is important to use only the minimum amount of water necessary and to perform continuous aspiration at the same time.^[16]^ Despite the procedure being performed under conscious sedation, excessive sedation may suppress the swallowing reflex, which may increase the risk of aspiration.^[18]^ Therefore, it is important to maintain an appropriate level of sedation according to the condition of the patient, who can be supplemented with oxygen, such as high-flow nasal cannula oxygen therapy if necessary to maintain tissue oxygenation.^[19]^

This study had a few limitations. First, while UEMR is a relatively simple technique that can be easily applied by experienced endoscopists, there may be an initial learning curve. Second, this study was a retrospective single-center study, which is subject to sample size limitations and possible selection bias. These limitations need to be overcome through future multicenter, large-scale prospective studies to further establish the efficacy and safety of UEMR. In addition, it is necessary to evaluate the applicability of UEMR to lesions of different sizes and locations and to analyze recurrence and survival rates through long-term follow-up. This will allow future comparative studies between UEMR and other endoscopic resection techniques to establish optimal treatment strategies.

5. Conclusion

In conventional EMR, which has been mainly applied to small lesions, the lesion is lifted through submucosal injection and then resected. However, this method is limited by the size and location of the lesion, and there are also technical difficulties in narrow spaces such as the esophagus. UEMR floats the lesion using water without submucosal injection, which induces a natural separation between the lesion and muscle layer, facilitating resection.^[8]^ In conclusion, this study suggests that UEMR is a safe and effective procedure with a high complete resection rate and low complication rate in the treatment of esophageal submucosal tumors. These results can serve as a basis for considering the use of UEMR in the endoscopic treatment of esophageal tumors in the future.

Author contributions

Conceptualization: Soo Bin Synn, Su Jin Kim.

**Data ** Soo Bin Synn, Jin Ook Jang, Woo Jin Kim, Cheol Min Lee, Cheol Woong Choi, Su Bum Park, Dae Gon Ryu.

**Formal ** Soo Bin Synn.

Investigation: Soo Bin Synn, Jin Ook Jang, Woo Jin Kim, Cheol Min Lee, Cheol Woong Choi, Su Bum Park, Dae Gon Ryu.

Methodology: Soo Bin Synn, Su Jin Kim.

**Project ** Su Jin Kim.

Supervision: Su Jin Kim.

Visualization: Soo Bin Synn.

**Writing – original ** Soo Bin Synn.

**Writing – review & ** Soo Bin Synn, Jin Ook Jang, Woo Jin Kim, Cheol Min Lee, Cheol Woong Choi, Su Bum Park, Dae Gon Ryu, Su Jin Kim.