Endoluminal surgery in the field of colon and rectal surgery includes advanced endoscopic techniques such as endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) to remove larger polyps and early T1 cancers. EMR involves piecemeal removal of a larger polyp, often sessile, while ESD consists of en bloc dissection of a lesion off the muscularis propria, usually submitted as a single specimen. Hybrid ESD consists of incising the mucosa around the lesion with a cutting knife and then snaring the lesion/polyp off the muscularis en bloc. The technique of ESD was first introduced in Japan in 1988 for the treatment of early gastric cancer1 and is technically more challenging to learn and master than EMR. The advantage of ESD in colon and rectal surgery is that it can be utilized for more advanced lesions such as superficial T1 colon cancers similar to how it is used in Japan to treat early gastric cancers. The learning curve for these procedures is steep and highly variable in the literature, estimated at 30-100 cases. 2, 3, 4

In Japan, a ESD training is based on a preceptor model with a stepwise progression to learn the technique.10, 11, 12 Trainees begin with didactic lectures followed by extensive observation of the procedure by master endoscopists. Trainees then perform ESD on an animal model followed by performing parts of the technique on patients under direct supervision. Once competent in the technique, the freshly trained endoscopist will start with smaller lesions in the stomach and progress to more advanced, technically demanding cases. A minimum of 20-30 gastric ESD cases are encouraged prior to attempting colorectal lesions.6, 7, 8, 9 This formalized training program can take 3-4 years to complete. Adoption of advanced endoscopic techniques in a new colorectal surgery practice can be challenging as there are few colon and rectal surgery training programs in the US in which these techniques are taught, unlike gastroenterology advanced fellowships. The few colorectal surgery fellowships that do teach EMR/ESD to the trainees certainly do not have the formalized training process that exists in Japan, but a trainee can learn the basics of the techniques with focused effort. How would a colon and rectal surgeon with no previous exposure just starting to practice embark on adopting endoluminal surgery?

Prior to learning the more advanced technique of ESD, the colorectal surgeon should become proficient in basic colonoscopy, snare polypectomy, and EMR.8 It is also important to understand the settings and nuances of electrocautery, especially the generator that is present at your institution. Some electrosurgical generators are well suited for performing EMR/ESD, while others are less than optimal. Management of complications such as bleeding or perforation are also a necessary skill set as these do occur, although rare, when utilizing advanced techniques such as ESD/EMR.

ESD courses aimed at teaching the technique are becoming more common through societies such as American Society for Gastrointestinal Endoscopy and American College of Gastroenterology as well as single institution and industry sponsored courses.9,12 Several of the courses have a hands-on experience using animal models, which can be crucial when trying to learn the technique. After completion of an ESD course, observation of live ESD cases by an expert endoscopist similar to what is done in Japan is highly recommended with either a colorectal surgeon or advanced gastroenterologist. However, even at high volume centers in the US that perform ESD, significant exposure to the technique is difficult during a short visit to the institution.12 A more intensive and rewarding observational experience would be to visit a high volume center in Japan, which has been a common pathway for physicians in the US to learn ESD. However, this can be fairly expensive and requires the physician to take off a significant amount of time from their practice. To supplement a trainee's education to learn ESD, self-learning software was developed in Japan to help teach the technique in Japan and France but is not commercially available.13 Colon and rectal surgeons are somewhat at a disadvantage for learning the technique as they do not have any gastric lesions on which to perform ESD prior to tackling the often more complicated and challenging colorectal lesions.

With all the information provided above, how should a colorectal surgeon go about adopting endoluminal surgery in their practice? I can give a first-hand account of how that was accomplished from my own experience, including several practical tips that are not written down in any publication. I previously had extensive experience with colonoscopy, polypectomy, and EMR over the past 20 years so I was starting the process of learning the advanced endoluminal techniques from a sound background. Not everyone needs 20 years of experience, but as recommended by the Japanese training programs, a colon and rectal surgeon should have a solid background in colonoscopy, polypectomy, and some EMR before moving on to more advanced techniques. I was fortunate to be at an institution with a master endoscopist (colorectal surgeon) who had a vibrant and busy practice of EMR/ESD, up to 4-6 per week. Every week for an entire year I spent first observing the procedures and then performing parts of the procedure under his supervision, which is similar to some of the training experience in Japan. I had planned to participate in an animal lab set up by one of the vendors for colon and rectal surgery trainees, but then moved to another facility at the same institution. I thus ventured out on my own to start an endoluminal surgery program as one did not exist at this new location.

My first step was to find out what equipment was present at the new institution before I even got there. I met with the manager of endoscopy as well as the director of the operating room. Building the program required obtaining all disposable tools necessary to perform ESD/advanced EMR. A list of the recommended items includes:

Cutting ESD knife

Different sized stiff snares

Coagulation grasper

Multiple sized hemoclips (especially 16 and 17 mm clips)

Lifting agent (inject the lesion and expand the submucosal layer)

I contacted the appropriate vendors and was able to obtain all the necessary disposable tools as above. When performing ESD is it important to use an electrosurgical unit (ESU) capable of sensing changes in tissue impedance to adjust the voltage and allow responsive adaptation for consistent and safe energy application to the lesion.14 These ESU's have multiple settings for cutting / coagulation and can be highly customized. When I started doing these advanced procedures, I was using a different ESU that did not measure changes in tissue impedance and encountered more bleeding. My institution subsequently purchased an ESU that measures tissue impedance, which improved bleeding and had more effective cutting during ESD. It is important to understand all the modes of the ESU (e.g., Highcut, Drycut, Endocut Q, Softcoag, Forcedcoag, Spraycoag) to maximize its utility when performing endoluminal surgery.14

Early on I also inquired on what kind of sedation was used at the new facility where I started performing ESD. Most of the procedures that I observed by the master endoscopist were performed under conscious sedation. All endoscopic procedures where I currently practice are performed under monitored anesthesia care (MAC) using IV propofol administered by a CRNA (under supervision of anesthesiologist). My experience has been that performing these procedures under MAC provides a more stable platform on which to perform these longer procedures, compared to screening colonoscopy. Most of the EMR/ESD that I have performed have been in the endoscopy suite. However, if there is an increased risk of having to perform a surgical resection, I will attempt the ESD/EMR in the operating room under general anesthesia. When first starting out, it might be beneficial to do all the advanced procedures in the operating room, especially if MAC anesthesia is not available in the endoscopy unit. How does one decide which patients are suitable for the endoscopy suite vs the operating room? As the majority of the more advanced polyps are referred by gastroenterologists, the color pictures from the previous colonoscopy should be reviewed to visualize the size, location, and potential morphology of the lesion, as these gross features could indicate an increased risk of cancer or conversion to surgical resection due to the size or difficult location. If the previous colonoscopy was performed from an outside facility, an attempt is made to obtain the color pictures. In addition, the prior colonoscopy report and pathology from the polyp should be reviewed, assuming it had been biopsied. Pathological features such as high-grade dysplasia or adenocarcinoma in situ could suggest an underlying malignancy, increasing the chances of requiring surgical resection. If the lesion was described as “firm” by the previous endoscopist even though the biopsies may have shown only an adenoma, the chance of surgical resection increases significantly.

It is important to have a dedicated team of staff to assist in performing these advanced procedures as a well-trained endoscopy technician can be instrumental in facilitating the successful completion of an EMR/ESD. These procedures are more complex with multiple steps and take more time than a simple polypectomy. A novice endoscopy technician may not have the skill set to assist with an EMR/ESD of a larger lesion. The master endoscopist that trained me always had the same, very experienced endoscopy technician assisting. While having the same experienced endoscopy technician assist with these advanced procedures is beneficial, it is also important to have several other staff who are capable of assisting for redundancy and backup.

Once the technique of ESD has been mastered on benign disease, it can be used to remove early T1 cancers in the colon and rectum. Many benign polyps can be removed by EMR and hybrid ESD without resorting to formal ESD, which is more time consuming and technically challenging. How then is it determined which colonic lesions should be removed by ESD due to the “suspicion” of the lesion containing an early cancer? In Japan, Kudo's classification using magnified chromoendoscopy looking at pit patterns has been used for years to identify malignant lesions in the colon.15 Types I (round pits) and II (stellate pits) are non-neoplastic tissue (normal or hyperplastic), Types IIIS, IIIL and IV are associated with adenomatous polyps, and Types VI is highly correlated with early superficial colorectal cancer. However, Type VN is associated with deep submucosal invasion (> 1000 μm), and these lesions should be treated with surgical resection. In the US, magnified colonoscopes are not common and therefore magnified chromoendoscopy has limited use.16

There have been several other classification systems described to help assess the risk of a polyp containing cancer. The Paris classification has been used widely to categorize polyps and assess the risk of cancer within a polyp.17, 18 Polyps are classified as polypoid (protrudes above the surface) or non-polypoid (non-protruding). Polypoid type polyps (type 0-I) can be further subdivided into pedunculated (type 0-Ip) or sessile (type 0-Is). Non-polypoid type polyps (0-II) are slightly elevated(0-IIa), flat (0-IIb), or depressed (0-IIc). Finally, excavated lesions (type 0-III) have the highest risk of colorectal cancer, and removal should not be attempted endoscopically. The risk of cancer in a polyp has been demonstrated to be proportional to the size of the polyp and presence of depression (type 0-IIc).17, 18 Another system of nomenclature has been developed to describe lateral spreading tumors, which are non-polypoid colorectal lesions measuring greater than 10 mm that extend laterally and not vertically.19,23 These lesions can be divided into granular (LST-G) and non-granular (LST-NG). Granular lesions are subdivided into homogenous (LST-G-H) and nodular mixed (LST-G-NM) subtypes while non-granular lesions (LST-NG) are classified into flat elevated (LST-NG-FE) and pseudodepressed (LST-NG-PD) subtypes. In a meta-analysis of 48 studies, Bogie et al found that LST-NG-PD had the highest risk of invasive cancer at 31.6%, while LST-G-H had the lowest risk at 0.5%.21

Rather than use magnified chromoendoscopy, narrow band imaging can be used to look at pit patterns. Ozgur et al recently published a series of patients showing that the Kudo classification using narrow band imaging (NBI) correlated with submucosal invasion (Table 1).22 No polyps in their study with Kudo I or Kudo II pit patterns had contained invasive cancer. On the other hand, Kudo III, Kudo IV, and Kudo V pit patterns were associated with invasive cancer 9.1 %, 36.4%, and 54.5% respectively with a 91% sensitivity for submucosal invasion with Kudo pit pattern IV-V.22 Lesion morphology consisting of the Paris classification and LST type were not predictive of invasion, which is contradictory to other published studies mentioned above. However, in Japan, the LST-NG-PD (pseudo-depressed) type has been associated with multifocal invasion,23 which would make this subset of lesions not suitable for EMR. Therefore, pit pattern analysis using high-definition colonoscopy with NBI and LST subtype can be used to select those lesions (Kudo IV-V, LST-NG) that might benefit from ESD or even surgical resection rather than EMR due to the risk of having invasive cancer. Lesions that have a Kudo pit pattern I-III can be removed using EMR as the risk of invasive cancer is low.23

After a higher risk polyp containing invasive cancer has been removed by ESD, there are certain criteria that determine if endoscopic management is adequate treatment. These criteria include: negative margins, Grade 1 tumor budding (low), moderate to well differentiated tumor, no lymphovascular invasion, and submucosal invasion < 1000 μm.24 If these criteria are met, the incidence of lymph node metastases is very low at 1-2%.25, 26 It is important that the lesion is removed en bloc as evaluation of the pathology, especially the depth of invasion and margins, is more difficult to assess with piecemeal removal by EMR.27 In addition, local recurrence rates are much higher following piecemeal resection (11.9%-22%) compared to en bloc removal (0.7%-3%).27, 28 Submucosal invasion depth of 1000 μm is used to differentiate superficial vs deep invasion. Lesions that have deep invasion (> 1000 μm) have a higher risk of lymph node metastases (up to 18%).29, 30

In order to facilitate pathological examination of the en bloc specimen, it can be pinned on a piece of cork or foam after removal (Figure 1). It is important for the pathologist looking at the specimen to assess for margins as well as high risk features such as poor differentiation, lymphovascular invasion, high tumor budding, or submucosal invasion > 1000 μm. Any of these features would indicate that endoscopic removal is not adequate and that surgical resection is recommended.23 Open lines of communication with pathology is crucial to ensure that all the above pertinent features are reported out. If the pathologist is not familiar with interpreting these high-risk features, an opinion on the pathology should be sought from another facility with the expertise in looking at these features, especially depth of submucosal invasion.