Therapeutic endoscopy plays a major role in the management of early gastric and esophageal cancers. Direct endoscopic treatment of early cancers or dysplasia can be broadly categorized into ablation therapy and endoscopic resection. In this review, the specific focus is on endoscopic resection of early gastric and esophageal cancer. Endoscopic resection is currently the standard treatment for these diseases in high-volume regions and it is increasingly gaining worldwide acceptance. Endoscopic resection is superior to biopsy for diagnosis of superficial GI tumors and can achieve similar results to surgery. Although ablation therapy may also cure early cancer by obliterating it, it cannot provide a pathological staging of the cancer, which is critical as this allows stratification and refinement of further treatment. However, endoscopic resection requires extensive training and may be associated with increased complications. Both the AG and ASGE have released statements supporting endoscopic resection of early esophageal and gastric cancers. The major advantage of endoscopic resection is its ability to provide pathological staging without precluding future surgical therapy. According to the WHO classification of GI tumors, early gastric and esophageal cancer is defined as cancer in which tumor invasion is confined to the mucosa or submucosa, irrespective of nodal status. The mucosa is subdivided into the epithelium, M2 or lamina propria, and M3, muscularis mucosa. If the cancer cells do not penetrate the lamina propria, the tumor is defined as carcinoma in situ or high-grade dysplasia. The submucosa is also subdivided into three layers according to depth. The main rationale for endoscopic resection of GI cancer is data that shows that very early cancer can be considered a local rather than systemic disease. Radical surgical resection with lymph node dissection is still the gold standard for operable esophageal and gastric cancer. However, analysis of hundreds of thousands of pathology reports, patient histories, and long-term survival data have shown that endoscopic resection in selected patients is comparable to surgery. After endoscopic resection, pathological assessment of depth of cancer invasion, degree of differentiation, and extent of lymphovascular invasion allows the risk of lymph node metastasis to be predicted using published data of patients with similar findings. The final staging can only be done through formal histological analysis, which only endoscopic resection can achieve. The goal of endoscopic resection is curative resection. If curative resection is achieved, no further treatment is needed. Curative resection is not synonymous with en bloc resection. If curative resection is not achieved, further treatment is usually required. Currently accepted indications for endoscopic resection of early gastric cancer are limited to differentiated-type adenocarcinoma without ulcerative findings, of which the depth of invasion is clinically diagnosed as T1a, and the greatest diameter is less than 2 cm. The absolute indications only represent a small portion of early gastric cancers. The rationale for this recommendation is based upon the knowledge that larger-sized lesions, or lesions with unfavorable histology type, are more likely to extend into the submucosal layer, and thus have a higher risk of lymph node metastasis. In addition, resection of large lesions has not been technically feasible until the development of EST techniques. Two factors should be considered for curability assessment, completeness of primary tumor removal, and no possibility of lymph node metastasis. The resection is judged as curative when all of the following conditions are fulfilled. Clinical observations have noted, however, that the accepted indications for endoscopic resection can be too strict and can lead to unnecessary surgery. More recently, expanded indications of EST based upon a large database involving more than 5,000 patients who underwent gastrectomy with D2-level lymph node dissection were proposed. These groups of patients were shown to have no or lower risks of lymph node metastasis than the risks of mortality from surgery. However, it should be mentioned that the data provided from these observations were surgical specimens sliced at 5 mm, and not 2 mm as required for EST specimens. Tumors of the following categories have a very low possibility of lymph node metastasis. Endoscopic resection for these tumors is still regarded as an investigational treatment, and not EMR, but EST should be employed. However, it should be noted that most expert endoscopists today almost routinely perform EST under the expanded criteria outside the clinical trial setting. For EST according to the expanded criteria to be considered curative, the tumor should be en bloc resected, have negative margins, and have no lymphovascular invasion. It should also meet one of the four criteria below. Resection that does not satisfy any of the above criteria is considered non-curative. Published long-term results of endoscopic resection for early gastric cancer is excellent, with disease-free 5-year survival at nearly 100% for tumors that met the expanded criteria, and disease-free 10-year survival at 99% for tumors that met the absolute criteria. The depth of invasion is the most important indicator of the risk for nodal metastasis in patients with early squamous esophageal cancer. Tumors invading only the epithelium or lamina propria have virtually no risk of nodal metastasis and are clearly indicated for endoscopic treatment. Cancer that infiltrates the muscularis mucosa or the upper third of the submucosa is considered only to be a relative indication, as the risk of nodal metastasis in this case is approximately 19%. Endoscopic therapy is only considered to be adequate if the chance of nodal metastasis is lower than the surgical mortality rate of the esophagectomy. Early esophageal adenocarcinoma is considered to have a lower risk of nodal metastasis compared with early squamous cell esophageal carcinoma. For tumors confined to the lamina propria, the overall chance of nodal metastasis is virtually 0%. Reports for lymph node metastasis for M3 cancers are largely retrospective and are based on surgical specimens sliced at 5mm intervals. The reported incidence of nodal metastasis in these lesions vary from 0 to 12%. Although there are no current guidelines in the absence of other risk factors, endoscopic resection of early esophageal adenocarcinoma confined to the mucosa is considered acceptable by most specialists. Although cancer confined to the SM1 layer has a relatively low risk of nodal metastasis, endoscopic resection is only considered as an investigational treatment. Long-term results for endoscopic resection of early esophageal cancer is limited compared to early gastric cancer. Reported overall in disease-free 5-year survival for mucosal cancer is excellent. However, additional data is needed to recommend endoscopic resection for early esophageal cancer deeper than the M3 layer. Gross tumor appearance on conventional endoscopy is commonly used to distinguish advanced and early gastric cancers. However, there is currently no standard criteria for distinguishing mucosal and submucosal cancer. The first report of endoscopic staging of mucosal and submucosal cancer showed that endoscopy could distinguish mucosal and submucosal cancer in 71.9% of cases. However, available data is relatively limited and the methodology of the investigations is questionable. A recent study used characteristic endoscopic criteria of pathologically confirmed mucosal and submucosal cancer to create a staging criteria. Using this criteria, the endoscopic staging of early gastric cancers was compared with the pathological staging of the resected specimen. The overall accuracy of endoscopic staging using this criteria was 78%, and the sensitivity and specificity of endoscopic staging for mucosal cancer was 85 and 74%, and 73 and 83% for submucosal cancer. EUS, either conventional or with a high-frequency mini-probe, are often used to predict the depth of invasion prior to endoscopic resection. Although EUS can provide valuable information, understaging or overstaging of the lesion is a noted problem. The accuracy of EUS relatively decreases for patients with large lesions, lesions of the depressed type, ulcerated lesions, undifferentiated cancer, and lesions located in the upper third of the stomach in the case of early gastric cancer. The accuracies of EUS using high-frequency ultrasound probes for the staging of early gastric cancer have been described to be as high as 95%, compared with 63% for conventional EUS. In previous reports, the overall accuracies of the depth of invasion by ultrasound probes were 65 to 86%. Overstaging of early gastric cancers with a 20 MHz probe occurred in 19 to 24% of patients due to peritumoral fibrosis, mimicking deeper invasion. For early esophageal cancer, the accuracy of prediction of invasion depth was similar to early gastric cancer. However, the accuracy for submucosal lesions was significantly lower than mucosal lesions. Unfortunately, EUS is a highly investigator-dependent examination, and current data is still limited. The esophageal intrapapillary capillary loop, or IPCL pattern, may help predict the infiltration depth of early squamous esophageal cancer. Magnifying endoscopy coupled with MBI allows detailed examination of the IPCL pattern in this patient with esophageal cancer confined to the mucosa. In submucosal cancer, near-total IPCL destruction and neovascularization of tumor vessels can be observed, and endoscopic resection should be avoided. Biopsies are essential for tissue diagnosis. However, they only represent a small portion of the lesion, and there are reports of discrepancies between the histology of biopsy specimens and that of resected lesions. Additionally, the inflammatory change following multiple biopsies often results in fibrosis of the submucosal layer, which may cause difficulty not only by increasing the rate of complications such as bleeding and perforation, but also increasing the rate of incomplete resection. Recently, a new imaging technique that is capable of examining the gastrointestinal mucosa microscopically has been developed. Convocal endomicroscopy allows real-time in vivo high-resolution and high-magnification imaging of the gastrointestinal epithelium, which is comparable to histopathology. Dysplasia typically shows an irregular cryptolumen and dark irregular thickened epithelium. Gastric adenocarcinoma shows a completely disorganized epithelium, fluorescein leakage, and dark irregular epithelium. Undifferentiated adenocarcinoma, such as signet ring cell cancer, may also be distinguished by the complete absence of discriminal glandular structures. Although data for confocal endomicroscopy of gastric lesions is still limited, preliminary results are promising. A recent study compared the accuracy of biopsies in confocal endomicroscopy using post-ESD histopathology as a reference. The overall accuracy of confocal diagnosis of gastric adenomas and adenocarcinomas was significantly higher than endoscopic biopsies, and the overall accuracy of confocal diagnosis of differentiated and undifferentiated adenocarcinomas also was higher than that of endoscopic biopsies. The inject-and-cut EMR technique was first applied to flexible colonoscopy in 1973. This method uses an electrocautery snare through a single-channel endoscope. The somnucosal cushion lifts the lesion to facilitate its removal and minimizes mechanical or electrocautery damage to the deep layers of the GI tract wall. The inject lifting cut or strip biopsy EMR technique uses the grasping forceps to lift the lesion and the electrocautery snare through two separate channels of a dual-channel endoscope. First, some mucosal injection is applied to lift the lesion. Then the snare is slipped over the forceps and around the lesion for resection. The counter traction allows larger lesions to be resected. CAP-assisted EMR was developed for the resection of early esophageal cancer and was directly applied for the resection of early gastric cancer. This technique utilizes a clear plastic cap that is connected to the tip of a standard endoscope. Caps are composed of clear plastic that may be soft or hard. The caps are cylindric and available with straight or oblique-shaped tips, both with outer diameters ranging from 12 to 18 millimeters. CAP-assisted EMR uses a cap affixed to the tip of the endoscope and a crescent-shaped electrocautery snare that must be opened and positioned on the internal rim at the tip of the cap. Usually after marking, some mucosal injection is performed. The snare is then positioned on the internal rim of the cap. The endoscope is then positioned immediately over the target lesion. Suction is used to retract the mucosa into the cap, and the snare is closed to capture the lesion. The lesion is then resected with a standard snare excision technique. Stepwise radical endoscopic resection uses standard EMR techniques for piecemeal resection of large areas. It has been largely used for eradication of Barrett's cancer and the surrounding remnant dysplasia. After coagulation marks are put on the area to be resected, resection is performed using standard or CAP-assisted EMR techniques. The recurrence rate after stepwise radical endoscopic resection has been reported to be much lower than that can be expected after monotherapy. However, symptomatic esophageal stenosis may occur in as much as half of the cases. Precut EMR combines both standard EMR and EST techniques for a larger resection. This patient was diagnosed with dysplasia on the gastric angle. First marking around the lesion is performed. After some mucosal injection, the periphery of the lesion is cut using an endosurgical knife. When using a single-channel endoscope, additional injection can be done before applying the snare. When circumferential incision is complete, the lesion is then removed using a snare. Before resection, both the proximal and distal ends of the lesion should be checked to make sure they are within the snare, and the snare can be gently shaken to make sure the resection is not too deep. The main advantage of pre-cut EMR over conventional EMR is the target lesion can be resected with greater precision, and incomplete resection can be avoided. EMR has the advantage of being relatively simple with a shorter learning curve. These techniques, however, cannot be used to resect lesions larger than 15mm in one piece. Specimens obtained following piecemeal resections are difficult for the pathologist to analyze, and they render a pathological staging inadequate. This is a major factor leading to high risk of recurrence when these techniques are used. In a recent meta-analysis compared with EMR for early gastric cancer, ESD showed considerable advantages regarding unblock resection rate, complete resection rate, and low goal recurrence even for small lesions, but it had the disadvantages of higher complication rates for perforation. A regular diameter forward-viewing endoscope is sufficient for most lesions. Multi-bending scopes may be useful for lesions at difficult locations, however there is insufficient evidence to recommend their routine use. An auxiliary water channel with a water pump or water jet system is extremely useful, especially when bleeding is anticipated. The electro-surgical unit uses the heat energy produced by a high-frequency electrical current passing through the tissue for cutting and coagulation. Most knives used for ESD are monopolar and require a conduction plate placed on the patient. Although a pure cutting current can be used, incision and dissection usually are performed with a blended current. There are various preset modes for cutting and coagulation. The endocut I mode was initially developed for needle-type knives, and endocut Q for snare EMR. However there are no clear recommendations, and many authorities use the endocut Q mode for incision, including incision with the IT knife. Coagulation is also available in various settings. Swift coagulation is useful for submucosal dissection, and soft coagulation is usually used for bleeding control with hemostatic forceps. Effect settings can be adjusted, and an increase in effect increases the proportion of coagulation. Additionally, most ESUs used for ESD have adjustable cut duration and intervals. Although there are no available data for the optimal ESU settings, we recommend endocut I or Q at effect 3 for incision. Swift coagulation at effect 4 and 40 watts for dissection, and soft coagulation at effect 6 and 80 watts for hemostatic forceps. Endo-knives are the most basic accessories for ESD. There are numerous commercially available knives, and even more in development. Knives can largely be categorized into tip-type knives and blade-type knives. Each knife has specific properties and may have advantages in certain circumstances according to the location and characteristics of the lesion. Various solutions are currently used for submucosal injection. The ideal agent should be inexpensive, readily available, non-toxic, and easy to inject, and provide a long-lasting submucosal cushion. Repeated submucosal injections can be required if the cushion dissipates before complete removal of the lesion. The addition of staining dye, such as indigo carmine or methylene blue, to the injection solution is frequently used to assist in identifying the deep margins during the resection process. Hemostatic forceps are essential for preventing and managing bleeding during ESD. Transparent hoods are available in varying sizes and shapes. These accessories can improve the endoscopic field of view and assist in resection by pushing away the resected flap or resected tissue and exposing the dissection plane during ESD. The semi-caliber hood provides a more focused view during ESD and has an inner groove to stabilize the knife. For safe and successful ESD, the endoscopist should be aware of the structure and anatomy of the area to be resected. The esophagus is a thin-walled organ with limited space for scope manipulation. The mucosa is relatively thick and the muscle layer is thin, which increases the risk of perforation. However, due to the relatively loose submucosal layer, bleeding occurs less than gastric ESD. Currently, curable resection for esophageal cancer is confined to mucosal lesions. Therefore, the target of dissection should be the superficial submucosal layer. The muscle layer of the stomach consists of three layers and is considerably thicker than the esophagus and colon. Although the risk of perforation is generally lower than thinner-walled organs, the anterior and posterior aspects of the lower body is relatively thin and is considered the weakest point in the stomach. Bleeding varies with location and bleeding tends to be more frequent and severe in the body compared to the antrum. The dissection target for gastric ESD is usually set deeper than esophageal ESD. The procedure of endoscopic submucosal dissection can be briefly summarized into four basic steps. 1. Identifying the margin of the lesion and placing marking dots around the circumference of the lesion. 2. Injecting various fluids into the submucosal layer to separate it from the muscle layer. 3. Circumferential incision of the mucosa surrounding the lesion. 4. Dissection of the submucosal connective tissue under the lesion until it is completely resected. Marking is the process of creating marks or dots on the normal mucosa outside of the lesion. Most units use chromoendoscopy such as acetic acid and indigo carmine to identify the margins before marking. Special methods such as various knives, most commonly needle type knives, hot biopsy forceps or argon plasma coagulation can be used to create marks. Marking should be performed at least 5mm outside of the margin. This is to create a safety margin and to prevent coagulation damage to the lesion. For esophageal lesions, Lugol's solution is usually used to identify the margins. We used swift coagulation at effect 4 and 40 watts for marking. During marking, the knife should be contacted lightly with the mucosa. Excessive contact will cause mucosal perforation, which will lead to leakage during submucosal injection. Imaging-enhanced endoscopy may also help identify the margins before resection. Narrowband imaging with magnification and autofluorescence imaging have both been used for this purpose. However, there is limited data to recommend their routine use. Margin biopsies before or during ESD can also be used to determine the resection margin. Before ESD in this patient, the lateral margins were unclear. Margin biopsies prior to ESD with standby frozen pathologic assessment showed that the suspicious margin was free from tumor involvement and ESD was performed. In one study comparing chromoendoscopy and frozen margin biopsies for a prediction of the lateral border, in early gastric cancers with obscure margins, frozen biopsies resulted in a higher complete resection rate. Injection is the process of creating a submucosal cushion prior to incision or endoscopic dissection of the submucosal layer. A 21-gauge needle is most commonly used. The needle is inserted into the mucosa at an angle of 30 to 60 degrees, rather than a right angle. If the angle is too acute or the injection is too deep, most of the injection fluid will leak laterally or into the deep layers. Injection should start at the distal portion or anal side of the lesion and progress towards the proximal or oral side of the lesion, so as not to obstruct the field of view. Incision is the process of creating a continuous cut around the lesion, just outside the marks. After some mucosal injection, an initial small hole or pre-cut in the mucosa is made at the intended starting point. The starting point usually begins opposite to the planned dissection point. In this case, a 90 knife is inserted at an angle of 30 to 45 degrees into the pre-cut mucosa. We used EndoCut Qmode, effect 4, at 40 watts for incision. Because of the insulated tip, the IT knife can perform mucosal incision even at right angles with minimal risk of perforation. However, the knife should never be forcefully pressed against the mucosa, and the endoscopist should be able to feel a slight bling sensation during incision. Tip-type or needle-type knives are also frequently used for incision. Incision should start at least 2 mm outside of the marks. The distal hood can help to stabilize the knife during incision by compressing and fixing the mucosa. During incision, mild to moderate bleeding can often be controlled with just the knife. We used swift coagulation, effect 6 at 40 watts for coagulation. Tip-type knives should never be pressed at an acute angle with the mucosa as the risk of perforation is higher for this type of knife. Incision of esophageal lesions is similar to gastric lesions, except that there is limited space to maneuver the endoscope. Incision usually begins at the opposite side of the planned dissection point. Due to the parallel orientation with the blade of the Aiti knife and the esophageal wall, care should be taken not to press downwards with the blade, as this may cause perforation. An angle of at least 30 degrees is preferred when using the Aiti knife. If it is difficult to turn during incision, a hook knife may help as the direction of the cut can easily be controlled. If the incision is superficial, or there is a degree of submucosal fibrosis, the submucosal layer will not be properly exposed after submucosal incision. This will affect the following dissection process in a manner that will be more difficult and time-consuming. Therefore, additional trimming on the incision line for optimal submucosal exposure should be performed. To prevent excessive bleeding and prolongation of the procedure, precoagulation of visible vessels is helpful. Hemostatic forceps can be used for mid- to large-sized vessels, and endo knives can be used to coagulate small vessels. To prevent damage to the deeper layers, prolonged coagulation should be avoided, and forceps should be slightly pulled back before coagulation. Dissection of the lesion from the submucosal layer is the most difficult and time-consuming step of ESD. Like incision, dissection can be carried out with various knives. Before dissection, additional submucosal injection can be done if needed. Dissection with the IT knife involves side-to-side lateral movement. For more information, visit www.osho.com Somnucosal exposure is of paramount importance and the hood should be used to push the mucosa and expose the somnucosa. Bleeding should be quickly controlled if possible to shorten the procedure time. The IT knife can be easily hooked upon the lateral margin of the submucosal connective tissue for wide and rapid dissection. During dissection, the endoscopist should always know where the dissection plane is. Generally, upwards dissection is considered safer than downwards dissection. Because the target of endoscopic resection of the esophagus is usually the mucosa, the dissection does not have to be as deep as the stomach. Compared to the stomach, dissection itself is relatively easier because there is less bleeding and the direction of dissection is uniform. Scissor type knives are used for dissection in this patient. Endocut Q-mode at effect 1 is used for cutting and coagulation is set at soft coagulation effect 4 at 60 watts. During dissection, to prevent injury and perforation, the knife is pulled upwards and back before cutting. Visible vessels can easily be coagulated with scissor-type knives without the need for hemostatic forceps. ESD can be performed with the endoscope in a straight position, or retroflexed. This animation and the following videos will show the standard position of the scope with relation to the lesion and the initiation points of incision and dissection. ESD in the antrum is usually performed with the scope straightened, however lesions on the lesser curvature side of the antrum can be approached with the scope retroflexed. ESD of lesions on the angle and body are usually performed with the endoscope retroflexed. In this case the distal end from the endoscopic view is the oral side, however it should be noted that the position of the scope is diverse and will differ with different clinical situations. If the endoscopist feels discomfort or the lesion cannot be approached easily, another position should be attempted. Lesions on the cardia are usually approached with the endoscope retroflexed. Some authorities, however, prefer dissection from both sides, first from the anal side and then from the oral side, with the endoscope straightened. For ESD, the dissection direction should be planned ahead. If possible, dissection should proceed in a manner that the flap is influenced by gravity. This will help to expose the somnucosa and facilitate dissection. Both injection and incision are started from the distal or anal side of the lesion. Contrary to incision, dissection starts at the oral side of the lesion. The IT knife is used for side-to-side lateral dissection. Due to the relatively thick wall and loose density of somnucosal vessels, the risk of perforation and bleeding is relatively low in the antrum. It is for these reasons that ESD of antral lesions is suitable for less experienced endoscopists. ESD of lesions located at the gastric angle or lesser curvature side of the antrum are usually done with the endoscope j-turned. If too much air is in the stomach, it may be difficult to maintain an optimal angle with the knife and the mucosal surface. In this case, air should be suctioned from the body. This will pull the lesion toward the endoscope. Dissection usually starts at the anal side of the lesion. However, depending on the situation, dissection can be started on both sides. During dissection, the endoscopist should sense a gradual feeling of pushing the flap in the upward direction. The muscle layer is easily exposed and due to the curved nature of the region, dissection should proceed with caution against perforation. Once ESD at the angle is fully experienced, lesions on the lesser curvature of the body are easily approached. Compared to the antrum, the mucosa is thicker and the submucosal vascularity is relatively abundant in the body. The anterior and posterior wall of the lower body has the thinnest muscle layer in the stomach. This is due to the absence of the internal oblique muscle. If not for the increased risk of perforation, the level of ESD difficulty is similar to the angle. Like the angle, incision starts at the oral side, with the scope retroflexed, and dissection starts at the anal side. The endoscopist must be extremely cautious of perforation, especially if the knife is not level with the wall or the lesion is not fully visualized. The cardia is abundant in some mucosal fat and vessels. An extremely high level of skill and concentration is required for ESD in this region, where even target biopsies may be difficult. Obtaining optimal endoscopic visualization may be difficult, and complicated endoscopic manipulation is usually required. Marking an incision starts at the oral side. If the lesion is adjacent or overlaps the EG junction, the oral side should be carefully examined with the scope straightened. Dissection begins at the anal side. However, if an adequate angle is difficult to acquire, dissection can also be performed from the oral side. Pre-ESD evaluation is extremely important for lesions that involve the pyloric ring. The lesion should be examined from both the stomach and the duodenal bulb. Perforation in this area is rare due to the thick muscular pyloric ring. After marking, both incision and dissection starts at the aspect of the pyloric ring. After dissection around the ring is completed, dissection is similar to that of the antrum. If the lesion is adjacent but does not overlap the pyloric ring, dissection can start at the oral side of the lesion, identical to dissection in the antrum. Esophageal ESD is principally the same as those in other areas of the GI tract. For marking around the lesion, dots are placed about 5 mm outside the lesion using soft coagulation mode. Some mucosal injection should begin at the anal side. Repeated 2 mm injections is normally sufficient. An incision in the mucosa around the lesion is started at the anal side. We used endocut I mode at effect 3. The anal half of the lesion, which is horseshoe-shaped, is completed first, followed by the oral half. Incision from the left wall side is preferable with consideration of gravity, as submerging the lesion in the collection of fluid can disturb the endoscopic view. Incision from the right wall side is preferred with consideration of gravity, Dissection of the submucosa begins from the oral end to the anal end using swift coagulation mode at effect 4 and 40 watts. Although esophageal ESD itself is not as difficult as some gastric ESDs, the adjacent thoracic organs and the risk of perforation are a burden to the endoscopist. Therefore, most ESD centers recommend a training period with gastric lesions before performing esophageal ESD. The success rate of ESD depends on the technical proficiency of the endoscopist and the condition of the tumor. Even for a skilled endoscopist, however, somucosal fibrosis can be an obstacle to success. Even after repeated injection, somucosal elevation is minimal and additional trimming is usually required. Dissection is time-consuming and difficult, and may require additional knives. Somnucosal fibrosis, which usually results from inflammation or tumor invasion, makes it harder to lift the tumor tissue from the muscle layer. This, in turn, lengthens the procedure time, creates risk of complications such as perforations, and reduces the success rate of complete on-block resection. In a recent study, the severity of endoscopic somnucosal fibrosis was associated with a lower on-block resection rate and with abundant immediate bleeding. Moreover, the greater the degree of somnucosal fibrosis, the longer the time taken for the ESD procedure, and the higher the frequency of complications such as perforation and immediate bleeding. After resection, the specimen should be slightly stretched and fixed upon a flat surface. The oral and anal sides, as well as the margins that are closest to the lesion, should be identified. During pathological assessment, the sections will be parallel to the line connecting the closest margins. Magnified view of the specimen with stereoscopy may provide additional information to the pathologist regarding the optimal cutting direction of the specimen and tumor involvement of the margins. Pathologic analysis should contain information regarding the histology, depth of invasion, presence of lymphovascular invasion, involvement of the horizontal and vertical margins, and presence of an ulcer. Immunohistochemical staining can provide additional information, including the mucin phenotype, which is related to prognosis, and more accurate assessment of lymphovascular invasion. ESD specimens are sectioned at 2 mm intervals, and mapping of the specimen gives valuable information for clinical decision making. Following ESD, serial x-rays should be taken to detect perforation. The patient is usually put on MPO for 1 or 2 days with IV PPIs to control epigastric pain. Follow-up endoscopy is required for high-risk patients and patients with signs of bleeding. If the patient is tolerable, a clear liquid diet can be started as early as the following day. In the case of gastric lesions, H. pylori should be tested and treated if necessary. For follow-up of gastric cancer after ESD, we recommend at least annual endoscopy and abdominal CT scans. Bleeding is the most common complication, occurring in up to 8% of patients undergoing standard EMR, and in up to 7% of patients undergoing ESD. Immediate bleeding, which can be brisk, appears more common with resections of tumors located in the upper third of the stomach. Most cases of immediate bleeding can be controlled with hemostatic forceps using soft coagulation mode. Precise targeting is required, and flushing water on the area after grasping the tissue with forceps helps to locate the bleeding focus before coagulation. Minor bleeding that occurs during incision or dissection can be easily controlled with endo knives using swift coagulation. When using the endo knife for hemostasis, excessive downward pressure should not be applied to the tissue. A slight lateral movement of the knife using coagulation is recommended. During incision, it may help to extend the incision line past the bleeding area. This will help to identify and expose the bleeding focus. When bleeding is severe enough to obstruct the endoscopic view, the distal attachment or cap can be used to compress the area of interest and locate the bleeding focus. Arterial bleeding can be severe and should be managed quickly. Hemostatic forceps are sufficient in most cases. However, endoclips may be required if the bleeding cannot be controlled. Perforation is one of the most feared complications of ESD. Reported rates of perforation range from 1 to 4%, even in experienced centers. Reported risk factors for perforation are ESD at the body, submucosal fibrosis, and piecemeal resection. If perforation is discovered during the procedure, endoscopic management with endoclips is usually successful in most cases. Following ESD and pathological analysis, patients can be stratified according to the achievement of curative resection. If the resection was non-curable, additional surgery is usually recommended. However, if the patient is unable to undergo surgery, other options, such as re-ESD or ablation therapy, have been attempted. The reported incidence of lymph node metastasis following gastrectomy of early gastric cancers with non-curable endoscopic resection ranged from 1 to 9%. In all reported cases, the lesions were over the expanded indications of ESD. Although the current literature clearly supports endoscopic resection for early stage GI cancers, precise prediction of invasion depth and the presence of nodal metastasis is not possible with current technology. Therefore, some patients will require additional surgery following ESD. Gastrectomy and lymph node dissection are performed as standard treatment for early gastric cancer exceeding the indications for endoscopic resection due to the risk of lymph node metastasis, which varies from 2 to 20%. This suggests that at least 80% of patients might potentially preserve their stomachs with minimal treatment. Investigators, mainly from Asia, have combined laparoscopic surgery with ESD or endoscopic full thickness gastric resection with the purpose of achieving complete resection of the tumor while preserving the organ and assessing pathological nodal status. ESN combines ESD and central node navigation surgery for complete resection of the tumor and assessment of pathological nodal status. ESN is performed under general anesthesia and markers such as endocyanide green are used to detect the central nodes and the regional lymphatic drainage. ICG is injected into the submucosa surrounding the lesion and within 20 minutes after injection, stained central nodes are observed intraperitoneally around the serosa and surrounding fat tissue. For more information visit www.osho.com Detected sentinel nodes are then removed and sent for intraoperative frozen analysis. According to the results of the sentinel nodes, either EST or standard gastrectomy is then performed. There are two studies that incorporated laparoscopic lymph node dissection with ESD, and only one study that used sentinel node navigation surgery. Nodal metastasis was detected in 9.5 to 15%, and both studies showed no recurrence during the follow-up period. With sentinel node navigation surgery, two patients were able to convert to standard gastrectomies without an additional operation. Standard nodes for early gastric cancer uses the ESD technique for full-thickness gastric wall resection. First marking dots are put around the lesion, and ICG is injected before laparoscopic lymph node dissection. After regional nodal dissection, circumferential incision using the standard ESD technique is performed. Then a gastric perforation is created on the incision line, and an IT knife is inserted into the hole for full-thickness resection. Full-thickness resection is performed under laparoscopic guidance to prevent injury to the adjacent organs and peritoneal bleeding. Final resection is performed with laparoscopy. After resection, the gastric wall opening is closed with laparoscopy. In one report of hybrid notes for early gastric cancer, the authors reported that hybrid notes was a feasible procedure that enabled complete and whole-layer excision of the lesion while obtaining regional lymph node assessment, thus encompassing the advantages of both ESD and laparoscopic surgery. However, in some cases, additional gastrectomy was needed due to failure to maintain the stomach shape after endoscopic full-thickness gastric resection and post-surgical leakage and ischemia. There was also a concern about potential peritoneal dissemination as the tumor surface was exposed to the peritoneal cavity. With the development of flexible laser fibers, laser ESD is being attempted in selected centers. In this patient, we used a thulium laser system with a wavelength of 2 micrometers. Because of the rapid vaporization with effective hemostasis, this system may be suitable for tissue resection without deep injury. After marking around the lesion, incision is started from the distal side of the lesion. Dissection with the laser is performed in a similar way to dissection with other endo-knives. Endoscopic resection improves the quality of life of patients with early GI cancer and dysplasia by preserving their organ function. Many lesions that required surgery in the past are now treated with EMR or ESD. Prior to resection, careful consideration of the ability of the performing endoscopist and the rationale of indications is needed to achieve a high rate of curative resection. Endoscopic skills for ESD can be acquired by source reading, participating in live demonstrations, and directly observing an expert's procedure. Before ESD, careful preparation of instruments and discussion with team members can guarantee the success of treatment. Small lesions, which can be treated with EMR, should not be selected as candidates for ESD.

therapeutic endoscopy

gastric cancer

esophageal cancer

ablation therapy

endoscopic resection

diagnosis

complications

curative resection

lymph node metastasis

mucosal lesions

pathological assessment