Endoscopic mucosal resection (EMR) of early gastric cancer, which has been proved to be safe and effective and is the established standard of care in Japan, has become increasingly established worldwide in the past decade. Endoscopic submucosal dissection (ESD) is superior to EMR, as it is designed to provide precise pathologic staging and long-term curative therapy based on an en bloc R0 specimen irrespective of the size and/or location of the tumor. However, ESD requires highly skilled and experienced endoscopists. The introduction of ESD to the Western world necessitates collaborations between Eastern and Western endoscopists, pathologists, and surgeons.
Key points
- •
ESD designed to provide precise pathologic staging and curability based on en bloc R0 specimen irrespective of the size and/or location of the tumor.
- •
ESD requires high technical skills on the part of the operator, is time consuming, and is associated with procedural related complications.
- •
Standardized ESD training system is urgently needed to disseminate safe and effective ESD technique to practices with limited ESD experience.
Videos of endoscopic submucosal dissection procedures on the stomach accompany this article
Principle of endoscopic resection as a definitive treatment
Early gastric cancer (EGC) is defined when the cancer invasion is confined to the mucosa or submucosa (T1 cancer), irrespective of the presence of lymph node metastasis. The presence of lymph node metastasis is a strong predictor on patients’ prognosis, and for this reason gastrectomy with lymph node dissection has historically been the gold standard for treatment of EGC.
The 5-year cancer-specific survival rates of EGC limited to the mucosa or the superficial submucosa were reported to be 99% and 96%, respectively. However, the traditional approach of radical gastric surgery is associated with significant morbidity and reductions in quality of life. Surgery to remove intramucosal gastric cancer whereby the incidence of lymph node metastasis is low (up to 3%) is therefore excessive for most patients. By comparison, surgery in the majority is appropriate when the cancer involves the deep submucosa, where the incidence of lymph node metastasis increases to as high as 20%. A stratification method to identify patients who have negligible risk for developing lymph node metastasis would thus optimize the selection of patients who can be cured by endoscopic resection and thus avoid the risks of surgery. The ideal patients for endoscopic resection rather than surgery are those who have a lower mortality risk from metastasis.
Precise stratification of patients with favorable prognosis should be underscored. Endoscopic ultrasonography has limited staging accuracy (80%–90%), and thus would result in unnecessary surgery in up to 20% of patients. Ablation endoscopic techniques may cure EGC, but do not provide a pathologic specimen for analysis, leaving the patient without proper staging or prognosis. Prior experience in the 1980s and 1990s suggests that pathologic staging is the best predictor of the risk for lymph node metastasis.
The most accurate method to stratify the patients’ prognosis for developing lymph node metastasis was reported by Gotoda and colleagues in 2000. In a study involving 5265 patients who had undergone gastrectomy with careful lymph node dissection and pathologic analysis, the risks of lymph node metastasis can be clustered to several pathologic findings of the involved mucosa and submucosa: macroscopic appearance, size, depth, differentiation of cancer, and lymphatic and vascular involvement. This seminal work provides one of the pillars of endoscopic resection of EGC.
The Paris classification of superficial neoplasias of the gastrointestinal tract provides another pillar. It allows standardization of the endoscopic appearance of EGC, which is then useful to estimate tumor depth and likelihood of the risk of lymph node metastasis. The en bloc resected specimen provides further information on size, depth, and differentiation of cancer, as well as lymphatic and vascular involvement. Thus the tumor can be accurately staged, the patient’s prognosis estimated, and the need for additional therapy assessed.
The major advantage of endoscopic resection is the ability to provide accurate pathologic staging without precluding future surgical therapy. After endoscopic resection, pathologic assessment of depth of cancer invasion, degree of cancer differentiation, and involvement of lymphatics or vessels allows the prediction of the risk of lymph node metastasis. The endoscopic submucosal dissection (ESD) technique was developed to extend the ability of endoscopic mucosal resection (EMR) to remove lesions larger than 2 cm en bloc, as EMR is limited to the resection of small tumors. It is also known that piecemeal resections of lesions larger than 2 cm lead to a high risk for recurrence of local cancer and inadequate pathologic staging. ESD allows for large en bloc resection regardless of tumor size, location, and/or submucosal fibrosis, thereby allowing precise pathologic staging. ESD is the most gratifying for patients with EGC because of its minimally invasive and curative potential, which is why it is increasingly used globally.
Principle of endoscopic resection as a definitive treatment
Early gastric cancer (EGC) is defined when the cancer invasion is confined to the mucosa or submucosa (T1 cancer), irrespective of the presence of lymph node metastasis. The presence of lymph node metastasis is a strong predictor on patients’ prognosis, and for this reason gastrectomy with lymph node dissection has historically been the gold standard for treatment of EGC.
The 5-year cancer-specific survival rates of EGC limited to the mucosa or the superficial submucosa were reported to be 99% and 96%, respectively. However, the traditional approach of radical gastric surgery is associated with significant morbidity and reductions in quality of life. Surgery to remove intramucosal gastric cancer whereby the incidence of lymph node metastasis is low (up to 3%) is therefore excessive for most patients. By comparison, surgery in the majority is appropriate when the cancer involves the deep submucosa, where the incidence of lymph node metastasis increases to as high as 20%. A stratification method to identify patients who have negligible risk for developing lymph node metastasis would thus optimize the selection of patients who can be cured by endoscopic resection and thus avoid the risks of surgery. The ideal patients for endoscopic resection rather than surgery are those who have a lower mortality risk from metastasis.
Precise stratification of patients with favorable prognosis should be underscored. Endoscopic ultrasonography has limited staging accuracy (80%–90%), and thus would result in unnecessary surgery in up to 20% of patients. Ablation endoscopic techniques may cure EGC, but do not provide a pathologic specimen for analysis, leaving the patient without proper staging or prognosis. Prior experience in the 1980s and 1990s suggests that pathologic staging is the best predictor of the risk for lymph node metastasis.
The most accurate method to stratify the patients’ prognosis for developing lymph node metastasis was reported by Gotoda and colleagues in 2000. In a study involving 5265 patients who had undergone gastrectomy with careful lymph node dissection and pathologic analysis, the risks of lymph node metastasis can be clustered to several pathologic findings of the involved mucosa and submucosa: macroscopic appearance, size, depth, differentiation of cancer, and lymphatic and vascular involvement. This seminal work provides one of the pillars of endoscopic resection of EGC.
The Paris classification of superficial neoplasias of the gastrointestinal tract provides another pillar. It allows standardization of the endoscopic appearance of EGC, which is then useful to estimate tumor depth and likelihood of the risk of lymph node metastasis. The en bloc resected specimen provides further information on size, depth, and differentiation of cancer, as well as lymphatic and vascular involvement. Thus the tumor can be accurately staged, the patient’s prognosis estimated, and the need for additional therapy assessed.
The major advantage of endoscopic resection is the ability to provide accurate pathologic staging without precluding future surgical therapy. After endoscopic resection, pathologic assessment of depth of cancer invasion, degree of cancer differentiation, and involvement of lymphatics or vessels allows the prediction of the risk of lymph node metastasis. The endoscopic submucosal dissection (ESD) technique was developed to extend the ability of endoscopic mucosal resection (EMR) to remove lesions larger than 2 cm en bloc, as EMR is limited to the resection of small tumors. It is also known that piecemeal resections of lesions larger than 2 cm lead to a high risk for recurrence of local cancer and inadequate pathologic staging. ESD allows for large en bloc resection regardless of tumor size, location, and/or submucosal fibrosis, thereby allowing precise pathologic staging. ESD is the most gratifying for patients with EGC because of its minimally invasive and curative potential, which is why it is increasingly used globally.
Indications for endoscopic resection
The traditional criteria for endoscopic resection of EGC were founded on the technical limitation of EMR to remove gastric lesions of less than 2 cm in diameter en bloc. The empirical indications for EMR include : (1) papillary or tubular (differentiated) adenocarcinoma, (2) less than 2 cm in diameter, (3) without ulceration within tumor, and (4) no lymphatic-vascular involvement.
Clinical observations have noted, however, that the empirical indications for EMR were too strict and had led to unnecessary surgery. Therefore, expanded criteria for endoscopic resection have been proposed, especially after large en bloc resection became technically achievable using ESD. The large number of patients included in the study reported by Gotoda and colleagues was instrumental in defining the expanded criteria, as its 95% confidence intervals (CI) were narrow in comparison with the early studies with wide upper CI attributable to small sample sizes. In addition, recent data showed that no lymph node metastasis was found in 310 patients with poorly differentiated adenocarcinoma and/or signet-ring cell EGC, less than 2 cm in diameter, without ulceration, and without lymphatics or vascular involvement (95% CI 0%–0.96%). Therefore, patients with these finding may also be treated with ESD alone.
The longitudinal results of patients who were treated by endoscopic resections have provided the ultimate proof of its safety and efficacy. The long-term outcomes after EMR for small differentiated mucosal EGC less than 2 cm in diameter have been reported to be comparable with those following gastrectomy. Patients who underwent ESD following the expanded criteria have long-term survival and outcomes similar to those of patients treated according to the traditional criteria. The 5-year survival rate was 92% in patients in the traditional criteria group and 93% in the expanded criteria group. There was no significant difference in overall survival between these groups (multivariable hazard ratio 1.10; 95% CI 0.67–1.81).
The current indication of endoscopic resection for patients with EGC is called the Expanded Criteria for Endoscopic Resection in EGC ( Table 1 ). The criteria are best used by comparing the risk for developing lymph node metastasis or distant metastasis against the risk of surgery, and considering patients’ morbidities and preferences. It is important to understand that the expanded criteria were developed to identify which patients have a low risk of lymph node metastasis. Patients meeting the criteria incur a risk of lymph node metastasis up to the upper limit of 95% CI. This risk is by no means nonexistent ( Fig. 1 ).
| Criteria | Incidence (No. with Metastasis/Total Number) | 95% Confidence Interval |
|---|---|---|
| Intramucosal cancer | 0/1230; 0% | 0–0.3 |
| Differentiated (well and/or moderately differentiated and/or papillary adenocarcinoma) type | ||
| No lymphatic-vessel involvement | ||
| Irrespective of ulcer findings | ||
| Tumor size <3 cm | ||
| Intramucosal cancer | 0/929; 0% | 0–0.4 |
| Differentiated type | ||
| No lymphatic-vessel involvement | ||
| Without ulcer findings | ||
| Irrespective of tumor size | ||
| Intramucosal cancer | 0/310; 0% | 0–0.96 |
| Undifferentiated (poorly differentiated adenocarcinoma and/or signet-ring cell carcinoma) type | ||
| No lymphatic-vessel involvement | ||
| Without ulcer findings | ||
| Tumor size <2 cm | ||
| Minute submucosal penetration (sm1) | 0/145; 0% | 0–2.5 |
| Differentiated type | ||
| No lymphatic-vessel involvement | ||
| Tumor size <3 cm |
Complications
Although delayed bleeding is thought to be the most common complication, occurring in up to 8% of patients undergoing gastric ESD, the risk of the procedure is primarily related to uncontrollable bleeding during ESD. Acute bleeding may obscure the visual field, leading to a higher risk of complications. Therefore, endoscopic hemostasis should be immediately performed step by step ( [CR] and [CR] ). Small vessels can be coagulated using the ESD knife (forced coagulation mode, 50 W). Larger vessels should be coagulated by specially designed hemostatic forceps (FD-410LR; Olympus Medical Systems, Tokyo, Japan) using soft coagulation (80 W).
After dissection has been completed, further hemostasis is performed on visible vessels to minimize delayed bleeding. The hemostatic forceps using soft coagulation are used to coagulate any visible small vessels. Excessive coagulation in the area of the exposed muscle layer of the ESD defect should be avoided because of the risk of delayed perforation. Delayed bleeding, manifested by hematemesis or melena at 0 to 30 days after the procedure, is treated by emergent endoscopy, performed after resuscitation, using similar techniques ( [CR] ). Delayed bleeding after ESD has been correlated with tumor location (lower part) and size.
Perforations are typically closed by endoclips ( [CR] and [CR] ). If pneumoperitoneum is significant, the patient may develop respiratory compromise or even shock ( Fig. 2 ). Thus, to prevent the abdominal compartment syndrome, decompression of the pneumoperitoneum must be immediately performed using a 14-gauge puncture needle with side slits after confirmation, using a 23-gauge needle syringe filled with saline ( [CR] ). Nasogastric suction is applied for 12 hours and broad-spectrum antibiotic is given for 2 days. Bacterial peritonitis is relatively rare because of the antibacterial effect of gastric acid. Diet is slowly advanced on day 3 or day 4 after ESD.
To prevent gastric perforations and better facilitate ESD, polyethylene glycol or sodium hyaluronate as an injection agent has recently been reported. These agents remain longer in the submucosa and create clearer dissection layer. No evidence of peritoneal dissemination and/or lymph node metastasis caused by gastric perforation has been reported. The use of CO 2 to insufflate the stomach during ESD is also extremely useful, as CO 2 is readily absorbed should perforation occur.
Diagnosis
Accurate diagnosis of the lateral and vertical margins of EGC is important before endoscopic resection. Similarly to the colon, for which patient preparation is critical, a clean upper gastrointestinal tract is also hugely important. Mucus and debris need to be removed before and during the examination. In Japan, mucolytic and defoaming agents are used. In Western countries, acetylcysteine may be used as a mucolytic and activated dimethicone as a defoaming agent.
Most EGC are nonpolypoid, and the depressed type is the most common. Several endoscopic findings suggest submucosal invasion in a depressed EGC: thickening, rigidity, disappearance of mucosal surface pattern, extensive redness, or nodularity within the depression, and swelling of converging folds. Whereas the larger depressed EGC are more likely to be submucosal, size is not predictive of submucosal invasion in the flat ones. Moreover, the prediction of submucosal involvement in the ulcerated type is more difficult because of the presence of fibrosis, which may mimic some of the findings described in the depressed lesions. Endoscopic ultrasonography is rarely used to estimate the depth of the lesion.
Indigo carmine chromoendoscopy is the most useful method to determine the lateral margin of EGC. The appearance of the lesion is readily seen and relatively easily understood. Its use allows the determination of the macroscopic appearance and size of the lesion. At the time of diagnosis, the lesion is biopsied at its worst-appearing portion to confirm and document the diagnosis. Multiple biopsies of the lesion are avoided because of the risk of inducing undue submucosal fibrosis, which would make endoscopic resection difficult. Normal surrounding mucosa is biopsied separately to confirm that the assessment of the lateral margin is correct. Lesions that meet the expanded criteria are candidates for endoscopic resection.
Narrow-band imaging (NBI) has increasingly been described as an alternative to indigo carmine image enhancement, especially for the diagnosis of small, depressed, or flat types of EGC. However, its application requires high magnification, which is not available worldwide. In addition, NBI is not applicable for determining the lateral extent of undifferentiated types of EGC. The vessel plus surface classification system, which has been well studied by Yao and colleagues, is based on the ability to visualize both the microvascular and microsurface patterns.
Training
The learning and application of these relatively complex endoscopic therapeutic techniques for early gastric cancer have been demonstrated across the world. Most Japanese experts set the level of expertise at 50 to 100 cases to become proficient in gastric ESD, and require a trainee to perform at least 30 gastric ESD cases under the supervision of an expert to gain basic proficiency in this technique. Fundamental skills and knowledge of the macroscopic diagnosis, indications, tools, and techniques, as well as the resection specimen for histologic preparation and interpretation, should be sought through self-education seminars and conferences. Participation in hands-on courses with isolated or live animal stomach and live demonstrations is vital in accelerating the learning curve. In comparison with the self-taught method, specialty centers that use a standard of practice to systematically provide training and competency under direct expert supervision have been more successful.
A standardized ESD training system is urgently needed to disseminate a safe and effective ESD technique to practices with limited ESD experience. A questionnaire survey of Japanese experts reported that ESD training should be conducted in a step-up approach after prior experience with conventional EMR. The panel concluded that preceptees should observe and attend ESD procedures as an assistant in at least 20 and 5 cases, respectively, so as to understand a wide variety of ESD procedures and strategies and to develop trouble-shooting abilities. The trainee should start with an antral lesion of less than 20 mm in diameter without ulcer, as it has the lowest risk of noncurative resection, and then progress to lesions in the distal and proximal stomach. Submucosal dissection has been shown to be more difficult than mucosal incision, mostly because of uncontrollable hemorrhage. As such, appropriate supervision by a highly experienced endoscopist is necessary in the early phase of learning, because a significant number of cases require the supervisor to complete the procedure after uncontrollable hemorrhage or perforation occurs. Choi and colleagues highlighted the difficulties in learning such novel complex techniques independently without the direction of an experienced mentor. These investigators reported an initial complete resection rate of 60% to 65% and an en bloc resection rate of 45% to 55% with 3 perforations. Of note, neither years of prior endoscopic experience, number of endoscopic procedures per year, nor prior experience with endoscopic retrograde cholangiopancreatography has been associated with a significant difference in the risk of perforation during training.
Western ESD experts to supervise ESD training are limited in number, and virtual simulators for ESD are not yet available. Thus the use of animal models to facilitate the early training of ESD is important to minimize the risk of higher complications at the beginning of the learning curve in humans. Proper use of ex vivo and in vivo animal models is performed in an animal facility under the direction of a veterinarian, with dedicated equipment and a standardized setup. Parra-Blanco and colleagues demonstrated an effective learning strategy of the Western endoscopist by training in ESD in the absence of expert supervision. Indeed, Parra-Blanco learned ESD from Japanese experts, but then trained in harvested pig stomachs and then in live pigs, and reported a 4.5% (95% CI 0.12%–22.8%) perforation rate. A European group assessed the impact on 18 experienced endoscopists (39% with basic ESD experience) who participated in a 2-day training course that included seminars and hands-on training with living pigs, and was supervised by experts in ESD. Each endoscopist performed a mean 4.1 ESD, mostly in the stomach (84%), with a perforation rate of 22%. Although technical maneuvers may be simulated using a pig model, anatomic differences between the pig and human stomach may pose limitations to training. Although compared with humans the pig model in gastric ESD exhibits less frequent bleeding, more difficult submucosal injection, and less fibrosis, the use of models allows endoscopists to ascend the learning curve in a relatively short time, and enhance the safety and efficacy of the patient experience.
Practical approach
The low prevalence of superficial gastric epithelial neoplasms has translated into very few opportunities for Western endoscopists to perform gastric ESD. The first dedicated ESD devices were approved in the United States less than 2 years ago. Furthermore, the choice of devices, endoscopes, and ancillary equipment for ESD available in the United States is different to that available in Japan. Finally, the technical expertise, training opportunities, and backgrounds of endoscopists embarking on ESD in the West differ significantly from those of their Eastern counterparts. This section reviews the technical aspects of gastric ESD from a Western perspective.
Endoscopes and Endoscope Accessories
Standard upper endoscopes with water-jet function are adequate for most cases. Water-jet irrigation is highly advantageous because it can facilitate identification of bleeding vessels. Some experts recommend the use of specialized endoscopes because of perceived advantages (therapeutic channel endoscope for better suction, double-channel endoscope to pass 2 devices simultaneously to save time on device exchanges, double-bending endoscope to facilitate access to some difficult-to-reach lesions). The authors favor the use of standard-size endoscopes because of their small diameter and ease of maneuverability.
A specialized soft, clear plastic distal attachment (cap) should be applied to all endoscopes for ESD. For each endoscope model there is a corresponding size of cap (JMDN 38819001; TOP Corp, Tokyo, Japan). The use of CO 2 for insufflation is highly recommended because it causes less luminal distention and patient discomfort. Furthermore, if there is a perforation the leaking CO 2 will rapidly be reabsorbed from the peritoneal cavity, thus decreasing the risk of respiratory compromise caused by increased intraperitoneal pressure. With the use of CO 2 , the ESD can be completed after the perforation has been closed using endoscopic clips.
Dedicated ESD Devices
A large number of devices for ESD are now available in Japan and are divided into 2 general categories: needle-knife type and scissors type. The needle-knife devices can be further subdivided into uncovered type and insulated type. Only a limited number of needle-type devices are approved by the Food and Drug Administration, and at present none of the scissors types is available in the United States.
There are 6 basic functions required from a device at the time of ESD: mucosal marking, submucosal injection, precutting (the very first small incision into the mucosa needed to engage all insulated-tip [IT] type devices), cutting of the circumferential incision, submucosal dissection, and hemostasis. The ESD devices available in the United States and the type of function that they can accomplish are summarized in Table 2 . It should be emphasized that although a knife can be used to perform a specific function extremely well, it does not necessarily make the device functionally suitable in all circumstances.
| Device | Marking | Injection | Precutting | Circumferential Incision | Submucosal Dissection | Hemostasis |
|---|---|---|---|---|---|---|
| IT2 | ✓ | ✓ | ✓ | |||
| Dual | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Hook | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Hybrid | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Coagulating forceps | ✓ |
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
