Endoscopic mucosal resection and endoscopic submucosal dissection in the West – too many concerns and caveats?

Local treatment of early gastrointestinal neoplasia promises to be curative and superior to surgery in terms of organ preservation, safety, duration of hospitalization, and cost-effectiveness. Prerequisites for a local approach are that these advantages outweigh the risk of lymph node metastases and recurrence of neoplasia. In addition the procedure-related morbidity and mortality should be low.

Numerous studies on histopathology of surgically resected specimens of early neoplasia have demonstrated that the vertical depth of tumor infiltration is the most important predictive parameter for lymph node metastases. According to these results, the risk can be negligible in the upper and lower gastrointestinal tract if a neoplasia is limited to the mucosa, with the exception of deep mucosal infiltration of esophageal squamous cell carcinoma [1]. In addition, the risk is also very low in well-differentiated gastric or colorectal adenocarcinoma with superficial invasion of the submucosa and no infiltration of lymphatic vessels. Unfortunately even modern imaging techniques such as endomicroscopy, high frequency ultrasound or optical coherence tomography do not allow an accurate preoperative determination of the depth of tumor infiltration. Therefore it is widely accepted that endoscopic resection is required to obtain specimens for histopathological evaluation and identification of more advanced tumor stages for which surgery should be considered. This approach frequently changes the diagnosis obtained by biopsies and may reveal submucosal invasion even in cases with a flat neoplastic lesion [2] [3] [4]. As histopathological results cannot be awaited during the procedure, a complete endoscopic resection of the neoplastic lesion is usually attempted. Only en bloc resection provides specimens for complete histological evaluation not only of the vertical and but also of the lateral tumor margins. Piecemeal resection usually allows the determination of the vertical invasion provided that no larger bridges of tissue are left between the resection areas. However lateral tumor-free margins cannot be evaluated by histology of several specimens. Therefore the completeness of piecemeal resection depends on careful endoscopic exclusion of any neoplastic areas remaining. Minor remnants at the tumor margins or between resections cannot always be avoided. However this limitation of piecemeal resection should have no impact on the identification of patients with an increased risk for lymph node metastases. The main drawback is the higher risk of residual neoplasia and local recurrences. This can be overcome by obligatory follow-up endoscopy within 6 – 8 weeks to verify or exclude neoplastic tissue remnants after scar formation. Remnants are usually small and can be easily removed or ablated. Biopsies should be taken before ablation (e. g. by radiofrequency or argon plasma gas coagulation) if there is any histological or endoscopic evidence of a more advanced neoplasia. ”Complete local remission“ was recently defined as histologically confirmed complete removal of neoplasia (R0 situation) and negative follow-up endoscopies or vertically tumor-free margins but laterally incomplete or indeterminate margins (R1/Rx) with two endoscopic check-up examinations without evidence of residual tumor even after local re-treatment [5]. In contrast to the problem of residual neoplasia, true recurrences after complete local remission (CLR) are probably rare and have to be differentiated from metachronous lesions.

In Western countries endoscopic mucosal resection (EMR) has been widely accepted for local treatment of mucosal neoplasia in Barrett’s esophagus and the colorectum. A large prospective trial demonstrated CLR in more than 95 % of patients with early Barrett’s neoplasia [6]. In another series on selected patients with Barrett’s esophagus and limited extension of mucosal adenocarcinoma the same endpoint was achieved in 99 of 100 patients in spite of a rate of en bloc resection of only 70 %, indicating that piecemeal resection was required in every third patient. Histology could confirm tumor-free vertical margins in all cases that allowed an appropriate determination of the risk of lymph node metastases. In contrast, histology demonstrated R0 resection in only one-third of the cases [5]. The frequent need for piecemeal resection may explain local recurrences in 11 % of the patients during a median follow-up period of approximately 3 years. Such lesions can frequently not be differentiated from metachronous neoplasia in Barrett’s esophagus. All of these lesions were removed endoscopically and there was no progress to more advanced cancer stages. The calculated 5-year survival rate was 98 %.

Numerous Western studies have demonstrated that EMR is also effective and safe for curative treatment of colorectal mucosal neoplastic lesions. It has to be frequently performed in a piecemeal fashion, particularly in cases of larger sessile or flat lesions but histology enables the identification of cases with adenocarcinoma and selection of those patients who will have to undergo surgery, particularly when the tumor infiltrates the submucosa to a depth greater than 1 mm. A recent retrospective study demonstrated a success rate of EMR in 95 % of 161 cases with larger sessile or flat neoplasia [7]. There was no case of perforation. About one-third of the patients had to undergo more than one endoscopic intervention but surgery for failed clearance was required in only 4.6 %.

The majority of studies on EMR were reported from tertiary referral centers. How do these results compare to a 1-year survey among all French gastroenterologists reported in this issue of Endoscopy [8]? One-third of the French gastroenterologists who perform EMR participated in the trial, which included 1202 patients and 1335 EMRs. Only 10 % of the resections were performed in the upper gastrointestinal tract, indicating a low case volume for the majority of participating gastroenterologists. En bloc resection rates were 62 % and 73 % for the upper and lower gastrointestinal tract, respectively, which correspond to published data. Carcinomas with submucosal infiltration were identified in 38 of all included patients (3.2 %). The depth of invasion was specified in 92 % of the cases. According to guidelines, elective surgery for more advanced tumor stages would have been necessary in only 1.6 % of all patients. The EMR-related complication rate was 5.3 %, and perforation occurred in only 0.5 % of the interventions. No significant association was seen between the case volume of endoscopists and either the procedure-related morbidity or en bloc resection rate. These short-term results of a survey can of course not be compared with those from a prospective clinical trial, due to lack of a protocol and an appropriate follow-up. However they confirm that the number of patients with early upper gastrointestinal neoplasia is limited in a Western country. EMR achieves high rates of a potentially curative treatment and can be safely performed even by endoscopist with a limited case volume. Only a minority of patients who undergo EMR have cancer with submucosal invasion and only a few cases have to undergo elective surgery according to histopathological results.

Can these promising results of EMR be further improved by endoscopic submucosal dissection (ESD)? In Japan ESD has become the treatment of choice for early gastric cancer and is increasingly also used for early esophageal and colorectal neoplasia. A recent meta-analysis demonstrated significantly higher rates of en bloc resection and curative resection for ESD compared with EMR [9]. However all of the analyzed randomized controlled trials were performed in Asia. Outcome of ESD may be less favorable in Western countries due to the limited expertise in this challenging and potentially hazardous technique. A survey among European endoscopists, who publish mainly in highly demanding medical journals, clearly indicated that ESD is still not common practice [10]. Only 20 centers provided data on gastric ESD, which was mainly performed by a single endoscopist. Each endoscopist had treated a mean number of only four cases during the previous year of the survey, with a total number of 510 cases of ESD. Histologically confirmed complete resection of early gastric neoplasia was achieved in 77 % of all cases. The rate of major complications was 13 %. The authors conclude that European gastroenterologists are still beginners in performing ESD but achieve a high rate of efficacy. The procedure-related morbidity seems to be higher than in most Japanese series but the authors assume that it could probably be reduced by proper training. These data are interesting and important but we do not know whether or not they are representative for Europe. They do not provide an intention-to-treat analysis and an appropriate follow-up. An adequate Western evaluation of ESD would require a prospective large-scale multicenter trial, which is currently not available, and would appear to be difficult to perform in view of the low number of cases of early gastric neoplasia in Western populations.

A panel of experts should be congratulated for reaching a consensus on several important general statements on ESD in Europe [11]. Most importantly, ESD should meet quality standards, should be performed following national guidelines, those issued by The European Society of Gastrointestinal Endoscopy (ESGE) or under institutional review board (IRB) approval, and all cases should be registered. The panel recommends structured training courses and a stepwise clinical approach starting with ESD in the rectum, then in the distal stomach, colon, proximal stomach, and finally in the esophagus. In contrast to Japan, the rectum and not the stomach is suggested as the initial target organ for training in patients. The main reason is not that ESD is easier to learn in the rectum (at least there is no proof for this assumption) but that the case volume of early rectal neoplasia is much higher than that for gastric neoplasia in Western countries. Does this justify ESD instead of the well established and safe technique of piecemeal EMR? In this issue of Endoscopy, experts from the University of Tokyo report on long-term outcomes of ESD for colorectal neoplasia in 310 cases [12]. En bloc resection was achieved in 90 % of the cases and the margins of the specimens were tumor-free in 75 %. However, do these excellent results from experts outweigh the difficulties and long duration of the procedure and a perforation rate of 5 %? Patients with lesions limited to the mucosa probably do not benefit from a better histology provided by en bloc resection compared with piecemeal resection because there is no risk of lymph node metastases even in those with high-grade intraepithelial neoplasia or ”mucosal cancer.“ In addition, piecemeal EMR would probably also reveal patients with deeper submucosal tumor infiltration who would require surgery. A clinically relevant advantage of en bloc resection could be a more precise histopathology of a complete resection of lesions with a submucosal invasion of less than 1 mm. This group of patients does not usually require subsequent surgery. However histology revealed superficial invasion of the submucosa in only 37 of all of the 310 lesions (12 %). Nearly half of these cases represented neoplastic scars or lateral spreading tumors of the nongranular type, which are extremely difficult to resect and require large expertise in ESD. In spite of the extensive skill of the group, a perforation rate of 13 % was registered when ESD was performed in neoplasia associated with scars. Nonsurgical treatment of this complication is usually successful in centers specialized in ESD, but it has to be shown that these results can be reproduced by beginners. Long-term outcome of this trial demonstrated a low recurrence rate of only 2 %, which is substantially lower compared with the majority of trials on colorectal EMR without consideration of re-treatment, which usually achieves definitive complete remission.

The esophagus could be another organ for appropriate training of ESD in Western countries in view of the increasing incidence of early Barrett’s neoplasia. However ESD is more difficult in the esophagus compared with the stomach because of the challenging anatomy and frequently associated fibrosis and scar formation related to gastroesophageal reflux disease (GERD-)induced inflammation. Recently presented interim analyses of two prospective trials showed that ESD can be safely performed by Western endoscopists after training in an experimental setting and participation in clinical studies on gastric ESD. The endpoint of one of these trials was en bloc resection of neoplastic lesions in Barrett’s esophagus [13]. This could be achieved in the majority of cases but lateral tumor-free margins were histologically confirmed in less than one-third of the cases, indicating difficulties in delineation of the margins before ESD. The aim of the other study was to achieve a complete resection not only of the neoplastic area but also of the non-neoplastic epithelium [14]. According to histology this approach failed in one-third of the cases. The wider extension caused strictures in 44 % of the patients. Complete response of neoplasia was achieved in all cases but was not different from a comparative group who underwent cap-assisted EMR.

In conclusion, recent studies from Asian countries reconfirm that ESD is superior to EMR in terms of higher rates of en bloc resection and histologically proven complete resection of early gastrointestinal neoplasia. In addition it is associated with a lower recurrence rate. However these advantages may have only a minor impact on the clinical outcome. Appropriate piecemeal EMR also identifies patients with an increased risk of lymph node metastases in whom surgery is indicated. The more frequent incomplete resection and higher recurrence rates of EMR can be managed by early follow-up endoscopies and re-treatment of usually minor neoplastic remnants. This approach seems to achieve comparably high rates of complete remission of neoplasia. In patients with early Barrett’s neoplasia, EMR can be followed by radiofrequency ablation of non-neoplastic areas to reduce the risk of recurrences or metachronous lesions without the risk of stricture formation, as seen after widespread endoscopic resection [15].

The level of evidence to show a superiority of ESD to EMR in terms of clinical outcome is still low. However ESD could become the method of choice for local treatment of neoplasia in Western countries because of its potential advantages, provided that the technique can be simplified and safety can be improved. This goal may be achievable by technical improvements and structured training. To date ESD seems to be justified only in selected cases such as early gastric cancer (as there are inappropriate data on EMR in the stomach), neoplastic lesions that are difficult to remove by EMR (e. g. because of submucosal fibrosis), and lesions that are suspicious for submucosal invasion and therefore require complete histological evaluation.

The dilemma for Western endoscopists is that these cases with the optimal indication are the most difficult to treat. On the other hand ”easy“ lesions (e. g. lateral spreading rectal tumors of the granular type) can also be well managed by piecemeal EMR probably with fewer complications. In view of these open questions and ethical issues it is doubtful that ESD can be appropriately evaluated by surveys or registries in Western countries. The aim should be to demonstrate that results from Japan can be reproduced in countries with a different group of patients and a limited skill in ESD, otherwise surgeons can easily argue in favor of radical surgery because a reference to studies from Japan alone is insufficient to justify a local endoscopic procedure without any appropriate published data from the Western hemisphere. To overcome this problem, centers with a high case volume of early gastrointestinal neoplasia and EMR should cooperate and initiate prospective, preferably controlled, trials on ESD under IRB approval. Enthusiasts of ESD who have only a few potential candidates for treatment should refer them to cooperating and participating centers. Many of these institutions will offer training courses and live workshops on ESD. In addition, Western and Asian centers should collaborate closely in terms of training, exchange of data, and initiation of international multicenter trials.

Competing interests: None

References

• 1 Fujishiro M. Perspective of the practical indications of endoscopic submucosal dissection of gastrointestinal neoplasms.  World J Gastroenterol. 2008;  14 4289-4295
• 2 Larghi A, Lightdale C, Memeo L. et al . EUS followed by EMR for staging of high-grade dysplasia and early cancer in Barrett’s esophagus.  Gastrointest Endosc. 2005;  62 16-23
• 3 Moss A, Bourke M J, Hourigan L F. et al . Endoscopic resection for Barrett’s high-grade dysplasia and early esophageal adenocarcinoma: an essential staging procedure with long-term therapeutic benefit.  Am J Gastroenterol. 2010;  105 1276-1283
• 4 Pech O, Gossner L, Manner H. et al . Prospective evaluation of the macroscopic types and location of early Barrett’s neoplasia in 380 lesions.  Endoscopy. 2007;  39 588-593
• 5 Ell C, May A, Pech O. et al . Curative endoscopic resection of early esophageal adenocarcinoma (Barrett’s cancer).  Gastrointest Endosc. 2007;  65 3-10
• 6 Pech O, Behrens A, May A. et al . Long-term results and risk factor analysis for recurrence after curative endoscopic therapy in 349 patients with high-grade intraepithelial neoplasia and mucosal adenocarcinoma in Barrett’s esophagus.  Gut. 2008;  57 1200-1206
• 7 Arebi N, Swain D, Suzuki N. et al . Endoscopic mucosal resection of 161 cases of large sessile or flat colorectal polyps.  Scand J Gastroenterol. 2007;  42 859-866
• 8 Heresbach D, Kornhauser R, Seyrig J A. et al . A national survey of endoscopic mucosal resection for superficial gastrointestinal neoplasia.  Endoscopy. 2010;  42 806-813
• 9 Cao Y, Liao C, Tan A. et al . Meta-analysis of endoscopic submucosal dissection versus endoscopic mucosal resection for tumors of the gastrointestinal tract.  Endoscopy. 2009;  4 751-757
• 10 Ribeiro-Mourão F, Piemental-Nunes P, Dinis-Ribeiro M. Endoscopic submucosal dissection for gastric lesions: results of an European inquiry.  Endoscopy. 2010;  42 814-819
• 11 Deprez P H, Bergman J J, Meisner S. et al . Current practice with endoscopic submucosal dissection in Europe: position statement from a panel of experts.  Endoscopy. 2010;  42 853-858
• 12 Niimi K, Fujishiro M, Kodashima S. et al . The long-term outcomes of endoscopic submucosal dissection for colorectal neoplasms.  Endoscopy. 2010;  42 723-729
• 13 Neuhaus H, Mayershofer R, Charton J P. et al . Endoscopic submucosal dissection (ESD) of early Barrett’s neoplasia (EBN) with a water-jet hybridknife (ESDH): first prospective clinical trial.  Gastrointest Endosc. 2010;  71 AB 125
• 14 Deprez P H, Piessevaux H, Aouattah T. et al . ESD in Barrett’s esophagus high grade dysplasia and mucosal cancer: prospective comparison with CAP mucosectomy.  Gastrointest Endosc. 2010;  71 AB126
• 15 Pouw R E, Wirths K, Eisendraht P. et al . Efficacy of radiofrequency ablation combined with endoscopic resection for Barrett’s esophagus with early neoplasia.  Clin Gastroenterol Hepatol. 2010;  8 23-29