Most colorectal cancer arises from adenomatous polyps. This gradual process may be interrupted by screening and treatment using colonoscopy and polypectomy. Advances in imaging platforms have led to classification systems that facilitate prediction of histologic type and both stratification for and prediction of the risk of invasion. Endoscopic treatment should be the standard of care even for extensive advanced mucosal neoplasm. Technique selection is influenced by lesion features, location, patient factors, and local expertise. Postprocedural complications are more common following advanced resection and endoscopists should be familiar with risk factors, early detection methods, and management.
Key points
- •
Endoscopic mucosal resection (EMR) is a safe, cost-effective and curative intervention for most advanced mucosal neoplasms of the colon.
- •
Lesion morphology can be used to stratify for the risk of submucosal invasion (SMI).
- •
En bloc EMR is usually limited to lesions less than or equal to 20 mm. Larger lesions are best treated by piecemeal EMR or, if early invasive disease is suspected, endoscopic submucosal dissection (ESD).
- •
Real-time macroscopic assessment of lesion type and risk of SMI may be valuable in therapeutic strategy selection.
- •
Recurrence is seen in 10% to 20% of EMR at first surveillance after piecemeal resection but is effectively treated endoscopically with long-term remission.
- •
In the colon, ESD has little if any long-term advantage compared with EMR.
- •
ESD is associated with a longer procedure time, increased risk of perforation, mandatory multiday hospitalization, and increased costs compared with EMR.
- •
Bleeding and perforation can be adequately managed endoscopically. Careful assessment of the postresection mucosal defect is a vital component of the procedure.
Introduction
Most colorectal cancer (CRC) arises from adenomatous polyps through a process of accumulated molecular abnormalities, leading to increased cytologic dysplasia and eventually cancer. The gradual nature of this process allows effective screening and intervention. Long-term follow-up clearly shows the effectiveness of colonoscopy and polypectomy in reducing the incidence and mortality of CRC.
More than 80% to 90% of polyps encountered during routine colonoscopy are less than 10 mm in size, do not contain advanced disease, and can be readily treated with conventional snare polypectomy by an appropriately trained endoscopist.
Approximately 5% of adenomatous polyps are flat and sessile lesions larger than 10 mm. They may show extensive lateral growth along the bowel surface before developing an invasive component. The term advanced mucosal neoplasia (AMN) may be used to describe flat and sessile lesions larger than 20 mm.
Treatment of colonic AMN may require the use of advanced polypectomy techniques such as endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), or hybrid techniques. Technique selection is influenced by lesion morphology, location, patient comorbidities, and local expertise. Prospective studies have shown that wide-field EMR is safe and effective. Compared with surgery, it is substantially more cost-effective and may confer a mortality benefit, particularly in elderly patients with comorbidities.
Postprocedural complications are most common following complex endoscopic resection. Bleeding (both immediate and delayed) and perforation are most commonly encountered. Also, incomplete resection leading to residual disease and luminal stenosis after circumferential resection may occur.
Introduction
Most colorectal cancer (CRC) arises from adenomatous polyps through a process of accumulated molecular abnormalities, leading to increased cytologic dysplasia and eventually cancer. The gradual nature of this process allows effective screening and intervention. Long-term follow-up clearly shows the effectiveness of colonoscopy and polypectomy in reducing the incidence and mortality of CRC.
More than 80% to 90% of polyps encountered during routine colonoscopy are less than 10 mm in size, do not contain advanced disease, and can be readily treated with conventional snare polypectomy by an appropriately trained endoscopist.
Approximately 5% of adenomatous polyps are flat and sessile lesions larger than 10 mm. They may show extensive lateral growth along the bowel surface before developing an invasive component. The term advanced mucosal neoplasia (AMN) may be used to describe flat and sessile lesions larger than 20 mm.
Treatment of colonic AMN may require the use of advanced polypectomy techniques such as endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), or hybrid techniques. Technique selection is influenced by lesion morphology, location, patient comorbidities, and local expertise. Prospective studies have shown that wide-field EMR is safe and effective. Compared with surgery, it is substantially more cost-effective and may confer a mortality benefit, particularly in elderly patients with comorbidities.
Postprocedural complications are most common following complex endoscopic resection. Bleeding (both immediate and delayed) and perforation are most commonly encountered. Also, incomplete resection leading to residual disease and luminal stenosis after circumferential resection may occur.
Indications/contraindications and limitations
Accepted indications and contraindications for advanced endoscopic resection in the colon are summarized in Box 1 . In general, all patients found to have colonic AMN should be considered for endoscopic therapy with an experienced clinician. Modeling using the (CR) Physiologic and operative severity score for the enumeration of mortality and morbidity (POSSUM) and the Association of Coloproctology of Great Britain and Ireland score (ACPGBI) scores applied to a large prospective multicenter AMN cohort has shown this to be associated with less morbidity and mortality. In addition, markedly reduced costs with a predicted difference of approximately 6 days’ hospital stay and approximately US$10,000 per patient savings were calculated compared with surgical resection.
Indications
Large sessile or laterally spreading adenomas or serrated lesions larger than 20 mm
Lesions in difficult locations: appendiceal orifice, ileocecal valve, lower rectum, behind folds
Polyps in difficult situations: following failed removal attempts, lesions over scars
Large thick-stalked pedunculated polyps
Large colonic lipomas
Lesions in chronic inflammatory bowel disease
Contraindications
Findings suggesting deep invasion (deep irregular depression, fold convergence, firm consistency, nonlifting, type Vn/Vi Kudo pit pattern, type IIIb Sano vascular pattern, type 3 Narrow-band Imaging International Colorectal Endoscopic [NICE] criteria)
Unavailability of supporting infrastructure (dedicated trained staff, tertiary level radiological and surgical support)
Inadequate experience and/or training
Endoscopic Versus Surgical Resection
Even extensive colonic lesions limited to the mucosa can be cured with endoscopic resection, because the unique absence of lymphatics in the colonic mucosa precludes lymph node metastasis (LNM).
Submucosal invasion (SMI) is the hallmark of CRC, and is associated with LNM in 1% to 16% of endoscopically attempted resections. The risk of LNM is related to the depth of invasion, the tumor grade, and the presence or absence of lymphovascular invasion (LVI). When invasion is limited to the submucosa (T1 lesions) further stratification into high-risk and low-risk groups can be performed. Low-risk features include SMI depth less than 1 mm, well-differentiated tumor grade, and absence of LVI. In such cases, endoscopic therapy may be considered adequate when resection is complete and the deep and lateral margins are negative. High-risk features include tumor budding, LVI, poorly differentiated tumor grade, and SMI depth greater than 1 mm. If found in an EMR specimen or predicted by endoscopic assessment, surgical resection with lymph node dissection is appropriate.
Real-time endoscopic prediction of SMI risk may be challenging. The hallmark of invasive disease is disorganization or loss of the pit pattern and/or microvascular pattern. Endoscopists may be alerted to this by a transition point in the surface structure of the lesion from the regular pits (type III or IV) to a disordered, less obvious pit structure (discussed later).
The decision to proceed with surgery or endoscopic therapy should also take into account factors such as patient comorbidities and lesion location. For example, surgery in the rectum compared with the right colon is more complicated and associated with higher morbidity, whereas EMR in the right colon is associated with an increased risk of delayed bleeding ( Fig. 1 ).
Endoscopic Mucosal Resection Versus Endoscopic Submucosal Dissection
Performing ESD in the colon differs from gastric ESD. A narrow elongated lumen, a more difficult and less stable scope position, and a thinner wall all contribute to a more challenging procedure with a higher risk for perforation. Consequently, colonic ESD requires careful consideration and should be reserved for specific indications. The Japan Colorectal ESD Standardization Implementation Working Group issued recommendations on the indications for colonic ESD. These recommendations are summarized in Box 2 . Most Western experts do not support the use of ESD for noninvasive disease.
- •
Large (>20 mm) lesion amenable to endoscopic treatment in which en bloc EMR resection is difficult
- •
Laterally spreading tumor-nongranular: particularly of the pseudodepressed type
- •
Lesions with Kudo type V pit pattern
- •
Carcinoma with known SMI
- •
Large depressed lesion
- •
Large elevated lesion suspected of harboring cancer
- •
Mucosal lesions with significant fibrosis
- •
Local residual early cancer after prior endoscopic resection
- •
Sporadic localized tumors in chronic inflammation, such as ulcerative colitis
The benefit of en bloc excision by ESD rather than piecemeal EMR is based on the potential for improved outcomes in 3 areas:
- 1.
More accurate histologic assessment; this may be of benefit with subtle SMI that could be missed or misinterpreted by histopathologists when assessing multiple piecemeal specimens. In practice, this has not proved to be clinically relevant.
- 2.
Reduced recurrence; however, recent long-term outcomes from large prospective EMR series show that recurrence is usually diminutive and easily treated in follow-up and does not affect the likelihood of long-term remission.
- 3.
The possibility of cure in low-risk SMI; based on large Japanese series to date, such cases are infrequent, comprising only approximately 10% of patients treated by ESD.
Thus in most cases an initial ESD strategy to treat AMN offers little additional benefit to patients, but exposes them to increased procedural risks and the health service to increased costs, decreased colonoscopy capacity, and mandatory multiday hospital admission for these patients.
Available data suggest that, compared with EMR, ESD has a higher rate of curative resection (91%–95% vs 79.6%). In a systematic review of 22 studies with 2841 treated lesions ESD en bloc resection rates of 96% with an R0 rate of 88% were observed. However, most of these studies were retrospective and thus subject to bias, because cases in which ESD was not possible or was technically unsuccessful are generally not reported. Moreover, recent prospective studies performed in Western countries show lower en bloc resection rates of 64% to 90% and R0 rates of 53% to 80%.
ESD is also associated with a longer procedure time (75–121 minutes vs 25 minutes), increased risk of perforation (3%–7.4% vs 1%–2%), increased risk of hospitalization (100% vs 7.7%), and increased costs compared with EMR.
Taken together, it seems that ESD confers little, if any, significant advantage compared with EMR in the colon and has not been proved thus far to be associated with better long-term outcome.
Proficiency and Safety
Performing safe and effective endoscopic resection of AMN requires selective expertise and should preferably be performed in tertiary referral centers with a high volume of cases. Physician training in EMR and ESD requires dedicated training programs that incorporate clinical, interpretive, and technical skills. Equally important is the availability of highly trained nursing staff and surgical, radiological, and anesthetic support.
Technique/procedure
Patient assessment and preparation
- •
Candidates for advanced endoscopic resection should fully consent to the procedure and its alternatives.
- •
A comprehensive medical history and medication list is essential. Newer generation anticoagulants and antiplatelets are increasingly used. These agents are highly potent, have a rapid onset of action, and physicians should be aware of the risks and recommendations regarding cessation and resumption of these medications in the extreme environment of advanced endoscopic resection ( Table 1 ).
Table 1
Recommendations on cessation and resumption of antiplatelets and anticoagulation medications in high-risk advanced endoscopic resection
Mode of Action
Elimination/Metabolism Route
Cessation: Preprocedure
Resumption: Postprocedure
High-risk Thromboembolic Condition
Low-risk Thromboembolic Condition
High-risk Thromboembolic Condition
Low-risk Thromboembolic Condition
Aspirin
COX inactivation
—
No
7 d before
NA
POD 1
Clopidogrel
ADP receptor antagonist
Hepatic
7 d before a
7 d before
POD 1
POD 1
Prasugrel
ADP receptor antagonist
Hepatic
7 d before a
7 d before
POD 1
POD 1
Warfarin
Vitamin K antagonist
Hepatic
5 d before b
5 d before
POD 1 b
POD 1
Dabigatran
Direct Thrombin inhibition
Renal
5 d before c
5 d before c
POD 2 d
POD 7 d
Rivaroxaban
Factor Xa inhibition
Renal
3 d before c
3 d before c
POD 2 d
POD 7 d
Apixaban
Factor Xa inhibition
Renal
3 d before c
3 d before c
POD 2 d
POD 7 d
Abbreviations: ADP, adenosine diphosphate; COX, cyclooxygenase; POD, postoperative day.
Data from Baron TH, Kamath PS, McBane RD. New anticoagulant and antiplatelet agents: a primer for the gastroenterologist. Clin Gastroenterol Hepatol 2014;12(2):187–95; and Veitch AM, Baglin TP, Gershlick AH, et al. Guidelines for the management of anticoagulant and antiplatelet therapy in patients undergoing endoscopic procedures. Gut 2008;57(9):1322–9.
a High-risk conditions always warrant discussion with the patient’s cardiologist (for clopidogrel, presence of coronary artery stents; for anticoagulants, prosthetic metal heart valve, atrial fibrillation plus mitral stenosis, <3 months following venous thromboembolism event, thrombophilia syndromes).
b For high-risk condition, start low-molecular-weight heparin (LMWH) 2 days after stopping warfarin; omit LMWH on day of procedure. On resumption of warfarin, continue LMWH until adequate International Normalized Ratio is reached.
c Recommendations are for patients with normal renal function. Impaired renal function warrants longer cessation of dabigatran and rivaroxaban.
d The ideal timing of resumption of newer anticoagulation agents following high-risk endoscopic procedures is not known. Expert opinions are presented here.
- •
The patient’s comorbidities need to be factored into the therapeutic process at all stages.
- •
Dedicated referral letters should include a description of the lesion location, size, Paris morphology, and color images.
- •
Biopsies are not mandatory because they rarely provide additional data relevant to treatment selection. Aggressive tunneling biopsies are best avoided because of their propensity to create submucosal fibrosis, which can require a more technically complicated procedure.
- •
Standard colonoscopic techniques and patient positioning are used to reach the lesion.
- •
The lesion should be positioned at 6 o’clock in the endoscopic field, preferably on the nondependent side. Ideally, the patient should also be positioned in a way that any fluid or resected specimens accumulate away from the lesion. Thus the working field is not obscured and, in the event of complications (bleeding or perforation), access is optimized and the risk of extraluminal contamination is minimized.
- •
Polyps in difficult positions (eg, around folds, ileocecal valve [ICV], flexures) may require multiple patient position and endoscope changes and ancillary devices such as a short distal cap attachment.
Technical preparations
Submucosal Injectate
The submucosal injection solution is composed of 3 main elements: the colloid solution, diluted epinephrine, and an inert dye (indigo carmine or methylene blue).
- •
The colloid fluid creates a cushion between the mucosa and muscularis propria (MP), thus reducing the risk of deep tissue entrapment in the snare, transmural thermal injury, and perforation. Normal saline (NS) is commonly used; however, in a randomized trial, the inexpensive colloid plasma volume expander succinylated gelatin (Gelofusine; Braun, Melsungen, Germany) was superior to NS, requiring significantly fewer injections, fewer resections, and an overall reduced EMR time.
- •
Diluted epinephrine (adrenaline, 1:100,000) in the injectate may be used to reduce minute bleeding, keeping the EMR field dry and delaying the dispersion of the submucosal injectate. It does not seem to significantly alter clinically significant post-EMR bleeding (CSPEB).
- •
An inert dye (80 mg of indigo carmine or 20 mg of methylene blue in a 500-mL solution) avid for the connective tissue of the submucosa is also added to the solution. This dye facilitates better delineation of the lesion margins, especially in Paris 0-IIa/IIb, nongranular (NG), or serrated lesions. In addition, the extent of the submucosal cushion is defined and the correct plane of resection can be confirmed at all times.
The Electrosurgical Unit Settings
Microprocessor-controlled electrosurgical generators delivering alternating cycles of high-frequency short-pulse cutting with more prolonged coagulation current are now most commonly used (VIO 300D; ERBE, Tübingen, Germany. ESG100; Olympus Medical, Tokyo, Japan). Tissue impedance is sensed via signals from the return electrode, adjusting power output in order to avoid deep tissue injury.
Carbon Dioxide Insufflation
The routine use of CO 2 for insufflation during endoscopic procedures has gained wide acceptance in recent years. Multiple studies comparing the use of CO 2 with air insufflation have consistently shown reduce postprocedural pain and bloating without any adverse effects. The use of CO 2 insufflation during EMR of large colonic lesions has been shown to significantly reduce postprocedural admissions for pain, which improves and clarifies postprocedural decision making, reducing unnecessary imaging and cross-team consultations.
Snares
A variety of snares are available for use in advanced colonic resection. The choice of a specific snare may be influenced by lesion size, morphology, and location; however, on many occasions it is a personal preference ( Fig. 2 ). Stiff serrated (spiral) snares are preferred to increase tissue capture. The 20-mm spiral snare (wire diameter, 0.48 mm) is mainly used for large en bloc or wide-field piecemeal resections. The 15-mm braided snare is also commonly used for piecemeal EMR. Additional snares of various sizes (10–20 mm and occasionally larger) and shapes (oval, round, and hexagonal) are occasionally required. Small thin wire snares (wire diameter, 0.3 mm) are used to remove tissue in difficult locations and situations (periappendiceal, submucosal fibrosis) and small residual tissue within and at the margin of the defect.
Approach
Lesion Assessment
Endoscopic assessment before lesion resection is essential. This assessment entails an overview evaluation of lesion morphology, followed by targeted interrogation of any suspicious areas. A thorough assessment can identify lesions with significant scarring or inaccessibility and may also identify lesions with possible SMI, which may dictate a change in the endoscopic approach or referral for surgical treatment.
Overview
The Paris classification of superficial neoplasia should be used for morphologic classification ( Fig. 3 ). Polypoid lesions are elevated more than 2.5 mm above the surrounding mucosa. They are divided into sessile (0-Is), pedunculated (0-Ip), or mixed (0-Isp) types. Nonpolypoid lesions may be slightly elevated (0-IIa), completely flat (0-IIb), or slightly depressed (0-IIc). Excavated lesions are labeled (0-III). Sessile lesions can be further categorized based on their surface topography into granular (G), nongranular (NG), or mixed morphologies ( Fig. 4 ).
Paris classification and surface topography may be used to predict lesion histopathology and stratify the risk of SMI. Low-risk lesions include homogeneous 0-IIa G types. The highest risk lesions are 0-IIa + c NG types. Intermediate risk lesions are heterogenous 0-IIa types with a depressed (IIc) or elevated (Is) component. Furthermore, the 0-Is nodule or 0-IIc depressed component in a flat lesion are the mostly likely sites to harbor SMI. Fig. 4 summarizes the relative risk of SMI in commonly encountered morphologic types. Recognition of these high-risk features should guide targeted surface assessment of the lesion and may ultimately dictate a change in the treatment strategy.
Focal interrogation
Focal interrogation is vital and involves assessing areas at risk of SMI, particularly nodules or depressed areas, evaluating the pit pattern and vascular patterns (VP). The mucosal pit pattern was initially described with the use of chromoendoscopy by Kudo and colleagues 2 decades ago. The use of high-definition white-light endoscopy may now achieve sufficient accuracy in most situations.
AMN most commonly display pit pattern types III (large or elongated pits) or IV (branching gyruslike pits). Type III predicts tubular adenoma (TA) and type IV tubulovillous adenoma (TVA). Type V is associated with cancer. Vi (irregularly mixed pits) suggests at least intramucosal cancer (World Health Organization 2010 now terms this high-grade dysplasia) and Vn (nonstructural pit pattern) is predictive of submucosal invasive malignancy.
Narrow-band imaging (NBI) and postprocessing imaging techniques, such as Fujinon intelligent chromoendoscopy and I-scan, highlight the mucosal VP. The modified Sano system classifies lesions by their mucosal microvasculature as seen with NBI. An organized brown capillary network surrounding the pits can be seen in tubular adenomas, whereas irregular complex branching capillaries or avascular areas suggest advanced histology and cancer.
Pit pattern and VP have also been shown to accurately predict deep SMI (>1000 microm). More recently, the Narrow-band Imaging International Colorectal Endoscopic (NICE) criteria, a composite score including lesion color, surface pattern, and capillary pattern, has been shown to accurately predict SMI in flat colonic lesions. Fig. 5 summarizes the various focal interrogation techniques.
Resection techniques
Endoscopic Mucosal Resection
Core injection principles
- •
An adequately performed injection is vital to successful resection. The injection should elevate the lesion into the lumen and toward the colonoscope, improving access.
- •
Excessive injection should be avoided because this may hinder adequate visualization and create excessive tension within the cushion, which makes snare capture of adequate tissue challenging. For large piecemeal EMR, we recommend an inject-and-resect technique with 1 to 3 resections per injection.
- •
Inadequate injection may be appreciated by ( Fig. 6 ):
- ○
Ongoing injection without tissue elevation or intraluminal fluid escape; this indicates transmural placement of the needle tip with extramural injection. If excessive, this may cause postprocedural pain. Draw back gently to find the correct plane.
- ○
Mucosal injection: the immediate appearance of a superficial blue bleb without lifting of the lesion.
- ○
Jet sign: when significant submucosal fibrosis is present, a jet of fluid exits the lesion at high pressure during injection.
- ○
A canyoning effect may occur as a result of central fibrosis. The lesion remains anchored in its original position, but the tissue of the perimeter elevates. This effect makes the lesion difficult to access, ensnare, and remove.
Fig. 6
Examples of inadequate submucosal injection.
- ○
- •
Poor mucosal lifting usually indicates one of 2 things: invasive disease or submucosal fibrosis caused by previous aggressive biopsy or resection attempts. In some NG LSTs it may be caused by lesion biology. Sometimes at the flexures or in the rectosigmoid in lesions with a Is component it may be caused by repeated tumor prolapse.
- •
Position the needle tip tangentially to the mucosal surface and gently touch the surface. Ask an assistant to commence the injection while simultaneously stabbing the mucosa with the needle tip. The correct plane is confirmed by an immediate elevation of the mucosa.
- •
Use a dynamic injection technique: pull back slightly on the injection catheter or colonoscope and slowly deflect the tip of the colonoscope up while maintaining the position of the needle tip in the submucosal plane. The mucosa may even be gently rotated into the lumen by torquing the endoscope shaft.
Core snaring principles
- •
The lesion should be removed in as few pieces as safely possible. Provided safety is ensured, en bloc resection should be considered for lesions less than or equal to 20 mm, which may allow a more accurate histologic assessment and reduce the risk of recurrent/residual tissue compared with piecemeal resection.
- •
Orient the lesion in the 5 to 6 o’clock position.
- •
Attempt to resect any large nodules (Is) first in 1 piece and submit them separately for pathologic analysis because these are most likely to harbor an invasive component ( Fig. 7 ).
