(a–f) Invasive pattern. Invasive pattern is defined as irregular and distorted crypts in a demarcated area as observed in Kudo’s type Vn and selected cases of Vi (e.g. deep submucosal invasive cancers), where surgical resection is the appropriate treatment. (a) The left side of this tumour is flat and depressed morphology as outlined by yellow dots. (b) Indigo-carmine dye spray demonstrates the tumour margin and surface structure clearly. (c) Crystal-violet staining of the right-side protrusion showed a regular-type IV pit pattern suggesting intramucosal neoplasia in this protrusion. (d) Crystal-violet staining of the left side of this tumour revealed an irregular distorted pit pattern corresponding to the yellow dots area and diagnosed as an invasive pattern suggesting submucosal deep invasion. (e) A stereomicroscopic image of the surgical specimen. The protruded area is on the left side and flat/depressed area lined by yellow dots is on the right side. (f) Histopathology confirmed submucosal deep invasion in the area of invasive pattern and intramucosal neoplasia in the type 4 pit area
JNET type 2B and 3. The Japan NBI Expert Team (JNET) classification consists of four categories of vessel and surface patterns – types 1, 2A, 2B and 3. Types 1, 2A, 2B and 3 correspond to histopathological findings of hyperplastic polyp/sessile serrated polyp (SSA/P), low-grade intramucosal neoplasia, high-grade intramucosal neoplasia/superficial submucosal invasive cancer and deep submucosal invasive cancer, respectively. (a) A slight depressed area showed an irregular vessel pattern (variable calibre and irregular distribution) and irregular surface pattern diagnosed as JNET type 2B suggesting high-grade intramucosal neoplasia or submucosal slight invasive caner. (b) Irregular nodular area showed loose vessel pattern and interruption of thick vessels and amorphous areas in surface pattern diagnosed as JNET type 3 suggesting deep submucosal invasive cancer. It is important to keep in mind that JNET type 2B lesions correspond to high-grade dysplasia/superficial cancer, which implies that there is a risk of SM invasion. The gold standard for differentiating SM1 and SM2 is crystal-violet staining. Therefore, pit pattern diagnosis is essential for choosing the appropriate treatment strategy for JNET type 2B lesions
The following morphological features may provide ancillary information that suggest the presence of deep submucosal invasion – (a) deep depression, (b) fold convergence, (c) irregular bottom of depression surface, (d) chicken skin appearance, (e) redness, (f) expansive appearance, (g) firm consistency, (h) irregular surface, (i) loss of lobulation and (j) thick stalk
Biopsies should be avoided before ESD as the malignant potential is often underestimated on biopsy specimens, and biopsy-induced fibrosis within the lesion may compromise submucosal lifting, thus making ESD more difficult, especially for flat tumours.
Preparation and Setting of ESD
Sedation and Bowel Preparation
ESD is a technically challenging procedure that is associated with significantly longer procedural time than EMR which has implications for patient preparation. Colorectal ESD may be performed with conscious sedation as it allows for the patient to be repositioned during the procedure to enlist the help of gravity to achieve counter-traction of the lesion. The routine use of carbon dioxide insufflation [8] is preferred as it minimises patient discomfort associated with prolonged insufflation and reduces the risk of severe complication, such as pneumothorax or pneumoperitoneum, even when perforation has occurred. Adequate bowel preparation is essential to facilitate optimal visualisation of the lesion as well as to mitigate the risk of peritoneal contamination in the event of a perforation.
Equipment and Accessories
The tools used in ESD have evolved significantly to meet the unique challenges posed by both the procedure and the site of resection. Hence familiarity with the equipment is essential for performing ESD.
Endoscope
An endoscope with an auxiliary water channel to produce a water jet is preferred to maintain visualisation when bleeding is encountered. An endoscope with a 3.2 mm instrument channel is preferred, if available, to accommodate ESD accessories.
Carbon Dioxide Regulator
The use of carbon dioxide insufflation is preferred during ESD. CO2 is more rapidly absorbed than nitrogen and oxygen, the main constituents of room air. Hence CO2 insufflation gives rise to less sustained luminal insufflation, which is associated with less patient discomfort and shorter procedural time, and is safe with deep sedation [8–12]. It is also believed that rapid CO2 reabsorption in the setting of CO2 insufflation may reduce the likelihood of tension pneumoperitoneum following a perforation. CO2 insufflation is generally safe in patients with pulmonary dysfunction. In patients with severe obstructive lung disease, CO2 retention may occur with longer procedure time; hence CO2 monitoring may be considered [13].
Electrosurgical Generators
Electrosurgical Generators (ESUs) with simple cutting and coagulation modes have been superseded by newer models that have multiple modes to be used on different lesion characteristics. These modes are specific to the manufacturer. Pure cutting and coagulation current modes are avoided due to the increased risk of bleeding and delayed perforation, respectively [14]. The newer ESUs provide a variety of electrosurgical waveforms that produce a variety of tissues effects that is useful for the various steps in ESD. The ESUs also contain microprocessors that adjust power according to tissue impedance to avoid deep tissue injury [15].
ESD Knives
Electrosurgical knives are essential tools for ESD as submucosal dissection differentiates ESD from other types of endoscopic resection. ESD knives fall into two broad categories—needle-knife and scissors type. The former can be further classified into those with a blunt tip and the tip-cutting knives [16]. The knives differ in their indication and methods of use and the endoscopist should be familiar with the use of the specific knives. These accessories and other ESD tools are described in details elsewhere in this book.
Distal Attachments and Other Counter-Traction Methods
A disposable transparent cap affixed to the distal end of the endoscope is an essential tool in ESD. It serves to provide counter-traction to maintain visualization during submucosal dissection by keeping the resected mucosal flap off the endoscope lens. Caps with holes on the side facilitate drainage of fluid and blood. Various other tools and methods have been devised to provide counter-traction though none have gained widespread use. The clip-line traction method using dental floss is useful primarily in the rectum and distal colon but less effective in the proximal colon [17, 18]. Other methods would include clip with sinker method, external grasping forceps, internal traction, double-channel scope and double-scope methods [19]. The SO clip is commercially available in Japan and is applicable to any location without the need for scope reinsertion.
The lack of a truly effective counter-traction method highlights the inherent limitation of the flexible endoscope – therapeutic accessories for an increasingly complex array of interventional procedures are delivered through an accessory channel that is also required for periodic insufflation, suction and water irrigation. Novel triangulation platforms with independent instrument channels that enable endoscopic resection to be performed with the comparative ease of laparoscopic surgery are currently under development [20, 21].
Technical Aspects of ESD
The basic steps in ESD are injecting fluid into the submucosa to elevate the tumour, incising the surrounding mucosa to gain access to the submucosa and dissecting the submucosa beneath the tumour to achieve tumour resection.
Conventional ESD Method
- 1.
Submucosal injection under the lesion and the surrounding normal mucosa
- 2.
An initial partial mucosal incision (i.e. one-quarter of the circumference)
- 3.
Submucosal dissection beneath the incision
- 4.
Repeat partial mucosal incision and dissection in a segmental fashion
- 5.
Completion of the en bloc resection.
(A) Diagram of conventional ESD. Before ESD, the depth of tumour invasion is estimated using a combination of magnified NBI and chromoendoscopy with indigo-carmine and crystal-violet staining for pit pattern diagnosis. (a) 0-IIa + IIc (laterally spreading tumour non-granular type, LST-NG) lesion in a transverse colon, and the estimated tumour size was 25 mm in diameter. Fold convergences were recognised and submucosal invasion is suspected by a conventional endoscopy. (b) An NBI image revealed the tumour margin clearly, and NICE classification might be type 2 suggesting an intramucosal neoplasia. (c) A magnified NBI revealed variable calibre and irregular distribution of vessel pattern/irregular or obscure surface pattern, therefore, diagnosed as JNET type 2B suggesting a high-grade intramucosal neoplasia or submucosal superficial invasive cancer. (d) An indigo-carmine dye spray revealed a tumour margin and surface structure clearly. (e) A magnified image after indigo-carmine spray revealed type IIIs and IIIl pit pattern in this relative depressed area. (f–h) After crystal-violet staining (f), IIIs pit (g) and Vi slight irregular pit (h) were observed in the centre, but the area of type Vi was relatively small; therefore, a type Vi (non-invasive pattern) suggesting intramucosal or submucosal superficial invasion was diagnosed. (B) Diagram of conventional ESD. (a) A retroflexed view of a LST-NG after indigo-carmine dye spraying. A short-type ST hood (Fujifilm Medical Co.) was attached to PCF260J (Olympus Medical Co.). (b) After submucosal injection under the lesion and the surrounding normal mucosa, an initial partial mucosal incision (i.e. one-quarter of the circumference) is made and submucosal dissection is performed beneath the incision. (c) Repeat partial mucosal incision and dissection in a segmental fashion similar to pocket-creation method. (d) After completing 50% of the dissection, a partial marginal resection is planed from anal side using a straightforward view. (e) Repeat submucosal dissection. (f, g) Completion of the en bloc resection. (h) A resected specimen was stretched out using fine needles for detailed histopathological analysis
Unlike gastric ESD, the perimeter of the colorectal lesion need not be marked by cautery prior to ESD as the borders of colorectal lesions are more distinct and readily identified. The lesion is lifted by injecting a submucosal injection agent to create a submucosal cushion. Suboptimal lifting may indicate either deep submucosal tumour invasion or the presence of fibrosis which in severe cases may preclude further ESD. The addition of dyes (e.g. indigo carmine) to the injectable enables clearer delineation of the submucosal plane of dissection though some experts advocate avoiding the addition of dyes to enable better visualisation of submucosal vessels during dissection.
Following lifting, a partial mucosal incision is made around the lesion to gain access to the submucosa to facilitate submucosal dissection. Instead of creating a circumferential mucosal incision , a partial marginal incision, either at the oral or anal end, is recommended before submucosal dissection is commenced. This is to ensure that the submucosal cushion is sustained without dispersion of fluid during dissection. The authors advocate initial submucosal dissection to be performed just beneath the mucosal layer to prevent perforation. Once the submucosal layer has opened sufficiently to allow adequate visualisation of the cutting area, dissection may be extended to the lower third of the submucosal layer. Repeated injections are made to maintain the submucosal fluid cushion to minimise the risk of perforation.
As the lesion is dissected in a segmental fashion, the position of the patient may be changed periodically to enlist the effect of gravity to achieve traction or counter-traction on the lesion. The lesion is repositioned by rotating the patient to use gravity to keep the field of vision clear of blood and water. This allows the endoscopist to have a clear field of vision and achieve traction. The use of conscious sedation in this instance facilitates changes in patient’s position.
Following resection, the base of resection is examined for bleeding and perforation. Unlike gastric ESD, the routine coagulation of visible vessels following colorectal ESD is minimised to prevent thermal injury.
Pocket-Creation Method (PCM) (Fig. 8.5)
- 1.
Submucosal injection
- 2.
Creation of tunnel entry with 20 mm mucosal incision on the anal side of the lesion
- 3.
Creation of a submucosal pocket by dissection under the lesion
- 4.
Lateral dissection on the dependent side (i.e. closer to gravity)
- 5.
Lateral dissection on the antigravity side (i.e. away from gravity)
- 6.
Completion of en bloc resection.
(A) PCM-1. (a) A LST-granular (nodular mixed type) in the transverse colon. (b) Submucosal injection to the anal side of this tumour. (c) Creation of tunnel entry with 20 mm mucosal incision on the anal side of the lesion. (d, e) Creation of a submucosal pocket by dissection under the lesion. (f) Lateral dissection on the dependent side (i.e. closer to gravity). (g) Additional submucosal dissection using a traction by short-type ST hood (Fujifilm Medical Co.). (h) Lateral dissection on the antigravity side (i.e. away from gravity). (B) PCM-2. (a) Creation of a submucosal pocket by dissection using IT knife nano under the lesion. (b, c) Lateral dissection using IT knife nano away from gravity. (d–f) Completion of en bloc resection
The PCM offers the following potential advantages. As only a minimal incision is made to introduce the knife into the pocket for subsequent dissection, the submucosal fluid cushion is sustained during ESD with minimal dispersion of injected fluid. The position of tip of the scope is stabilised within the pocket which facilitates tissue retraction allowing for dissection to proceed tangential to the muscle layer. This may be particularly helpful in navigating dissection across semilunar folds as these substantially alter the angle of the mucosal layer, thus predisposing to failure of en bloc resection and perforation [24]. Emerging data suggest PCM may be associated with higher en bloc resection rates and shorter procedure times compared to the conventional method although this requires confirmation from prospective trials [25].
Hybrid ESD Technique
In view of the technical difficulty of performing colorectal ESD, a hybrid EMR-ESD has been described that is characterised by a circumferential mucosal incision followed by snare resection of the lesion [26, 27]. However, compared to the conventional technique, the hybrid technique is associated with lower R0 and en bloc resection rates and has a similar rate of adverse events [28, 29].
Management of Complications
While ESD has been associated with a higher rate of complications than EMR, most ESD complications can be managed endoscopically, and only a small proportion of patients require surgery for complications. Hence the risk of ESD complications that require surgery is low (<1%). In addition, the risk of ESD complication for tumours larger than 20 mm in diameter may be comparable to EMR. However the complication rates are higher in Western countries compared to Asian countries (3.1% vs 0.8%) [29, 30].
Intraprocedural Bleeding
Intraprocedural bleeding is an expected occurrence during ESD. Hence its prevention and management are important aspects of performing colorectal ESD. Bleeding is often encountered during mucosal incision and submucosal dissection. This may be largely avoided by using appropriate electrosurgical settings during the pre-cut and dissection phases and coagulation of exposed vessels. Bleeding may be arrested with soft coagulation using the ESD knife and, if unsuccessful, with the use of haemostatic forceps. The connective tissue surrounding the vessel is dissected to isolate the vessel. The vessel is then either hooked by the ESD knife or grasped by the haemostatic forceps and coagulation current is applied. The use of clips is avoided as it may interfere with continued dissection of the tumour. Intraprocedural bleeding may be avoided by identifying larger non-bleeding submucosal vessels during dissection and coagulating them prophylactically. Intraprocedural bleeding rarely requires surgery for haemostasis.
Delayed Bleeding
Delayed bleeding has been reported in an average 2.7% and up to 13.9% of patients [29, 31]. The risk factors for delayed bleeding include lesion size, rectal location, presence of submucosal fibrosis and low-volume centres [31, 32]. The rate of bleeding with the hybrid technique is similar to the conventional technique [29]. Delayed bleeding may be treated endoscopically with deployment of clips. The use of thermal methods is less favoured so as to avoid further thermal injury to the muscle layer.
Perforation
Perforation rates of 4 to 10% have been reported with higher-volume centres reporting lower complication rates [33–36]. More recent data report a perforation rate of 1–2%. Tumour size, location and presence of submucosal fibrosis and endoscopist inexperience (<50 cases) are risk factors for perforation [34, 37–40]. While ESD has a higher perforation rate than EMR, the reported rate of ESD complication-related surgery is low at <1% as the majority of complications may be treated endoscopically [30]. Prediction models have been developed to stratify the risk of perforation that may be potentially useful in clinical decision-making [40, 41].
The risk of perforation is minimised by ensuring an adequate submucosal cushion with repeated injection during dissection. The incorporation of a contrast dye (e.g. indigo carmine), which stains the submucosa but not the muscularis propria, guides the plane of dissection and allows for identification of exposed muscle during dissection. The use of pure cutting or coagulation current is also avoided to minimise the risk of bleeding and delayed perforation, respectively [14]. Biopsies of the lesion before ESD should be avoided as the resultant fibrosis may predispose to perforation during ESD.
Prompt identification of perforation during ESD enables immediate treatment that reduces mortality and the need for surgery [42]. Upon identification of a perforation, the patient’s hemodynamic status should be promptly reviewed to identify a tension pneumoperitoneum. A tension pneumoperitoneum is an emergency characterised by the presence of hypotension and gaseous distension of the abdomen and is treated by immediate needle decompression.
Perforations that are recognised during ESD are amenable to primary endoscopic clip closure. A spectrum of endoscopically recognisable injuries to the muscularis propria that encompasses perforation has been described [43]. The presence of a hole with or without observed contamination, a specimen target sign or a mirror target sign would mandate prompt clip closure of the defect to avoid peritoneal contamination [44].
Successful endoscopic clip closure may be achieved by bearing in mind the following principles. Prior to clip closure, it may be advisable to reduce luminal gas insufflation to reduce the tension on the defect and clean the field by suctioning the fluid pool and repositioning the patient. It may be necessary to continue with dissection of the tissue adjacent to the defect to enable clip deployment without compromising subsequent completion of the dissection. During clip closure , the mucosa is suctioned into the opened clip, gentle pressure from the clip is applied and with further suction the lumen is deflated and the clip is closed. The lumen is re-insufflated to ensure adequate apposition of the defect edges before the clip is deployed [45].
In addition to clips, perforations have also been successfully closed with other devices such as the over-the-scope clips and suturing devices [46, 47]. However, the need for scope withdrawal and reinsertion may potentially compromise outcomes.
Delayed Perforation
Delayed perforation occurs in 0.1–0.4% of cases and results from thermal injury [32, 34]. Patients may present more than one or two days after the ESD. CT is performed to confirm the diagnosis and assess the degree of peritoneal contamination. The presence of peritoneal contamination resulting from egress of luminal contents usually precludes endoscopic closure, and surgery is usually required.
Electrocoagulation Syndrome
Electrocoagulation syndrome is a consequence of transmural thermal injury to the colon resulting in localised peritonitis. It is characterised by clinical features of inflammation (i.e. fever, raised inflammatory markers) and peritoneal signs in the absence of radiological evidence of perforation [48, 49]. Patients usually present within hours to days of the ESD. The incidence of electrocoagulation syndrome following ESD is 8.6–9.5% [50, 51]. It is more common following ESD compared to EMR. Lesion size (›30 mm) and colonic location increase the risk of electrocoagulation syndrome [52]. The vast majority of patients may be treated conservatively with bowel rest, intravenous fluids and antibiotics. Most patients improve within 24 hours without any sequelae and any worsening should prompt re-evaluation.
Strictures
Strictures following ESD are rare and have been reported in a very specific instance – following ESD of near circumferential colorectal lesions. In patients requiring more than 75% circumferential ESD, strictures have been reported in 3.8–19.7% of patients [53, 54]. These strictures may be treated with one or two times endoscopic dilatations.
Histological Assessment of ESD Specimen
Dissected specimens should be retrieved and handled with care to prevent fragmentation. If the resected specimen is large, a rectal overtube may be used to aid removal of the specimen without fragmentation [55]. Once the specimen is retrieved, it is stretched and fixed on the board. It is useful to determine the spatial orientation and cutting direction of the resected specimen.
Resections with clear vertical and horizontal margins, tumour budding grade I, submucosal invasion up to 1 mm, no lymphovascular invasion and no poorly differentiated or mucinous component are considered curative. A positive vertical margin implies deep invasion and is an indication for surgery.
ESD Outcomes
A recent meta-analysis reported a pooled R0 resection rate of 92.9% and en bloc resection rates of 91%. However, there was substantial variation in the rates reported in individual studies with significantly lower rates in non-Asian countries compared to Asian countries. The hybrid resection technique was associated with lower rates of R0 resection and en bloc resection than the conventional method [29]. The risk of recurrence following ESD is low but higher in Western compared to Asian countries (5.2% vs 1.1%) [29]. Recurrences are more likely with piecemeal resection, with incomplete resections with positive margins on histology, and in cases of non-curative resection [56]. Compared to EMR, ESD achieves higher rates of en bloc resection and curative resection. This is, however, at the expense of higher rates of perforation [34].
Conclusions
ESD has emerged as an effective treatment for superficial colorectal neoplasia but still remains a challenging procedure as demonstrated by the wide variation in clinical outcomes. While ESD enables surgery to be avoided in patients with superficial invasion, patients with clear morphological evidence of deep submucosal invasion should be sent for surgery instead. Optimal outcomes may be achieved by careful patient selection, the use of appropriate endoscopic instruments, training in techniques and the management of complications.
Full access? Get Clinical Tree
