The brief of this issue of Cancer Forum is to review information available since the 2005 publication of the National Health and Medical Research Council relating to risk management of individuals with previous adenomas or colorectal cancer. However, this can be abbreviated to the last three years, as Cancer Council Australia commissioned a review of colonoscopy in surveillance for colorectal cancer, which included adenoma and cancer follow-up. This has subsequently been endorsed by the National Health and Medical Research Council. Since then, there have been advances in some areas, although many questions remain and clinical judgement comes into play. In the current era of accountability, economic hardship and increasing demand, surveillance strategies should be proven effective and individualised, based on issues such as fitness, quality of life and personal preferences. International guidelines have aligned, although the simpler strategies specified in European guidelines are noted with interest. Despite clear recommendations, the lack of guideline use in routine practice is concerning and widespread promulgation of simple ‘aid-memoirs’ could help, along with incentives. Information supports risk related to multiplicity, size and histopathology of adenoma and cancer findings at the index colonoscopy. Quality issues relating to colonoscopy and pathology reporting are being driven through professional fora and training. The paradox of multiplicity and quality colonoscopy needs addressing in a patient-centred response. Risk-stratification and adjustment over time is likely to gain increasing importance. The serrated pathway, its biology and epidemiology, have attracted attention for the rapid progression and association with interval cancers. Practice points for the management of malignant polyps continue to be topical. The effectiveness of intensive follow up strategies following curative treatment for colorectal cancer remains unproven, although colonoscopic surveillance is still of value.
Before commencing on the issue of risk and what can be done to manage the risk, it is worth pausing to take stock of the evidence that the risk is modifiable. Risk assessment has little clinical relevance unless there are effective ways to modify that risk. Primary prevention, through reducing risk, has a role, and there is increasing evidence around strategies such as aspirin or calcium supplemental chemoprophylaxis, dietary modifications such as for red meat, fibre, cruciferous vegetables, and lifestyle factors such as exercise and healthy weight maintenance.1 Additionally, following the positive results for polyp burden reduction in familial adenomatous polyposis, Eicosapentaenoic Acid-Free Fatty Acid (EPA-FFA) is currently under study through a randomised control trial (RCT) in high risk adenoma patients.2,3 But colonoscopy with polyp detection and removal is the most likely, but not certain, strategy to prevent colorectal cancer.
Many commentators take it for granted that colonoscopic screening or surveillance reduces the incidence of and mortality from colorectal cancer (CRC) without critical evaluation. The non-randomised experience of colonoscopic surveillance in Lynch Syndrome is often quoted.4 A recent report from the Nurses Health Study and Health Professionals Observational Follow-up Study also reports reduced CRC incidence in participants having a negative colonoscopy (HR 0.44 95% CI 0.38 to 0.52), as well as a reduced mortality from CRC (0.32 95% CI 0.24 to 0.45). For both incidence and mortality, the benefit included protection from proximal colon cancer.5 However, by any good epidemiological standard, the answer would need to come from RCTs, where the intervention is colonoscopy at intervals (perhaps 10 years) versus a control group with no screening or, to be practical, standard screening advice in their setting. Reduced mortality from CRC associated with colonoscopy intervention would be the best endpoint. In fact, there have been no such trials published. Several long-term trials against different randomised control groups are under way: the Veteran’s Administration trial in the US is against Faecal Immunochemical Testing (FIT); a large Spanish trial is also against FIT testing; a New York trial against standard US screening advice (measuring participation only of people responding to an initial invitation); and an important Scandinavian trial where the control group has no screening (screening is not implemented or advocated at a population level in Scandinavia).6-9 The Spanish trial has published CRC incidence rates after the initial screening round and there were just as many CRCs detected in the FIT arm as the colonoscopy arm.8 This gives pause for thought on cost benefit (poor for colonoscopy) and reach into the population (poor for colonoscopy). Of note, those that did participate in the colonoscopy arm – which were substantially fewer than in the FIT arm – had as many CRCs as were detected in the larger proportion who accepted FIT testing. The advanced adenoma detection rate, however, was about three times higher in the colonoscopy arm, perhaps pointing to a longer term benefit of colonoscopy in preventing CRC within this trial.
There are RCTs demonstrating reduction in cancer mortality through the faecal occult blood test (FOBT) and in flexible sigmoidoscopy programs.10 The question is relevant given the complication rate and (albeit low) mortality associated with colonoscopy.
The lack of RCTs addressing cancer incidence and mortality through colonoscopy screening also impinges on the rationale for management of risk for adenoma patients. In adenoma follow up and indeed in general, the US National Polyp Study is often quoted as demonstrating that colonoscopy with adenoma removal prevents CRC.11,12 This trial randomised participants to a more (zero, one and three years) versus a less (zero and three years) intensive surveillance schedule – showing no difference in adenoma or advanced adenoma outcomes. It did not have a control group of ‘no colonoscopy’. The initial and later analyses did assess the cancer outcomes in comparison with population incidences of CRC, and historical groups of adenoma patients who did not have colonoscopy – pointing to the possibility that the participants did avoid CRC, as there were statistically fewer that developed within both trial arms compared with those control groups. It should be noted that many other long-term studies of adenoma patients in surveillance programs have not identified a reduced cancer incidence rate below the average incidence – though one assumes that the populations under study were above average risk for CRC to start with, given their propensity to form adenomas.
There is evidence that FIT testing, complementing scheduled colonoscopy in an adenoma and cancer surveillance program, can bring forward the time of detection of advanced adenomas and cancers.13 This has not been formally addressed in any national screening guidelines, but is implemented in some organised programs in Australia,13 including the authors’.
A comparison of the US Multi-Society Task Force on Colorectal Cancer guidelines,14 with the British Society of Gastroenterology Guidelines,15 and more recently European guidelines,16,17 has recently been published.18 The greatest deviation from the Australian Guidelines and worthy of note, are the European Society of Gastrointestinal Endoscopy’s guidelines that recommend returning screenees to the average-risk national screening program or a colonoscopy after 10 years if no screening program exists, in the low risk group (1-2 small adenomas with low-grade dysplasia), and an increase in interval from three years to five years after a normal follow up colonoscopy in the high risk group (3-4 adenomas, villous features or high grade dysplasia, or ≥10mm in size).19 Another strong recommendation, although backed only by low quality evidence, is that the endoscopist be responsible for providing a recommendation for the post-polypectomy surveillance schedule. Differences between US and Canadian guidelines have also been published, highlighting the standard of care for average risk (in low risk long-term adenoma follow-up), and differentiating between three or more, and 10 or more adenomas as do the Australian guidelines.19,20 The paper is worthy of review.20
Despite considerable investment in the development of guidelines, numerous groups have shown barriers to their implementation and ‘widespread ignorance of guidelines’.11 Most societies or national bodies have provided funding to develop algorithms. The British Society of Gastroenterology published a handy wall chart summarising their guidelines.15 Recently, such a wall chart presentation of the Australian 2011 Cancer Council/National Health and Medical Research Council (NHMRC) Guidelines for Colonoscopy in Surveillance for adenomas and CRC has been produced and is presented (figure 1 and 2). Although simple strategies such as this have been shown to be effective, barriers to the use of guidelines go beyond this.21-23 Access to relevant past information may not be readily available (e.g. previous colonoscopy results, histology and family history) and the current procedure details may be inadequate (colon completely examined, clearance confirmed, exact number and sizes of the polyps noted, histology findings). Clinicians must also deal with patient anxiety and the fear of litigation. Linkage of guideline use to key performance indicators, bonuses and indemnity could enhance wider uptake.
The Australian 2011 guidelines had some degree of complexity over frequency of surveillance colonoscopy, derived from the special consideration of risk associated with multiple adenomas. Different risks (and therefore follow-up intervals) were assigned to patients with 1-2 vs 3-4 vs 5 to 9 vs 10 or more adenomas.19 Whether this needs simplification to enable it to be better accepted in clinical practice, or whether there is sufficient justification to promote that complexity, is a matter for discussion. The logic and data around the complexity is clear, but the complexity in itself may dilute the impact of the guidelines overall.
Multiple adenomas – measured cumulatively or at the last colonoscopy?
A level of uncertainty exists in the literature on this question. The 2011 guidelines recognised the uncertainty, and followed the pragmatic option of accounting for adenomas only at the last colonoscopy, rather than attempting a cumulative history. Further predictive studies need to address this issue. Inherently, one would think that it is the cumulative number of adenomas over time which engages the risk for metachronous CRC most closely, as the timing and frequency of interventions to remove adenomas are somewhat incidental to the biological drive to multiplicity – and presumably its associated metachronous cancer risk. Nevertheless, this has not been systematically teased out in adenoma follow-up studies.
Cut and discard
The evolving practice to ‘cut and discard’ small polyps through cold snare guillotine techniques threatens the assessment of metachronous risk which, as we know, is most powerfully associated with multiplicity of adenomas of whatever size, over and above the other histological and polyp characteristics of size, villosity and dysplasia.24 Although we are advocates for ‘cut’, we are not advocates for ‘discard’. In Australian practice, there is no differential rebate for multiple polyp assessment (as there is in the US), so pathology costs are the same.
Multiplicity and adenomatous polyposis syndromes
Multiplicity of adenomas plays very importantly into decisions around mutational analysis of the APC and MUTYH genes, again information lost with a ‘discard’ policy. In our Familial Cancer Clinic, we carefully record on a spreadsheet the entire colonoscopic history of patients referred, to inform decision-making. We will consider mutational analysis with as few as five documented adenomas. The predictive value of mutational analysis is directly related to the multiplicity.
Size, histology and dysplasia
Size, histology and dysplasia are relatively easily measurable and accessible for the purposes of determining risk. Furthermore, their predictive value is consistent across many studies. The three factors are closely correlated, so much so that the British guidelines take only size into account, being immediately assessable at the time of colonoscopy. If villosity and high grade dysplasia are not included in prediction algorithms, leaving only size and multiplicity of adenomas to determine high risk for metachronous advanced lesions, it does reduce the size of the high risk group slightly, with a minor shift in Receiver Operator Characteristic curves.18
Surveillance tailored multifactorial risk
Risk algorithms, not favoured to date in the 2005 or 2011 Guidelines, may yet prove useful with access to easily computed and reliable algorithms even built into endoscopy surveillance management programs. More experience is needed with this approach.25
Another important theme relating to risk profiling is the number of adenomas and CRCs detected in relation to the quality of colonoscopy.26 Attention has focused on measurement of quality and surrogates for quality. This includes the time taken to withdraw the colonoscope (during which inspection for polyps takes place),27 adenoma detection rates,28 bowel preparation cleanliness, retroversion of the colonoscope in the right colon and rectum,29 and the thorny issue of missed cancers occurring at an interval after a colonoscopy.27-31 Whereas quality of colonoscopy is the subject of another paper in this issue, it does bear reinforcement that all of these parameters have a logical connection to quality colonoscopy and point to ways of implementing quality control systems in colonoscopy.32 Perhaps the most compelling data, now from two sources, is that a colonoscopist’s adenoma detection rate in routine screening colonoscopy is indirectly but tightly related to the incidence of CRCs occurring in the years after colonoscopy – the interval cancer rate. This has been evident in both Polish and US studies.7,31
The multiplicity paradox
The integration of the themes of risk associated with multiple adenomas, and the logical training and practice goal to increase adenoma detection rates, brings us to a paradox: those patients who are under the care of high quality colonoscopists with high adenoma detection rates will likely be found to have more polyps and adenomas, driving them under current guidelines (which are themselves, as noted, determined by multiplicity) to have even more frequent colonoscopies, inevitably towards points of diminishing return. On the other hand, individuals who are under the care of poor quality colonoscopists with low adenoma detection rates will be found to have few (or no) polyps, placing them in a ‘lower’ risk group, requiring less frequent colonoscopies on current guidelines – yet we know these people are the ones who develop the interval cancers. An anecdotal impression is that low quality colonoscopists compensate by offering frequent colonoscopies, outside guidelines. The answer to this dilemma must be to introduce quality control systems across all colonoscopy practices, including monitoring adenoma detection rates. With time, we may be able to introduce colonoscopy quality parameters into the guidelines such that the interval between colonoscopies can be discounted (lengthened) where good quality colonoscopy has been documented through a range of parameters relating to the procedure and the colonoscopist. Notwithstanding that a colonoscopist’s adenoma detection rate in US studies is calculated from the relatively homogeneous population of average risk patients undergoing screening colonoscopy (a population which is not within current Australian guidelines for clinical practice and is not reimbursable through Medicare), adenoma detection rates in other Australian settings can be used with some reliability. At the same time, there would need to be an economic incentive for the proceduralist to meet these standards (or disincentive if not). This should surely be in the patient’s interests and attractive to the payers. This would then address the paradox.
The 2011 NHMRC guidelines are equivocal regarding the need to maintain surveillance at the interval determined by the polyp and patient characteristics at the time of the index (the last) colonoscopy. With follow up colonoscopies showing no further polyps, can the interval be relaxed? In some situations the answer is clearly ‘no’. This would include the serrated polyposis syndrome discussed below, perhaps serrated polyps short of the syndrome, and the well characterised genotypically defined syndromes of Lynch Syndrome, familial adenomatous polyposis, MUTYH associated polyposis, Peutz Jeghers Syndromes (polyps grow quicker than adenomatous polyposis in the author’s experience) and juvenile polyposis. Debate on the velocity of carcinogenesis in MUTYH associated polyposis has been engaging.33 However, in the common adenoma patient, follow-up interval is less certain. In the Royal Melbourne Hospital-Flinders long-term experience (submitted for publication), there is a relatively high risk for advanced adenomas to be found within 18 months of an index colonoscopy, where an advanced adenoma is also identified and removed (we carefully reviewed the data to exclude patients from the analysis where the index advanced adenoma was not completely removed). With time, the risk did attenuate, but still there was a long tail of advanced adenoma detection that continued at a stable rate, suggesting an intrinsic continuing risk that needs to be addressed through a fixed frequency of colonoscopy – arguably three yearly from our data. This is supported from US experience. For small adenomas, the risk is small as reported in many series, such that the risk for metachronous cancer reverts to average risk or below average risk.6,7,27,34
The serrated pathway
The discovery and understanding of the serrated polyp pathway to CRC has been the focus of much attention since the last guidelines. There is now some evidence that identifies interval cancers in adenoma and other surveillance programs as being more likely to be associated with the serrated pathway, either through methylation of the MLH1 promotor, or more generally, having high CpG Island Methylator Phenotype status.35,36 Studies on antecedent polyps in these patients, especially as to their serrated architecture, are needed. Some evidence suggests that polyps pass through this pathway more rapidly than the more conventional microsatellite stable, APC gate-controlled pathway. Importantly, it would point to the need for more frequent surveillance in patients who have shown a propensity to develop sessile serrated polyps.37 A consensus meeting dedicated to serrated lesions recommended particular attention (increased frequency) to patients with three or more sessile serrated adenomas/polyps or traditional serrated adenomas, especially if large (every two years) and any with dysplasia.38 This question needs more data before implementing a change to the guidelines. The 2011 guidelines signalled an issue relating to this question, but did not spell out any alteration to the frequency of colonoscopy in follow-up for these patients, which are determined, as in conventional adenoma follow-up, by multiplicity and size of adenomas, with villosity and dysplasia also implicated through the definition of an advanced adenoma. Advanced adenomas in the current guidelines attract a three year interval for colonoscopy.
Serrated polyposis syndrome
Serrated polyposis Syndrome (previously known as Hyperplastic polyposis) is increasingly being recognised by colonoscopists. It is defined by five serrated polyps proximal to the sigmoid colon, with two one cm or over in size, or 20 (some say 30) serrated polyps spread throughout the colon. The third definition is any serrated polyps in a first degree relative of a patient with serrated polyposis. This remains tantalisingly without a genetic predisposition identified, whereas all other multiple polyposis syndromes have had their germline predisposition identified. Perhaps this is not surprising, as Mendelian inheritance is not commonly seen in the families of patients with serrated polyposis syndrome. The colonoscopist needs to treat this syndrome respectfully: although the absolute risk of CRC is not well defined, it is undoubtedly high.39 Most colonoscopists have experienced interval CRCs occurring during surveillance of these patients, even within the recommended two year interval. Although this could be due to the inherent difficulty in detecting the subtle, flat and sessile serrated polyps with their indiscernible margins in the right colon (though perhaps flagged through its mucus cap), the evidence around the real possibility of a rapid pathway through diffuse methylation of suppressor genes or other mechanisms needs constant scrutiny. The high risk of CRC in the first degree relatives of patients with the serrated polyposis syndrome needs addressing in surveillance.40
Little new information has emerged to change the recommendations for management of malignant polyps, which balances the risk of surgical intervention (after malignant polypectomy) versus the risk of nodal metastases with ultimate progression within the lifetime of the patient.41 Attention has been given to the importance of pathology reporting for decision-making. The recent publication by the Royal College of Pathologists of Australia of a structured reporting protocol for polypectomy and local resections of the colon and rectum are likely to be beneficial.
This section addresses the risk of metachronous CRC in patients who have already developed CRC and the role that colonoscopy plays in managing this. A more comprehensive analysis of the contemporary literature is available, which points to the limited benefit of surveillance after CRC resection, duration of follow-up, intensity and methods of follow-up, cost-effectiveness, and identifying RCTs in progress further addressing the question.42 Colonoscopies should be done with the same quality in cancer follow-up as in adenoma follow-up.
The main change introduced in the 2011 guidelines was the introduction of a colonoscopy at one year after resection. Although the need for peri-operative total colonoscopy to seek synchronous cancers overlooked either due to incomplete index colonoscopy due to obstructing lesions, or other considerations, has long been recognised, the importance of a routine colonoscopy at 12 months from follow-up studies was brought to the fore in the 2011 guidelines. This holds true and may, incidentally, have a message for patients with advanced adenomas at index colonoscopy as well – notwithstanding the National Polyp Study noted above. Perhaps not surprisingly, the risk of metachronous adenomas and cancers is generally lower after cancer resection, than in adenoma follow-up. Counterintuitive? Probably not, as the resection reduces the epithelial mass available for adenomas and cancers to develop.
The metachronous risk of CRC after segmental oncological resection in Lynch Syndrome is now very clear: it is high – up to 60% at 40 years. Thus there is a strong rationale for suspecting, then diagnosing (preferably molecularly) and counselling patients with Lynch Syndrome to undergo extensive colonic surgical resection prior to resection of the index cancer or other advanced lesion in the colon. At a minimum, in the appropriate circumstance such as an early age onset index colon cancer, immunohistochemistry on the cancer should be done as part of the diagnostic work up. This information should usefully help decision-making around the surgical approach. Family history of cancer and the pattern of loss of expression in the cancer would all play into this decision-making.
Evidence is accumulating on risks for metachronous adenomas and cancers in patients with adenomas or CRC. Risk reduction through appropriate colonoscopic surveillance has been described in the 2011 NHMRC Clinical Practice Guidelines for Surveillance Colonoscopy, However, implementation of these guidelines has been limited by lack of resources to promote the guidelines in clinical practice, except for their publication on the NHMRC website. This will be addressed in part by the algorithmic depiction of the guidelines now available, and published here, for dissemination at points of service, be it general practice, endoscopy services in private and practice and through dedicated and managed follow-up programs. Further, the need and implications of quality practice in colonoscopy, especially with respect to adenoma detection rates, will need leadership and buy in by the endoscopic community and professional bodies.
Points of continuing clinical research attention include systems to integrate cumulative adenoma detection in patients into risk and surveillance planning, the biology of the serrated pathway with its implications for surveillance scheduling, and further attention to early follow-up risk in patients with advanced adenomas.