Pragmatic Approach to Complex Colon Polyps

Fig. 4.1

Paris endoscopic classification of superficial neoplastic colorectal lesions

Such morphological classifications serve not only to standardise descriptions of LSLs but also convey important information regarding SMI risk within the LSL. For example, data from a large, prospective multicentre cohort study showed granular 0-IIa LSLs harbour a very low risk of SMI (approximately 1%), granular 0-IIa + Is LSLs harbour an intermediate risk of SMI (approximately 7%) and non-granular 0-Is LSLs harbour a high risk of SMI (approximately 15%) [5, 11]. Figure 4.2 summarises the risk of SMI within LSLs according to their morphology and location within the colon [11].

../images/442926_1_En_4_Chapter/442926_1_En_4_Fig2_HTML.png — Fig. 4.2

Risk of occult SMI in LSLs according to morphology and location. (Adapted from Burgess et al. *Gastroenterology* 153(3), 2017; 732–742.e1)

Surface Assessment

Assessment of a lesion’s surface pit pattern using the Kudo classification may distinguish non-neoplastic from neoplastic polyps [12]. This classification was originally developed using chromoendoscopy and magnifying colonoscopy, but its principles can also be applied using high-definition colonoscopes with magnification and NBI or FICE. According to this classification, Kudo type I appear as roundish pits; Kudo type II are stellar or papillary pits; Kudo type III-S are small, roundish, tubular pits (smaller than type I) and Kudo type III-L are roundish, tubular pits (larger than type I); Kudo type IV are branch-like or gyrus pits; Kudo type Vi are irregular pits; and Kudo type Vn are non-structural pits with an amorphous structure.

Polyps with type I or II pits are considered benign (e.g. normal, hyperplastic or inflammatory), whereas polyps with type III are usually tubular adenomas, type IV usually contain villous histology, type Vi are indicative of superficial SMI and type Vn are associated with deep SMI. For the purpose of determining suitability of ER, it is important to exclude the presence of any type V pit pattern, although diminutive focal SMI may sometimes still be present within an LSL with a seemingly intact surface pit pattern. Other classification systems that assess surface vascular pattern (Sano and NBI International Colorectal Endoscopic [NICE] classification) [13, 14] are also validated to assess the histology of colorectal polyps.

Taken together, stratification of SMI risk within an LSL is best determined by assessing both its surface pit pattern and gross morphology. This is supported by findings from a recent large prospective multicentre study of 2277 LSLs of ≥20 mm size, where particular features that should alert endoscopists to an increased risk of covert SMI included recto-sigmoid located LSLs (odds ratio [OR] 1.91), lesions with Paris 0-IIc and 0-Is morphologies (OR 1.80 and OR 2.73, respectively), non-granular surface morphology (OR 2.80), increasing lesion size (OR 1.12 per 10 mm increase) and Kudo pit pattern V (OR 14.2) [11].

Lesion Biopsy

Routine lesion biopsy to determine its histology and/or exclude presence of SMI prior to ER is generally discouraged. Lesion biopsy, particularly over flat (0-IIa or 0-IIb) areas, may cause significant submucosal fibrosis (SMF). This can hinder adequate expansion of the submucosal space following injection and impede subsequent ER. Biopsies from 0-Is nodules, which are more likely to harbour SMI, may result in less SMF; but this has not been confirmed in any large study. Nonetheless, biopsies are prone to sampling error, as they are not representative of an entire lesion and therefore may potentially underestimate its histological grade. As such, lesions without obvious SMI after careful endoscopic assessment are best referred for ER without biopsy. However, if SMI is suspected, targeted biopsies to confirm invasive disease can be considered prior to further management.

Endoscopic Mucosal Resection

EMR is the recommended first-line therapy for colorectal LSLs ≥20 mm in size (Figs. 4.2 and 4.3). On an intention-to-treat basis within tertiary centres, EMR was technically successful for >95% of lesions with >90% avoiding surgery on long-term follow-up [5, 15]. Large prospective studies have also demonstrated the safety of EMR [5, 16]. Economic modelling studies have shown EMR treatment of LSLs is significantly more cost-effective compared with surgery, saving between 7000 and 13,000 USD per patient with reduction in hospitalisation length of stay by 2.8 days [4, 17]. Nonetheless, every case should be assessed based on the risks and benefits of resection and in the context of a patient’s overall health and comorbidities. The following paragraphs discuss the technique of EMR, emphasising several key technical aspects that have been shown to improve complete resection, reduce recurrence and lower rates of adverse events.

../images/442926_1_En_4_Chapter/442926_1_En_4_Fig3_HTML.jpg — Fig. 4.3

EMR of 40 mm LSL in the proximal ascending colon. (a) A granular, Kudo IV, 40 mm Paris 0-IIa + Is LSL in the proximal ascending colon. The lesion is in the dependant side, and the patient is repositioned to improve access. (b) Submucosal injection with formation of a fluid cushion beneath the lesion. (c) Snare resection of the 0-Is component is performed to allow access to the remaining lesion. (d) Piecemeal resection proceeds progressively from one side of the lesion to the other. (e) Completion of EMR with blue-stained submucosal tissue indicating absence of deep injury. A non-bleeding arteriole is visible within the resection defect and does not require prophylactic treatment. (f) Snare tip soft coagulation applied to resection margins. This has been shown to reduce adenomatous recurrence

Patient Preparation

All patients need to be consented to the procedure and informed of its risks and treatment alternatives. Referrals should include description of the lesion size, location and Paris morphology including colour images. The endoscopist should be aware of a patient’s full medical history and medication list, in particular any regular antithrombotic agents. We suggest that all antiplatelet agents, including aspirin, be withheld for 7 days so that their effects are not active at the time of ER. Similarly, anticoagulants should also be withheld for sufficient duration, but this period varies according to medication and the patient’s renal function. Endoscopists should refer to published guidelines [18, 19] or seek the opinion of a haematologist if unsure. Temporary cessation of these medications may substantially increase the risk of thrombosis, and some cases (e.g. recent coronary stent or significant thromboembolism) are best discussed with the patients’ treating specialty physician before proceeding.

Once the lesion is located at colonoscopy, it should be positioned at 6 o’clock, ideally on the non-dependant side of the lumen. Correct positioning is important as this avoids obscuration of the working field by luminal contents, optimises access and minimises extra-luminal contamination in the event of a perforation. Carbon dioxide is routinely used for insufflation in all endoscopic procedures and has been shown to significantly reduce post-procedural admissions for pain after EMR [20]. Lastly, for polyps in difficult positions (e.g. ileocaecal valve, extending over folds or flexures), multiple patient position changes and use of a short distal attachment cap may assist access, positioning and resection of the lesion. Use of an adult colonoscope where possible may help reduce procedure time as its larger working channel (3.7 mm) compared with a paediatric colonoscope (3.2 mm) allows more specimens to be suctioned directly into a polyp trap, rather than requiring retrieval using a net. However, for difficult-to-access positions, especially where retroflexion views are required, the increased flexibility of the paediatric colonoscope may be advantageous.

Submucosal Injectate and Snares

The submucosal injectate includes a colloidal solution, epinephrine diluted to 1:100,000 and an inert dye (80 mg indigo carmine or 20 mg methylene blue per 500 mL solution). Fluid injection into the submucosa creates a cushion between the mucosal polyp and muscularis propria (MP). This reduces the risk of entrapment of deeper tissues by the snare, thereby avoiding deep mural injury (DMI) after resection. Use of colloid solutions such as succinylated gelatin (Gelofusine; Braun, Melsungen, Germany) is preferred as it is associated with reduced number of injections and resections and procedural duration compared with normal saline [21]. Diluted epinephrine reduces intra-procedural bleeding (IPB) and delays dispersion of the submucosal injectate, but does not alter the incidence of clinically significant post-procedural bleeding (CSPEB) [22]. The inert dye is avid for the submucosal connective tissue, which facilitates identification of lesion margins and the resection plane.

Snares of different sizes, shapes and wire types are often required to complete EMR. Specific snares perform better in different situations, influenced by factors such as lesion size, location and morphology. Stiff spiral or braided snares (0.48 mm thickness wire of 15, 20 and 25 mm diameter) are designed to increase tissue capture and are suited for piecemeal EMR. In most cases, the 15 mm snare provides adequate balance of efficacy and safety, but the choice of snare is usually a personal preference. Small, thin-wire snares (0.3 mm thickness wire of 10 mm diameter) are useful for difficult-to-remove tissue (e.g. peri-appendiceal or submucosal fibrosis) as well as small residual tissue within and at the margin of the EMR defect.

Resection Technique

Submucosal Injection

The needle tip should be primed with submucosal injectate, be placed tangentially to the lesion and be touching but not penetrating the mucosa. Begin by asking the assistant to commence injection whilst simultaneously penetrating the needle tip into the submucosa. The correct plane is confirmed by immediate elevation of the polyp, indicating expansion of the submucosa (Video 4.1). Injection and resection is best commenced in the most difficult-to-access aspect of the LSL. Generally, this is the caecal side of the LSL that is behind or below a mucosal fold. By first elevating this side, the lesion is moved toward the scope aiding its access and resection. Sometimes, retroflexion of the scope, particularly for LSLs located in the right colon, may assist injection and elevation of its difficult-to-access side.

Submucosal injection is performed using a dynamic technique by simultaneously pulling back slightly on the injection catheter and gentle rotation of the colonoscope with slow upward tip deflection whilst keeping the needle tip within the submucosa. This method ensures a more even distribution of the injectate into the submucosa and also allows the endoscopist to manipulate the direction of lift to move difficult-to-reach parts of an LSL into a more accessible location. Endoscopists should be wary to avoid over-injection as excessive tension within the fluid cushion may hinder visualisation and prevent adequate tissue capture by the snare.

Poor lesion elevation may be due to SMF or deep SMI. SMF may result from previous biopsy or ER attempts, reaction to submucosal injection of tattoo particles or prolapse of a large sessile (Paris 0-Is) lesion over a flexure [23]. When SMF is present, a ‘jet sign’, where a jet of fluid exits the lesion at high pressure during injection, or ‘canyoning effect’, where the centre of the lesion remains fixed in its original position but the peripheries elevate, may occur [5, 23]. Inadequate elevation may also result from transmural injection. When this is suspected, gently pull back the needle, whilst continuing injection will assist locating the submucosal plane. Conversely, superficial injection into only the mucosa will result in the immediate appearance of a superficial blue bleb without lesion elevation. In the setting of invasive disease, complete absence of lesion elevation typically only occurs when there is deep SMI with resulting obliteration of the submucosa by tumour infiltration .

Snare Resection

Correct snare placement is essential to ensuring effective resection and procedural safety. As mentioned, the lesion should be first optimally orientated, at 6 o’clock on the endoscopic view. Begin resection on one side of the lesion (usually the most difficult-to-access side) and systematically perform piecemeal resection across the lesion, aiming to resect the lesion in as few pieces as possible. LSLs ≤20 mm are best removed en bloc if safe to do so, as such a specimen improves histological assessment and is associated with less recurrence compared with piecemeal resection.

Open the snare fully over the lesion, taking care to align it along the same plane as the tissue. Then angle down firmly with the up-down dial onto the fluid cushion while gently aspirating gas. The aspiration of gas reduces colonic wall tension and allows the tissue to ‘fall’ into the opened snare. When closing the snare over the edge of a lesion, aim to include a 2–3 mm rim of normal mucosa to ensure complete resection and avoid small residual islands of adenoma at the margins (Video 4.1). Once adequately positioned, gradually close the snare whilst anchoring the catheter into the mucosa. Tight closure of the snare helps to exclude MP from the tissue within the snare. Concern for premature resection is not required as it is generally not possible to transect tissue of more than 10 mm with a braided or spiral snare without the use of diathermy.

It is recommended that the endoscopist take control of the snare for tissue transection as sensory feedback provides information regarding the safety and efficacy of excision. Safe tissue capture is confirmed by three manoeuvres:

Mobility: movement of the snare catheter quickly back and forth should result in independent and free movement of the tissue relative to the underlying colonic wall.
Degree of closure: the snare should close fully (a distance of no more than 1 cm between thumb and fingers). If the endoscopist is unsure, the snare can be partially opened and tented into the lumen to allow any inadvertently captured MP to drop away, before repeating snare closure.
Transection speed: transection should occur quickly. The snare is kept tightly closed while the foot pedal is depressed in short pulses. Usually one to three pulses are sufficient to transect the polyp tissue. A longer transection raises concern of MP entrapment, SMF or SMI.

Microprocessor-controlled electrosurgery is essential for safe and effective EMR. The microprocessor senses tissue impedance and adjusts power output in order to avoid deep tissue injury . Fractionated current that alternates cutting and coagulating cycles is preferred (e.g. Endo Cut mode Q, effect 3, cut duration 1, cut interval 6; ERBE, Tübingen, Germany). Low-voltage coagulation (Soft Coag, effect 4, maximum 80 W; ERBE, Tübingen, Germany) may be used for coagulation of bleeding vessels using the tip of the snare by gently placing this over any points of bleeding (STSC, snare tip soft coagulation) (Video 4.2). Following resection, adequate irrigation of the mucosal defect allows assessment for residual adenomatous tissue and evidence of DMI and assists identification and control of IPB.

The edge of the defect is used as a guide for the next resection. The inside edge of the snare should be aligned along the defect margin, the snare placed over an area of adenoma, and the steps of tissue capture as outlined above repeated. Using the defect edge as the starting point for the subsequent piecemeal resection allows for a systematic approach to EMR (Video 4.3). This method also reduces the risk of adenoma islands forming with the defect, which can be difficult to completely remove, especially when they are small. STSC is then performed by gently touching the snare tip to the EMR margin following completion of resection and has been shown to reduce adenomatous recurrence at the EMR scar from 21% to 6% at first surveillance colonoscopy (RR 0.28, P < 0.001) without an increase in incidence of delayed bleeding or perforation (Video 4.5) [24].

Lastly, LSLs containing a nodule (e.g. Paris 0-IIa + Is lesions) are best treated by first removing the 0-IIa component, followed by resection of the 0-Is component in one piece. The 0-Is component should be submitted separately for histological assessment as these are more likely to harbour SMI. Lesions other than those definitely located within the caecum, adjacent to the ileocaecal valve or in the low rectum, with suspicion of harbouring SMI, should have two to three endoscopic tattoos placed at different points 2–3 cm distal to the resection site [23]. The tattoo serves to assist localisation at surgery or identification of the post-EMR scar for surveillance purposes.

Aftercare

Patients undergoing advanced ER should be closely observed following their procedure. A two-stage management algorithm as shown in Fig. 4.4 is suggested. Patient discomfort and abnormalities in vital signs may herald perforation or significant bleeding and necessitate immediate medical assessment. Most abdominal pain due to benign causes will be responsive to simple analgesia such as intravenous acetaminophen. CT abdomen should be obtained for pain that persists or if associated with clinical signs of peritonitis.

../images/442926_1_En_4_Chapter/442926_1_En_4_Fig4_HTML.png — Fig. 4.4

EMR of 60 mm LSL from the descending colon. (a) 60 mm granular, Paris 0-IIa, Kudo IV LSL in the descending colon. (b) Injection and resection of the lesion is commenced at one side of the lesion and progresses toward the other. (c) Completed EMR with homogeneous blue matt appearance of dye-stained submucosa indicating absence of DMI. (d) Snare tip soft coagulation is applied to the resection margins to treat any nonvisible residual adenoma and reduce risk of adenomatous recurrence

Patients who appear well and are asymptomatic with normal vital signs after a period of observation can be moved to second-stage recovery. Following a successful trial of clear liquids and period of observation, patients may be discharged home on the same day with instructions to represent to the hospital should they develop symptoms of rectal bleeding, fevers or abdominal pain. Those who become unwell during the period of observation are admitted and further investigations organised. A suggested management algorithm of patients following complex colorectal ER is shown in Fig. 4.5.

../images/442926_1_En_4_Chapter/442926_1_En_4_Fig5_HTML.png — Fig. 4.5

Suggested management scheme following advanced resection of colorectal LSLs. (Adapted from Klein A and Bourke MJ *Gastroenterology Clinics of North America* 2015)

Repeat colonoscopy is advised following the index procedure to assess presence of residual or recurrent adenoma. Typically, guidelines suggest that this be performed at 4–6 months, although recent evidence from prospective multicentre data suggests LSLs <40 mm, without IPB nor high-grade dysplasia, have a 91% negative predictive value for residual or recurrent adenoma at first surveillance colonoscopy [25]. For lesions satisfying all these criteria, first surveillance colonoscopy could be delayed until 18 months after index EMR; otherwise, colonoscopy should be performed after 4–6 months.

Variant Techniques

Piecemeal Cold Snare Polypectomy

The main disadvantage to ER using electrosurgery is the risk of delayed bleeding and perforation. Cold snare polypectomy (CSP) is not associated with these risks, but this technique can only resect small polyps (≤10 mm) en bloc. For larger polyps, utilisation of CSP in a piecemeal fashion with or without submucosal injection has the potential to achieve complete excision whilst mitigating many of the adverse events of EMR. The data is so far limited, but available studies of piecemeal CSP for selected sessile adenomatous and serrated polyps 10–19 mm in size show good clinical efficacy with very few significant adverse events, and larger, prospective and comparative studies are awaited [26, 27].

Underwater EMR

Underwater EMR (UEMR ) is performed in a decompressed colonic segment devoid of insufflated air, in which water is infused to assist visualisation. Submucosal injection to lift polyps is not required in this setting as the MP layer remains both circular and distant from the mucosal and submucosal layers, which ‘float’ within the lumen. As such, large pieces of dysplastic mucosal tissue can be removed by snare resection and electrosurgery. Since its first description in 2012, multiple studies have reported its efficacy and safety. Recently, a retrospective comparative study found UEMR was associated with higher rates of complete macroscopic resection and lower recurrence at first surveillance colonoscopy and required fewer procedures to reach curative resection compared with conventional EMR [28]. However, these findings have not been validated in a randomised controlled trial, and it remains an alternate technique to conventional EMR, for which a broader and more robust evidence base is available.

Polyps for Special Considerations

LSLs in the Anorectum

Unlike elsewhere in the colon, lymphovascular drainage of the distal rectum within 5 cm of the anal verge and the anus enters directly into the systemic circulation, bypassing the reticuloendothelial system of the portal system which has a major role in sequestering enteric pathogens. As a result, resection of LSL located in the anorectum is associated with significant bacteraemia, and prophylactic intravenous antibiotics are recommended for EMR of larger LSLs (≥30 mm). In addition, somatic sensation supplies the dentate line, and addition of a long-acting local anaesthetic such as ropivacaine 0.5–0.75% to the submucosal injectate is effective for post-procedural analgesia [29]. In this setting, cardiac monitoring by electrocardiography is required. Lastly, haemorrhoidal vessels are thick walled and are resistant to snare entrapment as long as there’s adequate submucosal lift and generally present no additional bleeding risk during EMR [29].

Peri-appendiceal and Ileocaecal Valve LSLs

EMR of LSLs involving the appendiceal orifice is challenging but can generally be successfully completed provided that <50% of the valve circumference is involved and the proximal (deep) margin within the appendix is adequately visualised. A small (10 mm), stiff thin-wire snare is useful for removing adenomatous tissue from within the orifice. ER around and within the appendiceal orifice may precipitate appendicitis, and administration of a prophylactic intra-procedural intravenous antibiotic effective against enteric pathogens followed by a short 5–7-day oral course should be considered.

EMR of LSLs involving the ileocaecal valve (ICV) has higher risk of failure (OR 3.38, 95% CI 1.2–9.52, P = 0.021) [5]. Lesion access can be improved by use of a transparent distal cap attached to a paediatric colonoscope, which is preferable over an adult scope in this setting due to its smaller retroflexion radius, which may assist access to the inferior lip, and better overall manoeuvrability. The cap helps to deflect mucosal folds, stabilise the endoscope and visualise the distal ileum and ICV lips. Extensive proximal extension of adenomatous tissue into the narrow terminal ileum and LSLs that involve both the superior and inferior lips is challenging for complete EMR [30]. The EMR technique of LSLs involving the ICV is similar to that used in other colonic locations, albeit with a few notable adaptions. A small (10 mm), stiff thin-wire snare is preferred when working with the distal ileum due to space constraints. Submucosal injection should be conservative as excessive amounts can hamper endoscopic visualisation and lesion access. Careful attention should be paid to the anterior and posterior angles of the ICV, where residual tissue can be missed [30]. Lastly, the ICV submucosa is relatively adipose, and commonly underlying fat is exposed following resection, but is not necessarily a sign of deeper injury (Fig. 4.6).