Endoscopic submucosal dissection (ESD) is an innovative endoscopic technique that allows for en bloc resection of superficial gastrointestinal tumors regardless of their size. This allows for the precise histopathological assessment of specimen margins and reduces the risk of recurrence compared with endoscopic piecemeal mucosal resection (EPMR) on long-term follow-up.^1,2,3 In Korea and Japan, ESD is routinely performed in the endoscopic treatment of superficial gastric neoplasias,^4,5,6 and is becoming the standard treatment for colorectal neoplasias such as early colorectal cancer or benign colorectal tumors ≥2 cm for which en bloc resection with conventional endoscopic mucosal resection (EMR) is difficult. In a comparative study of ESD versus EMR for colorectal tumors ≥2 cm,^7,8,9 there was a 2% recurrence rate in the ESD group versus 14% in the EMR group, which also required additional EMR.

However, there are some limitations in performing colorectal ESD, as colorectal ESD is technically more difficult than conventional EMR and has a high risk of complications such as perforation and bleeding. Furthermore, serious peritonitis can develop from colonic perforation and as a result of secondary contamination by colonic bacteria and feces.^5,7,10 Therefore, accurate preprocedural diagnosis of the lesion and selection of the appropriate treatment from a precise diagnosis are important. According to a recent guideline,¹¹ when endoscopic treatment is performed for a large colorectal tumor, en bloc resection such as ESD is the principal approach. However, piecemeal resection is also acceptable except in cases of definitive submucosal invasion. In addition, depending on the skill of the endoscopist or the current practice in the hospital, there can be variations in the selection of the treatment method.

Colorectal ESD was introduced in Korea in 2003 and is now being routinely performed at major teaching hospitals. The outcomes of colorectal ESD in Korea are comparable to those in Japan.^4,12 However, the indications for colorectal ESD are still controversial. To establish colorectal ESD as a standard treatment, consistent preprocedural diagnosis of the colorectal lesion and selection of appropriate treatment options are required. In order to establish a standard indication for colorectal ESD, we sought to investigate the preferred criteria of ESD experts for colorectal ESD in their practice.

The results consisted of 33 questions and answers that explored the indications of endoscopists for selecting a treatment method in colorectal tumors ≥2 cm (Figs 1,2,3,4,5,6). The colorectal ESD indication for colorectal tumors <2 cm is also presented in Supplementary Fig. 1.

With the development of various endoscopic tools, devices, increased experience, and growing expertise in ESD, colorectal ESD is now routinely performed in Korea. Originally used for gastric tumors and not for colorectal tumors, there are, however, important differences between superficial gastric tumors and superficial colorectal tumors. These include gastric tumors usually being carcinomatous, whereas colorectal tumors are mostly benign lesions.^6,15 The rate of local recurrence after piecemeal resection of intramucosal tumors of the colorectum has increased;^10,16 however, even in such cases, residual or recurrent tumors can be treated with repeated endoscopic resection or surgery. In this regard, en bloc resection of colorectal tumors is less absolute than that of gastric tumors. At the same time, the advances in magnifying endoscopy with dye-spraying and image-enhancement endoscopy have enabled the preprocedural assessment of histology and invasion depth of tumors with high accuracy.^17,18,19

In this study, we surveyed the criteria of endoscopists for indicating colorectal ESD in their practice. From the survey, lesions such as LST-NG or tumors with type V-I pit pattern or type III-a NBI pattern, or tumors with a depressed area or shallow submucosa-invasive cancer cases were favored for colorectal ESD. These lesions have a significantly higher possibility of submucosal invasion and require en bloc resection for an accurate pathological evaluation. Other types of lesions chosen for colorectal ESD include tumors with partial submucosal fibrosis, tumors with partially nonlifting shallow submucosal invasion, tumors that are sporadically localized in the background of chronic inflammation, and cases of local residual early cancer after endoscopic resection. These types of lesions are included in the indication for ESD because the lesions are technically difficult to treat with conventional EMR. The selection criteria for colorectal ESD from our survey are almost the same as the indications for colorectal ESD in Japan.¹¹ In addition, more benign-looking tumors tended to be included in clinical practice for colorectal ESD, as LST-G nodular mixed type, tumors with types III and IV pit pattern, and serrated tumors with type II NBI pattern are mostly benign lesions confined to the mucosa. In practice, the indications by Korean endoscopists for colorectal ESD are broader than those in the recent guidelines.^11,20

Comparisons of outcomes for colorectal ESD versus EMR for large colorectal tumors have been reported in several clinical studies.^10,21,22 For large colorectal tumors, ESD had higher en bloc resection rates and lower recurrence rates than EMR, but had higher complication rates and required a longer procedure time. Although the local recurrence rates of the EMR group were higher than those of the ESD group, 89.8% of recurrent cases were successfully treated with additional EMR. Only 10.2% of recurrent cases required additional surgical resection.^10,22 Therefore, EPMR for large colorectal tumors is an acceptable treatment method. Significantly, when preprocedural diagnosis is performed precisely, recurrent lesions will usually be adenomas and additional endoscopic resection will be successful without surgical treatment.

Before performing colorectal ESD or EMR, diagnosis based on image-enhanced endoscopy and magnifying endoscopy allows for the distinction of a carcinoma from an adenoma and for estimating the invasion depth of the tumor.^23,24 The diagnostic accuracy rate of discrimination between adenoma and carcinoma was 70% to 90% for pit pattern diagnosis with magnifying observation, and a similar rate was obtained by using image-enhancement endoscopy such as NBI.^17,18,19 There are two subtypes of LST, namely LST-G and LST-NG. The rate of submucosal invasion in LST-NG was significantly higher than that of LST-G. In LST-G nodular mixed type, submucosal invasion may exist in a large nodule or an area of a type V-I pit pattern.²⁵ Importantly, EPMR without breaking the carcinomatous portion into fragments should be applied. For LST-NG cases, the lesion should be removed en bloc with ESD because there is a higher potential for invasion and fibrosis, and the specimen obtained from the complete en bloc excision can be pathologically evaluated in detail.^20,25,26,27

From our survey, endoscopists tend to conduct colorectal ESD for apparent adenomatous LST lesions in their practice. Although en block resection such as ESD is desirable if applicable, colorectal ESD is technically difficult, time consuming, and has a high risk of complications such as perforation. The objective of endoscopic treatment for adenoma is elimination of the lesion; therefore, EPMR is a good alternative treatment method for eliminating an obvious adenoma. Through precise preprocedural diagnosis, prearranged EPMR with minimal fractions is an appropriate treatment method for adenomatous LST lesions. Endoscopists should not perform superfluous ESD merely to improve their own ESD skills or for other reasons of self-interest, as this could unnecessarily place the patient at a risk of complications. To establish a standard indication for colorectal ESD, it is important to balance the need for completeness of treatment versus patient safety and simplified treatment.