Endoscopic Mucosal Resection: An Overview of the Value of Different Techniques

Publisher Website
Google Scholar
Abstract
Two articles in this issue of Endoscopy advocate the use of endoscopic mucosal resection (EMR), in a new treatment technique for intramucosal gastric tumors using an insulated-tip diathermic knife [ 1 ], and a novel approach to endoscopic colorectal mucosal resection using a three-channel outer tube and multiple forceps in an experimental assessment [ 2 ]. EMR, which is a recently introduced therapeutic method for resection of flat or depressed lesions of the gastrointestinal tract, has come to play an increasingly important role in the treatment of early cancer in the gastrointestinal tract. Most importantly, long-term studies which demonstrated EMR outcomes similar to those of surgery have led to the acceptance of EMR as a standard procedure [ 3 ]. Why would endoscopists consider EMR in the first place? If a lesion is to be removed endoscopically, it could be removed by either ablation or resection. Ablation may be technically easier and faster than resection, but it is usually less desirable because a specimen cannot be obtained for histological confirmation. Which lesions are suited for EMR? This procedure is potentially curative in early cancers of the gastrointestinal tract with a negligible risk of lymphatic metastasis. It has been shown in the surgical literature that the risk of lymphatic invasion increases with progressive infiltration of the mucosa and submucosa. The risk is almost nil when invasion is limited to the mucosa, very low (around 10 %) if the muscularis mucosa is invaded, and increases to rates as high as 40 % when the submucosa is invaded, depending on whether the superficial (sm1) or deeper (sm2) portions are involved [ 4 ]. Based on these findings, Japanese authors have recommended that stage m1 and m2 early cancers represent an absolute indication for mucosal resection, while m3 and sm1 cancers should represent a relative indication. A retrospective study of the 1312 solitary early gastric cancers resected at the National Cancer Center Hospital between 1962 and 1991 revealed that in mucosal cancers the incidence of metastases was extremely low at 0.64 % (3/462) in intestinal type, regardless of size, though several lesions of the diffuse type still showed metastases despite their small size [ 5 ]. In general, lymph node metastasis of early gastric cancer showed a rate of 0 % to 3 % in mucosal cancer and 9 % to 19 % in cancer invading the submucosa [ 6 ]. Various factors which should be considered prior to the performance of EMR, other than infiltrative depth of lesion, have been suggested, such as the endoscopic appearance of early cancers in the gastrointestinal tract, their size, and their differentiation grade on histological examination. Many endoscopists have established various criteria for application of EMR. Generally, the absolute indications for early gastric cancer have been those accepted by the Japanese Gastroenterological Endoscopy Society: (i) elevated-type intramucosal cancer less than 20 mm in size; (ii) depressed-type mucosal cancer without ulceration, less than 10 mm in size; and (iii) intestinal-type adenocarcinoma. In the esophagus, the criteria for EMR are as follows: superficial esophageal cancer of type IIa, IIb, or IIc, not more than 2 cm in size, and intraepithelial cancer or m1 cancer (cancer with minimal invasion into the muscularis mucosa). In a large survey of more than 2000 patients who had surgery, endoscopic treatment, or both, the incidences of lymph node metastases were 0 % and 3 % in patients with m1 and m2 stage cancers, respectively [ 3 ]. The role of EMR in the treatment of early cancer of the esophagus is not yet clear because of the limited number of reports published to date. In the colon, the criteria for EMR are as follows: well differentiated adenocarcinoma of type Is, IIb, IIa, or IIc, not more than 1 cm in size, m cancer (cancer confined to the mucosa) or smIa cancer (cancer with minimal invasion to submucosa), and without any invasion to lymphatic channels or vessels [ 7 ] [ 8 ] [ 9 ] [ 10 ] . There is a general consensus in the specification of the indications for EMR, although follow-up procedures and management of incomplete resections and recurrent cancers are still subject to differing opinions. According to data from The Cancer Institute Hospital in Tokyo, which involves more than 10 000 resected gastric cancer cases, IIc-type mucosal cancer of less than 10 mm is associated with no lymph node involvement, while IIc-type mucosal cancer of less than 20 mm has lymph node involvement in 0.4 % of cases. No lymph node metastasis is found with IIa-type mucosal cancer of less than 20 mm and IIa-type submucosal cancer of less than 20 mm also has no lymph node involvement. However, submucosal cancer of less than 30 mm shows lymph node involvement in 28.6 % of cases [ 11 ]. The differentiation in the depth tumor infiltration between mucosal and submucosal cancer in IIa-type elevated lesion is relatively problematic. In poorly differentiated carcinoma some risk of lymph-node metastasis exists even in small lesions, therefore those lesions should be generally excluded from the application of mucosal resection. With regard to EMR indications concerning the size of the lesion, some trials are now underway using piecemeal resection. When piecemeal resection is carried out, the residual portion between the first and second resected areas must not be left because it could easily develop local recurrence. Divided resection is gradually gaining general support in Japan, although it is still controversial. Amano et al. recommended an extended indication for mucosal resection in early gastric cancer as follows: all well-differentiated adenocarcinomas, mucosal cancers less than 30 mm in size without ulcer or ulcer scars; mucosal cancer less than 20 mm in size with ulcer or ulcer scar; sm1 cancer less than 20 mm in size without ulcer or ulcer scar; and poorly differentiated cancer less than 10 mm in size [ 4 ]. The rate of...

This publication has 0 references indexed in Scilit: