The Vienna classification should be used to classify the histopathology of tumors [13]. ESD results in a complete removal of mucosal and submucosal layers and since lymph node metastasis is related to the depth of invasion in addition to differentiation, Japanese investigators have developed criteria for ESD versus surgical management of tumors based on tumor differentiation, ulceration, depth of invasion, vascular involvement, and lymph node invasion (Tables 4.2 and 4.3). The major indications for ESD include SCC of esophagus involving lamina propria, high-grade dysplasia (HGD), and intramucosal cancer (IMC) in Barrett’s esophagus, including en bloc resection for accurate local staging of T1a lesion, well-differentiated early gastric cancer, and laterally spreading tumor (LST) of colon that cannot be removed by EMR. Practically, ESD provides the endoscopist with the ability to remove large superficial tumors in a single piece including lesions that may not be completely removable by EMR, such as ulcerated lesions, lesions with submucosal fibrosis, recurrent neoplasms, non-lifting lesions, and potentially lesions with very early submucosal invasion.

Some disadvantages of ESD include significant training requirements, prolonged procedure time, need for specialized equipment, need for an assistant during the entire procedure, and significantly higher complication rates.

Esophageal: The role of ESD in squamous cell carcinoma of the esophagus is well established in Asian studies. ESD en bloc resection rates of 100% with local recurrence of 1% have been reported compared to 53 and 10%, respectively, for EMR [14]. Involvement of muscularis mucosa with carcinoma increases the risk of lymph node metastasis to about 10%, and these patients will have higher recurrence rates if treated with ESD. Suggested indications for ESD for squamous cell cancer includes T1a lesions limited to lamina propria. Once the lesion invades the muscularis mucosa and infiltrates the superficial submucosa (up to 200 μm), the risk of local lymph node metastasis approaches 15%, and these patients represent a relative indication for ESD after the evaluation for local lymph nodes.

Patients with high-grade dysplasia and intramucosal cancer in the setting of Barrett’s esophagus are candidates for ESD. Another indication for ESD is early adenocarcinoma lesions that are well differentiated, with less than 500-μm extension into the submucosa. Studies demonstrate 90% en bloc resection rates with lower R0 resection rates but with over 95% long-term cure rates [15]. When ESD is performed in the setting of Barrett’s esophagus, it should be followed by radiofrequency ablation of any remaining Barrett’s epithelium [16]. Based on a systematic review demonstrating a 1–2% lymph node metastasis in setting of intramucosal cancer in Barrett’s, these patients are candidates for ESD given the mortality of 1–2% associated with esophagectomy [17].

Gastric: The role of ESD is best established for early gastric cancer. In large series from Asia en bloc, resection rates of >95% and R0 resection of >93% have been documented, with local recurrence of 1% and 5-year survival of 96–100% [18, 19]. A meta-analysis of outcomes of ESD versus EMR for early gastric cancer demonstrated higher rates of en bloc resection (92% vs. 52%) and R0 resection rates (82% vs. 42%) for ESD compared to EMR, although all-cause mortality did not differ between the two resection techniques [20]. Indications for ESD include lesions of any size if they have dysplasia, intramucosal well-differentiated adenocarcinoma without ulceration (<2 cm absolute indication, >2 cm expanded indication), and other expanded indications in the Japanese literature. These expanded indications include intramucosal well-differentiated adenocarcinoma with ulcer and <3 cm in size, intramucosal poorly differentiated adenocarcinoma <2 cm in size, and well-differentiated adenocarcinoma <3 cm in size and with superficial submucosal invasion, i.e., sm1 < 500 μm [21].

Colorectal: ESD in the rectum is performed by most endoscopists trained in ESD, as the rectum has a thick wall and retroflexion is easily performed. ESD for laterally spreading tumors outside of rectal area in the colon is difficult and mainly performed in Asia. A systematic review of >2800 patients, including 11% of patients with submucosal cancer, found an R0 resection rate of 88%. When compared to EMR, ESD results in less local recurrence (2% vs. 12% for EMR) [22, 23]. In one of the largest series, experienced Japanese endoscopists achieved en bloc and curative resections in 88 and 89% of 1111 colorectal tumors. Perforation occurred in 5% of procedures, with post-procedure bleeding in 1.5% [24]. Colorectal ESD is indicated for large adenomas that cannot be completed removed by EMR. Laterally spreading tumors (LST) are classified as granular and non-granular. Non-granular lesions have a higher likelihood of submucosal invasion. ESD may be particularly useful for larger 0-IIc lesions that do not lift with submucosal injection, lesions with scaring, and residual tumor with scar following prior EMR. ESD may be adequate resection for well-differentiated adenocarcinomas with invasion of the submucosal of <1000 μm, in the absence of lympho-vascular invasion, as the risk of lymph node metastasis in this setting is extremely low [25].

Colonic ESD outside the rectum, particularly in the right colon, is extremely difficult and challenging because of the thin wall of the colon with high risk of perforation and bleeding. Additionally, most lesions can be managed with piecemeal EMR technique with similar outcomes. Thus, colonic ESD (outside the rectum) should only be attempted by a skilled endoscopist with extensive experience performing ESD in other locations.