Colorectal endoscopic submucosal dissection: when and by whom?

Endoscopic submucosal dissection (ESD) was developed in Japan more than a decade ago for the treatment of early gastric cancer (EGC) [1]. The drivers for this remarkable innovation were the heavy burden of disease and the desire to develop a minimally invasive and definitive therapy for the numerous EGCs detected in the screening program. Embracing simple oncological principles, the aim was to realize a technique to achieve reliable en bloc excision of appropriately selected gastric neoplasms with free lateral and deep margins. Although no randomized trial exists or is ever likely to be conducted, in EGC it has steadily become the standard of care throughout large parts of the world, particularly where disease prevalence is high, due to its greater safety, markedly reduced cost, and similar therapeutic outcomes compared with surgery in cases where the risk of lymph node metastasis is low [2] The justifications for the use of ESD for EGC are multifaceted and hard to argue against, particularly as there are no other effective endoscopic therapies and surgery carries a significant risk of peri-operative morbidity and long-term digestive dysfunction.

Subsequently, ESD has been applied to premalignant and superficially invasive neoplasms in other sites including the esophagus, duodenum, and colon. Here the benefits are not so obvious. Its use in early squamous cell cancer of the esophagus is logical, and large experiences have reported favorable results [3]. In the duodenum, the results have been poor: the risk of perforation exceeds 20 % and as the majority of duodenal lesions are noninvasive adenomas, the benefit of en bloc excision is marginal, particularly when weighed against the risk of major and potentially life-altering complications [4]. The most controversial area is the application of ESD in the colon.

In Western countries, EMR is established as the method of choice for curative treatment of sessile and flat colorectal adenomas or laterally spreading tumors (LSTs) > 15 – 20 mm in size. Numerous studies have shown that EMR can be safely, effectively, and efficiently performed after a short learning period [5] [6] [7]. Major adverse events are infrequent, mean procedure time is appproximately 25 minutes, and more than 95 % of patients are discharged on the same day [5]. The main disadvantage of this widely available technique is that lesions with a diameter of > 20 mm can usually only be resected in a piecemeal fashion. This creates the two problems for which EMR is often criticized. First, it complicates histopathological assessment of the resected specimen(s) and complete excision cannot be confirmed by the pathologist. However, for benign noninvasive adenomas this is not a neccessity or requirement. Second, it increases the risk of recurrence at first follow-up due to incomplete resection. Provided scheduled surveillance is adhered to, then from an oncological point of view, neither limitation appears to be relevant in patients with any form of noninvasive lesion, even in long-term follow-up [8].

These limitations can usually be overcome by ESD, which allows en bloc resection of early colorectal neoplasia and is by far superior to EMR if curative resection is considered to be the primary goal of treatment [9]. However, ESD is technically demanding, time consuming, requires multiday post-procedural admission, and causes more complications than EMR. Under consideration of these pros and cons, colorectal ESD has been accepted as the first-line approach only for a selected group of patients in Japan [10]. The criteria include large lesions with well-established predictive factors for early cancer with minor submucosal invasion, such as nongranular LSTs and those classified as Paris type 0-IIc or 0-IIa plus IIc and/or Kudo pit pattern type V.

A recent Japanese multicenter trial of ESD for colorectal neoplasia in 1111 cases demonstrated an en bloc resection rate of 88 % [11]. Histology revealed adenocarcinoma in 19 % of these selected patients. Surgery was recommended in every second case because of adverse histological features associated with an increased risk of lymph node metastasis, predominantly deep submucosal invasion. Thus, the main oncological advantage of ESD over EMR, curative resection for early cancer, was realized in only 10 % of all patients, those with well-differentiated, superficial submucosal invasive adenocarcinoma without lymphovascular invasion. This limited clinical benefit should be weighed against the technical difficulties and long duration of the procedure (mean 116 minutes), and a perforation rate of 5 % despite being performed in experienced large-volume centers. Long-term follow-up showed a recurrence of only 2 %, which is substantially lower than EMR. This may allow extension of follow-up intervals and reduce re-interventions for removal of residual neoplasia. However, studies of EMR for early colorectal lesions show that the clinical relevance of “recurrences” is low because they are usually related to residual noncancerous neoplasia after piecemeal resection. These lesions can be easily detected and removed at the first follow-up colonoscopy at 4 – 6 months after EMR, and these patients then have a very low risk of relapse at subsequent colonoscopies [8].

In spite of a high number of patients with early colorectal neoplasia in Western countries, so far only two prospective single-center series on ESD have been reported [12] [13]. A large German referral center included 82 patients with large sessile lesions in the rectum (87 %) and the sigmoid (13 %) [12]. The rates of successful ESD, en bloc resection, and R0 resection were 92.7 %, 81.6 %, and 69.7 %, respectively. The R0 resection rate significantly increased and the procedural duration decreased with experience during the study period. Resection was incomplete in 13 of 14 resected cancers, so that additional surgery was recommended. Assuming these cancers would have been identified by piecemeal EMR, only the single patient with curative resection benefitted from ESD. Perforation and bleeding were registered in 1.3 % and 7.9 % of the patients, respectively, and none of these complications required a surgical intervention. During a median period of 2 years, residual neoplasia was registered in 6 of 65 patients. The risk was significantly higher after piecemeal resection compared with en bloc resection (41.7 % vs. 0 %). The high rate of residual neoplasia after failure of en bloc resection may indicate that there is no advantage of ESD over EMR if resection in a single piece cannot be achieved.

An Italian pilot study enrolled 40 patients with LSTs > 3 cm in diameter for ESD [13]. En bloc resection was achieved in 90 % of the cases and failures underwent piecemeal resection. The R0 resection rate was 80 %. Histopathology of the resected specimen showed cancer with submucosal invasion in two cases. R0 resection had failed in both of these cases and surgery was required. Perforation in one case and delayed bleeding in two cases could be managed endoscopically. Only one patient, in whom R0 resection had not been achieved, had residual neoplasia at follow-up endoscopies.

In view of the limited data from Western countries, the French prospective multicenter study presented in this issue of Endoscopy is important because it provides additional information on the pros and cons of ESD for rectal tumors [14]. A total of 45 cases were enrolled in 9 tertiary referral centers. The spectrum of indications was widely selected and included mainly those with no predictive parameters for cancer. Even small lesions with a diameter of 10 – 20 mm met the inclusion criteria, although they are usually amenable to en bloc EMR. The procedures were all performed under general anesthesia by experienced endoscopists who had participated in animal training programs and performed at least 10 human rectal ESD procedures. The inclusion of an external safety and data monitoring board was important, as it was necessary to suspend the study after the first 25 patients due to the high rate of complications, and three centres were closed due to low enrollment. The main aim was to achieve en bloc and curative resection of the neoplastic area by means of ESD. The study protocol allowed the combination of ESD and EMR in cases of “technical inability to continue dissection.” ESD alone succeeded in only 24 of the 45 cases, which corresponds to a technical success rate of 53 %. The combined approach of ESD plus EMR in the remaining 21 difficult cases achieved an en bloc resection rate of 24 %. The overall en bloc and R0 resection rates were 64 % and 53 %. Histology revealed cancer in only three cases. Curative resection had been achieved in one of these cases, and the other two patients required surgery. One-third of the patients had major complications, with perforation and bleeding occurring in eight and six cases, respectively. The mean duration of hospitalization in those with perforation was 5 days. Analysis of the learning curve showed that the en bloc resection rate increased from 52 % to 82 %, and the perforation rate decreased from 34 % to 0 % by the end of the study. Long-term outcomes indicated no residual neoplasia in 93 % of the cases at the first follow-up at 3 months. One patient with residual adenoma and high grade dysplasia required surgery. The rate of complete local remission was 88 % at 1 year.

How do these results compare to other studies on colorectal ESD and EMR? Despite a favorable selection of patients, the efficacy and safety of ESD was inferior to previous Western results of ESD [12] [13]. The procedure succeeded in only every second patient. Even in combination with snare resection, the R0 resection rate was limited to 53 %. The learning curve showed that the en bloc resection rate can be increased to 82 % with increasing experience of the individual endoscopists. These results correspond to data from highly specialized European referral centers but are still inferior to Japanese data. Even under the condition that ESD performed by skilled endoscopists achieves en bloc resection, the extent to which patients gain a clinical benefit compared with those in whom piecemeal EMR is performed is not clear. Surgery may become unnecessary in patients with adenocarcinoma stage T1 in whom curative resection is achieved and histology shows a well-differentiated cancer with no infiltration of lymphatic vessels and vertical invasion of the submucosa of < 1000 µm. However, this advantage over piecemeal EMR was only relevant for a single patient in each of the three European trials [12] [13] [14].

The second potential advantage of ESD over EMR is the lower recurrence rate. However, this difference seems to be only significant at the first follow-up examination because of a higher rate of remaining neoplasia after piecemeal EMR. The French trial re-confirms the results of previous studies that the risk is low after ESD, although the one case with remaining adenoma and high grade dysplasia indicates that a short-term follow-up examination is certainly needed. EMR also has this limitation, where recurrence is not uncommon. However, at the second follow-up (e. g. at 18 months after the initial intervention), the majority of patients were in complete remission from colorectal neoplasia [8]. Thus far, it has not yet been shown that ESD is superior to EMR in terms of this important parameter of clinical outcome.

These limited advantages of colorectal ESD over EMR have to be balanced against several major drawbacks. These data show that ESD is more time and resource consuming, more hazardous, and requires multiday hospital admission, even when performed by Japanese experts who have honed their ESD skills on a large volume of EGCs. The main advantage is curative resection in low-risk, early invasive adenocarcinoma, but overall this pathology is rare in Western cohorts and infrequent even in large Japanese series; thus, the vast majority of patients referred with advanced lesions for tertiary endoscopic management will gain no clinical benefit from an upfront ESD strategy, but will be subjected to all of the procedural risks.

From an ethical and clinical point of view, colorectal ESD should be primarily limited to lesions with an increased probability of early submucosal invasion ([Fig. 1]) and selected EMR failures (e. g. due to difficult cases of recurrences). Patients can be centralized to institutions that are specialized in advanced diagnostic and therapeutic endoscopy of early neoplasia in the upper and lower gastrointestinal tract. They have access to a high number of well-selected cases and can provide appropriate training programs for a limited number of endoscopists. Patients should be enrolled in carefully designed studies for further evaluation of the role of colorectal ESD in Western countries.

Fig. 1 30-mm, 0-IIa, nongranular laterally spreading tumor of the mid-rectum. Narrow-band imaging shows loss of the pit pattern centrally (Vi), suggestive of superficial invasive carcinoma. The patient was treated by endoscopic submucosal dissection (ESD) to achieve an en bloc excision. Histology showed a well differentiated 6 × 6 mm focus of invasive adenocarcinoma with superficial submucosal infiltration (150 μm) arising within a severely dysplastic tubular adenoma. This elderly patient opted for no further treatment and it is likely that cure has been achieved by ESD alone in this carefully selected case.

• References

• 1 Yamamoto H, Kawata H, Sunada K et al. Successful en-bloc resection of large superficial tumors in the stomach and colon using sodium hyaluronate and small-caliber-tip transparent hood. Endoscopy 2003; 35: 690-694
• 2 Neuhaus H. ESD around the world: Europe. Gastrointest Endosc Clin N Am 2014; 24: 295-311
• 3 Takahashi H, Arimura Y, Masao H et al. Endoscopic submucosal dissection is superior to conventional endoscopic resection as a curative treatment for early squamous cell carcinoma of the esophagus (with video). Gastrointest Endosc 2010; 72: 255-264
• 4 Bourke MJ. Endoscopic resection in the duodenum: current limitations and future directions. Endoscopy 2013; 45: 127-132
• 5 Moss A, Bourke MJ, Williams SJ et al. Endoscopic mucosal resection outcomes and prediction of submucosal cancer from advanced colonic mucosal neoplasia. Gastroenterology 2011; 140: 1907-1918
• 6 Knabe M, Pohl J, Gerges C et al. Standardized long-term follow-up after endoscopic resection of large, non-pedunculated colorectal lesions: a prospective two-center study. Am J Gastroenterol 2014; 109: 183-189
• 7 Tutticci N, Bourke MJ. Advanced endoscopic resection in the colon: recent innovations, current limitations and future directions. Expert Rev Gastroenterol Hepatol 2014; 8: 161-177
• 8 Moss A, Williams SJ, Hourigan L et al. Long term adenoma recurrence following wide field endoscopic mucosal resection (WF-EMR) for advanced colonic mucosal neoplasia is infrequent: results and risk factors in 1000 cases from the Australian Colonic EMR (ACE) study. Gut 2014; In press
• 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 Tanaka S, Terasaki M, Kanao H et al. Current status and future perspectives of endoscopic submucosal dissection for colorectal tumors. Dig Endosc 2012; 24: 73-79
• 11 Saito Y, Uraoka T, Yamaguchi Y et al. A prospective, multicenter study of 1111 colorectal endoscopic submucosal dissections (with video). Gastrointest Endosc 2010; 72: 1217-1225
• 12 Probst A, Golger D, Anthuber M et al. Endoscopic submucosal dissection in large sessile lesions of the rectosigmoid: learning curve in a European center. Endoscopy 2012; 44: 660-667
• 13 Repici A, Hassan C, Pagano N et al. High efficacy of endoscopic submucosal dissection for rectal laterally spreading tumors larger than 3 cm. Gastrointest Endosc 2013; 77: 96-101
• 14 Rahmi G, Lepilliez V, Coumaros D et al. Endoscopic submucosal dissection for superficial rectal tumors: prospective evaluation in France. Endoscopy 2014; 46: 670-676