Esophageal Endoscopic Submucosal Dissection

Daniel S. Strand, MD

Andrew Y. Wang, MD

Endoscopic submucosal dissection (ESD) is a therapeutic endoscopic technique that enables the en bloc resection of superficial epithelial neoplasia within the alimentary track. Developed by Japanese endoscopists following the observation that piecemeal endoscopic mucosal resection (EMR) of large lesions resulted in appreciable rates of recurrence, the goal of ESD is the en bloc removal of superficial luminal neoplasms, including in the esophagus.¹,²,³ Although adoption of ESD among Western endoscopists has been slowed by a lack of formalized training and the slow release of specialized endoscopic tools, there is significant enthusiasm for this procedure as it offers the ability to provide a curative R0 resection for large dysplastic lesions and early esophageal cancers.⁴,⁵

INDICATIONS

ESD can be used to treat large dysplastic lesions or superficial carcinomas in the esophagus that are well or moderately differentiated and ideally confined to the mucosa (m1-m2 for squamous carcinoma and m1-m3 for Barrett adenocarcinoma).⁶ Such lesions confer a low risk for lymph node metastasis (LNM) and may be amenable to endoscopic cure. ESD may play a role in the cure of esophageal cancers that invade into the superficial submucosa, but the increased risk of LNM must be considered and discussed with the patient. ESD has been reported for both esophageal squamous cell carcinomas (SCCs) and Barrett adenocarcinomas.

Squamous Cell Carcinoma

Because squamous cell carcinoma develops via a field effect, margins can be difficult to discern and ESD—not EMR—should be the first option for treating superficial esophageal squamous dysplasia. Localized SCC that invades no further than the muscularis
mucosae (i.e., T1a lesions, specifically those confined to the m1 and m2 layers) is the ideal target for esophageal ESD, given a low risk for LNM.⁷ M3 (deep mucosal) and early T1b (submucosal involvement of <200 µm) SCCs with no evidence of LNM by endoscopic ultrasound (EUS) or cross-sectional imaging can be considered for ESD. These lesions may have an increased risk of LNM up to 15%, depending on size. Lesions with deeper submucosal invasion have an unacceptably high rate of LNM, and ESD should be avoided in patients fit for surgery.⁸

Esophageal Adenocarcinoma

Similar to SCC, the principal indication for ESD in esophageal adenocarcinoma (EAC) arising from Barrett esophagus is a T1a lesion limited to the mucosa or lamina propria, as the risk of LNM is predicted to be less than 2%.⁹ This level of LNM is rational, because surgical esophagectomy has a similar risk of overall mortality (2%) and significant additional morbidity. Extension of this indication to include patients with superficial submucosal (sm1/early T1b, <200 µm) involvement may be acceptable in some settings, particularly for those patients who are unsuitable candidates for surgery.⁸

CONTRAINDICATIONS

1. Invasive esophageal carcinoma with extension beyond the superficial submucosa.
2. Radiographic or EUS evidence of locoregional or distant spread (N1 or M1 or greater disease).
3. EAC with known high-risk features: lymphovascular invasion or poor differentiation.⁸
4. Patient is medically unfit for therapeutic endoscopy, procedural sedation, or general anesthesia.
5. Patient has an uncorrectable coagulopathy.
6. Very small lesions (<10 mm) that may be reliably resected en bloc via EMR techniques¹⁰ (relative contraindication).

PRE-ESD LESION ASSESSMENT

Establishing the depth and local extent of neoplasia is the major determinant of whether or not to perform ESD for superficial esophageal cancers, provided that the means and experience required to remove the lesion by ESD is present. Therefore, careful preprocedure assessment of any lesion considered for ESD is critical.

1. Endoscopic visual assessment: High-definition white-light inspection and advanced endoscopic imaging are important for detection and assessment of prospective lesions prior
to ESD. Advanced imaging techniques include dye-based chromoendoscopy, optical enhancement technologies (e.g., narrow-band imaging), and methods such as optical coherence tomography or confocal endomicroscopy. While review of these approaches is beyond the scope of this chapter, advanced endoscopists performing ESD should be well versed in use of dye-based and electronic chromoendoscopy. Use of endoscopes capable of optical zoom magnification can offer a diagnostic advantage, in particular when evaluating intrapapillary capillary loops (IPCLs) as a means of diagnosing and staging SCCs. Careful mucosal endoscopic examination of esophageal lesions is essential for delineation of margins, as well as identification of highly dysplastic or malignant foci, which must be removed.¹¹ For squamous lesions, we found the topical application of a dilute 0.5% Lugol iodine solution (Safecor Health, Woburn, MA) to the esophageal mucosa prior to lesion marking very useful. This will delineate the extent of SCC, as dysplastic or malignant cells will not uptake the dilute iodine, whereas normal squamous cells will. Barrett esophagus can be surveyed after the application of acetic acid, which may facilitate the endoscopic diagnosis of high-grade intraepithelial neoplasia or cancer with a sensitivity of up to 96%.¹¹
2. Endoscopic ultrasound: EUS can be considered prior to esophageal ESD but is not mandatory. The ability of EUS to accurately discriminate endoscopic resectability (T1 vs T2) is imperfect, even when high-frequency EUS miniprobes are used. Up to 25% of lesions may be understaged by EUS, while up to 12% may be overstaged when compared to pathology.¹² Despite tempered enthusiasm for EUS in T-staging, EUS using a dedicated radial or linear echoendoscope is valuable for nodal staging, and linear EUS and fine-needle aspiration should be performed if suspicious lymph nodes are identified.