36 Advanced Colorectal Polyps and Early Cancer Resection
David James Tate and Michael John Bourke
The incidence and mortality 1 of colorectal cancer can be significantly decreased by colonoscopy and polypectomy. 2 Most colorectal polyps are small (< 10 mm) and can be easily treated by endoscopists with the relevant degree of training using cold snare polypectomy 3 , 4 or conventional electrosurgical polypectomy where necessary. ACPs are generally considered as those greater than or equal to 20 mm and these larger lesions have a much greater frequency of advanced histology and invasive cancer, are more challenging and hazardous to completely remove endoscopically, and require special techniques to safely achieve this. Some smaller lesions also fall into this group due to morphologic features suggesting advanced histology (e.g. depressed component). Approximately 2% of colorectal lesions are flat, 20 mm or larger, and termed laterally spreading lesions (LSLs). 5 Even very large LSLs limited to the mucosa are resectable endoscopically due to the lack of lymphatic drainage from this area. LSLs were traditionally managed surgically, but a growing body of evidence supports the similar efficacy 6 and durability 7 of ER versus surgery and more recently the superior cost-effectiveness 8 and safety profiles of ER, particularly in dedicated tertiary centers. Once SMIC has developed, there is still a possibility for ER if certain criteria are met, although this is more controversial. ▶Box 36.1 lists the indications for endoscopic therapy of advanced colorectal polyps. Over the last decade, high-quality prospective studies have emerged that provide an evidence base for the technique and safety of ER for ACPs. Two techniques exist: endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD). The main benefit of ESD is en bloc resection providing accurate histopathologic assessment of submucosal invasive cancer (SMIC), but this is at the expense of significantly greater complications. EMR is the main technique used in Western centers to resect large LSLs in the colon. Complications of EMR are infrequent and, in the vast majority of cases are controlled endoscopically; these include intraprocedural bleeding (11.3%), clinically significant postendoscopic bleeding (CSPEB, 6%), and perforation (1.3%). Adenoma recurrence rates of 10 to 20% are reported in high-volume centers, but this is easily resected endoscopically at surveillance procedures. Novel techniques promise to reduce the rate of recurrent adenoma and further predict those lesions that will recur. Novel procedural techniques in ER are promising, but require validation in prospective, multicenter randomized trials.
Box 36.1 Indications for endoscopic therapy of advanced colorectal polyps
Indications for endoscopic therapy of advanced colorectal polyps
LSLs with no features suggestive of deep SMIC (e.g., those with Kudo II–O/III/IV, Sano II, NICE type II morphology)
Large pedunculated lesions arising from the mucosa
Previously attempted LSLs or LSLs in difficult locations (e.g., periappendiceal, ileocecal valve, anorectal junction)
Indications for surgery for advanced colorectal polyps
LSLs with features of deep SMIC (Kudo Vi/n, Sano IIIb vascular pattern, NICE type III 19 , nonlifting, Paris 0–IIa + c or Paris 0–III morphology)
Potentially LSLs with previously failed endoscopic resection at a specialized tertiary center
LSL, laterally spreading lesions
SMIC, submucosal invasive cancer
NICE, Narrow-Band Imaging International Colorectal Endoscopic
36.2 Technical Aspects and Preparation
ER of colonic ACPs requires training, adequate case volume to maintain skills, tertiary-level radiology and surgical support and a histopathologist with a dedicated interest in colorectal neoplasia.
36.2.1 Patient Preparations
Outcomes are optimized when patients are managed within an advanced tissue resection network. Research infrastructure is also highly desirable as many important clinical questions remain unanswered. Prospective monitoring of procedural and clinical outcomes with benchmarking against accepted standards is a minimum requirement. 9 Regular clinical meetings between stakeholders with discussion of interesting and challenging cases facilitate best patient care.
Seamless referral pathways that facilitate rapid and accurate transfer of data are preferable. This should include detailed imaging, a description of the lesion and comorbidities of the patient including anticoagulant medications and the indication for their use. 10 Biopsy prior to the referral of a colonic LSL is not necessary unless invasive cancer is strongly suspected; extensive photodocumentation provides more useful information. Biopsy commonly results in submucosal fibrosis, increasing the complexity of ER. Endoscopic tattoo placed on the opposite wall to the lesion for ER is encouraged to mark lesions that may be difficult to locate later.
Informed consent is vital. Complications of endoscopic mucosal resection (EMR) include deep injury to the colonic wall (deep mural injury [DMI]), bleeding, postprocedural pain, serositis, and recurrent or residual adenoma (RRA). For EMR of large colonic LSLs, rates of 1.3% 6 for colonic perforation, and up to 7% for post-EMR bleeding are quoted. Pain after EMR is uncommon and usually self-limiting, but must be reported. Patients are given contact details such that timely advice may be given should complications occur once they have left the endoscopy unit.
36.2.2 Techniques of Endoscopic Resection
Two established techniques exist for the ER of large LSLs: EMR and ESD. EMR has been refined extensively since its inception in 1977. 11 The technique involves expansion of the submucosal layer with a chromoinjectate, and placement of a snare over the target lesion. Closure of the snare with the addition of microprocessor-controlled fractionated current transects the tissue and the submucosal cushion provides a heat sink and safety barrier against ensnaring and damaging deeper structures. Lesions up to 20 mm in size can be removed en bloc, with larger lesions usually requiring piecemeal resection.
ESD was initially developed in Japan for the en bloc excision of early gastric cancer, avoiding the morbidity associated with surgery. The technique involves expansion of the submucosal plane with chromoinjectate, use of an endoscopic knife to incise the margin of the lesion (incision phase), and then separation of the lesion from the deeper structures in the submucosal plane (dissection phase) using various types of electrosurgical current and the endoscopic knife.
The main advantages afforded by ESD over EMR are derived from en bloc resection of the target lesion. This results in reduced recurrence in short- and medium-term follow-up, possible cure in low-risk submucosal invasive cancer, and a superior specimen for histologic assessment. Long-term follow-up studies of EMR for LSL greater than 20 mm, however, show that if the initial EMR was technically successful, then after two follow-up procedures at intervals of 4 and 12 months more than 98% of patients are free of recurrence and considered cured.
ESD can be used to treat LSLs with potential superficial SMIC to achieve cure. However, in large Japanese series, the number of such patients is approximately 10% and thus a universal ESD strategy does not offer a true benefit to the majority. 12 Moreover, this benefit only applies if both patient and physician decide and agree that surgery is not necessary despite submucosal invasive cancer.
These benefits come at a cost. ESD is technically challenging, significantly more time consuming, and is associated with a significantly higher rate of complications in comparison to EMR; even in expert Japanese centers, perforation rates are higher (5.7 vs. 1.4%) and mean procedure durations significantly longer (65.9–108 minutes vs. 29–30 minutes). 13 In addition, multinight hospital stay is mandated for all lesions removed by ESD whereas it is required for fewer than 5% 6 of those removed by EMR. However, the strongest argument against a universal ESD approach for all LSLs is that it does not decrease the rate of additional surgery after ER. In Japanese centers, an ESD-only approach was associated with a significantly higher rate of surgery (9.9 vs. 5.8%) than EMR. 13
36.2.3 Equipment Required
Historically, the submucosal injectate for EMR was normal saline (NS), however, it results in a nonsustained mucosal lift and does not delineate the lesion margin. The optimal submucosal injectate contains three constituents:
A colloid solution, for example, succinylated gelatin (e.g., Gelofusine; B. Braun, Sempach, Switzerland) has been shown to be superior to NS in a double-blind randomized trial, 14 resulting in significantly fewer injections and resections and a halving of the procedure time. Other solutions have also been described including hyaluronic acid, dextrose solution, and hydroxyethyl starch.
The addition of an inert dye to the injectate (chromoinjectate) allows for accurate delineation of the lesion margin; this is particularly useful for lesions with serrated or nongranular morphology to ensure complete resection. It also shows the extent of the submucosal cushion, the safe zone for EMR. Methylene blue and indigo carmine (e.g., 80 mg of indigo carmine or 20 mg of methylene blue in a 500-mL solution) are in commonplace use. They are avid for the submucosal areolar tissue and create a relatively homogeneous “blue mat” appearance when resection is within the submucosal plane (▶Fig. 36.1).
Dilute adrenaline (1: 100,000 solution) added to the injectate reduces intraprocedural bleeding maintaining a clean EMR field and delays the dispersion of the submucosal injectate. It may also reduce the rate of CSPEB. 15
The use of a microprocessor-controlled electrosurgical generator capable of delivering fractionated current in short cutting bursts interspersed with longer coagulation pulses is essential for safe ER. These are now commonplace in tertiary endoscopy units and include, for example, ERBE VIO 300 (ERBE, Tübingen, Germany) or Olympus ESG 100 (Olympus, Tokyo, Japan). The return electrode senses tissue impedance and modifies current delivery to achieve the desired result.
Insufflation of Carbon Dioxide
The superiority of carbon dioxide insufflation over air during gastrointestinal (GI) endoscopy is firmly established, particularly with regard to decreased postprocedural pain, flatus, and bowel distension. 16 During EMR for large LSLs, carbon dioxide insufflation resulted in significantly less postprocedural admissions for pain in a large prospective series. 17
Snares for Endoscopic Resection
A complete suite of snares of various sizes, shapes, stiffness, and wire diameters is necessary. Snares with a thinner wire diameter provide greater current density and owing to this and their narrower caliber, more swiftly transect the target tissue. The workhorse for EMR of LSLs is the 20-mm “spiral” snare (0.48-mm wire diameter). This snare has a series of serrations covering the wire and facilitates the capture of normal tissue at the lesion margin. Increasingly, there are a range of alternatives available and there is renewed interest in the performance characteristics of various snares for particular indications. Commonly, in the right colon we use a 15-mm braided snare; the smaller size is possibly safer. Different wire stiffness, snare shape (oval, hexagonal), and performance characteristics can facilitate removal of a variety of potentially refractory lesions. Stiff thin-wire snares (0.3 mm) can allow tissue capture of residual adenoma at resection margins or previously attempted adenoma associated with significant submucosal fibrosis (▶Fig. 36.2). Larger snares may be required for pedunculated or bulky exophytic lesions.
36.3 Lesion Assessment
Techniques to interrogate LSLs in real time prior to ER and determine the presence and degree of SMIC have been developed in recent years.
Overview assessment of the lesion is performed using high- definition white light (HD-WL). Lesion morphology should be described using the Paris classification. 18 This describes a flat lesion with less than or equal to 2.5-mm elevation above the surrounding mucosa as 0–IIa, with a central depression as 0–IIc, and a completely flat lesion as 0–IIb. 0–Is lesions are broad based but elevated greater than 2.5 mm above the surrounding mucosa. Combinations of these terms exist (▶Fig. 36.3). 0–Ip and 0–Isp lesions are polypoid and semipolypoid, respectively, and classified separately. Sessile lesions are also labelled based on their surface morphology as granular (G) or nongranular (NG) (▶Fig. 36.4).
Focal interrogation of the lesion is then performed with HD-WL and image enhancement techniques such as narrow-band imaging (NBI) 19 (Olympus) or FICE (Fujifilm Medical, Saitama, Japan). The intention is to assess the pit pattern and vascular pattern of the lesion. Areas suspicious for SMIC, which are commonly demarcated, are identified and interrogated in turn, possibly with the addition of magnification (▶Fig. 36.5).
Three systems exist for such assessment (▶Table 36.1). Kudo 20 described five types of pit pattern; types III and IV indicate noninvasive disease and are suitable for ER. Experts suggest that the combination of chromic dye and magnification is required to assess the Kudo pit pattern. The Sano classification 21 is based on the capillary pattern as observed under NBI. Three types are recognized. Sano type II indicates noninvasive disease suitable for ER. Recently, the NBI International Colorectal Endoscopic (NICE) 19 criteria have been described. These are based on a combination of color, vessel, and surface pattern. The score has been reported to have a sensitivity of 94.9% and negative predictive value of 95.9% when any one of the three criteria demonstrate evidence of submucosal invasion, with substantial interobserver agreement (Kappa value 0.70).
Combining Paris classification, surface morphology, and lesion location is very useful for stratifying the risk of SMIC. Distal colonic lesions in general have a higher risk of SMIC. In large Japanese and Western 6 , 22 studies, 90% of lesions greater than or equal to 20 mm presenting for endoscopic resection are granular and greater than 75% are Paris 0–IIa class. In the absence of Kudo V pit pattern or a depressed “c” component, both strong predictors of invasive disease, the risk of SMIC varies with LSL morphology 23 (▶Fig. 36.6). In granular lesions, the risk is 0.9% for 0–IIa, 2.9% for 0–Is, and 7.1% for 0–IIa/Is lesions, with a greater risk in the distal colon. Nongranular lesions are at higher risk for SMIC with a 4% risk for 0–IIa NG lesions, 12.8% for 0–IIa + Is NG, and 16.7% for 0–Is NG. Again, there is generally a greater risk in the distal colon. This information can inform a targeted approach to ER, employing en bloc excision in the case of those lesions with a predicted high risk for SMIC.
36.4 Resection Technique
36.4.1 Endoscopic Mucosal Resection
An adequate submucosal cushion is required for safe EMR, and is used to improve access to the lesion. A deficient cushion risks deep injury to the colonic wall; an excessive cushion creates tension and can impede snare capture and visibility during the resection (▶Fig. 36.7).
The technique for injection is as follows:
Position the lesion at 6 o’clock in the colonoscopic view.
Place the tip of the catheter at 30 to 45 degrees tangential to the lesion.
Ask the assistant to extend the needle.
Ask the assistant to commence injection while simultaneously stabbing the mucosa with the needle tip. Confirmation that the submucosal plane has been entered is with immediate and swift elevation of the lesion.
Dynamic movement of the catheter during injection can be used to elevate the lesion to the required orientation with the needle tip anchored in the submucosal cushion. The most inaccessible portion of the lesion can often be exposed in this manner.
If the lesion does not lift, consider the following:
Extramural needle placement: Gentle pulling back on the injection catheter will usually locate the submucosal plane.
Needle tip remaining within the colonic lumen: Recognized by visualization of chromoinjectate spilling into the lumen.
Nonlifting lesion: This can be caused by submucosal fibrosis from a previous resection attempt, previous biopsy, or lesion biology (SMIC). The appearances of canyoning (lifting of the surrounding normal mucosa without lifting of the lesion) and the “jet sign” (rapid ejection of the chromoinjectate from the lesion) are confirmatory signs associated with nonlifting.
Intramucosal injection is another type of failed injection. This can be recognized by immediate bleb-like elevation of the mucosa with no simultaneous lifting of the lesion. The bleb can be punctured with the needle tip and the injection repeated.