Mucosal Resection of the Esophagus

Fig. 2.1

Example of Barrett’s esophagus using high-definition white-light endoscopy

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig2_HTML.jpg — Fig. 2.2

Example of Barrett’s esophagus under narrow-band imaging (NBI)

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig3_HTML.jpg — Fig. 2.3

Visible lesion under white-light endoscopy using near focus

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig4_HTML.jpg — Fig. 2.4

Example of nodularity under white-light endoscopy

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig5_HTML.jpg — Fig. 2.5

Example of nodularity using narrow-band imaging (NBI) with near focus

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig6_HTML.jpg — Fig. 2.6

EMR for diagnosis and management of nodular Barrett’s. Footnote: (a) Nodular lesion under white-light endoscopy. (b) Nodular lesion under NBI. (c) Ligation-assisted EMR of nodular lesion. (d) Resection specimen. (e) Low-magnification histology demonstrating intramucosal carcinoma. (f) Lymphovascular invasion by intramucosal carcinoma

From a resection standpoint, EMR only allows for local removal of a lesion, highlighting the importance of proper selection of patients. Specifically in terms of BE, only patients with high-grade dysplasia or EAC limited to the mucosa (M1, confined to the epithelium; M2, confined to the lamina propria; or M3, invading the muscularis mucosae) with minimal risk of lymph node metastasis should be referred for EMR. The risk of lymph node metastasis for tumors limited to the mucosa has been reported as 0–3% [20].

Contention lies in EAC with submucosal involvement as esophagectomy has been considered the standard of care for quite some time given the high risk of lymph node involvement, but as will be discussed later in this chapter, certain superficial submucosal cancers may be amenable to EMR.

Currently, under the National Comprehensive Cancer Network guidelines , endoscopic therapy is recommended for patients with lesions limited to the epithelium (Tis or HGD), lamina propria, or muscularis mucosa (T1a). Endoscopic therapy can also be considered for lesions involving the superficial submucosa (T1b) – in lieu of esophagectomy – in the absence of lymph node metastasis, lymphovascular invasion, or poorly differentiated tumors. Discussion with the surgeon regarding the risk of esophagectomy vs. the risk of lymph node metastasis should also be undertaken [21]. The American Society for Gastrointestinal Endoscopy (ASGE) guidelines, for comparison, recommend endoscopic resection of all visible lesions (strong recommendation, moderate quality of evidence) while recommending against surgery in patients with high-grade dysplasia/intramucosal carcinoma (strong recommendation, very low quality of evidence) [22].

Squamous Cell Carcinoma

Esophageal EMR was one of the first endoscopic techniques described in removing early squamous cell carcinomas (SCCs) in the esophagus [23]. EMR remains a popular treatment option for SCC that is confined to the mucosa, which requires early detection given the ease of metastatic spread of SCC owing to the relatively thin wall of the esophagus as well as its rich lymphatic network. More common in Asia, it is associated with a poor prognosis, and EMR is indicated in SCC M1 and M2 lesions (confined to the epithelium or lamina propria) [24]. In M3 (invading the muscularis mucosae) or SM1 (invasion to the superficial third of the submucosa) lesions, EMR can be considered if there is no evidence of lymph node involvement. Similarly, ESD has shown complete resection rates of 78–100% of M1 and M2 lesions with low recurrence rates (0–2.6%), but like EMR is generally not utilized in SCC lesions with lymphovascular invasion or submucosal invasion >200 μm [25].

While ESD is not the focus of this chapter, it is important to note when comparing EMR to ESD in esophageal mucosal cancers 20 mm or less, ESD has been found to provide an en bloc resection rate of 100%, whereas for EMR it was 87% (cap assisted) and 71% (two-channel technique) [26, 27]. The curative resection rate for ESD was also 97%, significantly higher than either EMR technique (71% for cap and 46% for two-channel). Therefore, the European Society of Gastrointestinal Endoscopy (ESGE) recommends ESD as the preferred method of endoscopic resection of esophageal squamous cell carcinoma (ESCC) given the higher en bloc resection rate and superior histological assessment [28]. EMR can be used for lesions smaller than 10 mm if en bloc resection can be assured. For early EAC and HGD, EMR is still the mainstay method of endoscopic resection; but ESD can be considered for lesions larger than 15 mm, poorly lifting tumors, and lesions with high-risk features for submucosal invasion.

The National Comprehensive Cancer Network guidelines recommend endoscopic therapy for patients with lesions limited to the epithelium (Tis or HGD), lamina propria, or muscularis mucosa (T1a). For patients with T1b lesions, esophagectomy is recommended [21].

EMR Technique in the Esophagus

At its core, EMR involves the removal of a lesion using a snare, with or without electrocoagulation (Table 2.1). There have been a variety of techniques that have been developed for resection of esophageal lesions, accounting for the relatively flat nature of many of these lesions. Prior to EMR, our practice is to thoroughly visualize the lesion using high-definition white-light endoscopy (WLE) and narrow-band imaging (NBI). While using virtual chromoendoscopy, we then mark the boundaries of the target lesion, usually 3–5 mm from the lesion border (Fig. 2.7), using various modalities of coagulation such as argon plasma coagulation (APC) or simply using a snare tip with soft coagulation. We recommend these markers, particularly in lesions requiring multiple resections, owing to the difficulty of recognizing the margins of the lesion during the actual resection given the effect of coagulation, bleeding, and submucosal injection in obscuring the working field. After the resection, confirming resection of the entire lesion is done by ensuring that all markers are no longer visible. It is important to note that the ASGE recommends against routine complete endoscopic resection of the entire Barrett’s segment, but instead supporting resection of the visible lesion followed by ablation of the remaining Barrett’s segment [22]. Should the segment of Barrett’s contain diffusely nodular BE, however, we will perform EMR of the entire segment.

Table 2.1

Steps for esophageal endoscopic mucosal resection

1. Visual inspection of lesion using high-definition white-light endoscopy and narrow-band imaging

2. Marking of borders of lesions (with snare tip or argon plasma coagulation) under visualization with WLE or narrow-band imaging

3. Resection of lesion

Injection assisted: Submucosal injection for lifting of target area, followed by snare resection

Ligation assisted: Band ligation of target area followed by snare resection

Cap assisted: Submucosal injection followed by suctioning of target area inside cap, followed by enclosing the snare and then resection

4. Thorough inspection of resection site to ensure no residual tissue, bleeding, or deep injury/perforation

5. Treatment of any bleeding sites or closure of defects if needed

6. Pinning of specimen onto Styrofoam board

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig7_HTML.jpg — Fig. 2.7

Example of using coagulation markers to mark boundaries of visible lesion

Injection-Assisted EMR

The original technique for EMR incorporates the use of solution injection into the submucosal space below the lesion to effectively create a safety barrier for resection. This injection creates a “lifting” and separation of the lesion from the underlying muscularis propria, which facilitates snaring of the intended target. Furthermore, this injection minimizes damage from electrocautery or physical forces to the muscularis propria and therefore minimizes the risk of perforation. While saline was first used as the injection solution, a wide variety of injection solutions (particularly viscous solutions) have been developed to facilitate lifting owing to the quick dissipation of saline into the adjacent space [29]. While not comprehensive, other commonly utilized solutions include dextrose 50% (an inexpensive, easily available hypertonic solution), succinylated gelatin (clear, inexpensive, safe colloid solution), hydroxyethyl starch (safe and inexpensive solution which can maintain submucosal lifting longer than saline), sodium hyaluronate (a highly viscoelastic but expensive solution), hydroxypropyl methylcellulose (HPMC, a readily available and viscoelastic solution), and hyaluronic acid (HA, a glycosaminoglycan with high viscosity that is expensive) [30]. Additionally, dilute epinephrine (typically 1:100,000) can be added into the injection solution to reduce the bleeding during the resection which helps maintain better visualization throughout the procedure. Varying amounts of injection solution can be added at the endoscopist’s discretion depending on the size of the lesion.

Ligation-Assisted EMR (Video Included)

Perhaps the simplest and most widely used of the EMR techniques, ligation-assisted EMR, also known as multiband mucosectomy (MBM), involves the use of a band ligation device [31, 32]. Similar to an esophageal varices banding device, a band ligation system includes a distal attachment cap for the endoscope with a trigger cord and control handle that goes through the working channel. This transparent cap contains six rubber bands, and once the ligation cap is maneuvered to be directly over the desired lesion, the lesion is suctioned into the cap using the endoscope’s suction system . A single band is then released with clockwise rotation of the control handle, with a distinct releasing sensation felt by the endoscopist signaling deployment of the band (Fig. 2.8). This band will help create a pseudopolyp, and submucosal injection is not routinely required as the esophageal muscle layer retracts when ligated by the rubber band. An electrocautery snare can then be inserted through the working channel to cut the lesion either above or below the rubber band (Figs. 2.9 and 2.10). Piecemeal resection for larger lesions can be performed by repeating this process until complete resection is performed. The most commonly used ligation devices for this technique are the Duette Multi-Band Mucosectomy Kit (Fig. 2.11, Cook Medical, Winston-Salem, NC, USA) which incorporates a 7 Fr hexagonal snare and the Captivator EMR Device (Boston Scientific, Natick, MA, USA) which also utilizes a 7 Fr hexagonal snare. There is no consensus on the type, and settings of diathermy used for resection are widely variable in practice. In our practice, we mostly use blended current using Endo Cut Q (effect 3, cut duration 1, cut interval 6) and forced coagulation (effect 2, 50 watts).

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig8_HTML.jpg — Fig. 2.8

Ligation-assisted endoscopic mucosal resection

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig9_HTML.jpg — Fig. 2.9

Snare used underneath rubber band

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig10_HTML.jpg — Fig. 2.10

Mucosal defect after ligation-assisted resection

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig11_HTML.jpg — Fig. 2.11

Duette Multi-Band Mucosectomy Kit (Cook Medical, USA)

Cap-Assisted EMR

In this technique, a transparent cap is attached to the distal end of the endoscope with a variety of caps available, with either a straight or oblique shape (Fig. 2.12). A submucosal injection is often used to create a cushion and aid in suctioning. A crescent-shaped electrocautery snare is passed through the biopsy channel and opened inside the cap. The snare is then positioned within the internal circumferential ridge at the tip of the cap. Once the snare is in good position, the cap is placed over the lesion, followed by suctioning of the lesion into the cap. Once complete suctioning of the lesion into the cap is achieved, the snare is closed with electrocoagulation, effectively removing the lesion. Piecemeal resection can be used for larger lesions by repeating this process until the entire lesion is resected (Fig. 2.13).

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig12_HTML.jpg — Fig. 2.12

Cap-assisted resection kit displaying oblique cap (Olympus, Japan)

../images/442926_1_En_2_Chapter/442926_1_En_2_Fig13_HTML.jpg — Fig. 2.13

Mucosal defect after piecemeal cap-assisted endoscopic mucosal resection

Once EMR is performed, the specimen is typically pinned down using thin pins on a Styrofoam board. This pinning helps preserve the orientation of the specimen for accurate histologic analysis and is crucial for determining the depth of invasion and whether negative horizontal and vertical margins have been obtained [33].

Safety and Efficacy

EMR used for the indications previously described is an effective, safe, and durable therapy [34–37]. EMR has been widely studied in the management of Barrett’s esophagus patients with high-grade dysplasia and intramucosal EAC . Pech et al. evaluated 1000 patients who received EMR for Barrett’s-associated mucosal adenocarcinomas using the cap-assisted and ligation-assisted techniques and found that complete remission was achieved in 96.3% of patients with only 2 deaths related to the cancer during a mean follow-up period of 56.6 months [37]. Major adverse events such as bleeding and perforation occurred in 1.5% of patients, and all adverse events were able to be managed endoscopically. Similarly, in the EURO-II trial, Phoa et al. found that EMR followed by radiofrequency ablation (RFA) resulted in complete eradication of intestinal metaplasia (CE-IM) and neoplasia (CE-N) rates of 87% and 92%, respectively, in patients with high-grade dysplasia or EAC. In terms of high-grade dysplasia, a meta-analysis by Tomizawa et al. compiled data on studies where EMR was used on the entire Barrett’s segment [38]. Reporting on 8 studies involving 676 patients, CE-IM and CE-N occurred in 85% and 96.6% of patients, respectively. Furthermore, Haidry et al. examined long-term outcomes from a UK registry of 500 patients with Barrett’s-associated neoplasia, finding a significant increase in CE-IM (56–83%) and CE-N (77–92%) with an increase in EMR (48–60%) for visible lesions [39]. Given the high stricture rate (37.4%) of circumferential EMR, ablation methods such as radiofrequency ablation (RFA) are most frequently used in the management of flat (non-nodular) high-grade dysplasia, but EMR still plays an important diagnostic and therapeutic role in the management of nodular lesions, which may harbor areas of high-grade dysplasia or adenocarcinoma.

In terms of SCCs, Yamashina et al. examined 402 patients with mucosal and submucosal SCC who were treated endoscopically at a single center, of which 194 patients received EMR (median lesion size of 20 mm) as their resection method (208 patients received ESD) [40]. All patients had complete local remission for all lesions after a single procedure. During a mean follow-up period of 50 months, 5-year survival rates of 90.5%, 71.1%, and 70.8% were found for SCC limited to the epithelium/lamina propria, muscularis mucosa, and submucosa, respectively. Cumulative 5-year metastasis rates were 0.4%, 8.7%, 7.7%, and 36.2% for tumors limited to the epithelium/lamina propria and muscularis mucosa and with submucosal invasion to 0.2 mm and submucosal invasion more than 0.2 mm, respectively. Adverse events included perforation (0.2%), bleeding (0.2%), and strictures (13.2%). Yoshii et al. retrospectively examined 44 patients who underwent EMR for T1a (54.6%) and T1b (45.4%) SCC [41]. Within a median follow-up period of 51 months, two patients (4.5%) died from primary SCC, while four patients (9.7%) developed lymph node metastases. Adverse events included perforation (2.2%) and strictures (20%). Extrapolating from this data, it appears that EMR for SCC is safe and effective when limited to use in superficial lesions and can be considered for submucosal use if patients are not candidates for esophagectomy or unwilling to undergo surgical options.

The variety of EMR techniques prompts the question as to which technique is more effective. In terms of Barrett’s-associated neoplasia, ligation-assisted EMR has become the more popular method, while in early squamous cell neoplasia, cap resection represents the most widely used technique. As mentioned above, the cap-assisted method is more technically demanding, particularly with piecemeal resections as submucosal lifting and repositioning of the snare is required for each resection. Zhang et al. performed a randomized controlled trial comparing the two methods (n = 42 in both groups) for squamous cell neoplasia and found that while complete endoscopic resection was achieved in all lesions, the procedure time was significantly shorter with the ligation method (11 vs. 22 minutes) and associated costs were lower in the ligation method [31]. Pouw et al. also compared the two methods in a randomized controlled trial for resection of high-grade dysplasia and intramucosal EAC, finding that ligation-assisted EMR was faster (34 minutes vs. 50 minutes) and cheaper (€240 vs. €322) with no significant difference in depth of resection [42]. Thus, ligation-assisted EMR appears to be quicker and easier, but selection of the EMR technique will rest on the preference of each endoscopist.

An alternative technique to EMR is ESD, which was originally designed for gastric lesions, but can be used for esophageal lesions as well. While this technique will be described in detail in other chapters, ESD utilizes an endoscopic knife to create a circumferential incision after submucosal injection to allow for an en bloc resection (Fig. 2.14) [1, 43–45]. Few studies have compared ESD with EMR directly, but a randomized trial by Terheggen et al. compared the efficacy and safety of ESD (n = 20) and EMR (n = 20) in patients with Barrett’s-associated neoplasia [46]. Although the ESD group did have a higher rate of margin-free resection, there was no difference in complete remission from neoplasia at 3 months, and a recurrent case of esophageal adenocarcinoma was seen once in the ESD group. As expected, ESD was also significantly longer than EMR (54 minutes vs. 22 minutes); and the only cases of perforation occurred in the ESD group (n = 2), although this was not statistically significant. Guo et al. performed a meta-analysis comparing both procedures for superficial esophageal cancers (total of 8 studies involving 1080 patients, all in Asia) and found that while ESD had a higher en bloc and curative resection rate, it also had a higher operative time and perforation rate with no differences in stricture and bleeding rates [44]. Therefore, individual expertise in each procedure will likely play a significant role in determining which technique to use in the treatment of superficial esophageal lesions.