The technique and instruments of endoscopic polypectomy were developed in Japan[10]. Since then, the strip biopsy was described as an extension of endoscopic snare polypectomy in 1984[11]. This technique uses a double channel endoscope. After submucosal injection under the lesion, a snare is inserted through one channel and is used to remove the lesion, while a grasper, inserted through the other channel, is used to lift the lesion. In 1988, Endoscopic mucosal resection(EMR) after circumferential pre-cutting was described[12]. In this technique, cutting around the lesion with a needle knife is done after submucosal injection using hypertonic saline mixed with diluted epinephrine, and then the lesion is removed by a snare. In 1992, EMR using transparent cap (EMR-C) was developed[13]. This technique uses a transparent plastic cap that is connected to the tip of an endoscope. After submucosal injection, the lesion is sucked into the cap while a specialized crescent shaped snare, which is deployed at the tip of the cap, is closed. With this technique intramucosal cancers 2 cm or less in diameter can be safely removed[14]. After that, EMR with ligation (EMR-L) was described[15]. In this technique, a standard variceal ligation device is used to capture the lesion. After sucking the lesion into the cap, lodged band is deployed underneath the lesion. The banded lesion is then resected by a snare. EMR-L is also safe and effective treatment modality for selected EGCs[16]. As described above, EMR-C and EMR-L are relatively simple and effective treatment modality for EGC. However, it is apt to fragment tumors that are larger than 1.5-2.0 cm in diameter, which results in incomplete histological diagnosis and possibly in an increased risk of local recurrence[17-19]. To overcome this drawback ESD, which is particularly effective for en bloc resection of tumors regardless of their size, was developed in late 1990s.

Endoscopic submucosal dissection (ESD) permits en bloc resection of larger lesions than that can be treated with EMR[20-26]. This technique consists of several steps (Figure 1)[27,28]. The marking around the lesion is done and circumferential mucosal pre-cutting is performed after submucosal injection. To distinguish clearly between the muscle layer and the submucosal layer and allow better hemostasis, normal saline mixed with diluted epinephrine and indigo-carmine is often used as submucosal injection solution. The injection is repeated a few times until the target mucosa is sufficiently raised. After lifting of the lesion, submucosal layer under the lesion is dissected with lateral movement using various knives. Several knives have been developed and used for ESD, which include needle knife, insulation-tipped diathermic knife (IT knife), hook knife, flex knife, triangle tip knife, flush knife, splash knife, IT-2 knife and dual knife[6,28-30]. After resection of the lesion, visible vessels in the artificial ulcer is treated with hemostatic devices to prevent delayed bleeding.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Endoscopic submucosal dissection procedure. A: On the lesser curvature side of the antrum, a 1.2-cm type IIa + IIc early gastric cancer is observed; B: Indigo-carmine is sprayed along the extent of tumor to aid visualization. C: Marking outside the lesion; D: After injection of saline mixed with diluted epinephrine (1:100000) and indigo-carmine into the submucosal layer, circumferential mucosal pre-cutting is performed using a knife; E: After dissection of the submucosal layer, an artificial ulcer is observed; F: Fixation of the tissue specimen.

The standard criteria[8] for selection of patients with EGC who are appropriate for the endoscopic treatment are below: (1) well or moderately differentiated adenocarcinoma and/or papillary carcinoma; (2) confined to the mucosa; (3) smaller than 2 cm for superficially elevated type lesions; (4) smaller than 1 cm for the flat and depressed type lesions; (5) without ulcer or ulcer scar; and (6) without venous or lymphatic involvement[34]. The rationale for this guideline is based on the knowledge that patients meeting the criteria are expected free from lymph node (LN) metastasis[35].

Expansion of the criteria for selection of patients with EGC who are appropriate for the endoscopic treatment has been proposed in Japan from clinical observations that the too strict absolute indication leads to unnecessary surgery[36]. From the surgical data involving 5265 patients who underwent gastrectomy for EGC, Gotoda et al[7] were able to further define the risk of LN metastasis in certain groups of patients with EGC and showed four groups with a low risk of LN metastasis: (1) differentiated intramucosal adenocarcinoma without lymphovascular invasion less than 3 cm in diameter, irrespective of ulcer findings; (2) differentiated intramucosal adenocarcinoma without lymphovascular invasion and ulcer findings, irrespective of tumor size; (3) undifferentiated intramucosal cancer without lymphovascular invasion and ulcer findings smaller than 2 cm in diameter; and (4) differentiated adenocarcinoma with minute submucosal penetration (SM1, cancer invasion into the upper third of the submucosa) but without lymphovascular invasion smaller than 3 cm in diameter. These results have allowed the development of expanded criteria for endoscopic treatment for EGC[8] (Table 1). In the study by An et al[37], predictive factors for LN metastasis in EGC with submucosal invasion were identified and possibility of EMR was addressed in highly selected submucosal cancers with no lymphatic involvement, SM1 invasion, and tumor size < 1 cm. In a recent study be Lee et al[38] to compare the therapeutic outcomes of conventional and expanded indications of ESD for differentiated EGC, the conventional indication group and expanded indication group did not differ with regard to the rates of local recurrence (0.7% vs 0%), metachronous recurrence (3.6% vs 3.3%) or cumulative disease-free survival. Survival outcome was similar in the subgroups classified by tumor depth and size.

The risk of LN metastasis is known to increase in undifferentiated cancer due to lymphovascular invasion[39]. In the analysis of 3843 patients who underwent gastrectomy with LN dissection for solitary undifferentiated EGC, none of the 310 intramucosal cancers 20 mm or less in size without lymphovascular invasion and ulcerative findings was associated with LN metastases[40]. Recently, favorable outcomes of endoscopic resection have been reported in selected patients with undifferentiated mucosal cancer or minimal submucosal invasion cancer (SM1)[41-43]. However, these are all single center retrospective studies. Therefore, large scale, prospective studies are warranted to confirm the feasibility of ESD for undifferentiated gastric cancer.

To expand the indication of endoscopic treatment to submucosal invasion (SM1) differentiated EGC, the histological heterogeneity of gastric cancer is the important issue to be addressed. Based on morphological features and histological background, gastric carcinoma is divided into differentiated and undifferentiated type or intestinal and diffuse type[44,45]. Gastric cancer shows remarkable heterogeneity in histological pattern, cellular phenotype, and genotype[46]. In a retrospective study to compare the clinicopathologic features of node-positive and node-negative differentiated submucosal invasion differentiated gastric cancers, histological heterogeneity was the independent risk factor for LN metastasis[47]. Thus, it is recommended to apply the endoscopic treatment to the differentiated EGC without histological heterogeneity when it is considered in submucosal invasion cancer.

Bleeding is the most common major complication of endoscopic treatment for EGC. Most bleeding occurs during the procedure or within 24 h[48]. Bleeding is divided into immediate (intraoperative) bleeding during procedure and delayed bleeding after procedure. Significant immediate bleeding occurs more often in the upper and middle thirds of the stomach than in the lower third of the stomach because of the larger diameter of the submucosal arteries in the upper and middle thirds of the stomach[49]. However, bleeding can be successfully treated in most cases through coagulation of the bleeding vessels, or placement of metallic clips for severe bleeding. In terms of delayed bleeding, the incidence rates were reported to range from 0%-15.6% in the recent review involving 28 studies with at least 300 ESD cases for EGC[50]. Delayed bleeding is associated to tumor location, larger tumor, recurrent lesion, macroscopic type (flat or depressed), old age (≥ 80 years) and longer procedure time[51-55]. At first, delayed bleeding was reported to occur more frequently after ESD for lesions in the lower and middle thirds of the stomach compared to the upper third of the stomach[55]. However, the reason remains unclear. In the recent study involving 1000 cases of ESD for early gastric neoplasms, delayed bleeding occurs more often in upper portion of the stomach than in lower portion (28.6% vs 13.8%, P = 0.003)[51]. In relation to antiplatelet drugs, there is a controversy in the risk of bleeding after ES. Two Korean retrospective studies have reported conflicting results on the risk of bleeding after ESD for gastric neoplasms[56,57].

Perforation is less common major complication of endoscopic resection for EGC than bleeding and has been reported to range from 1.2% to 5.2%[50]. Perforation is diagnosed when mesenteric fat or intra-abdominal space is directly observed during the procedure (frank perforation) or free air is found on a plain chest X-ray after the procedure without a visible stomach wall defect during the procedure (micro-perforation). Immediately recognized small perforations can be successfully treated without surgery with a combination of endoscopic clipping and broad spectrum antibiotics[58-60]. However, large perforations would require immediate surgery. In cases of micro-perforation, management is not well established. In a retrospective study by Jeong et al[61], 13 cases (3.18%) of micro-perforation after EMR for gastric neoplasms were reported. Among them, 11 cases were successfully treated only with fasting, nasogastric tube drainage and broad spectrum antibiotics. In severe pneumoperitoneum, respiratory deterioration and/or shock could occur. Decompression of the pneumoperitoneum must be performed with a puncture needle in such cases[60]. Instead of air insufflations, CO₂ insufflation during procedure could minimize such pneumoperitoneum caused by a perforation[62,63].

Stenosis after gastric ESD has been reported to range from 0.7% to 1.9%[50]. In a retrospective study, stenosis occurred with 17% of cardiac resections and 7% of pyloric resections[64]. Circumferential extent of the mucosal defect of > 3/4 and longitudinal extent > 5 cm were related to stenosis with both cardiac and pyloric resections. However, all affected patients (n = 15) were successfully treated by endoscopic balloon dilation.

Aspiration pneumonia after gastric ESD has been reported to range from 0.8% to 1.6%[50]. However, the risk of aspiration pneumonia appears to increase in sedation with continuous propofol infusion with intermittent or continuous administration of an opiod[65,66]. In addition, longer procedure time (> 2 h), male gender and old age (> 75 years) are associated with occurrence of aspiration pneumonia after ESD.

Pain after endoscopic resection is usually mild and dull in nature and can be controlled by proton pump inhibitor and opioids[36].