Management of the malignant colorectal polyp: <scp>ACPGBI</scp> position statement

A colorectal polyp is an abnormal protrusion of the mucosa into the bowel lumen that is classified by histopathological examination (Fig. 1). Adenomas are a common finding during colonoscopy in symptomatic patients and in asymptomatic individuals undergoing screening. It is estimated that the prevalence of large-bowel adenoma is 21–28% in 50–59-year-old subjects, increasing to 40–45% in 60–69-year-old subjects and rising further to 53–58% in people over the age of 70 1. Adenomas are important as they are the precursor lesion of most colorectal cancers arising through the adenoma–carcinoma sequence 2. This association is supported by epidemiological, clinical and molecular research 3. Whilst most polyps removed are small, it is well recognized that the risk of malignant transformation increases with increasing polyp size. There is compelling evidence that removing adenomas from the colon substantially reduces the risk of a patient developing colorectal cancer 4. The term 'malignant polyp' refers to an adenoma that appears benign macroscopically but in which there is invasion through the muscularis mucosae into the submucosa. Such a focus of carcinoma is detected on histological examination. A malignant polyp is therefore an early carcinoma. It accounts for 0.75–5.6% 5-9 of large-bowel polyps removed in general diagnostic colonoscopy practice. The wide discrepancy is explained by differences in the study populations, being lower where all polyps removed are histologically assessed and higher in series that only include larger polyps. In the Nottingham Bowel Cancer Screening Trial, 1466 patients underwent colonoscopy because of a positive faecal occult blood-test result. Of these, 710 (48%) were found to have an adenoma and a further 73 (5%) had an adenoma containing a focus of cancer 10. A similar incidence of malignant polyps has been found in the National Bowel Cancer Screening Programme. Of the first 1-million individuals screened, just over 17 000 had a positive faecal occult blood-test result. Of these, 1574 (9%) had cancer of which 155 (10%) were malignant polyps 11. In recent years, greater use of diagnostic colonoscopy has been accompanied by a rise in the number of malignant polyps removed endoscopically 12. The management of a malignant polyp following endoscopic removal is difficult because the possibility of residual malignant cells within the bowel wall or positive regional lymph nodes varies from patient to patient, depending on a number of prognostic factors. The evidence base for management of these lesions is poor and is mostly based on data from symptomatic patients 6, 13, 14. Advising patients on the course of action after removal is difficult. It includes surveillance only, where the risk of residual disease is deemed to be low, or major surgical resection for those with a higher risk. However, the level of risk is often difficult to calculate. Furthermore, the presence of comorbidity and the site of the lesion should also be considered when discussing further management with the patient. This position statement deals with a number of matters relating to the management of patients with a malignant polyp. It is divided into four sections. The first section considers colonoscopy and includes a discussion of endoscopic techniques for the removal of larger adenomas, which are more likely to be malignant. The influence of polypectomy technique on histopathological assessment and the regimens for postresection endoscopic surveillance are discussed. The second section reviews the histopathology of adenomas and polypoid carcinomas and considers important prognostic indicators. The third section deals with how the prognostic indicators influence the risk of residual cancer in the bowel wall or lymph nodes to estimate the likelihood of recurrence if no further treatment is undertaken. The traditional advice for a 'high-risk' adenoma was to advise a radical 'cancer' segmental colectomy, including the lympho-vascular pedicle. The fourth section discusses the balance of the risk of morbidity and mortality following endoscopic resection against the risk of recurrence and how this risk is communicated to the patient. Finally, the role of endoscopic and radiological imaging in the assessment and follow up of malignant polyps is considered, particularly where radical surgery is not performed. This also includes a discussion of the optimal follow-up protocol. Searches of the Cochrane Database, PubMed, MEDLINE and Embase were performed using keywords relevant to each section of the position statement. They were mostly limited to English-language articles. Additional publications were retrieved from references cited in articles identified from the primary search. All evidence was classified according to an accepted hierarchy of evidence, and recommendations were graded from A to C on the basis of the level of associated evidence and/or noted as Good Practice and/or part of the National Institutes of Clinical Excellence/ Scottish Intercollegiate Guidelines Network (NICE/SIGN) recommendation or Rapid Technology Appraisal (Table 1)15. Modern endoscopic practice is safe, thorough and offers extensive opportunities for identification, diagnosis and definitive management of most benign colorectal neoplasms. A small proportion of these will be early cancers, and the diffusion of colorectal cancer screening has led to more of these being discovered 16. The purpose of this section is to illustrate how, with proper location and characterization of colorectal neoplasms, malignant lesions can be detected and treated optimally by endoscopy alone when appropriate. This should reduce the number of 'surprise' malignant polyps and the clinical dilemmas they cause. It will reduce the necessity for surgery for inadequately treated polyps. Surgery can then be targeted on lesions with a high risk of local recurrence, lymph node metastasis and colorectal cancer death 17. The prerequisite for colonoscopy is a safe, complete examination of the entire colon and rectum. Great strides have been made in the completeness and accuracy of UK colonoscopy since the survey carried out by Bowles et al. 18 showed variable performance and unsatisfactory caecal intubation and complication rates and low levels of colonoscopy training. Re-audit following investment in training, accreditation and setting of quality assurance standards has demonstrated great improvements in examination quality, diagnostic accuracy and the safety of colonoscopy in the UK 19. Quality colonoscopy optimizes the chance of finding pathology. White light examination (WLE) alone is usually sufficient to identify colonic abnormalities, but tandem colonoscopy shows that this may miss 22% of all adenomas and 2.1% of adenomas over 1 cm in diameter 20. Most malignant polyps are larger pedunculated or protruding sessile lesions. These are usually easy to see. An important minority of neoplasms is not truly polypoid, but flat or even depressed. These are harder to spot but carry a much greater risk of malignancy. The challenge is to find these lesions. Good bowel preparation, meticulous mucosal washing, insufflation and careful inspection are critical 21. The flexures and inner angles of colonic folds are sites where there is a higher risk of lesions being missed and they require more careful inspection. Modern endoscopes have better bending sections, allowing retroflexion in the caecum and rectum to view the caecum beyond the ileocaecal valve and distal rectum. Right-sided colonic lesions are easier to miss because they tend to be flatter. Missed right-sided lesions may explain why colonoscopic screening has failed to make the hoped-for impact on the detection of right-sided colonic cancer 22. Patient repositioning during the examination, with the inspected flexure uppermost to open up collapsed areas, increases luminal distension 23 and adenoma detection 24. The time spent inspecting the mucosa (the withdrawal time) should be sufficient to allow a thorough mucosal view. The influence of a careful examination technique is highly significant for detection. Barclay et al. 25 found that colonoscopists who took longer than 6 min in the withdrawal and inspection phase had a higher detection rate for any neoplasm (28.3% vs 11.8%) and for advanced neoplasms (6.4% vs 2.6%) than those with a shorter inspection time. The adenoma detection rate (ADR) is a surrogate marker of the quality of colonoscopy and can therefore be used as a comparative measure in studies. Kaminski et al. 26 used the Polish bowel cancer screening data set to validate ADR as a quality indicator. They showed that an individual endoscopist's ADR is associated with the subsequent interval cancer risk, with a lower ADR predicting significantly higher interval cancer risk. Findings The quality of colonoscopy can be monitored using the adenoma detection rate and is enhanced by excellent preparation and meticulous technique. During withdrawal, inspection for more than 6 min and patient repositioning increases the adenoma detection rate, which is a reproducible quality marker (Level IIa). Recommendations Attention to preparation and endoscopic technique, especially the inspection (withdrawal) phase, will increase the quality of the examination and detection of pathology (Grade B). Polyps may be missed, either because they are subtle by being small, flat or depressed, or because they hide behind folds and flexures. Dye spray chromoendoscopy uses contrast reactive dyes, such as indigocarmine, to enhance mucosal features. At concentrations of 0.1–0.8%, indigocarmine fills cavities, pits and grooves in the mucosa and even flat, small polyps will be detected. Chromoendoscopy has been shown to increase the ADR but it is awkward and time consuming for use in routine practice. Other vital or absorptive dyes, such as crystal violet or methylene blue, are actively absorbed into the intestinal crypts, where they stain the convex portions but not the grooves. Very detailed magnification views are possible, but these dyes are slow to absorb and messy to use. A Cochrane collaborative review comparing white light and chromoscopic detection of colorectal neoplasia showed significantly enhanced detection of neoplasms using chromoendoscopy 27, 28. Most of the extra adenomas seen were diminutive, and more patients were found to have multiple polyps. Histology of these extra adenomas showed mostly low-grade dysplasia: this, plus the time to perform dye spray, has prevented pancolonic chromoscopy from becoming routine. Selective application of dye spray to areas of subtle mucosal change is valuable and can detect a higher number of neoplasms and help differentiate neoplastic from non-neoplastic lesions 29. Chromoendoscopy has a valuable place in detecting dysplasia in ulcerative colitis surveillance 30. Findings Dye spray chromoendoscopy enhances detection of colonic pathology, differentiation of neoplastic and non-neoplastic lesions and of dysplasia in inflammatory bowel disease (Level I). Recommendations Endoscopists should selectively use dye spray chromoendoscopy as it enhances detection and differentiation of colonic pathology (Grade A). Optical and processor-based technologies are near-instant methods to examine the mucosa to delineate polyps. There are no clear data to indicate that these methods increase the adenoma-detection rate but they have proven ability to characterize polyps once found 31. Narrow band imaging (NBI; marketed by Olympus), multiband imaging (MBI) (marketed as FICE – flexible spectral imaging colour enhancement; Fujinon) and i-scan (Pentax) manipulate the wavelengths of light used to examine the mucosa. NBI uses a real-time optical filter to select two or three restricted wavelengths of light to emphasize the mucosal microvasculature and identify vascular alterations associated with pathological conditions. FICE and i-scan are postprocessor technologies that recreate the image as per the desired wavelengths to enhance mucosal surface patterns. Autofluorescence imaging (AFI; marketed by Olympus) uses tissue-component responses to specific short and ultraviolet wavelengths, the characteristics of which are different in neoplasms. Because these technologies are based on different endoscope platforms, UK endoscopists rarely have a choice of technology. There is doubt whether NBI, FICE, i-scan or AFI improve the ADRs compared with good white light colonoscopy (WLC) in average-risk patients. Rex and Helbig 32 found no additional benefit with NBI compared with an expert using WLE. Adler et al. 33 found a nonsignificant trend for NBI for adenoma detection. In higher-risk patients under surveillance for hereditary nonpolyposis colorectal cancer (HNPCC), East et al. 34. showed significantly improved adenoma detection – typically subtle flat adenomas. A recent Cochrane review found no evidence that NBI is better than high-definition WLC at detecting adenomas, but NBI was better than standard-definition WLC and equal to high-definition WLC 35. A study by Kuiper et al. 36 compared adenoma-detection rates and polyp characterization using high-definition white light, AFI and NBI. There was no improvement in detection using these technologies over WLE but they proved sensitive and specific in differentiating neoplastic and non-neoplastic lesions. It is this property – and the speed of switching mode – that makes these systems valuable. Training and practice are critical for these techniques to be clinically valuable in everyday use; training programmes are being developed 37. Findings A careful, expert white light examination of the colon can be augmented by selective or targeted chromoscopy and/or optical enhancement to examine suspicious areas (Level I). Recommendations Techniques that enhance surface and vascular patterns of colonic lesions should be used in routine practice. Endoscopists should learn to interpret these imaging methods (Grade B). Cap colonoscopy – fitting a standard endoscope with a disposable hood or cap, can be used to improve exposure of hidden mucosa. The capped endoscope can be flexed against haustral folds, flattening them. A better view of the mucosa beyond the fold can then be obtained. Westwood et al. 38 recently reviewed published experience with this technique and found an increase in polyp detection and caecal intubation rates. As yet there is no reliable way to accurately predict malignant change in a polyp, but there are features of polyp size, shape, consistency, surface and vascularity that should alert the endoscopist to possible malignancy. Combining sophisticated imaging modalities may eventually provide an 'optical biopsy' 39. Knowledge and accurate use of these descriptive methods allows malignant risk stratification. The Erlangen Group 40 examined 11 188 adenomatous polyps in a European series from 1978 to 1993. Using multivariate analysis they related malignant risk to a number of features – both within the patients themselves (age and sex) and related to the multiplicity, site, size and histological type of polyps. Polyps < 5 mm in diameter carry negligible risk of malignancy, whereas those with a diameter of more than 25 mm carry a considerable risk (Table 2). There are problems with estimating size in vivo. A useful guide is that an open standard biopsy forcep width is 8 mm, while a closed forcep width is 2.5 mm. Endoscopists must practise taking such measurements. The site of a polyp within the colon is also a risk factor where proximal colonic polyps are, size for size, at greater risk of containing malignancy 41 (Table 3). The malignant risk for adenomas in the right colon (proximal to the splenic flexure) was higher than that for similar-size left-sided or rectal polyps. Increasing use of positional imaging technology allows more reliable description of lesion position in the colon – which can otherwise be inaccurate. Simple pattern recognition and experience are important. Malignancy is more likely when the contour is irregular, when there is ulceration or when the consistency of the polyp (when probed gently) is hard or when the stalk broadens 42. These classical signs are not always evident, and more sophisticated classifications have been developed. Lesions called 'flat' are rarely completely flat. The Paris Classification defines 'flat' as < 2.5 mm in height above the mucosa, which is the width of closed, standard endoscopic biopsy forceps. The category not specifically classified in the Paris Classification is the lateral spreading tumour (LST); in Europe and USA these are carpet adenomas. LSTs are flat adenomas larger than 10 mm in diameter that extend circumferentially and laterally rather than vertically. They may have a granular (LST-G) or a nongranular (LST-NG) surface. Nodules and depressed areas are seen within these lesions. They have a malignant potential that is not predicted solely by size but rather by the presence of nodules or depressed areas within them. The cancer risk in LST varies between 7% (LST-G) and 14% (LST-NG): the Paris Classification defines these as type 0-IIa 47. Table 4 shows the frequency of lesions classified by the Paris system related to both their size and the rate of submucosal invasion. The data are from Kudo, using the Paris system and include colon and rectal lesions 45. Classical protuberant lesions (0-Ip and 0-Is) are common, and size influences invasive risk: lesions of 5 mm or less are associated with negligible risk, but for lesions over 20 mm the risk of malignancy is high. Recognition of depression (type 0-IIc) in colorectal lesions is critical as this is often associated with invasive cancer, even when the lesion is small (< 10 mm). These true depressed lesions are rare but grow rapidly, become advanced at an early stage of the evolution of their growth and are seldom suitable for endoscopic resection. Initial clues are irregularities in mucosal appearances such as 'pinkness, minute depressions and/or haemorrhagic spots' 44. Because they are subtle, dye spray chromoendoscopy with indigocarmine is invaluable to demarcate them from background innominate grooves and delineates the surface, edge and any areas of depression. The colonoscope technologies NBI and FICE perform similar et al. chromoendoscopy with indigocarmine plus magnification of the – a not in the Polyps with more submucosal invasion – and – to have an depression with of the depressed surface and two or more Such are subtle in expert are highly of invasion they are difficult to into standard practice. a time there was doubt that flat polyps in because they were not being It is clear they and are being detected. The a series of and found adenomas these were a number of flat polyps and a small number of true depressed lesions with significant malignant Table 5 is from their data and in a from a that not only flat and depressed polyps but also that they can be detected in a UK and it the small, but risk of early cancer in polyps < 1 cm in as well as the greater risk with flat lesions of more than 1 cm in lesions are rare but et al. data from in where of cancers found by colonoscopy were of flat and small mm in The Bowel Cancer Screening all polyps to be classified by an endoscopist using the Paris Findings size and polyp all influence assessment of malignancy in a The Paris Classification is both descriptive and (Grade Recommendations Endoscopists should estimate size of polyps and use the Paris Classification to the of malignancy (Level A). inspection of the surface of polyps can further predict Chromoendoscopy using indigocarmine, with colonoscopy can flat or depressed lesions and identify the pattern of polyps that predict pathology. The Classification of patterns is shown in pattern and are non-neoplastic or polyps can to the adenoma and should be treated patterns and are most likely to be benign adenomas with a low risk of submucosal invasion. patterns indicate a high risk for invasion into at the submucosa. The pattern can further be divided into pits of and pits are This is only with pattern may be on the surface of a benign lesion but submucosal invasion can also has the likelihood of malignancy. Using this and related pattern with to of polyps. Table 6 their The ability to identify patterns the endoscopist to predict malignant change within a polyp and select et al. used a of and colonoscopy with chromoendoscopy to differentiate non-neoplastic and neoplastic lesions with and – it is not always to use pattern training and vascular pattern and mucosal pattern assessment with NBI allows diagnostic differentiation of non-neoplastic polyps from neoplastic polyps and will detect malignant change NBI detailed of the – the of this pattern has pattern to where type shows malignant In the type is into carcinoma and submucosal and with submucosal invasive cancer both magnification and considerable endoscopists use NBI to identify a neoplastic lesion and then dye spray to characterize the pattern rather than on the East et al. have demonstrated the of NBI with magnification in neoplasia in and in ulcerative NBI magnification can be as as the NBI system to be However, UK endoscopists not have to FICE is with and by detecting surface patterns can differentiation of polyp A is in a study by et al. FICE was shown to similar to NBI when using more a subsequent showed that FICE neoplastic and non-neoplastic polyps magnification et al. compared FICE with chromoscopy in small polyps and showed both modalities to have good and to neoplasia and Other methods of surface and lesion examination, as well as endoscopic are research or not sensitive or specific to be endoscopic into this in contrast with which is a routine in neoplastic particularly in with endoscopic Optical and endoscopy are being 39. A recent review and of that this offers diagnostic accuracy to colonoscopic histopathology in colorectal neoplasia This offers the possibility of in real-time optical biopsy in the is a by to enhance lesions difficult to by WLE. There as on in colorectal neoplastic characterization of malignant change or any a polyp can if there is submucosal with invasion A lesion to to the mucosa that or endoscopic resection will not be to clear the et al. in a small of patients found to with invasion – early cancers that were all or tumour invasion of the third and of the of the However, et al. used a with either or as the and compared this with endoscopic They found the to lower and accuracy compared with endoscopic for invasion vs and vs They that a lesion will be difficult to and of invasion more difficult to A of the is that submucosal makes a further at endoscopic at a more difficult by submucosal should be as a diagnostic to a for for as it makes less likely Findings characterization of polyp size, and surface pattern can predict histopathology of the lesion and allow of the risk of malignant change and of invasion. The submucosal invasion (Level I). Recommendations All colonoscopists should be with and use the Paris enhancement by chromoendoscopy and either NBI or FICE are to lesions being considered for advanced polypectomy techniques (Grade A). There is no in treatment for discussion with It is a to that optimal treatment is – and to an when is part of this of and/or with dye spray or techniques are for these an endoscopist can the patient or for of are becoming in the UK for such and and have proved in the USA and in There are a of techniques and these are to the type of polyp The should always where possible, an of the lesion in polypectomy is the of polyp management because the of lesions are size, position and can make this malignancy is within a pedunculated polyp the should be to the bowel to resection the and are polypectomy will be to in polyps may be and a of techniques are including application and polyps may be treated with polypectomy alone with that are but submucosal to enhance resection is routine. Most polypectomy of pedunculated malignant polyps if the from the invasive to is mm or more It offers the rate of local recurrence and and submucosal and mucosal resection or a submucosal with to the mucosa up in the submucosal of a over the entire lesion and the It allows flat or depressed lesions to be removed the lesion may be and using Modern in their and endoscopists must be with these when or left-sided lesions. The not only the but also as a There is no on which to use as a is but more such as or are to to only a is often to and dyes such as indigocarmine or methylene have been used to the of the the edge of the lesion and the of muscularis and can be in et al. 70 and et al. The of the lesion in and are critical for local