Utility of underwater endoscopic mucosal resection combined with a protruding anchor by saline injection for superficial non-ampullary duodenal tumors: a retrospective study in Japan

Article information

Clin Endosc. 2025;.ce.2024.181

Publication date (electronic) : 2025 March 12

doi : https://doi.org/10.5946/ce.2024.181

Yoshie Nomoto¹, Satoshi Shinozaki^1,2, Yoshimasa Miura^1,3, Hiroyuki Osawa¹, Yuji Ino¹, Tomonori Yano¹, Nikolaos Lazaridis⁴, Hironori Yamamoto^,1

¹Division of Gastroenterology, Department of Medicine, Jichi Medical University, Shimotsuke, Japan

²Shinozaki Medical Clinic, Utsunomiya, Japan

³Division of Gastroenterology and Hepatology, Department of Medicine, Nihon University School of Medicine, Tokyo, Japan

⁴Royal Free Unit for Endoscopy, The Royal Free Hospital and UCL Institute for Liver and Digestive Health, London, UK

Correspondence: Hironori Yamamoto Division of Gastroenterology, Department of Medicine, Jichi Medical University, 3311-1 Yakushiji, Shimotsuke, Tochigi 329-0498, Japan E-mail: ireef@jichi.ac.jp

Received 2024 July 4; Revised 2024 September 19; Accepted 2024 September 30.

Abstract

Graphical abstract

INTRODUCTION

Superficial non-ampullary duodenal tumors (SNADETs) are rare; however, the endoscopic resection of SNADETs is increasing in Japan.1 This rise could be attributed to the early detection of SNADETs, which has been facilitated by the development and evolution of image-enhanced endoscopy and heightened endoscopist awareness. Despite these advancements, endoscopic submucosal dissection (ESD) of SNADETs is associated with high perforation rates (6%–50%) and moderate R0 resection rates (38%–100%).2 In contrast, endoscopic mucosal resection (EMR) does not have adequate R0 resection rates, and piecemeal resection undoubtedly elevates the rate of local recurrence. This is noteworthy because repeat resection becomes more challenging endoscopically due to scar formation and the inherently thin duodenal wall. Furthermore, underwater EMR (UEMR) for SNADETs was reported in 2013 as a novel strategy for endoscopic resection. During this procedure, the gastrointestinal lumen is filled with water, and snare resection is performed without submucosal injection.3 As a result, UEMR for SNADETs <20 mm in size has significantly higher en-bloc resection rates and shorter procedure times compared to EMR.4

Anatomically, unlike the colonic haustra, the duodenal wall forms a linear muscularis. Thus, the duodenal fold does not contain the muscularis. This anatomical variance allows the convoluted folds to flow into the water-filled duodenal lumen and facilitates the resection of the superficial lesion without muscularis involvement, which contributes to a decreased perforation risk even without submucosal injection. Therefore, UEMR has emerged as the first-line therapy for SNADETs, preferred over both EMR and ESD.5

Despite the development and widespread use of UEMR for SNADETs, R0 resection has not yet been achieved in all cases.6 To address this challenge and improve the R0 resection rate while minimizing local recurrence, we developed a new UEMR method, facilitated by a protruding anchor created by saline injection into the distal duodenal fold (UEMR-A).7 The aim of this study was to evaluate the usefulness and safety of UEMR-A for SNADET resection compared to conventional UEMR (UEMR-C).

METHODS

UEMR-A procedure

Initially, a mucosal injection of 2 to 3 mL normal saline was performed into the anal side fold (not the flat area between folds) just after the target lesion. This injection was later used to anchor the tip of the snare and secure the anal side margin. Mucosal markings were not placed prior to the procedure. Afterward, all gas and air were aspirated from the duodenal lumen, which was then filled with water to create the underwater environment needed for resection. The snare tip, inserted through the accessory channel, was positioned on the protruded mucosa, serving as a cushion and enhancing stability. The rotatable snare was then opened and adjusted to cover the target lesion with sufficient margins on both the left and right sides. Even during snare closure, the protrusion enabled stable snare manipulation. Subsequently, snare resection was performed with electrocauterization. The deployment of endoscopic clips assisted in mucosal defect closure while remaining underwater (Fig. 1).

Fig. 1.

Procedure of underwater endoscopic mucosal resection aided by a protruding anchor created by saline injection into the distal duodenal fold (UEMR-A). (A) A superficial non-ampullary duodenal lesion measuring 7 mm in diameter. (B) A protrusion is created by injecting 2 to 3 mL of normal saline into the anal side fold. (C) A submucosal protrusion serves as an anchor for the snare tip and secures the anal side margin. (D) The snare tip is attached to the protrusion after water immersion, ensuring it does not easily miss the target even after the snare is opened. (E) The snare tip is moderately fixed, facilitating movement at various angles. Snaring can be performed under clear visualization. UEMR-A, UEMR aided by a protruding anchor created by saline injection into the distal duodenal fold.

RESULTS

Baseline characteristics and outcomes

There were no differences between the UEMR-A and UEMR-C groups regarding lesion size, location, or involvement of a fold (Tables 1, 2). Notably, more procedures were performed by expert endoscopists in the UEMR-C group than in the UEMR-A group. Moreover, the procedure time for UEMR-A was significantly shorter than that for UEMR-C, despite the additional time needed for submucosal injection. Additionally, the R0 resection rate in the UEMR-A group was significantly higher than that in the UEMR-C group, despite most procedures in the UEMR-A group being performed by non-experts. Furthermore, the horizontal margins were significantly clearer in the UEMR-A group compared to the UEMR-C group.

Table 1.

Baseline characteristics

Table 2.

Endoscopic resection and histopathology

For tumors smaller than 15 mm, the R0 resection rate was significantly higher in the UEMR-A group than in the UEMR-C group. For tumors located on the anterior wall, the R0 resection rate was also higher in the UEMR-A group compared to the UEMR-C group. No tumors exhibited submucosal invasion. In the UEMR-A group, the R0 resection rate performed by an expert was 82.4% (14/17), which was comparable to that of non-expert endoscopists at 83.8% (31/37). No intraprocedural bleeding or perforation was observed; however, delayed bleeding occurred in one patient in the UEMR-C group.

Factors associated with R0 resection

Factors affecting the R0 resection rate were examined using multivariate analysis to control for extraneous variables (Table 3). The use of UEMR-A was identified as the only positive significant factor for R0 resection in both univariate and multivariate analyses. Tumor size, involvement of a fold, and resection performed by an expert did not appear to influence R0 resection rates.

Table 3.

Factors associated with an R0 resection

DISCUSSION

This retrospective comparative study revealed that UEMR-A has a significantly higher R0 resection rate compared to UEMR-C. Anchoring with a mucosal protrusion facilitates precise snare placement by allowing the tip of the snare to be fixed in the duodenum. In the UEMR-C technique, the extent of the resection area is sometimes uncontrollable due to the influence of peristalsis and the fixed duodenum on the retroperitoneum. In contrast, UEMR-A creates an artificial protrusion on the anal side with a spring effect, allowing for adequate lateral margins through precise opening of the snare. This fixation of the snare tip by the protrusion prevents an excessive anal side margin and facilitates visualization of the oral and lateral side margins. Even in non-expert hands, the UEMR-A technique facilitated R0 resection rates similar to those performed by experts. This is the first original report demonstrating the superiority of UEMR-A over UEMR-C.

UEMR, developed by Binmoeller et al.,8 was disseminated worldwide over the last decade, a phenomenon referred to as the “underwater revolution”. In this context, endoscopic ultrasound studies have demonstrated that the underwater method can induce circumferential convolution of the mucosa. Nevertheless, convolution of the muscularis has not been identified, suggesting that UEMR is a relatively safe procedure. In contrast, ESD is considered an established strategy for resecting superficial esophageal, gastric, and colorectal tumors. However, duodenal ESD is considered a high-risk procedure for perforation and requires sophisticated techniques.9 Furthermore, although duodenal piecemeal EMR for SNADETs is a viable option, a thinner duodenal wall makes salvage ESD difficult and carries an extremely high risk of perforation in cases with local recurrence. Therefore, R0 resection on the first attempt is more important than resection in other parts of the gastrointestinal tract. Overall, considering the perforation risk and advanced endoscopic skills required for duodenal ESD, we believe that ESD should not be considered the standard treatment option for SNADETs by non-expert endoscopists. In light of these factors, UEMR-A is easier to perform than UEMR-C and can facilitate en-bloc resection regardless of the location of the lesion and the experience of the endoscopist, as long as the lesion size is less than 15 mm.

UEMR-C for SNADETs can be a challenging therapeutic procedure due to several factors, including the slippery duodenal mucosa, limited visual field, bowel peristalsis, and retroperitoneal fixation. As a result, these factors can prolong procedural times and potentially extend unnecessary margins. To address these difficulties, the UEMR-A technique has been developed specifically to tackle the issue of the slippery duodenal mucosa. Notably, there is an anatomical difference in the submucosal structure between the duodenal folds and the flat areas. This difference allows for solution injection to be dispersed in the thin submucosal layer between folds; however, injecting into the folds can create an appropriate protrusion.10 Thus, submucosal saline injection during UEMR-A should be performed on the “fold area” at the anal side of the SNADET to minimize the dispersion of the injected solution and to create a suitable soft protrusion with a spring effect.

To further enhance procedural efficacy, the protrusion created during UEMR-A is sufficiently soft that the tip of the snare can be moderately fixed, allowing push-and-pull movements in any direction. Moreover, the flexibility of the snare tip on the protruded mucosa facilitates the opening and maneuvering of the snare onto the duodenal mucosa to obtain adequate horizontal margins. Furthermore, since the snare tip is not solidly fixed, its location can be finely tuned and adjusted. This moderate fixation of the snare tip during UEMR-A helps shorten the procedure time compared to UEMR-C. This fixation also influences the curability of the endoscopic resection. Notably, this study revealed that there were no differences in R0 resection rates between UEMR-A and UEMR-C for lesions located on the papilla or the posterior wall, which align with the direction of the working channel. In contrast, lesions located on the opposite side of the papilla or the anterior wall require rotational maneuvers in the narrow lumen, making resection more difficult. In these latter cases, the R0 resection rate in the UEMR-A group was higher than that for in UEMR-C group. Therefore, UEMR-A can overcome the difficulties associated with lesion location, ultimately facilitating a high R0 resection rate.

During endoscopic procedures, the closing of the snare can be complicated by the slippery duodenal mucosa and excessively convoluted lumen, leading to the resection of a larger portion of normal mucosa at the anal side margin, which may elevate the risk of intraprocedural or delayed bleeding and perforation. However, these risks can be mitigated using UEMR-A. In addition, to secure appropriate lateral margins, attempts at snare closing can be repeated using UEMR-A, as the moderately fixed tip of the rotatable snare on the artificial protrusion allows for push-and-pull maneuvering. By applying a slight push during the closing of the snare, the shape of the snare can expand laterally, enabling repeated adjustments to secure negative horizontal margins, even in the slippery duodenal mucosa between the folds. Consequently, UEMR-A provides reliably negative horizontal margins for tumors larger than 10 mm in size that extend over two folds.

Any SNADETs <20 mm are generally indicated for UEMR.11,12 However, the R0 resection rate for this procedure has been found to be unsatisfactory, as indicated by the results of a large Japanese multicenter study (56%),6 which is notably lower than the rate observed in the current UEMR-A group (83%). Notably, Takatori et al.13 reported the usefulness of UEMR combined with partial submucosal injection, achieving a 96% en-bloc resection rate. This technique lifts wide-range mucosa with a submucosal injection into the anal side of the lesion, which facilitates careful observation and resection while viewing the anal side. In contrast, our method creates an anal-side artificial protrusion with a spring effect, allowing for an appropriate lateral margin by precisely opening the snare when it is pushed against the protrusion. This fixation of the snare tip by the protrusion prevents excessive resection of the anal side margin and aids in visualizing the oral and lateral side margins, resulting in a suitably sized resected specimen. However, as described in Table 2, the R0 resection rate for tumors ≥15 mm was not satisfactory, even in the UEMR-A group (57%); therefore, duodenal ESD should be considered for tumors of this size.

The strengths of this study include the following: first, this is a novel procedure that requires a simple step involving a single submucosal saline injection. Second, the fixation of the snare tip is adjustable, allowing for changes in its location as needed. The snare tip is moderately fixed and does not move easily due to the spring effect, even when the snare is pushed or pulled. Third, this method does not require any dedicated devices, making it cost-effective. However, our study has some limitations. Initially, this was a single-center retrospective observational study and not a prospective one. Additionally, the two groups did not run completely in parallel; the majority of the UEMR-C procedures were performed chronologically earlier than the UEMR-A procedures. Lastly, the level of experience among endoscopists was not the same in the two groups.

In conclusion, UEMR-A could improve the R0 resection rate compared to UEMR-C. Interestingly, despite the need for additional time for the injection, the overall procedural time was significantly shorter than that for UEMR-C. UEMR-A also minimized excessively large lateral margins. Therefore, we recommend UEMR-A for the resection of SNADETs <15 mm.

Article information Continued