Endoscopic submucosal dissection for early gastrointestinal malignancies: Current state and future perspectives

Abstract

Endoscopic submucosal dissection (ESD) has emerged as a pivotal therapeutic modality for early gastrointestinal (GI) cancers, providing a minimally invasive approach with curative potential. This technique enables the en bloc resection of neoplastic lesions confined to the mucosa and submucosa, thereby preserving organ function and reducing the need for more radical surgical interventions. ESD provides diagnostic clarity and enhances patient survival rates when performed by skilled practitioners in the early stages of GI cancers such as esophageal, gastric, and colorectal carcinomas. This article examines the indications, procedural advancements, technical considerations, and outcomes associated with ESD in early GI cancers. The challenges and complications that can arise are also highlighted. Additionally, we discuss the evolving role of novel techniques and adjunctive therapies to improve safety and efficacy. As the field progresses, ESD remains a cornerstone in managing early GI cancers, offering patients a promising option for organ preservation and long-term survival.

Key Words: Chromoendoscopy; Narrow band imaging; Hyperplastic; Submucosal; Polypoid; En bloc resection; Early gastrointestinal cancer; Endoscopic mucosal resection; Endoscopic submucosal dissection

Core Tip: This article provides a comprehensive overview of endoscopic submucosal dissection for early esophageal, gastric, and colon cancers, detailing technical approaches, common pitfalls, and current evidence. This review also discusses the management of common procedural complications. Gastroenterologists can use this guide to improve outcomes and optimize the application of endoscopic submucosal dissection in managing early gastrointestinal cancer.

INTRODUCTION

Endoscopic submucosal dissection (ESD) and endoscopic mucosal resection (EMR) are the primary endoscopic approaches and have replaced surgery as the standard therapy for early gastrointestinal (GI) cancers. Studies have documented the efficacy and safety of ESD and EMR. However, even with appropriate techniques, en bloc resection in EMR is limited by the snare diameter, necessitating piecemeal removal for larger lesions[1]. This piecemeal approach hinders the accurate assessment of an R0 resection, limiting the use of EMR for treating early-stage malignancies. In addition, some lesions, particularly those with submucosal invasion or scarring from prior procedures, will not lift with submucosal injection, hindering the applicability of EMR[2]. Conversely, ESD involves electrosurgical knives rather than snares for dissecting the tumor at the muscular level. This approach enables the treatment of larger tumors and those with deep submucosal invasion, which are limitations for snare-based methods. At the same time, ESD is a complex and time-consuming procedure, which is only available in tertiary or quaternary referral centers and requires extensive training.

DEVELOPMENT OF ESD

ESD developed as an advanced form of EMR that involves injecting hypertonic saline mixed with epinephrine into the submucosa to aid in endoscopic resection[3]. Eponyms such as exfoliating EMR, cutting EMR, and EMR with cutting incision were initially used to describe specific EMR techniques. ESD was coined in 2003 as a new treatment modality for removing GI lesions, offering an option between EMR and surgery.

Specialized electrosurgical tools, such as the IT-knife, triangle-tipped knife, and a transparent cap, have been developed to facilitate ESD[4]. ESD comprises three steps: Elevation of lesions from the underlying muscle layer by injecting fluid at the submucosal plane, circumferential cutting of the surrounding mucosa, and subsequent dissection of the connective tissue below the lesion (Figure 1)[5]. It enables complete resection of large lesions as the shape and size of resection can be easily controlled. It also involves submucosal dissection, and is applicable to ulcerative non-lifting neoplasms. These are the salient advantages of ESD over EMR. However, ESD is a time-consuming procedure associated with a higher risk of bleeding and perforation-related complications, and requires the involvement of two or more assistants, unlike EMR. Despite these limitations, ESD has emerged as the standard treatment for early gastric cancer (GC) and esophageal cancer. However, technical challenges hinder its effectiveness for colorectal cancer.

Figure 1 Steps of endoscopic submucosal dissection. A: Endoscopic view of a rectal sessile polypoidal lesion; B: Narrow band imaging showing a Japanese Narrow Band Imaging Expert Team type 2B lesion; C: Marking around the lesion with a dual-knife; D: Submucosal injection with methylene blue; E: Submucosal dissection with a dual-knife; F: Base of the lesion after resection and one clip applied over a visible vessel to achieve hemostasis.

Endoscopic evaluation before ESD

Accurate identification of surface and vascular patterns is essential for distinguishing between adenomatous polyps and carcinoma in situ. There are multiple techniques for detecting lesions with high-grade dysplasia based on endoscopic morphology. The Paris classification was the first system developed for endoscopically classifying neoplastic lesions and remains the most widely used (Table 1). Although initially developed for lesions in the colon, its application has since extended to gastric lesions. The Paris classification of lesions has been shown to correlate with the presence of submucosal invasion (Table 2)[6]. In a study of 479 patients, 0-Is lesions showed a low malignant potential (7.5%), whereas 0-IIc or IIa + IIc lesions had a significantly higher malignant potential (31.8%)[7].

Table 1 Morphological classification of polyps.

Endoscopic appearance	Paris class	Description
Protruded lesions	Ip	Pedunculated polyps
	Isp	Semi-pedunculated polyps
	Is	Sessile
Flat elevated lesions	0-IIa	Flat elevation of mucosa
	0-IIa/c	Flat elevation with central depression
Flat lesions	0-IIb	Flat mucosal change
	0-IIc	Mucosal depression ≤ 1.2 mm
	0-IIc/IIa	Mucosal depression with raised edge
Excavated lesions	0-III	Mucosal depressions > 1.2 mm

Table 2 Submucosal invasion in different Paris classification lesions.

Paris classification	Submucosal invasion, %
0-Ip	5
0-Is	34
0-IIa	4
0-IIb	0
0-IIc	61

Narrow band imaging (NBI) is another common endoscopic technique for characterizing colonic lesions. Two NBI classifications, the NBI International Colorectal Endoscopic classification and the Japanese NBI Expert Team (JNET) classification, are currently used to characterize these lesions (Table 3[8] and Table 4). The JNET classification comprises four categories, namely, 1, 2A, 2B, and 3[9]. Distinguishing between type 2B and type 3 Lesions poses a challenge even for experienced endoscopists. As per the existing literature, the JNET classification has a sensitivity of 44.9% to 53.8% for correctly identifying type 2B lesions[10,11].

Table 3 Narrow band imaging International Colorectal Endoscopic classification.

Endoscopic appearance	NICE type
Lesions with a color similar to the surrounding mucosa may have a net-like vascular pattern or lack a specific vascular pattern	1
Lesions with a brownish hue with brown vessels surrounding white structures	2
Lesions with a dark-to-brown appearance, disrupted or missing vessels, and an irregular surface pattern	3

NICE: Narrow band imaging International Colorectal Endoscopic.

Table 4 Japanese Narrow Band Imaging Expert Team classification.

JNET type	Vessel pattern	Surface pattern	Most likely histology
1	Invisible	Regular dark or white spots, similar to surrounding normal mucosa	Hyperplastic polyp/sessile serrated polyp
2A	Regular caliber and regular distribution	Regular	Low grade intramucosal neoplasia
		Tubular/branched/papillary
2B	Variable caliber, irregular distribution	Irregular or obscure	High grade intramucosal neoplasia
3	Loose vessel areas, interruption of thick vessels	Amorphous areas	Deep submucosal invasive cancer

JNET: Japanese Narrow Band Imaging Expert Team.

ESOPHAGEAL CANCER

ESD is an excellent therapeutic option for early-stage esophageal cancer, achieving a high rate of en bloc resection and relatively low local recurrence rate. The outcomes of ESD are significantly better those of EMR. NBI and chromoendoscopy facilitate the detection of early esophageal malignancies. NBI helps detect lesions by highlighting the pattern of small, dot-like vessels and the lesion’s background color compared to the surrounding healthy tissue[12]. The pink color sign, which appears 2-3 minutes after iodine staining, can distinguish between malignant and nonmalignant lesions with a sensitivity and specificity of 88% and 95%, respectively[13].

ESD is a treatment option for esophageal cancers limited to the mucosa’s lamina propria (whether noncircumferential or < 50 μm circumferential), and non-circumferential lesions invading submucosa up to < 200 μm deep[14]. The anatomical characteristics of the esophagus, such as its narrowness, the surrounding structures vertebrae, aorta, and trachea) that can compress it, and its thin muscle layer, make performing a safe ESD quite challenging[15].

Outcomes of esophageal ESD

In 2005, Oyama et al[16] reported a 95% en bloc resection rate with no perforation or local recurrence and only a 6% incidence of mediastinal emphysema. In a multicenter retrospective cohort study in Japan, ESD led to an en bloc resection rate of 96.7% [95% confidence interval (CI): 94.4%-98.1%] and R0 resection rate of 84.5% (95%CI: 80.5%-87.8%). Perforation, bleeding, esophageal stricture, and postoperative pneumonia occurred in 5.2% (95%CI: 3.3%-7.9%), 0%, 7.1% (95%CI: 4.9%-10.2%), and 1.6% (95%CI: 0.7%-3.5%) of patients, respectively[17].

In a multicenter study, en bloc resection and complete resection rates for esophageal squamous cell carcinoma were 100% and 69.8%, respectively. Postprocedural bleeding and perforation occurred in 4.8% and 1.6% of cases, respectively, whereas stenosis occurred in 23.8% of patients[18]. In another study, ESD achieved a 96.7% R0 resection rate and caused stenosis in 11.5% of patients, with no reports of bleeding or perforation[19].

Complications of esophageal ESD

Perforation: Mediastinal emphysema or mediastinitis can occur from perforation during esophageal ESD, and in severe cases, may rapidly compromise the respiratory and circulatory systems. Intraprocedural perforation during ESD has been reported in about 7% of cases[16,20]. These were successfully managed with clip closure, polyglycolic acid sheets, and fibrin glue. Delayed perforation after esophageal ESD is rare. Five cases of delayed perforation have been reported after esophageal ESD, with a median onset time of 6 days. In these cases, the mucosal defects exceeded half of the esophageal circumference and required invasive treatment including sub-total esophagectomy, empyema drainage, or temporary stent placement[21].

Mediastinal emphysema: Instances of mediastinal emphysema can occur even without perforation and may lead to subcutaneous emphysema in severe cases. Early detection of mediastinal emphysema is possible using imaging techniques, such as X-rays and computed tomography scans within 1 hour of a procedure, with 1.7% and 31% of cases identified, respectively[22]. The majority of these occurrences are minor and do not affect the clinical course. Therefore, routine computed tomography thorax scans are not recommended to detect mediastinal emphysema following uneventful procedures. Additionally, esophageal ESD should always be performed with carbon dioxide, which is quickly absorbed by the body’s tissue, to minimize the risk of mediastinal emphysema[23].

Stricture: Esophageal strictures have been reported in more than 50% of post-ESD defects, especially those involving more than 75% or the entire circumference[24]. A variety of prophylactic approaches for stricture exists, including local steroid injections, oral steroids, temporary stent placement, polyethylene glycolic acid sheets, and cultured oral mucosal epithelial cell sheets[25-27].

Technique: The esophagus has a narrow lumen, making gravity traction difficult. Moreover, ongoing dissection causes the lesion to shift distally, hindering traction and adequate visualization. Thus, improvements in techniques are essential for safe and successful esophageal ESD.

Clip-with-line method: Oyama et al[16] devised the clip-in-line method. Since then, it has been widely used as a simple and effective traction method. In a randomized controlled trial, the median procedure duration for traction-assisted ESD was significantly shorter (45 minutes) compared to conventional ESD (61 minutes) (P < 0.001). Also, traction-assisted ESD was relatively safer with no perforations reported, whereas traditional ESD had a perforation rate of 4.3%[28]. Due to its minimal cost, this method is recommended in Japanese guidelines for the endoscopic treatment of early esophageal cancer.

Tunnel method: For lesions involving more than two-thirds of the circumference, performing conventional ESD is quite difficult and associated with adverse events. The tunnel method offers a more efficient dissection process for large lesions compared to traditional methods, leading to a higher R0 resection rate[29]. In a meta-analysis, the tunnel method was associated with a notable decrease in the amount of time required for dissection during surgery (standardized mean difference = 1.52; 95%CI: 1.09-0.83; P < 0.001), significantly higher R0 rate [odds ratio (OR) = 2.29; 95%CI: 1.54-3.46; P < 0.001] and en bloc resection rate (OR = 3.98; 95%CI: 1.74-9.12; P = 0.001). At the same time, this technique also led to fewer complications such as post-procedure bleeding (OR = 0.38; 95%CI: 0.18-0.83; P = 0.02) and injuries to deeper tissue layers (OR = 0.44; 95%CI: 0.28-0.70; P < 0.001)[30].

Underwater ESD: This technique utilizes a water delivery function to fill the lumen with water via a water jet system. It maintains a good field of view and a wider gap in the mucosal incision space, despite the presence of submucosal fibrosis to some extent. Endotracheal intubation or an overtube is recommended to prevent aspiration[31].

GCA

Early GC (EGC) is a type of GC where the tumor invasion is limited to the mucosa or submucosa (classified as T1), regardless of lymph node involvement. The Japanese EGC Association initially proposed ESD as the treatment for nonulcerated well-differentiated EGCs that are confined to the mucosa (T1a) and are smaller than 20 mm, because these lesions have a low risk of spreading to the lymph node[32]. Subsequently, ESD was considered for: (1) Nonulcerated well-differentiated EGCs of any size; (2) Well-differentiated EGCs < 30 mm with superficial submucosal invasion (submucosal first layer; < 500 mm below the muscularis mucosae); or (3) Ulcerated well-differentiated EGCs < 30 mm[33].

Outcomes after gastric ESD

A Korean multicenter study reported en bloc resection and complete resection rates of 95.3% and 87.7%, respectively. Incidences of delayed bleeding, significant bleeding, and perforation were 15.6%, 0.6%, and 1.2%, respectively[34]. A Japanese study assessed the outcomes of ESD for patients with EGC with expanded indications, including nonulcerated lesions of any size and ulcerated lesions ≤ 3 cm. The en bloc resection and curative resection rates were 94.9% and 94.7%, respectively. After follow-up for a median period of 30 months (6 months to 89 months), the 5-year overall rate was 94.7% and the disease-specific survival rate was 100%[35].

Adverse events

Pain, bleeding, and perforation are some of the common complications associated with ESD[36]. Pain following endoscopic resection is usually mild and well-controlled by antisecretory medications (i.e. proton pump inhibitors and opioids). Bleeding, the most common complication, can be immediate or delayed. Immediate bleeding tends to occur for tumors located in the upper third of the stomach, because many large vessels are located in the vicinity. The rate of immediate bleed ranges from 1.8% to 15.6% according to Korean studies[34,37]. Delayed bleeding is defined as hematemesis or melena within 30 days of procedure and is more likely to occur with lesions in the proximal stomach, those > 40 mm, recurring lesions, and macroscopic flat type lesions[38]. An interesting case report described an elderly woman with a large gastric polypoidal lesion. She underwent pre-procedural superselective embolization of its blood supply, followed by ESD 1 day later. The procedure was performed successfully with no bleeding (Figure 2), and histopathology of the en bloc-resected specimen confirmed intramucosal carcinoma[39], illustrating a novel prophylactic approach for highly vascular lesions.

Figure 2 Novel prophylactic approach for highly vascular lesions. A: Sessile polypoidal lesion in the stomach; B: Endoscopic ultrasound showing increased vascularity; C: Pre-procedural super selective microcatheter angiogram (arrows indicate tumor blush) followed by coil embolization of the left gastric artery; D: Post-endoscopic submucosal dissection achieving a clear base.

Minor complications include the development of subsequent strictures and aspiration pneumonia. Strictures are prone to developing at the cardia and pylorus following the resection of lesions that are ≥ 1 cm at the gastroesophageal junction and pylorus, 75% or more circumferential resection, and greater than 5 cm longitudinal dissection. Strictures can be treated by balloon dilatation[40]. Elderly patients are at risk of aspiration pneumonia after ESD. This risk can be minimized by frequently removing gastric fluid and avoiding overdistension of the stomach during the ESD.

COLORECTAL CANCER

Endoscopic screening for adenoma has achieved remarkable success in detecting early-stage colorectal cancer[41]. However, the efficacy of screening depends not only on the adenoma detection rate but also on whether the endoscopic resection is complete. As such, advanced adenomas occur even after polypectomy[42], accounting for 27% of interval cancers at previous polypectomy sites[43]. Moreover, complete polyp resection accounts for up to 19% of interval cancers, which is especially challenging for sessile or flat lesions more than 20 mm. Recurrence rates of about 30% have been reported after piecemeal endoscopic resection[44]. Unlike EMR, ESD is associated with higher en bloc resection and lower recurrence, even for larger, flat, or sessile lesions. ESD is a demanding procedure with several shortcomings, including the need for technical expertise, longer procedure times, and a higher perforation rate.

The 2019 American Gastroenterology Association Clinical Practice Update recommends ESD for colorectal lesions exhibiting: Kudo V-type pit patterns, depressed morphology (Paris 0-IIc), complex architecture (0-Is or 0-IIa + Is), rectosigmoid location, nongranular laterally spreading adenomas up to 20 mm, granular laterally spreading adenomas measuring 30 mm or more, or residual/recurrent colorectal adenomas[45].

A meta-analysis by Fujiya et al[46] showed that ESD was associated with significantly higher odds of achieving curative resection (OR = 4.26) and a lower likelihood of recurrence (OR = 0.08) compared to EMR. The risk of delayed bleeding was similar for both ESD and EMR procedures, with an OR of 0.85; however, the rate of perforation was higher with ESD, showing an OR of 4.96 compared to EMR. In a recent meta-analysis by Zhao et al[47], the en bloc resection rates for ESD and EMR were 95% and 93.2%, respectively. Complete resection rates were much better for ESD than EMR (i.e. 71.9% and 42.8%, respectively), whereas the bleeding rates were similar (4.2% vs 3.5%). Perforation rates were 1.8% and 2.4% for EMR and ESD, respectively. EMR had a significantly higher recurrence rate than ESD (15.9% vs 0.5%)[47]. A retrospective case-controlled study examined 373 colorectal tumors measuring ≥ 20 mm; of these tumors, 145 were treated with ESD and the remaining 228 were treated with EMR. The recurrence rates were 2% and 14% for ESD and EMR, respectively. Perforations occurred in 6.2% of patients undergoing ESD and 1.3% of patients undergoing EMR. Delayed bleeding was reported in 1.4% of patients undergoing ESD and 3.1% of those undergoing EMR[48].

Adverse events

Bleeding and perforation are the two most frequent complications associated with colonic ESD. The risk of intraprocedural bleeding may be minimized by adequately injecting into the submucosal plane, employing meticulous dissection techniques, and proactively coagulating visible vessels. Immediate or delayed bleeding is usually managed with endoscopic hemostasis, and in a few instances, with embolization. Patients with larger lesions (> 40 mm) or proximal colon lesions, those on dual antiplatelet or anticoagulant medications, or those on hemodialysis are at increased risk of post-procedure bleeding. The risk of perforation is higher with associated submucosal fibrosis (OR = 2.9, 95%CI: 1.83-4.59), tumors in the ascending colon and cecum (OR = 2.35, 95%CI: 1.58-3.50), and larger sized tumors (OR = 2.1.7, 95%CI: 1.47-3.21)[49]. In a study by Fujihara et al[50], prophylactic closure of the defect using through-the-scope or over-the-scope clips was performed in 27 of 68 patients who underwent ESD for colorectal tumors. The closure group reported lower incidences of abdominal pain without delayed bleeding or perforation, whereas the nonclosure group had two instances of delayed bleeding and one of perforation.

Pre-procedure evaluation

Esophageal tumors: Performing endoscopic ultrasound (EUS) before ESD is controversial. The overall accuracy for T and N staging by EUS is 90%[51]. A study by Lee et al[52] found EUS to be 86.7% accurate for T1 Lesion determination and 83.3% accurate for lymph node staging. In a study by Pech et al[53] involving 179 patients, EUS accurately identified T stage in 74% of patients, with a sensitivity and specificity of 82% and 91%, respectively. According to Puli et al[54], EUS had 81.6% sensitivity and 99.4% specificity for staging T1 tumors, 81.4% and 96.3% for T2 tumors, 91.4% and 94.4% for T3 tumors, and 92.4% and 97.4% for T4 tumors, respectively, suggesting higher accuracy in staging advanced cancer than early cancer. Initial studies highlighted the crucial role of EUS in selecting suitable candidates for endoscopic resection; however, subsequent research revealed the shortcomings of EUS. Bergeron et al[55] demonstrated that EUS correctly staged tumor depth in 39% of T1a and 51% of T1b tumors. Amid these conflicting reports, the European Society for GI Endoscopy conditionally recommends that EUS be considered for superficial esophageal carcinomas with suspicion of submucosal invasion or lymph node metastases[56].

Gastric tumors

Conventional endoscopy can accurately evaluate the depth of invasion (mucosa vs submucosa) in early EGC in 72%-84% of cases[57]. There have been inconsistent reports comparing the accuracy of EUS with conventional endoscopy. In a study by Mouri et al[58], the accuracy of EUS in assessing the depth of invasion was 99% for mucosal lesions and 87% for lesions in the mucosa-submucosa junction and extending to the proximal submucosa (submucosal first layer). Lee et al[59] retrospectively studied 393 patients with well-differentiated EGC and concluded that the proportion of patients receiving appropriate treatment was similar between the endoscopy- and EUS-based groups (75.3% vs 71.5%; P = 0.184).

EUS can detect submucosal fibrosis, a condition that often hinders successful ESD. Hirasawa et al[60] analyzed 26 cases of incomplete resection due to severe scarring. The authors considered surgery preferable to ESD for EGC that penetrated > 5 mm into the submucosa. In a study by Kikuchi et al[61], 110 patients with EGC were divided into two groups based on EUS findings: Group P, with few blood vessels in the submucosa or < 4 small vessels per field of view; and group R comprising the remaining patients. Although EUS did not predict the incidence of perforation or exacerbation of anemia, the R group had a significantly longer mean procedure time (105.4 minutes vs 65.5 minutes; P < 0.001) and significantly increased incidence of deeper layer injury and hemoclip usage. Thus, in the absence of a unanimous consensus, the endoscopist can perform a diagnostic EUS before ESD.

Colorectal cancer

EUS has shown varying degrees of success in accurately staging early colorectal cancer, with studies demonstrating a sensitivity range of 57%-91% over the years[62]. One meta-analysis in 2009 revealed that EUS demonstrated a pooled sensitivity and specificity of 87.8% and 98.3% respectively, for detecting T1 rectal lesions[63].

Urban et al[64] concluded that high-frequency ultrasound probes guided the treatment of early colorectal cancers depending upon the endoscopic morphology. Lesions with low-risk morphology could be treated based on their endoscopic appearance. High-frequency probe ultrasonography changed the treatment approach in about 42% of high-risk lesions, specifically those that were depressed or had an invasive pit pattern. However, this technique sometimes led to unnecessary surgery for lesions that could have been handled endoscopically. Thus, a pre-procedure EUS should only be performed when there is suspicion of submucosal fibrosis. According to a study by Makino et al[65], interruption of the third layer by more than 5 mm predicts incomplete ESD, suggesting that surgery is a more appropriate treatment option. Also, EUS is an unreliable tool for nodal staging[51]. Thus, for rectal malignancies, magnetic resonance imaging of the pelvis is routinely conducted before endoscopic resection to rule out lymph node involvement that might necessitate neoadjuvant chemotherapy and surgical resection.

Real-world perspective

The main advantages of ESD compared to surgery are that it is less invasive, less expensive, and preserves physiological function; however, complete resection is not possible in certain complex cases. Thus, the complete resection rate is not 100%, even in high-volume centers. Subsequent surgery is inevitable in these cases, so a systematic approach is needed to carefully select patients with early GI cancers who can benefit from an endoscopic approach. Additional surgery, including lymph node dissection, is recommended if endoscopic resection of the tumor is not found to be curative after microscopic examination. An ESD is considered noncurative if cancer cells are found at the lateral or deep margins, if the cancer has invaded the deep layer, if lymphatic vessels are invaded, if the cell type is undifferentiated, or if any combination of these factors is present (Figure 3 and Table 5)[66].

Figure 3 Algorithm for the management of early gastrointestinal cancer. ¹Bleeding, perforation; ²Vide Table 1. GI: Gastrointestinal; ESD: Endoscopic submucosal dissection.

Table 5 Factors for which additional treatment after endoscopic submucosal dissection should be recommended.

Esophagus	Stomach	Colon and rectum
Tumors in contact with or invading the muscularis mucosa	Positive lateral margins	Positive vertical margins at the site of submucosal invasion
Tumors invading the submucosal layer	Deep submucosal invasion, regardless of positive vertical margins (> 500 μm)	Vascular or lymphatic invasion
	Vascular or lymphatic invasion	Depth of submucosal invasion greater than 1000 μm
	Diffuse type histology	Vascular or lymphatic invasion

CONCLUSION

For endoscopists with sufficient experience, ESD is technically feasible and safe for the resection of early GI cancers. Given the excellent technical and oncologic outcomes, ESD should be considered for lesions meeting the absolute criteria for endoscopic resection. Additionally, ESD may be a decent treatment modality for lesions, fulfilling the expanded number of candidates unfit for extensive surgery. However, most of the publications on ESD outcomes and safety are from Eastern countries, especially Japan. As ESD is an established endoscopic resection in Western countries, additional research will be valuable to explore differences in tumor behavior among patients and risk factors for lymph nodal metastases specific to Western patients.