Endoscopic Removal of Colorectal Lesions: Recommendations by the US Multi-Society Task Force on Colorectal Cancer

Abstract

Colonoscopy with polypectomy reduces the incidence of and mortality from colorectal cancer (CRC).1,2 It is the cornerstone of effective prevention.3 The National Polyp Study showed that removal of adenomas during colonoscopy is associated with a reduction in CRC mortality by up to 50% relative to population controls.1,2 The lifetime risk to develop CRC in the United States is approximately 4.3%, with 90% of cases occurring after the age of 50 years.4 The recent reductions in CRC incidence and mortality have been largely attributed to the widespread uptake of CRC screening with polypectomy.5 The techniques and outcomes of polyp removal using colonoscopy, however, had historically remained understudied and thus, practice widely varied. Reports have shown that residual tissue after polypectomy that is judged to be “complete” by the endoscopist is common, ranging from 6.5% to 22.7%.6 The significant variation in incomplete resection rates among endoscopists has highlighted the dependence of polypectomy effectiveness on operator technique. A pooled analysis from 8 surveillance studies that followed participants with adenomas after a baseline colonoscopy suggested that although the majority (50%) of post-colonoscopy colon cancers were likely due to missed lesions, close to one-fifth of incident cancers were related to incomplete resection.7 Polypectomy techniques have expanded in parallel with advances in endoscopic imaging, technology, and tools. Optimal techniques encompass effectiveness, safety, and efficiency. Colorectal lesion characteristics, including location, size, morphology, and histology, influence the optimal removal method. For example, the applications of cold snare polypectomy for small lesions, which can remove adenomatous tissue en bloc with surrounding normal mucosa, and endoscopic mucosal resection (EMR) for large and flat lesions, which utilizes submucosal injection to lift the lesion before snare resection, have evolved to improve complete and safer resection. The primary aim of polypectomy is the complete and safe removal of the colorectal lesion and the ultimate prevention of CRC. This consensus statement provides recommendations to optimize complete and safe endoscopic removal techniques for colorectal lesions (Table 1), based on available literature and experience. The recommendations from the US Multi-Society Task force (USMSTF) on the management of malignant polyps, polyposis syndromes,8 and surveillance after colonoscopy and polypectomy9 are available in other documents. Table 2 summarizes abbreviations and definitions of terms utilized in these recommendations.Table 1.: Statements of Best Practice in This DocumentTable 1-A.: Statements of Best Practice in This DocumentTable 1-B.: Statements of Best Practice in This DocumentTable 2.: Abbreviations, Terms, and DefinitionsTable 2-A.: Abbreviations, Terms, and DefinitionsTable 2-B.: Abbreviations, Terms, and DefinitionsMethods Process The USMSTF is composed of 9 gastroenterology specialists who represent the American College of Gastroenterology, the American Gastroenterological Association, and the American Society for Gastrointestinal Endoscopy. We developed the guidance statements by consensus process through e-mail correspondence and multiple joint teleconferences. The final manuscript was reviewed and approval by the governing boards of the 3 respective societies. Literature Review We performed a systematic review of the literature based on a defined search by a medical librarian of the Ovid Medline, Embase, and Cochrane databases from 1946 to December 2017, as well as reviews of manual references and scientific meeting abstracts of the American College of Gastroenterology, American Gastroenterology Association, American Society for Gastrointestinal Endoscopy, and United European Gastroenterology Week from 2014–2017. The search was limited to human studies without any language restriction. We framed the search strategy using key words (Appendix 1, https://links.lww.com/AJG/B415) from formatted question statements (Appendix 2, https://links.lww.com/AJG/B416). We reviewed and synthesized high-quality studies to generate statements and, when not available, relied on lower-quality evidence and expert opinion. Grading of Recommendations, Assessment, Development, and Evaluation Ratings of Evidence: Level of Evidence and Strength of Recommendation The USMSTF group rated the quality of the evidence for each statement as very low quality, low quality, moderate quality, and high quality based on the GRADE (Grading of Recommendations, Assessment, Development, and Evaluation Ratings of Evidence) methodology (Table 3).10Table 3.: Grading of Recommendations Assessment, Development, and Evaluation Ratings of Evidence and Strength of RecommendationsWe provide a recommendation as strong or conditional according to modified GRADE criteria.11 Wording of recommendations was based on the strength of recommendation: “recommend” was used for strong recommendations and “suggest” was used for conditional recommendations. Section I: Lesion Assessment Statement 1: Lesion Assessment and Description The macroscopic characterization of a lesion provides information to facilitate the lesion’s histologic prediction, and optimal removal strategy. We recommend the documentation of endoscopic descriptors of the lesion, including location, size in millimeters, and morphology in the colonoscopy procedure report. (Strong recommendation, low-quality evidence) We suggest the use of the Paris classification to describe the surface morphology in order to provide a common nomenclature (Conditional recommendation, low-quality evidence) We suggest that for non-pedunculated adenomatous (Paris 0-II and 0-Is) lesions ≥10 mm, surface morphology should be also described as granular or non-granular lateral spreading lesions. (Conditional recommendation, low-quality evidence) We recommend photo documentation of all lesions ≥10 mm in size before removal, and suggest photo documentation of the post-resection defect (Strong recommendation, low-quality evidence). We suggest proficiency in the use of electronic- (eg, narrow-band imaging [NBI], i-scan, Fuji Intelligent Chromo Endoscopy or blue light imaging) or dye (chromoendoscopy)-based image enhanced endoscopy techniques to apply optical diagnosis classifications for colorectal lesion histology. (Conditional recommendation, moderate-quality evidence) We recommend proficiency in the endoscopic recognition of deep submucosal invasion. (Strong recommendation, moderate-quality evidence) The macroscopic characterization of a colorectal lesion, including its location, size, and shape, combined with the real-time assessment of the suspected histopathology and estimation of the depth of invasion provides information about whether a lesion is amenable to endoscopic resection. In this document, we review key components to the macroscopic characterization of colorectal lesions. A more detailed description of the macroscopic assessment of lesions with submucosal invasion, and a decision-making guide to their optimal management is provided in separate MSTF document on Endoscopic Recognition and Management Strategies for Malignant Colorectal Polyps. Paris Classification The Paris classification has been the most used international endoscopic classification of colorectal lesion morphology (Figure 1).12 Although studies have shown only moderate agreement among Western experts using the Paris classification, the application of a minimal standard terminology of colorectal lesions provides the first step in stratifying which lesions are more likely to contain advanced pathology and informs their removal strategy.13,14 In the Paris classification, there are 2 macroscopic types: (1) type 0, the superficial lesions; and (2) types 1–5, the advanced cancers.Figure 1.: Paris Endoscopic Classification of superficial neoplastic lesions in the colon and rectum.Paris Classification Superficial Lesions, Type 0 The classification of type 0 lesions is based on the distinction between polypoid (type 0-I); and non-polypoid, (type 0-II). The polypoid type consists of pedunculated (type 0-Ip), and sessile (type 0-Is) lesions. The non-polypoid type 0-II lesions are divided by the absence (superficially elevated [type 0-IIa] and flat [type 0-IIb]) or the presence of a depression (type 0-IIc). The non-polypoid, excavated (type 0-III) lesions are rare in the colon. Although depressed (0-IIc) lesions are uncommon (1%–6% of non-polypoid lesions), their risk of submucosal invasion is the highest: the overall risk is reported to be 27%–35.9% compared with 0.7%–2.4% in flat (0-IIa) lesions. More than 40% of small (6–10 mm) depressed (0-IIc) lesions contain submucosal invasive cancer; virtually all large (>20 mm) depressed (0-IIc) lesions have submucosal invasion.15–18 Lateral Spreading Tumors Non-polypoid lesions 10 mm or larger in diameter are referred to as laterally spreading tumors (LSTs). They have a low vertical axis and extend laterally along the colonic luminal wall. The morphologic subclassifications of LSTs facilitate the endoscopic removal plan, as they inform about submucosal fibrosis or the risk of submucosal invasion. Granular-type LSTs have a nodular surface and are composed of the homogeneous even-sized (LST-G-H) and mixed (LST-G-NM) nodular subtypes. Non-granular type LSTs have a smooth surface and are comprised of the flat elevated (LST-NG-FE) and pseudodepressed (LST-NG-PD) subtypes (Figure 2).19 LST-G-H have the lowest risk (0.5%; 95% confidence interval [CI], 0.1%–1.0%), whereas LST-NG-PD have the highest risk of submucosal invasion (31.6%; 95% CI, 19.8%–43.4%).19Figure 2.: Lateral spreading lesions. Non-polypoid lesions ≥10 mm in diameter are referred to as laterally spreading tumors (LSTs). They have a low vertical axis and extend laterally along the luminal wall. LSTs are morphologically subclassified into granular type (LST-G) (A, B), which have a nodular surface, and non-granular type (LST-NG), which have a smooth surface (C, D). This macroscopic distinction is important to facilitate the endoscopic removal plan as it provides information about the risk of cancer or submucosal fibrosis in order to anticipate the technical ease or difficulty of the removal. Overall, LSTs were found to contain submucosal invasion (SMI) in 8.5% of the cases (95% CI, 6.5%–10.5%; I 2 86.8%; 26 studies) and high-grade dysplasia in 36.7% of the cases (95% CI 30.3%–43.2%; I 2 91.9%; 23 studies). Non-granular LSTs more often contained SMI than granular LSTs: 11.7% vs 5.9% (OR, 1.89; 95% CI, 1.48–2.42).Optical Diagnosis Endoscopic prediction of the histologic class of a polyp may influence the resection approach to ensure complete removal. A number of studies, including several meta-analyses, have shown that optical diagnosis of colorectal lesions is feasible in routine clinical practice and comparable to the current reference standard, histopathology.20,21 The endoscopist’s level of confidence in the optical diagnosis of a colorectal lesion is an important factor in its application to clinical practice. Although the majority of lesions have typical endoscopic features that enable a high confidence prediction of histology, in lesions that lack clear features, optical diagnosis performance may be decreased. For example, in a meta-analysis of 28 studies on optical diagnosis of colorectal lesions, the highest performance of real-time optical diagnosis of colorectal polyps was achieved when the diagnosis was made with high confidence—the area under the hierarchical summary receiver-operating characteristic curve was 0.95 (95% CI, 0.93–0.97) for polyps of any size, and 0.92 (95% CI, 0.92–0.96) for diminutive (≤5 mm) ones. This compares to the overall area under the hierarchical summary receiver-operating characteristic curve of 0.92 (95% CI, 0.90–0.94). The Narrow Band Imaging International Colorectal Endoscopic (NICE) classification provides a validated criterion for the classification of type 1 (serrated class lesions–hyperplastic and sessile serrated lesions) and type 2 (adenomas), as well as those with deep submucosal invasion (type 3), using real-time NBI during colonoscopy22,23 (Figure 3). Its application has been shown to be useful in assessing the most clinically relevant approaches: leave hyperplastic diminutive lesions of the rectum and sigmoid colon, remove all adenomas anywhere in the colon and any serrated lesions proximal to sigmoid colon and >5 mm, and biopsy and refer to surgery lesions with deep submucosal invasion. Using this classification, experienced endoscopists have achieved 93% concordance of surveillance intervals made by real-time optical diagnosis and pathology, and a >90% negative predictive value for rectosigmoid lesions when assessments were made with high confidence.21 A feature that has been associated with conventional adenomas is a valley in the surface topography that appears red in white light and brown in NBI relative to the rest of the polyp surface. Although insensitive (<50%), the valley sign was highly specific (>90%) for conventional adenoma in diminutive (≤5 mm) lesions, suggesting it to be a valid predictor of adenomatous histology in diminutive colorectal lesions.24 Other endoscopic classifications of colorectal lesions using newer technologies warrant further investigation.Figure 3.: Optical diagnosis of colorectal lesions, NICE classification. The diagnostic criteria for colorectal lesions using NBI as recommended in the NICE classification. The use of confidence levels (high or low) in making an optical diagnosis is important in its implementation in clinical practice. *Can be applied using colonoscopes with or without optical (zoom) magnification. **These structures (regular or irregular) may represent the pits and the epithelium of the crypt opening. ***In the World Health Organization classification, sessile serrated polyp and sessile serrated adenoma are synonymous. Sessile serrated polyps often demonstrate some dark, dilated crypt orifices. ****Type 2 consists of Vienna classification types 3, 4, and superficial 5 (all adenomas with either low- or high-grade dysplasia, or with superficial submucosal carcinoma). The presence of high-grade dysplasia or superficial submucosal carcinoma may be suggested by an irregular vessel or surface pattern, and is often associated with atypical morphology (eg, depressed area).The subtle endoscopic appearance of large sessile serrated lesions—predominantly flat in shape with indistinct borders—has been associated with high rates of incomplete removal compared to conventional adenomas (31% vs 7.2%), with even higher rates (47.6%) in large lesions.6 A mucous cap may be present in some sessile serrated lesions and facilitate detection. The WASP (Workgroup Serrated Polyps and Polyposis) criteria added 4 sessile serrated lesion features (ie, clouded surface, indistinctive borders, irregular shape, and dark spots inside crypts) to the NICE classification (Figure 4)25 and showed that high confidence assessment of lesions could accurately (91%) distinguish sessile serrated lesions from non–sessile serrated lesions.26 Within a serrated lesion, areas with a distinct surface pattern change (with NICE Type 2 features) or a nodular component are suggestive of cytologic dysplasia.27 Identification of higher-risk lesions may influence endoscopic therapeutic strategy, pathology awareness, and surveillance recommendations.28Figure 4.: Morphologic features of sessile serrated lesions. Sessile serrated lesion–like features are defined as (A) a clouded surface, (B) indistinctive borders, (C) irregular shape, or (D) dark spots inside the crypts. These morphologic features are used to differentiate between sessile serrated lesions and hyperplastic lesions in the type 1 NICE polyps. The presence of at least 2 sessile serrated lesion–like features is hereby considered sufficient to diagnose a sessile serrated lesion.Unfavorable histologic features of colorectal lesions, such as lymphovascular invasion, tumor budding, or poor differentiation, are not feasible to endoscopically predict before resection. However, the vertical depth of invasion of submucosal cancers can be estimated based on the morphologic appearance using high-definition endoscopy without magnification. Lesion morphology, such as Paris classification 0-IIc and 0-IIa + 0-IIc, non-granular surface particularly pseudodepressed subtype, NICE type 3,23 and Kudo pit pattern V,29 as well as white spots (chicken skin appearance), redness, expansion, firmness, and fold convergence,30 are associated with submucosal invasive carcinoma (Video 1, https://links.lww.com/AJG/B420, https://links.lww.com/AJG/B421). The NICE type 3 and Kudo Vn patterns are specific for deep (>1000 μm) invasion. Deep submucosal invasion in a non-pedunculated lesion is associated with a substantial risk of residual cancer in the bowel wall or lymph nodes after any form of endoscopic resection. Therefore, the presence of these features should be followed by cold biopsy of the portion of the lesion demonstrating the features, tattoo of the area, and referral to surgery. Non-pedunculated lesions with superficial (<1000 μm) submucosal invasion are candidates for endoscopic resection. However, there are no endoscopic features that are sensitive in predicting superficial submucosal invasion. Non-granular morphology, particularly when associated with depression (Paris 0-IIc) or bulky (Paris 0-Is) shape, is associated with an increased risk of superficial invasion. When feasible, en bloc endoscopic resection, followed by pinning of the retrieved specimen to a flat surface (eg, cork, foam) and sectioning of the lesion perpendicular to the resection plane, allows accurate pathologic measurement of the depth of invasion. Specimens from lesions with endoscopic features suspicious for advanced histology, submucosal invasion, or cancer should be submitted in individual bottles for pathologic analysis. Section II: Endoscopic Removal Techniques Statement 2: Lesion Removal The primary aim of polypectomy is complete removal of the colorectal lesion and the subsequent prevention of CRC. Endoscopists should employ the safest, most complete, and efficient resection techniques based on available evidence. Polypectomy techniques vary widely in clinical practice. They are often driven by physician preference based on how they were taught and on trial and error, due to the lack of standardized training and the paucity of published evidence. In the past decade, evidence has evolved on the superiority of specific methods. Although more recent practice surveys suggest an increased uptake in the use of cold snare removal techniques for diminutive and small colorectal lesions and EMR for large colorectal lesions, considerable heterogeneity in management techniques persist.31–34 In a large survey of gastroenterologists and surgeons, physician specialty was strongly associated with management strategies. For example, surgeons were most likely to recommend surgical resection of complex benign colorectal lesions compared with gastroenterologists who were the least likely.13 Alarmingly, surgery for non-malignant colorectal lesions remains common practice.35–37 In the United States, colectomy for benign colon lesions has significantly increased over the last 14 years, representing one-quarter of colectomy procedures.38 One study showed rate increases from 6% in 2000 to 18% in 2014, for a mean (SD) lesion size of 27 (17) mm.39 This practice trend has occurred despite professional society and guideline recommendations for endoscopic removal as the first-line treatment. Endoscopic removal of benign colorectal lesions is more cost-effective than surgery, and is associated with lower morbidity and mortality.40,41 Data analyzed from a National Surgical Quality Improvement Program from 2011 through 2014, including 12,732 patients who underwent elective surgery for non-malignant colorectal lesions, showed a 0.7% 30-day mortality rate and 14% risk of major postoperative adverse events—with 7.8% readmissions, 3.6% redo surgeries, 1.8% colostomies, and 0.4% ileostomies.42 By comparison, the 30-day mortality associated with endoscopic resection of large colorectal lesions was only 0.08% in a review of 6440 patients,43 and zero in a prospective study of 1050 advanced colorectal lesions.44 Therefore, endoscopists should employ techniques that reflect the safest, most complete or effective, and most efficient resection techniques based on available evidence. A suggested management algorithm is presented in Figure 5.Figure 5.: Algorithm for the management of colorectal (≤5 mm) and small mm) lesions We recommend cold snare polypectomy to remove diminutive (≤5 mm) and small mm) lesions due to high complete resection rates and (Strong recommendation, moderate-quality evidence) We recommend the use of cold polypectomy to remove diminutive (≤5 mm) lesions due to high rates of incomplete resection. For diminutive lesions mm, cold snare polypectomy is or polypectomy may be (Strong recommendation, moderate-quality evidence) We recommend the use of biopsy for polypectomy of diminutive (≤5 mm) and small mm) lesions due to high incomplete resection and (Strong recommendation, moderate-quality evidence) (≤5 mm) lesions colorectal lesions are diminutive (≤5 that size, they are high-grade dysplasia or cancer removal using cold polypectomy has been associated with high rates of incomplete resection, ranging from to Although polypectomy is for complete removal compared to standard more than 1 is standard vs The of the mucosal surface and from the first biopsy may with and subsequent assessment of the of resection The use of enhanced imaging such as of the defect has not of resection, should be limited to diminutive lesions mm) and only to those when resection in a is The risk of incomplete removal of diminutive lesions can be with the use of cold snare polypectomy techniques The cold snare polypectomy is a more complete polyp removal it can a of normal the polyp as the snare is (Figure 2, This allows for en bloc lesion and of the without A systematic review and meta-analysis of 3 prospective studies on cold resection techniques for diminutive (≤5 mm) lesions showed a significantly lower incomplete polyp removal rate with the cold snare compared to cold polypectomy 95% CI, without heterogeneity and reported no adverse These superiority of cold snare polypectomy to other cold polypectomy techniques have been in a and are for lesions polypectomy technique. (A) colon lesion in white (B) Lesion characterization as a diminutive colon adenoma with type 2 NICE features using (C) the lesion at 5 in with the the snare the on the proximal of the lesion and (D) the snare it has normal surrounding tissue and the as the snare is according to the size of the the snare is in a and the to apply the polyp should not be or during to some on the snare with during in order to of the snare from the and of the the snare is it normal have the normal the lesion can be The snare is and the snare is The normal can be at the of the lesion in the cold snare specimen In the 5 the polyp remains in for efficient snare is and snare polypectomy is a effective, and efficient polypectomy for diminutive (≤5 mm) colorectal lesions compared to polypectomy A recent trial on mm colorectal lesion removal showed significantly lower incomplete polyp removal rates with cold snare compared to polypectomy cases of or occurred in either although the rate of tissue to the pathologic specimen was higher in the polypectomy group than cold snare polypectomy group of vs of parallel trial of lesions mm in size showed complete resection rates for cold snare polypectomy comparable to those for snare polypectomy based on from the resection after endoscopic occurred only in the snare polypectomy group of the majority of the lesions were diminutive mm) in of lesions in the snare polypectomy group and of lesions in the cold snare polypectomy mm) lesions for small lesions have been highly among The Study the of incomplete even for small lesions.6 They a incomplete resection rate for lesions mm by snare technique. snare and snare resection are distinct resection to the submucosal than polypectomy using and thus, the risk of and (Video 3, have shown the of cold snare vs snare polypectomy for small lesions and a compared to snare with incidence of and prospective study of patients who underwent colonoscopy 3 after cold snare polypectomy for lesions mm high rates of complete resection adenoma based on assessment and studies have shown sufficient resection and depth using cold snare including in the majority of have not defined the optimal snare for effective cold snare A study of a cold snare mm 9 mm shape, compared to a conventional snare mm, 10 mm diameter shape, showed significantly higher complete resection of small lesions with cold snare vs conventional snare vs particularly for lesions mm in The of specific snare on cold snare polypectomy outcomes further snare polypectomy has been shown to be a more efficient removal for lesions ranging from mm in size compared to cold or snare The procedure was significantly using cold snare polypectomy compared to cold techniques by an of (95% CI, of cold snare polypectomy have reported rates between and Non-pedunculated mm) lesions We suggest cold or snare polypectomy (with or without submucosal to remove mm non-pedunculated lesions. (Conditional recommendation, low-quality evidence) Optimal for removal of sessile lesions mm However, EMR should be considered for non-polypoid and serrated lesions in the to size have shown that using conventional polypectomy techniques for non-polypoid lesions ≥10 and serrated lesions proximal to the sigmoid colon a for complete endoscopic removal. The lesion are often and the tissue may be to with a A recent study of patients with proximal serrated lesions with a mean size of mm showed low rates of 95% CI, during a mean of when by This is in to a incomplete resection rate reported when by conventional polypectomy Non-pedunculated mm) lesions We recommend EMR as the of large mm) non-pedunculated colorectal lesions. Endoscopic resection can provide complete resection and the higher and associated with surgical treatment. (Strong recommendation,

Topics & Concepts