Endoscopic submucosal dissection (ESD) is useful for submucosal tumors (SMTs) within the superficial submucosal layer, but perforation frequently occurs during ESD for SMTs located at the deeper layer. Endoscopic resection for small esophageal SMTs is acceptable, although candidates for endoscopic removal are rare. Laparoscopic assistance will be effective for minimally invasive endoscopic local resection for certain types of gastric SMT. Endoscopic mucosal resection with a ligation device would be better than ESD for rectal carcinoid in terms of simplicity and effectiveness.
Key points
- •
Endoscopic submucosal dissection (ESD) is useful for submucosal tumors (SMTs) within the superficial submucosal layer, but perforation frequently occurs during ESD for SMTs located at the deeper layer.
- •
Endoscopic resection for small esophageal SMTs is acceptable, although candidates for endoscopic removal are rare.
- •
Laparoscopic assistance will be effective for minimally invasive endoscopic local resection for certain types of gastric SMT.
- •
Endoscopic mucosal resection with a ligation device would be better than ESD for rectal carcinoid in terms of simplicity and effectiveness.
Introduction—expansion of indication for ESD
Endoscopic submucosal dissection (ESD) has been gaining acceptance worldwide over the past decade due to an enthusiastic introduction by experts as well as clinical evidence showing clinical merits of this technique. The prominent advantage of ESD is that gastrointestinal epithelial lesions can be resected endoscopically in an en bloc fashion regardless of size or the existence of scar formation. ESD has allowed patients with early-stage gastrointestinal neoplasms to receive curative resection without possible organ loss. Owing to the minimally invasive nature of endoscopic intervention, surgery has gradually been replaced by this promising technique.
The ESD procedure is composed of the 2 following phases: circumferential mucosal incision and dissection of the connective tissue just below the lesion under direct visualization. If a resected lesion has a sufficient lateral margin from the circumferential incision and a sufficient vertical margin through subtumoral dissection, complete resection can theoretically be accomplished, implying that ESD can be applied not only to epithelial neoplasms but also to nonepithelial tumors. Indeed, reports of attempts and successes of ESD for submucosal tumors (SMTs) including carcinoid tumors have been increasing over recent years. In this review, the current application of ESD for gastrointestinal tumors other than cancers is summarized, after the characteristics of SMT and carcinoid tumor are reviewed.
Introduction—expansion of indication for ESD
Endoscopic submucosal dissection (ESD) has been gaining acceptance worldwide over the past decade due to an enthusiastic introduction by experts as well as clinical evidence showing clinical merits of this technique. The prominent advantage of ESD is that gastrointestinal epithelial lesions can be resected endoscopically in an en bloc fashion regardless of size or the existence of scar formation. ESD has allowed patients with early-stage gastrointestinal neoplasms to receive curative resection without possible organ loss. Owing to the minimally invasive nature of endoscopic intervention, surgery has gradually been replaced by this promising technique.
The ESD procedure is composed of the 2 following phases: circumferential mucosal incision and dissection of the connective tissue just below the lesion under direct visualization. If a resected lesion has a sufficient lateral margin from the circumferential incision and a sufficient vertical margin through subtumoral dissection, complete resection can theoretically be accomplished, implying that ESD can be applied not only to epithelial neoplasms but also to nonepithelial tumors. Indeed, reports of attempts and successes of ESD for submucosal tumors (SMTs) including carcinoid tumors have been increasing over recent years. In this review, the current application of ESD for gastrointestinal tumors other than cancers is summarized, after the characteristics of SMT and carcinoid tumor are reviewed.
SMT
A submucosal tumor is defined as a neoplastic lesion located within the gastrointestinal wall derived from tissue layers other than the epithelium. Representative SMTs are gastrointestinal stromal tumor (GIST), leiomyoma, lipoma, schwannoma, and carcinoid tumor. The incidence of each tumor varies by the organ. Because it is covered with normal epithelium, endoscopic diagnosis is sometimes difficult. Therefore, the term “SMT” can include nonneoplastic lesions as well (eg, cyst, ectopic pancreas). Some epithelial tumors or cancers grow toward the inside of the submucosal layer, which then mimics SMT and can lead to misdiagnosis. There is a consensus that GIST is the most relevant type of SMT. It is thought to be a mesenchymal tumor of mesothelial origin, especially originating from the interstitial cells of Cajal, and it has high likelihood of c-kit abnormality. Because it is difficult to distinguish between benign and malignant GIST even by pathologic assessment, GISTs are usually treated as tumors having malignant potential. In clinical settings, a treatment strategy is decided based on the risk of aggressive behavior of the tumor.
Endoscopic ultrasonography (EUS) is indispensable for accurate preoperative diagnosis. EUS offers various information about the tumor (ie, the size, location, shape, and echogenicity). Therefore, an extramural compression or cyst, mimicking an SMT, can be excluded from SMT that needs to be treated. Nevertheless, it is still difficult to distinguish GIST from other mesenchymal tumors (leiomyoma or schwannoma). When EUS shows a hypoechoic, well-demarcated SMT derived from the muscularis mucosa or the mucularis propria, endoscopic ultrasound-guided fine-needle aspiration is recommended to obtain histologic diagnosis.
In terms of treatment, a potentially malignant SMT (ie, GIST) would be a candidate for resection. Principally, any resection technique would be acceptable if en bloc resection is achieved, but endoscopic local resection is more preferable because of its minimally invasive nature. In the “pre-ESD” era, however, endoscopic en bloc local resection of SMT was technically challenging with relatively higher risk of complications. These procedures were done only by some endoscopists in the manner of modified endoscopic mucosal resection (EMR) to obtain histologic assessment of the tumor. Through the emergence of ESD, the possibility of en bloc resection increased, even for SMTs. Accordingly, ESD has been applied to SMT nowadays, although highly advanced endoscopic skills are required.
Carcinoid tumor
Carcinoid tumors, one of the neuroendocrine tumors, are most frequently found in the gastrointestinal tract, especially in the appendix and the small intestine. It arises from a deeper layer in the mucosa and easily invades the submucosa, rupturing the muscularis mucosa. Growth of this tumor is generally indolent, but it is treated as a potentially malignant tumor. Indeed, angiolymphatic infiltration or metastasis to distant organs often occurs as it progresses.
In clinical settings, small carcinoids located in the rectum, the duodenum, or the stomach would be good candidates for endoscopic resection, especially EMR. There is a certain amount of consensus that endoscopic resection in the rectum is acceptable for a carcinoid 10 mm or less in size that is confined within the submucosal layer because of a low likelihood of metastasis. On the other hand, gastric carcinoids are divided into 4 categories according to the clinicopathological features : type 1, arising multifocally in patients with autoimmune gastritis; type 2, arising multifocally in patients with multiple endocrine neoplasia type 1; type 3, arising sporadically and solely in patients without underlying diseases; and type 4, poorly differentiated neuroendocrine carcinomas, arising solely and growing aggressively. Because small carcinoids except for type 4 show indolent growth, endoscopic resection is recommended as the first choice for histologic investigation. Similarly to other neoplasms, complete resection without angiolymphatic invasion is essential to achieve curative resection.
ESD for esophageal submucosal tumor
Prevalence of SMTs in the esophagus is relatively low compared with the stomach. In most cases, esophageal SMTs are found incidentally, small and often only intermittently symptomatic, and may be considered as a candidate for endoscopic removal. In esophageal SMTs, leiomyoma and granular cell tumors are frequently seen, followed by GISTs, whereas carcinoids are quite rare. Because the number of esophageal SMTs is small and the clinicopathological behavior is still largely unknown, no established consensus regarding the treatment of esophageal SMTs exists. Therefore, the indication of treatment mainly depends on the decision at each institution as well as the patient’s preference. Several studies have shown the feasibility of endoscopic resection of esophageal SMTs, and that the optimal size for the resection is suggested to be 1 to 2 cm, considering the durability of transoral retrieval and the possibility of postoperative stricture ( Table 1 ).
Authors/Reference No./Year | Target | Average Size (cm) | Method | N | Complete Resection (%) | Perforation (%) |
---|---|---|---|---|---|---|
Chiu et al, 2006 | SMT | 1.0 | ESD | 1 | 100 | 0 |
Shi et al, 2011 | SMT-MP | 1.3 | ESD | 30 | 93.3 | 6.7 |
Liu et al, 2012 | SMT-MP | 2.1 | ESD | 14 | 92.9 | 7.1 |
Gong et al, 2012 | SMT-MP | 1.7 | ESTD | 8 | 100 | 0 |
Inoue et al, 2012 | SMT | 1.5 | ESTD | 3 | 100 | 0 |
Ye et al, 2013 | SMT | 1.8 | ESTD | 15 | 100 | 20 |
From the viewpoint of the principle of ESD, endoscopic removal is easier and safer in esophageal SMTs limited to the submucosal layer than located in the muscularis propria or deeper ( Fig. 1 ). Esophageal SMTs within the submucosal layer can be resected using conventional ESD technique, although endoscopists should be aware that the dissecting layer is thinner than that for epithelial lesions. On the other hand, ESD for SMT arising from the muscularis propria has a high possibility of perforation. Accordingly, preoperative assessment for the depth of SMT by EUS is mandatory for estimation of the origin of tumor and thus for the risk of ESD, as Shim and Jung have proposed that treatment strategy should be primarily determined by the depth of the lesion.
Park and colleagues demonstrated the feasibility of endoscopic enucleation with an insulated-tip knife with the purpose of curative resection as well as precise histologic assessment. One drawback, however, was that complete resection by histologic assessment may be difficult due to a burning effect on the surface of the tumor because the mucosa just above the lesion has to be cut longitudinally. In this regard, ESD would be simpler than the reported technique because the area that should be dissected is limited to the plane just below the tumor, which may be the reason the number of reports about the feasibility of ESD is larger than that of endoscopic enucleation.
As technical skills of ESD advance, ESD for esophageal SMTs arising from the muscularis propria has been reported. The perforation rate was relatively high (a maximum of 12.9%) compared with conventional ESD, but in most cases, it could be well treated endoscopically by clipping. Due to the invasiveness of surgery, ESD for esophageal SMTs at the deeper layers can be considered for one of the treatment modalities if secure endoscopic maneuverability and methods would be assured.
One concern that should be taken into consideration in esophageal ESD for SMT is postoperative stricture. Endoscopic enucleation by submucosal tunneling is a newly invented technique that prevents postresection stricture by preserving the covering mucosa of the SMT. Because the mucosa and the muscularis mucosae remain, the esophageal lumen can stay free of stricture. This concept was derived from per-oral endoscopic myotomy for achalasia. In per-oral endoscopic myotomy, the circular muscle bundles under the submucosal tunnel are cut longitudinally to dilate the esophageal lumen. Because the muscularis mucosa is untouched, stenosis of the lumen can be avoided. Endoscopic enucleation using the submucosal tunneling technique also requires highly advanced endoscopic skill but is promising as a less-invasive endoscopic surgery if the target lesion is small enough (less than approximately 3 cm) to retrieve through the submucosal tunnel.
ESD for gastric submucosal tumor
Gastric SMTs are not rare, and they are estimated to increase in accordance with increased use of esophagogastroduodenoscopy and improvement of diagnostic modalities. GISTs are most common among SMTs of the stomach and also are frequently considered for resection. Laparoscopic or open surgery is regarded as the standard approach; however, endoscopic resection also became available and is considered feasible for smaller lesions after establishment of ESD technique.
In the similar way of the history of development in ESD, various attempts at endoscopic removal of SMTs in gastrointestinal tract have been started from the resection of gastric submucosal lesions. Lee and colleagues reported the feasibility of ESD in 12 gastric SMTs arising from muscularis propria without perforation, although the complete resection rate was not very high (75%). They also suggested that EMR using the cap method was helpful for remnant SMTs after ESD. However, en bloc resection is still desirable for accurate histologic analysis as well as the purpose of curative resection. For SMTs arising from the muscularis propria especially, en bloc resection may lead to perforation. Some reports have shown the safety of ESD for SMTs with sacrificed complete resection rate, whereas other reports revealed feasibility of complete resection by ESD with a risk of complication ( Table 2 ).
Authors/Reference No./Year | Target | Average Size (cm) | Method | N | Complete Resection (%) | Perforation (%) |
---|---|---|---|---|---|---|
Lee et al, 2006 | SMT-MP | 2.1 | ESD | 12 | 75 | 0 |
Hoteya et al, 2009 | SMT-MM | 3.8 | ESD | 9 | 100 | 0 |
Bialek et al, 2012 | SMT | 2.5 a | ESD | 37 | 81.1 | 5.4 |
Li et al, 2012 | SMT-MP at EGJ | 1.8 | ESD | 143 | 94.4 | 4.2 |
Zhang et al, 2012 | SMT-MP at EGJ | 1.6 | ESD | 68 | 95.6 | 10.3 |
Zhou et al, 2011 | SMT | 2.8 | EFTR | 26 | 100 | — |
Schlag et al, 2013 | SMT | 1.5 | EFTR | 14 | 100 | — |