• SCREENING
• COLORECTAL NEOPLASIA
• COLORECTAL NEOPLASM
• COLORECTAL PATHOLOGY
• COLORECTAL CANCER SCREENING

Serrated polyposis syndrome (SPS; formerly ‘hyperplastic polyposis syndrome’) has emerged as by far the most common colorectal polyp syndrome.1 ,2 SPS is characterised by some combination of large and/or numerous serrated class lesions and is associated with an increased risk of colorectal cancer (CRC).3 A genetic basis for SPS has not yet been discovered, and many cases do not have a family history of SPS or even CRC. Thus, any current definition of SPS must be clinical and based on colonoscopic findings and the pathology of polyps removed during colonoscopy. Most clinicians accept the WHO criteria for diagnosis of SPS as reasonable working criteria for clinical studies as well as clinical practice. The WHO proposes three criteria, and satisfaction of any of the three criteria is sufficient to diagnose SPS (box 1). It is critical that clinicians understand that ‘serrated class’ lesions can include hyperplastic polyps (HPs) or sessile serrated polyps (SSPs; also known as sessile serrated adenoma) or even the relatively rare ‘traditional serrated adenoma’ when counting lesions to determine whether the WHO criteria have been met in a patient. Criteria 1 and 3 (box 1) constitute the two phenotypes of SPS that are most relevant to the discussion of colonoscopic control and prevention of CRC in patients with SPS. The first job of every colonoscopist with regard to SPS is to know the current criteria for SPS and apply them during colonoscopy so that the diagnosis of SPS is made and appropriate management started.

Patients with an established diagnosis of SPS are recommended to undergo colonoscopy annually.4 However, colonoscopists dealing with substantial numbers of patients with SPS quickly learn that patients with SPS appear to have a very broad range of polyp burden. For the patient who is initially found to have multiple serrated lesions 2–3 cm or more in size, annual colonoscopy is not enough to bring the polyp burden down to acceptable levels. In other patients with a modest number of predominantly small and diminutive lesions, annual surveillance seems too great a burden. Thus, SPS management seems ripe for some risk stratification with regard to colonoscopy surveillance. Can we identify a subgroup of patients with SPS at higher risk of developing CRC? Can we define in absolute terms the corresponding risk of developing CRC?

In this issue of Gut, two series including 730 patients with SPS identified risk factors for CRC in patients with SPS and described CRC incidence in SPS during surveillance.5 ,6 In both the Spanish and Dutch-British cohorts, patients with SPS were included irrespective of the clinical presentation, as expected when considering the low prevalence of SPS. The inclusion of symptomatic patients explains the high rate of already prevalent CRC at the time of SPS diagnosis.5 ,6 The reported 16–29% CRC prevalence indicates that, in a substantial proportion of patients, SPS diagnosis was delayed until or after the time of CRC diagnosis.5 ,6 This high prevalence rate is not a proxy for the rate of incident CRC in SPS. Interestingly, the recto-sigmoid location of approximately 50% of SPS-associated CRC was somewhat unexpected, suggesting either a relevant role for the adenoma-carcinoma sequence in patients with SPS (these patients often have synchronous conventional adenomas) or a less innocent role for recto-sigmoid serrated lesions that are not emphasised in the current WHO definitions of SPS.

In both studies, the prevalent CRC cases were exploited to create an internal comparison between SPS with and without CRC.5 ,6 The clinical, endoscopic and histological features associated with increased CRC risk in patients with SPS in the two studies were consistent with the current understanding of the molecular and clinical features of how serrated lesions, and a hyper-methylated colon generally, progress to cancer.7 For example, dysplasia within serrated polyps was a clear risk factor for CRC in both series,5 ,6 although it was restricted to high-grade dysplasia in the only series where dysplasia was graded.5 In the hierarchy of serrated lesions, SSPs with cytological dysplasia are generally considered a more advanced lesion than an SSP without cytological dysplasia (figure 1). Proximal location and multiplicity of serrated lesions were also identified as risk factors for CRC, albeit with some differences between the two cohorts. In the Dutch-British cohort, patients who met the WHO type 1 and type 3 criteria for SPS, which are proxies for proximal location and multiplicity, respectively, had higher CRC risk.6 In the Spanish cohort, in which HP and SSP were treated separately, it was the number of proximal SSPs that was associated with CRC risk.5 In addition, the paradox of an inverse relationship between smoking and CRC risk—clearly shown in one series6—was supported by a suggestive trend in the other.5

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Endoscopic (A) and histologic appearance (B) of a sessile serrated polyp with cytological dysplasia. In A the dysplastic portion (red box) has the endoscopic and histologic appearance of a conventional adenoma. The remainder of the lesion (black box) has the endoscopic and histologic appearance of a sessile serrated polyp.

These findings are potentially important, indicating that a subgroup of patients with SPS with distinctly advanced clinical, endoscopic or histological features may be at higher risk of CRC. However, some caution is needed in interpreting these findings. First, it is reasonable to assume that most SPS cases diagnosed with an already prevalent CRC were not under endoscopic surveillance. Therefore, endoscopic clearing of all the relevant lesions might well have removed the excess risk attributed to the identified risk factors. Second, risk stratification would imply that patients with SPS with and without the identified risk factors represent two distinct phenotypes. However, they might merely be earlier and later phases of the same phenotype. Third, the clinical, histological and endoscopic SPS risk factors identified in the studies could have been affected by suboptimal interobserver agreement or awareness of the disease.1 ,8 Large multicentre studies involving colonoscopy are always affected by variable detection skills of colonoscopists, and these studies were likely affected by the lack of central histology review.5 ,6 For instance, the degree of dysplasia in serrated lesions has been graded in the Spanish cohort, although currently not recommended by WHO classification. Fourth, the retrospective and heterogeneous inception of the cohorts, as well as possible selection bias, limit the generalisability of the data. Such factors might account for unexplained differences between the studies such as the association of concurrent advanced adenomas with CRC in the Dutch cohort but not in the Spanish one.5 ,6

In both study cohorts, substantial numbers of patients entered intensive endoscopic surveillance with clearing of all relevant lesions, and surgery was performed in up to 30% of SPS cases.5 ,6 The 1.5–1.9% absolute 5-year CRC risk estimated in both cohorts was much lower than expected,9 representing only a 3–4-fold increase compared with the general population.10 This is a critical result of the studies since it indicates that SPS patients without CRC who are followed at dedicated centres have a long-term risk of CRC that is reassuringly low, emphasising the importance of early recognition and treatment of SPS and the feasibility of colonoscopic surveillance.

The low overall risk of CRC during surveillance in patients without CRC could support less intensive surveillance in a subgroup of patients with SPS, depending on the strength of association between identified risk factors and CRC risk. Unfortunately, these associations were weak, with ORs ranging between 1.04 and 2.3 and large CIs.5 ,6 A twofold CRC increase associated with a hypothetical risk factor in the SPS population would change the 5-year CRC risk from 1% to 2%, and such a small difference might be considered to marginalise the clinical impact of the risk factor.

Despite the limitations of these studies, the low risk of CRC during SPS surveillance in these studies and the consistency of the identified CRC risk factors in the same studies with already established opinion about CRC risk factors in serrated class lesions (large size, dysplasia in SSP, multiplicity, proximal location) suggest that experts from centres managing patients with SPS could reach consensus for risk-stratified management of SPS pending the results of new studies. In figure 2, we give an example of what a management algorithm might look like. Effective management would presume that centres could achieve high-quality bowel preparation, use endoscopists with high detection rates for serrated lesions and conventional adenomas, and use effective serrated lesion resection techniques. Lesions resected should be sorted by size and location and counted so that the impact of previous colonoscopic resections can be determined at serial examinations. Serial examinations should show progressive reduction in the size and number of lesions, and the rapid disappearance of large lesions and lesions with dysplasia. It is quite possible that centres will differ in their effectiveness in preventing cancer in SPS, and that those differences will be a function of the detection skills of colonoscopists, the willingness and ability of endoscopists to perform long procedures with multiple endoscopic mucosal resections, and their ability to convince patients with SPS of the seriousness of SPS and the need for adherence to recommended follow-up.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Possible algorithm of clinical and endoscopic management of patients with serrated polyposis syndrome (SPS) based on risk stratification. EMR, endoscopic mucosal resection; SSP, sessile serrated polyps.