Superficial non-ampullary duodenal epithelial tumors (SNADETs) are defined as tumors originating from the non-ampullary region of the duodenum and consist of dysplastic glandular epithelium. SNADETs are uncommon, with an estimated prevalence of 0.3% to 4.6%.^1,2 Recently, the detection rate of SNADETs has been increasing owing to the widespread use of health checkup endoscopy, development of endoscopic imaging technology, and increased awareness of this disease.³ Approximately 60% of SNADETs occur in patients with familial adenomatous polyposis (FAP), and sporadic occurrence is rare.⁴ The natural history of SNADETs is not well understood owing to their rarity; nevertheless, SNADETs are considered precancerous lesions because the adenoma-adenocarcinoma sequence has been reported in the duodenum.⁵ Therefore, early diagnosis and management are essential.

Previously, the traditional treatment strategy for SNADETs was radical surgical resection, such as pancreaticoduodenectomy. However, endoscopic resection is currently regarded as the first-line treatment for SNADETs owing to the high risk of morbidity and mortality associated with duodenal surgical resection.⁶ Endoscopic resection methods include cold snare polypectomy (CSP), conventional endoscopic mucosal resection (cEMR), underwater EMR (uEMR), and endoscopic submucosal dissection (ESD).⁷

The duodenum has specific anatomical features, causing it to be one of the most dangerous and difficult areas in the digestive tract for endoscopic procedures.⁸ Furthermore, most endoscopists have limited experience in endoscopic resection in the duodenum, contrasting with the extensive experience in such procedures in the colon; therefore, outcomes of endoscopic resection for SNADETs vary by country and institution.⁸ For these reasons, strategies for endoscopic resection of SNADETs have not yet been standardized. Recently, the European Society of Gastrointestinal Endoscopy (ESGE) issued the first clinical practice guidelines for SNADETs.⁹ Herein, we reviewed the current status and recent advances in endoscopic resection for SNADETs.

1. Anatomical features of the duodenum

Endoscopic resection in the duodenum is more difficult and has a higher risk of adverse events compared with other organs owing to specific duodenal anatomical features, including: (1) narrow and tortuous duodenal lumen, which restricts the reverse method of manipulation; (2) rich blood supply; (3) abundant Brunner’s gland in the submucosal layer, leading to difficult mucosal lifting; (4) a thin proper muscular layer, which is associated with a high risk of perforation; (5) poor endoscope maneuverability owing to the long distance of the duodenum from the mouth and its C-loop structure; and (6) exposure to bile and pancreatic juice from the duodenal papillae, which cause more severe and fatal adverse events compared with endoscopic resection in other organs.¹⁰

2. Indication for endoscopic resection of SNADETs

Similar to the colorectal adenoma-adenocarcinoma sequence, duodenal adenoma is the precursor of duodenal carcinoma.⁵ Okada et al.¹¹ reported that approximately 21% of low-grade dysplasia progressed to high-grade dysplasia or noninvasive carcinoma over a 6-month follow-up, and high-grade dysplasia diagnosed at the first biopsy and a lesion diameter of ≥20 mm were significant predictive factors for progression to adenocarcinoma. Patients with FAP are known to have a high prevalence of duodenal adenomas, and such adenomas have been reported to slowly progress to cancer.¹² Sporadic non-ampullary duodenal cancers occur either by the adenoma-adenocarcinoma sequence or de novo.⁵

Determining which SNADET should be treated is crucial to reduce the risk of progression to invasive cancer. The ESGE recommends that all duodenal adenomas should be considered for endoscopic resection since they have a high likelihood of progressing to invasive carcinomas.⁹ Endoscopic resection plays a role as a treatment and diagnostic modality in SNADETs. First, endoscopic forceps biopsies have low diagnostic accuracy for SNADETs, whereas the diagnostic accuracy of endoscopic biopsy is 68% to 74%, and 20.3% of biopsy-proven adenomas are histologically diagnosed as carcinomas after resection.^3,13 We previously reported a histopathologic discrepancy rate of 19.0% between endoscopic forceps biopsies and endoscopic resection specimens in SNADETs.¹⁴ In addition, preoperative endoscopic forceps biopsies may induce scarring and fibrosis in the submucosal layer, causing difficulty in subsequent endoscopic resection and necessitating conversion to more invasive methods, increasing the risk of adverse events.^13,15 The ESGE suggests that if endoscopic features are suggestive of superficial duodenal adenoma, the use of biopsy for histological assessment should be limited prior to endoscopic resection.⁹

In addition to duodenal adenoma, early duodenal cancer without lymph node metastasis (LNM) is a good candidate for endoscopic resection.¹⁶ No established definition of early duodenal cancer or indications for endoscopic resection of early duodenal cancer regarding invasion depth and LNM risk exist. A previous study reported the lack of LNM in 34 surgically resected intramucosal duodenal cancers; however, LNM was observed in five of 12 submucosal duodenal cancers.¹⁷ Nishio et al.¹⁸ reported that among 42 patients with duodenal cancer who underwent lymph node dissection, LNM was found in none of the all 15 patients with pTis-T2.¹⁸ However, only two cases with pT1b and one with T2 were included in that study. Therefore, the authors suggested that lymph node dissection could be omitted in Tis-T1a duodenal cancers. The evidence remains weak owing to the low incidence and small number of reported cases of SNADETs; nonetheless, endoscopic resection of intramucosal non-ampullary duodenal cancer may be an ideal replacement treatment modality for surgical resection. Accordingly, all patients with duodenal cancer should undergo abdominal computed tomography before endoscopic resection to evaluate the presence of lymph node or distant metastases.

1. Cold snare polypectomy

CSP is an endoscopic procedure that uses only a snare with mechanical strangulation, without submucosal injection or electrocautery (Fig. 1). Tappero et al.¹⁹ first described CSP of colorectal polyps, and it has become a standard procedure for small colorectal lesions <10 mm.^20,21 Endoscopists have recognized the cold approach as a potential therapeutic option for SNADETs, based on the development of the cold snaring technique for colorectal lesions and the lower rate of adverse events compared with hot snare techniques.²² Regarding its advantages, CSP is easy and safe, with a low incidence of adverse events. Hamada et al.²³ reported that 332 duodenal adenomas were resected using CSP in 10 FAP patients, without adverse events, except for one intraprocedural arterial bleeding that occurred during resection and was easily managed using hemoclips; most of these lesions were sized <10 mm. Okimoto et al.²⁴ reported that intraprocedural bleeding occurred in 30 of 46 lesions (65.2%) during CSP and all cases were successfully managed using hemoclips. Other studies reported a lack of perforation or delayed bleeding after CSP (Table 1).^24-26

Figure 1. Cold snare polypectomy of a duodenal epithelial tumor. (A) A 6 mm white, slightly elevated lesion (0-IIa) located in the second portion of the duodenum. (B) Narrow-band imaging. (C) After snaring the entire lesion along with the surrounding mucosa, resection was performed without a high-frequency device. (D) The lesion was completely resected.

Table 1 . Summary of Cold Snare Polypectomy Outcomes for Duodenal Epithelial Tumors.

Year	Author	No. of lesions	Tumor size, mm	En bloc resection	R0 resection	Immediate		Delayed		Recurrence
						Perforation		Bleeding	Perforation
2017	Maruoka et al.25	30	4 (2−6)	29 (96.7)	17 (68.0)	0		0	0	0
2018	Hamada et al.23	332	NA	NA	NA	0		0	0	NA
2022	Takizawa et al.26	21	8 (3−10)	17 (81.0)	NA	0		0	0	1 (5.5)
2022	Okimoto et al.24	37	4 (2−7)	36 (97.3)	26 (70.3)	0		0	0	1 (2.7)
2022	Kato et al.29	187	5.5±2.4	149 (79.1)	64 (40.5)	0		1 (0.5)	0	NA

Data are presented as median (range), number (%), or mean±SD..

NA, not applicable..

Nevertheless, CSP may have inferior curability to hot snare techniques because of the absence of electrocautery, which can eradicate neoplastic tissue around the resected site. A previous study on colorectal polyps reported that CSP had a lower resection depth and a higher incomplete resection rate than cEMR.²⁷ Therefore, CSP should not be used for suspected carcinoma lesions. Nevertheless, despite the low rate of complete resection, several studies of CSP for SNADETs have reported a low incidence of recurrence.^24-26 Notably, recurrence rates are significantly associated with lesion size and not with the resection technique.²⁴

Therefore, CSP can be indicated for SNADETs <10 mm, although the ESGE recommends CSP for adenomas <6 mm.⁹ In particular, CSP can be appropriate for FAP patients with numerous and small duodenal polyps because it is simple and safe.²³

2. Conventional endoscopic mucosal resection

According to the ESGE guidelines, cEMR is the first choice of endoscopic resection for SNADETs.⁹ cEMR involves submucosal injection to create a lift between the lesion and the muscularis propria, allowing for subsequent resection using a snare combined with electrocautery (Fig. 2). cEMR is familiar to endoscopists, and most of them can perform this procedure without difficulty.²⁸ cEMR increases the chance of complete resection of lesions. A high-volume multicenter study showed that the en bloc and R0 resection rates of cEMR were inferior to those of ESD; however, the adverse event rate of cEMR was significantly lower than that of ESD.²⁹ Basically, en bloc resection rates vary with lesion size. Lesions ≤20 mm can be removeden blocusing cEMR at approximately 80% to 90% (Table 2).^29-38 However, en bloc resection rates of cEMR decrease in lesions >20 mm.^29,36 In these cases, ESD can be considered as the first treatment to achieve a high en bloc resection rate; however, considering the high incidence of adverse events with ESD in the duodenum, 20- to 30-mm sized lesions could be acceptable for piecemeal cEMR. Piecemeal cEMR tends to increase the recurrence rate;³⁶ nevertheless, Nonaka et al.³⁷ reported no residual recurrence during a median follow-up period of 51 months in a population with a high rate of piecemeal cEMR. Furthermore, most local recurrences can be treated by another endoscopic procedure.^29,36 The decision to use cEMR for lesions >30 mm should be carefully considered owing to considerable major adverse events and recurrence after cEMR.³⁹ Although these adverse events have a lower incidence than those of ESD, they cannot be completely avoided. Several studies have reported higher adverse event rates after endoscopic resection in the duodenum than in other areas of the digestive tract, with bleeding and perforation occurring in approximately ~12% and ~3% of cases, respectively (Table 2). In summary, SNADETs ≤20 mm can be safely and reliably removed using cEMR.

Figure 2. Conventional endoscopic mucosal resection of a duodenal epithelial tumor. (A) A 10 mm white, slightly elevated lesion (0-IIa) located in the second portion of the duodenum. (B) A saline solution containing small amounts of epinephrine and indigo carmine dye was injected beneath the lesion. (C) Snare resection was performed using a high-frequency device. (D) The lesion was completely resected. (E) The defect was completely closed by clipping. (F) The resected specimen.

Table 2 . Summary of Conventional Endoscopic Mucosal Resection Outcomes for Duodenal Epithelial Tumors.

Year	Author	No. of lesions	Tumor size, mm	En bloc resection	R0 resection	Immediate		Delayed		Recurrence
						Perforation		Bleeding	Perforation
2015	Nonaka et al.37	113	12 (3−50)	71 (62.8)	38 (33.6)	0		14 (12.4)	0	0
2017	Hoteya et al.31	55	9.4±5.3	43 (78.2)	33 (60.0)	1 (1.8)		4 (7.3)	0	2 (3.6)
2018	Yahagi et al.30	146	9.8±6.3	139 (95.2)	123 (82.2)	1 (0.7)		2 (1.4)	0	NA
2018	Tomizawa et al.36	166	20 (7−55)	88 (53.0)	NA	0		8 (4.8)	0	32 (23.0)
2019	Hara et al.33	136	9 (7−14)	121 (89.0)	93 (68.4)	0		0	0	4 (2.9)
2020	Kiguchi et al.32	167	9.9±4.1	161 (96.4)	133 (79.6)	0		1 (0.6)	0	NA
2020	Kuroki et al.34	157	10.7±7.2	152 (96.8)	141 (89.8)	1 (0.6)		4 (2.5)	4 (2.5)	2 (1.3)
2021	Hirasawa et al.35	59	11.0±7.1	50 (85.8)	42 (71.2)	1 (1.7)		1 (1.7)	0	NA
2022	Cho et al.38	58	12 (4−20)	39 (67.2)	36 (62.1)	0		0	1 (1.7)	0
2022	Kato et al.29	1,324	11.3±7.7	1150 (86.8)	784 (61.2)	10 (0.8)		35 (2.6)	2 (0.2)	NA

Data are presented as median (range), number (%), or mean±SD..

NA, not applicable..

Cap-assisted EMR or EMR with a ligation device has been proposed as an alternative to cEMR, considering the flat morphology of duodenal lesions (Fig. 3).^40,41 Kimoto et al.⁴¹ reported a cap-assisted EMR study on 228 SNADETs, with high en bloc and R0 resection rates (99.6% and 97.4%, respectively), a low adverse event rate (3.1%) without perforation, and no recurrence. This technique would be feasible and effective for SNADETs, particularly for small sizes (≤10 mm).

Figure 3. Endoscopic mucosal resection of a duodenal epithelial tumor using a ligation device. (A) A 6 mm slightly elevated lesion (0-IIa) located in the second portion of the duodenum. (B) A saline solution containing small amounts of epinephrine and indigo carmine dye was injected beneath the lesion. (C) The lesion was aspirated into the ligation device, and then the elastic band was deployed. (D) Snare resection was performed using a high-frequency device. (E) The lesion was completely resected. (F) The resected specimen.

Recently, a prospective study evaluated the efficacy and safety of thermal ablation on the defect margin after EMR (EMR-T) to reduce residual or recurrent adenomas in duodenal lateral spreading adenomas (≥10 mm).⁴² EMR was performed for all visible lesions, followed by ablation of defect margins by snare-tip soft coagulation, aiming to create a 2 to 3 mm rim of the completely ablated denatured tissue. EMR-T significantly reduced the recurrence rate compared with cEMR (2.3% vs 17.6%, p=0.01).⁴² In addition, precutting EMR, which is a technical modification of ESD, has been applied to SNADETs. This procedure involves making a circumferential incision along the margin of the lesion using a dual knife, causing the sufficiently elevated lesion to be fully captured by a snare.⁴³ As a result, precutting EMR was comparable to ESD for SNADETs, demonstrating a lower intraoperative perforation rate and shorter procedure time compared with ESD.⁴³

3. Cold snare endoscopic mucosal resection

cEMR, including submucosal injection and electrocautery, is associated with a risk of delayed bleeding, post-polypectomy syndrome, and perforation owing to thermal injury. Cold snare-EMR (CS-EMR) is an attractive option because it does not require electrocautery, reducing these risks. Furthermore, unlike CSP, which is indicated for SNADETs <10 mm, CS-EMR can be performed for large SNADETs. A recent meta-analysis including 1,137 sessile serrated colorectal lesions >10 mm demonstrated that CS-EMR had a significantly lower rate of delayed bleeding than cEMR.⁴⁴ Therefore, performing the CS-EMR technique for large SNADETs seems reasonable.

Wang et al.⁴⁵ reported that CS-EMR had lower intraprocedural and postprocedural bleeding rates than cEMR in 50 duodenal adenomas >15 mm. However, intraprocedural perforation occurred in two cases of 30- and 60-mm-sized adenomas treated using CS-EMR. Capturing large amounts of mucosa for cold snare excision with failed cutting and amputating the ensnared tissue against the tip of the endoscope may create a shearing force on the relatively fixed but thin duodenal muscle layer, resulting in perforation.⁴⁵ A systematic review and meta-analysis of CS-EMR for non-ampullary duodenal polyps showed a significantly lower rate of delayed bleeding than cEMR.⁴⁶ Other studies also reported a low incidence of adverse events after CS-EMR (Table 3).^45,47-49 Recurrence rates after CS-EMR for large SNADETs are as 12% to 46%.^45,47,48 Furthermore, most residual or recurrent lesions can be removed successfully using endoscopic resection, without surgical referral. A large tumor size is correlated with higher recurrence.^45,47,48 Considering low rates of adverse events, CS-EMR might be another treatment option for large duodenal adenomas in some situations such as high risk for delayed bleeding.

Table 3 . Summary of Cold Snare Endoscopic Mucosal Resection Outcomes for Duodenal Epithelial Tumors.

Year	Author	No. of lesions	Tumor size, mm	Immediate		Delayed		Recurrence
				Perforation		Bleeding	Perforation
2015	Choksi et al.49	15	24 (10−60)	0		1 (6.7)	0	NA
2022	Dang et al.48	39	20 (10−70)	0		0	0	18 (46.2)
2022	Repici et al.47	33	31.5±9.7	0		0	0	4 (12.1)
2023	Wang et al.45	50	30 (19−40)	2 (4.0)		2 (4.0)	0	10 (24.4)

Data are presented as median (range), number (%), or mean±SD..

NA, not applicable.

4. Underwater endoscopic mucosal resection

uEMR is a relatively new EMR technique involving filling the duodenal lumen with water instead of submucosal injection (Fig. 4). Binmoeller et al. first reported uEMR for colorectal polyps in 2012⁵⁰ and for duodenal adenoma in 2013.⁵¹ Duodenal folds (Kerckring folds) do not contain muscularis propria.⁵² The muscularis propria remains flat under water immersion along the long axis and maintains the circular shape during endoscopic ultrasound examination, whereas the mucosal and the submucosal layers tend to float in a water-filled lumen.⁵³ Unlike cEMR using air, water immersion can maintain the thickness of the duodenal wall, reduce thermal injury, and cause superficial lesions to float up, simulating protruded lesions (buoyancy effect). Therefore, the lesion can be easily snared and removed, with a low risk of muscularis propria entrapment by snaring.⁵³

Figure 4. Underwater endoscopic mucosal resection of a duodenal epithelial tumor. (A) A 10 mm white, slightly elevated lesion (0-IIa) located in the second portion of the duodenum. (B) After complete air deflation, the lumen was filled with water. (C) After snaring the entire lesion along with the surrounding mucosa, resection was performed with a high-frequency device. (D) The lesion was completely resected. (E) The defect was completely closed by clipping. (F) The resected specimen.

Omitting submucosal injection causes uEMR to be a low-labor procedure and shortens the procedure time compared to cEMR.^54-56 Prophylactic suturing after uEMR is easier than that after cEMR because post-uEMR ulcers are small, and the surrounding mucosa is soft without submucosal injection.⁵⁷ Furthermore, uEMR reduces the influence of fibrosis on the submucosa by eliminating injection, enabling the resection of lesions with scars. Even a small forceps biopsy can cause severe submucosal fibrosis in the duodenum, causing difficulty in subsequent endoscopic resection and necessitating conversion to more invasive methods in some cases.¹³ Kiguchi et al.³² reported that uEMR was less frequently converted to ESD compared to cEMR. Several studies reported that uEMR showed favorable efficacy and safety (Table 4).^32,35,57-60 However, uEMR achieved decreased en bloc resection and increased recurrence rates for lesions >20 mm, similar to cEMR.^29,61

Table 4 . Summary of Underwater Endoscopic Mucosal Resection Outcomes for Duodenal Epithelial Tumors.

Year	Author	No. of lesions	Tumor size, mm	En bloc resection	R0 resection	Immediate		Delayed		Recurrence
						Perforation		Bleeding	Perforation
2018	Yamasaki et al.57	31	12.0±7.3	27 (87.1)	19 (61.3)	0		0	0	1 (3.2)
2020	Kiguchi et al.32	90	10.3±4.1	78 (86.7)	60 (66.7)	0		2 (2.2)	0	NA
2020	Iwagami et al.61	162	10 (2−40)	110 (67.9)	NA	0		2 (1.2)	1 (0.6)	7 (4.5)
2021	Hirasawa et al.35	67	8.1±3.7	62 (92.5)	51 (76.1)	0		3 (4.5)	0	NA
2021	Furukawa et al.59	28	8 (2−20)	27 (96.4)	20 (71.4)	0		0	0	0
2022	Kato et al.29	579	11.0±7.5	455 (78.6)	316 (56.0)	3 (0.5)		12 (2.1)	1 (0.2)	NA
2022	Yamasaki et al.58	166	9.8±4.7	149 (89.8)	111 (66.9)	0		2 (1.2)	0	4 (2.7)
2023	Tanaka et al.60	96	8.1±4.3	90 (93.8)	65 (67.7)	0		1 (1)	0	NA
2023	Hashiguchi et al.63	25 (uEMR). 22 (uEMR-SIM).	8.4±3.1. 7.2±4.4.	22 (88.0). 22 (100).	12 (48.0). 20 (90.0).	0. 0.		0. 0.	0. 0.	0. 0.

Data are presented as mean±SD, number (%), or median (range)..

NA, not applicable; uEMR, underwater endoscopic mucosal resection; uEMR-SIM, uEMR with submucosal injection and marking..

Conversely, continuously maintaining water in the duodenal lumen is difficult. Large-volume water injections are associated with the risk of aspiration pneumonia.⁵⁷ Furthermore, endoscopic visibility may be reduced owing to intestinal peristalsis, bile, and blood. The use of sterile water heated to 37°C and antiperistalsis agents is helpful to overcome this drawback.⁵³ Recently, gel immersion endoscopic resection has been reported as an alternative method to uEMR.⁶² Continuously maintaining gelatinous liquid in the duodenal lumen, without additional infusion, is easier than maintaining water; therefore, gel immersion endoscopic resection demonstrates a significantly shorter procedure time than uEMR.⁶²

Another disadvantage of uEMR is that the floating center can hinder the visualization of the anal side of the lesions, increasing the risk of positive horizontal margins. Recently, a partial submucosal injection technique combining uEMR for SNADETs has been reported; the difficult side (typically the anal side) of the lesion is locally injected to recognize sufficient margins before resection.⁶⁴ A study of partial submucosal injection technique combining uEMR for 30 SNADETs reported en bloc resection and R0 resection rates of 97% and 83%, respectively.⁶⁴ A technique using uEMR with submucosal injection and marking has also been introduced; markings around the lesion and submucosal injection to expand, rather than lift, the lesion were performed before water immersion.⁶³ This resection method achieved a significantly higher en bloc resection rate than cEMR and a significantly higher R0 resection rate than uEMR.

5. Endoscopic submucosal dissection

ESD has been introduced for SNADETs >20 mm, with en bloc resection rates >90% (Table 5).^29-31,35,65 ESD is also considered for lesions that cannot be removed using cEMR, such as non-lifting lesions or those for which en bloc resection is required (Fig. 5).⁶⁶ ESD has the advantage of a high en blocresection rate (>90%), even for lesions >20 mm.^30,65,67 Many previous studies have shown that duodenal ESD can reach a high R0 resection rate of 83.7% to 96.0%^29-31,65 and a low recurrence rate of ~1.0%.^29,31 However, ESD has not been widely accepted as a standard treatment owing to the high risk of adverse events: bleeding rates are up to 14.3%, perforation rates are 1.7% to 28.6%, and emergency surgery rates are up to 2.5% (Table 5).^29-31 Moreover, ESD outcomes vary by country and institution. Most duodenal ESD studies have been published in Asia (Japan and Korea), and data from Western countries are limited. The largest series from Europe showed relatively low en bloc and R0 resection rates and a high recurrence rate (29.7%, 19.4%, and 14.7%, respectively) compared with previous studies from Asian expert centers.⁶⁸ In addition, there is no differences in long-term outcomes and survival between cEMR and ESD.^31,68-70

Figure 5. Endoscopic submucosal dissection of a duodenal epithelial tumor. (A) A 30 mm, slightly elevated lesion (0-IIa) located in the second portion of the duodenum. (B) Circumferential marking was performed, and a saline solution containing small amounts of epinephrine and indigo carmine dye was injected beneath the lesion. (C, D) Mucosal incision and submucosal dissection were performed. (E) The lesion was completely resected. (F, G) The resected area was completely closed using an endo loop and clips. (H) The resected specimen.

Table 5 . Summary of Endoscopic Submucosal Dissection Outcomes for Duodenal Epithelial Tumors.

Year	Author	No. of lesions	Tumor size (mm)	En bloc resection	R0 resection	Immediate		Delayed		Recurrence
						Perforation		Bleeding	perforation
2017	Hoteya et al.31	49 (large). 25 (small).	31.3±12.6. 11.6±4.1.	48 (98.0). 25 (100).	41 (83.7). 24 (96.0).	14 (28.6). 6 (24.0).		7 (14.3). 4 (16.0).	1 (2.0). 0.	0. 0.
2018	Yahagi et al.30	174	27.4±16.1	171 (98.3)	148 (85.1)	24 (13.7)		9 (5.2)	3 (1.7)	NA
2021	Hirasawa et al.35	64	21.4±10.2	63 (98.4)	61 (95.3)	12 (18.6)		3 (4.7)	2 (3.1)	NA
2022	Kato et al.29	1,017	20.6±13.5	964 (94.8)	790 (78.7)	95 (9.3)		48 (4.7)	23 (2.3)	56 (0.4)
2024	Seya et al.65	100	17 (5−76)	100 (100)	93 (93.0)	1 (1.0)		3 (3.0)	4 (4.0)	1 (1.0)

Data are presented as mean±SD, number (%), or median (range)..

NA, not applicable..

Therefore, duodenal endoscopic resection should focus on safety rather than outcomes of en bloc and R0 resection rates. We suggest that ESD should be performed in high-volume centers, and the lesion should be carefully selected. The pocket-creation method of ESD for SNADETs was introduced to overcome difficulties of ESD, including difficult locations and high rates of perforation. En bloc resection rate in the pocket-creation method group was 100%, even for lesions in the duodenal angles, and perforation was less frequent than in the conventional ESD group.⁷¹

The ESGE recommends that the high rate of adverse events, such as immediate or delayed bleeding or perforation with duodenal endoscopic resection, may be reduced by mucosal defect closure techniques, such as endoscopic clipping or over-the-scope clip (OTSC) clipping, and by noncontact hemostatic measures.⁹ Several studies have shown favorable outcomes of prophylactic procedures to prevent delayed bleeding. A retrospective study of 50 patients with duodenal cEMR reported that prophylactic argon plasma coagulation therapy on the resection defect may lower the risk of delayed bleeding.⁷² Another study including 37 duodenal adenomas treated with cEMR reported a significantly lower delayed bleeding rate in patients treated with prophylactic clipping or prophylactic argon plasma coagulation than in the no prophylaxis group.⁷³ Nonaka et al.³⁷ also showed decrease in the delayed bleeding rate from 32 % to 7 % with prophylactic clipping. Therefore, prophylactic procedures should be considered to prevent delayed bleeding.

Furthermore, mucosal defect closure techniques can help prevent delayed perforation. Kato et al.⁷⁴ reported that complete mucosal defect closure after duodenal ESD significantly decreased delayed adverse events. Closure of post-ESD defects is another strategy to enhance the safety of duodenal ESD. Various closure methods have been reported to date. Endoscopic clipping is the most convenient and standard method.⁷⁵ Small defects can be closed using clips (Fig. 4E). A large defect can be closed by combining an endo loop and clips; opening an endo loop along the defect margin, deploying clips on both edges of the defect, and tightening the endo loop (Fig. 5F and G).⁷⁴ String clip suturing is also available for large defects and involves deploying a clip with string at the distal edge of the defect and placing a second one at the opposite side to anchor the string.⁷⁶ In addition, endoscopic suturing is another option for closing large mucosal defects after endoscopic resection.⁷⁷

Recently, an OTSC system was introduced, which enables full-thickness defect closure by its strong holding and grasping forces. Complete defect closure by OTSC prevents delayed perforation,⁷⁸ and additional use of conventional clip after OTSC is useful to reduce the risk of delayed bleeding.⁷⁹ Covering the wound with PGA (polyglycolic acid) sheets and fibrin glue can be used as an alternative to suturing to prevent delayed perforation after duodenal ESD.^80,81 The mucosal defect is covered with several PGA sheets, and fibrin glue (fibrinogen and thrombin) is subsequently applied using each spray tube. This constitutes a simple and easy procedure compared with clip closure and can remain on the defect site for more than a week.

Several endoscopic procedures are available for the treatment of SNADETs; however, treatment strategies for SNADETs have not been standardized owing to their rarity. Accordingly, the final decision is frequently made based on the endoscopist’s personal experience. Adverse events that occur after endoscopic resection are challenging. Therefore, more consensus guidelines are required to establish a standard strategy for SNADETs.

SNADETs ≤10 mm in size are candidates for CSP, cEMR, or uEMR, based on previous reports (Table 6). Among these lesions, suspected carcinoma lesions should not be treated using CSP because of their low curability. cEMR or UEMR can be considered for lesions sized 10 to 20 mm. Piecemeal EMR (including CS-EMR) and ESD are the options for lesions >20 mm. In particular, ESD or surgical resection should be considered for suspected carcinoma lesions >30 mm. The treatment plan should be selected on a case-to-case basis, considering the balance between the risk of adverse events and necessity of en bloc resection.

Table 6 . Recommendation for Endoscopic Resection of Duodenal Epithelial Tumors.

Based on histology and endoscopic findings	Tumor size	Treatment method
Lesions of suspected adenoma	≤10 mm. 10−20 mm. >20 mm.	Cold snare polypectomy. Underwater EMR. Conventional EMR. Underwater EMR. Conventional EMR. Piecemeal EMR. ESD.
Lesions of suspected carcinoma	≤20 mm. 20−30 mm. >30 mm.	Underwater EMR. Conventional EMR. ESD. Piecemeal EMR. ESD. Surgery.

EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection..

G.H.K. is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.