Saturday 1 December 2012
Hyperplastic polyposis syndrome (HPS); hyperplastic polyposis syndromes; hyperplastic polyposes
Hyperplastic polyposis syndrome was originally conceived as a rare and benign condition that was important only insofar as it could be confused with FAP.
When a relatively relaxed definition was introduced, HPS was estimated to affect about 1/2000 subjects, making it more common than FAP.
The aim of the relaxed definition was to distinguish HPS from the common occurrence of multiple hyperplastic polyps clustering around a rectal cancer but at the same time to acknowledge the fact that relatively small numbers of polyps may be clinically important, particularly if the polyps are large and proximally located.
This definition suggested a lower limit of 30 small pan-colorectal polyps or as few as five polyps proximal to the sigmoid colon provided that at least two of these were >1 cm.
It was assumed that subjects with proximal hyperplastic polyps meeting the above definition would also have distal polyps and so would have more than five polyps overall.
It is now evident that small numbers of hyperplastic polyps, albeit relatively large, may be restricted to the proximal colon. Nevertheless, the risk of malignancy in such subjects may not be negligible.
At first, this seemed to be an exception to the rule since none of seven cases of hyperplastic polyposis from St Mark’s Hospital, London, was associated with colorectal cancer.
This differing cancer risk with respect to case reports versus series of cases has continued to the present time.
The 207 examples of hyperplastic polyposis reported up to 2006 have been single cases or small series of up to 4 cases (totaling 36 cases), and larger series comprising five or more cases (totaling 171 cases).
In these two groups, the frequency of patients with one or more colorectal cancer is 25/36 (69%) versus 63/171 (37%), respectively.
These numerical differences are explained by the fact that single case reports were published because the clinical or pathological features were considered to be unusual.
Examples of such notable features include numerous and/or large hyperplastic polyps, proximally located polyps, transitions to dysplasia, the presence of one or more colorectal cancers, and the early age of onset of colorectal cancer.
By contrast, larger series of cases have generally adopted definitions of hyperplastic polyposis in which polyps may be not be especially numerous, large, or located in proximal colon.
The high frequency of colorectal cancer in single case reports of hyperplastic polyposis is clearly influenced by ascertainment bias.
However, the use of relatively non-stringent diagnostic criteria for hyperplastic polyposis in the larger case series may have led to an under-estimation of the risks of malignancy.
The description of six patients with multiple ‘serrated adenomatous polyps’ of whom four patients had one or more colorectal cancer was central to our understanding of the heterogeneity of hyperplastic polyposis.
These cases had been initially diagnosed as hyperplastic polyposis, but the polyps were noted to differ from typical hyperplastic polyps and to display features of an adenomatous nature.
The polyps in question were large, sessile, showed exaggerated serration and crypt dilatation, and were hypermucinous.
Despite lacking the traditional cytological features of adenomatous dysplasia, these polyps were regarded as neoplasms with malignant potential. These lesions were subsequently shown to occur sporadically and were named ‘sessile serrated adenomas’.
In an editorial accompanying the paper on serrated adenomatous polyposis, it was suggested that there might be two forms of "hyperplastic polyposis":
One form would comprise "small, typical hyperplastic polyps", albeit multiple, and have little associated risk of malignancy. Many of the cases in the series described by Williams et al. and Ferrandez et al. would belong to this latter category.
The other form ("serrated adenomatous polyposis") would feature multiple and large sessile serrated adenomas, "traditional serrated adenomas", "mixed polyps" (serrated polyps with a dysplastic component indicative of neoplastic progression), and even conventional adenomas. This latter type of hyperplastic polyposis would be associated with a significant risk of colorectal cancer.
The demonstration of clinical or phenotypic heterogeneity accompanied by a consistent pattern of genetic signatures would highlight hyperplastic polyposis as a set of different conditions.
Rashid et al. grouped patients with hyperplastic polyps into three subsets:
13 subjects (including 3 from 1 family) with hyperplastic polyposis (>20 hyperplastic polyps),
5 subjects with less than 20 hyperplastic polyps (1–14) but at least 1 measuring 1 cm in diameter or more,
and 5 subjects with multiple but less than 20 hyperplastic polyps and none measuring 1 cm.
This study did not distinguish sessile serrated adenomas but included them with hyperplastic polyps.
KRAS mutation was most common in hyperplastic polyps from subjects with multiple hyperplastic polyps (16%) and was not found in any hyperplastic polyps from patients with large hyperplastic polyps.
Loss of heterozygosity (LOH) in chromosome 1p32-36 occurred in 13% of hyperplastic polyps but only in polyps from patients with hyperplastic polyposis. Five of 13 patients with hyperplastic polyposis had HPs with 1p LOH, implicating 16/77 (21%) polyps from these patients.
DNA microsatellite instability (MSI) at low- (MSI-L) or high-level (MSI-H) was uncommon in hyperplastic polyps (4.6%) but occurred in 22.0% of serrated polyps with dysplasia (traditional serrated adenomas or mixed polyps). One patient with hyperplastic polyposis and a colorectal cancer had MSI-H in multiple lesions (three of seven hyperplastic polyps and three of four adenomas).
Other studies have documented the existence of MSI-L or MSI-H in serrated polyps occurring both sporadically and in the context of hyperplastic polyposis.
This molecular finding is more common in "serrated polyps with dysplasia" ("mixed polyps" or "traditional serrated adenomas") and is therefore more likely to be associated with progression than with initiation of the "serrated adenoma pathway".
An important observation in the study by Rashid et al. was the absence of both KRAS mutation and 1p LOH in hyperplastic polyps from subjects with large polyps.
Since the report by Rashid et al. was published, it has become clear that the most frequent genetic alterations in "large hyperplastic polyps" or "sessile serrated adenomas" include mutation of BRAF and DNA methylation, implicating tumor suppressor genes such as p16INK4a, p14ARF, RASSF1, RASSF2, RASSF5, and MST-1, and the DNA repair genes MGMT and MLH1.
Extensive DNA methylation has even been described within the normal mucosa of subjects with hyperplastic polyposis.
The existence of two principal types of hyperplastic polyposis was raised over 10 years ago, but there have been few attempts to resolve hyperplastic polyposis into fully characterized entities with differing implications for clinical management.
Molecular classifications based on mutational and methylator patterns not only serve to objectify this exercise but may provide clues to the all-important underlying etiology.
There is increasing evidence that mutation of oncogenes such as BRAF and KRAS in isolation is not initially pro-tumorigenic but may in fact result in replication arrest and cell senescence or apoptosis.
In both serrated polyps and colorectal cancers, BRAF mutation is associated with high-level CpG island methylator phenotype (CIMP-high) while KRAS mutation is associated with CIMP-low and.
CIMP-high and CIMP-low may differ both qualitatively as well as quantitatively.
It is likely that DNA methylation is of crucial importance in silencing tumor suppressor genes such as p16INK4a (cell cycle regulator) and RASSF1 (pro-apoptotic) and thereby directing the effects of oncogene mutation away from senescence and apoptosis towards the tumorigenic characteristics of proliferation and immortality.
In order for the early tumorigenic potential of their mutated forms to be realized, BRAF may be more dependent upon CIMP-high synergies while KRAS may require only CIMP-low.
In patients with HPS, there appears to be concordant mutation of either BRAF or KRAS. However, the fundamental molecular distinction between these subsets may depend on background predispositions to CIMP-high versus CIMP-low.
It remains to be shown whether or not such a pro-methylator tendency is established prior to mutation of BRAF or KRAS or develops in the crypts of predisposed subjects following these mutations.
The finding of extensive methylation in the normal mucosa in HPS subjects with a severe phenotype has suggested the possibility of a genetic predisposition to DNA methylation.
Alternatively, methylation could occur as a purely reactive phenomenon in the mucosa surrounding colorectal cancers to account for satellite-like clusters of hyperplastic polyps. Transfection of mutant BRAF into a ‘normal’ colonic cell line led to methylation of MLH1.
It is possible that DNA methylation precedes mutation of BRAF but is subsequently augmented following mutation in a co-dependant manner.
Conventional adenomas lacking evidence of glandular serration may co-occur with serrated polyps in HPS.
This raises the possibility that a predisposition to DNA methylation could influence the initiation and progression of adenomas by increasing the likelihood of inactivating relevant tumor suppressor genes.
This fits with the fact that APC is methylated in a subset of colorectal cancers but not in association with CIMP-high.
It is also possible that other potential tumor suppressor genes that are known to be methylated, such as MCC, could play a role in the adenoma–carcinoma sequence.
The pathologist will encounter HPS through the examination of endoscopically removed polyps or a surgical specimen.
In the case of the former, the pathologist depends heavily on the expertise of the endoscopist who will need to recognize the inconspicuous pale and relatively flat "serrated polyps".
Surgical specimens will generally include one or more cancers and large polypoid lesions that include conventional adenomas.
Either because they are small or are large but sessile, hyperplastic polyps may be overlooked or given insufficient attention.
For all these reasons, hyperplastic polyposis is likely to be under-diagnosed.
Even when a polyposis syndrome is suspected, there may be confusion with other types of polyposis.
Hyperplastic polyposis may occasionally be familial and occurs quite frequently in association with familial clusters of colorectal polyps and cancer.
This may lead to confusion with Lynch syndrome, particularly when one or more of the colorectal cancers shows high-level DNA microsatellite instability (MSI-H).
Such a finding should not come as a surprise given the fact that approximately one third of colorectal cancers associated with HPS are MSI-H.
Hyperplastic polyposis may be viewed as the most severe phenotype within an over-arching ‘serrated polyp syndrome’ that is clearly distinct from Lynch syndrome.
Colorectal polyposes: from phenotype to diagnosis. Jass JR. Pathol Res Pract. 2008;204(7):431-47. PMID: 18541388