Recent studies have outlined the serrated polyp pathway of colorectal neoplasia, which is characterised by mutation of BRAF (V600E), DNA hypermethylation and microsatellite instability (MSI).1^–4 Carcinomas with defective DNA mismatch repair and MSI are believed to arise via this pathway.

The polyps characteristic of this pathway share the morphological feature of serration, which appears to result from decreased apoptosis.4 Serration is a saw-tooth glandular profile seen prototypically in hyperplastic polyps (HPs). The polyp most recently described in this family is the sessile serrated adenoma (SSA, also called sessile serrated polyp and serrated polyp with abnormal proliferation).1^–3 5 This lesion does not have “adenomatous” dysplasia, and so it needs to be distinguished from the serrated adenoma as originally described by Longacre and Fenoglio-Preiser; the latter is now often referred to as the traditional serrated adenoma (TSA) to make this distinction clear.1 3 6 For lesions that combine a component with hyperplastic polyp-like (or SSA-like) features and a separate component resembling a traditional adenoma, the term mixed hyperplastic adenomatous polyp (MHAP) can be used.6 7 It has been shown that the V600E BRAF mutation is found in both components of any particular MHAP.8

This diagnostic classification of colorectal polyps has been defined relatively recently, and the principal aim of this study was to determine the prevalence of these lesions in a large, unselected series of patients undergoing endoscopy. We also performed BRAF mutation analysis on a subset of lesions to correlate with our morphological diagnoses.

METHODS

The case records of Southern.IML Pathology were searched using SNOMED codes to identify all neoplastic and hyperplastic polypoid lesions of the colorectum for the calendar year 1 April 2005 to 31 March 2006. Only endoscopic biopsies were included. The population served by Southern.IML is predominantly urban and typical of Eastern Australia. The study was approved by the University of Wollongong/Illawarra Area Health Service Human Research Ethics Committee.

All slides were retrieved from the archives and every case was reviewed histologically. All serrated polyps (ie, HPs, TSAs, SSAs and MHAPs) were reviewed by at least two of the authors (NJC, HM, KLT) to obtain a consensus diagnosis. For each category of polyp, if a patient had more than one, a polyp was chosen at random for analysis. Thus, the results represent the number of patients with each type of polyp rather than the total number of polyps.

The criteria used for diagnosis were those identified by Torlakovic et al.5 Adenomas without serrated features were classified as non-serrated adenomas; this category corresponds to “conventional” tubular, tubulovillous and villous adenomas, and these polyps showed “traditional” cytological dysplasia. SSAs lacked “traditional” cytological dysplasia but showed architectural features indicating disordered growth (dilated crypt bases, growth of crypts parallel to muscularis mucosae, asymmetrical crypt bifurcations, irregular crypt branching). Other criteria for a diagnosis of SSA were mitotic activity in the upper crypt, goblet cells or gastric foveolar-type cells in the crypt base, dystrophic goblet cells or serration at the crypt base, eosinophilic cell change, and pseudostratification of surface nuclei.1 3 We classified polyps as SSAs provided we could identify diagnostic features even if they were present only focally. In contrast, HPs were characterised by crypts that were uniformly narrow in the lower third and lacked the features of SSA. Serrated polyps showing “conventional” cytological dysplasia were classified as TSAs. These lesions often show eosinophilic cell change.3 Polyps in which there were areas resembling HP or SSA, together with clearly separate areas resembling non-serrated adenoma or TSA, were classified as MHAPs.1 7

The patient records were examined to determine the age and sex of the patient, the clinical details supplied with the specimen, and the site of the polyps. Lesions proximal to the splenic flexure were classified as right-sided, while lesions of the splenic flexure, descending colon, sigmoid colon and rectum were classified as left-sided. The material was unselected; approximately 95% of specimens were obtained at colonoscopy, and the remainder were obtained at sigmoidoscopy.

We subjected all cases of SSA, TSA and MHAP with available tissue to BRAF mutation analysis. We also selected some HPs and non-serrated adenomas for BRAF analysis for comparison. For this part of the study, we further classified the HPs according to the types defined by Torlakovic et al.1 5

Pyrosequencing for the detection of the BRAF V600E mutation

The detection of the V600E mutation in BRAF was performed as previously described.9 Briefly, the QIAamp DNA Mini Kit (Qiagen, Valencia, California, USA) was used to extract DNA from sections (4 μm) cut from formalin-fixed paraffin-embedded tissue blocks. DNA was amplified using primers (forward: 5′-GAAGACCTCACAGTAAAAATAG-3′ and reverse: 5′-biotin-ATAGCCTCAATTCTTACCATCC-3′) on a Tetrad 2 DNA Engine (MJ Research, Waltham, Massachusetts, USA) to produce 122 bp BRAF amplicons. The sequencing primer 5′-AGGTGATTTTGGTCTAGCTACAG-3′ was used with nucleotide dispensation order CATCGATC to interrogate the sequence 5′-A/TGA/GAATC-3′. For pyrosequencing, single-stranded DNA templates were immobilised on streptavidin-coated Sepharose high-performance beads (Amersham Biosciences, Little Chalfont, UK), incubated at 90°C and allowed to anneal to 0.4 μM sequencing primer at room temperature. Pyrosequencing was performed using PyroGold Reagents (Biotage, Uppsala, Sweden) on the Pyromark ID instrument (Biotage). Controls in which the sequencing primer or template were omitted were used to detect background signal. Pyrogram outputs were analysed by the PyroMark 1.0 software (Biotage) using the allele quantification (AQ) software to determine the percentage of mutant versus wild-type alleles according to relative peak height. DNA from a healthy normal control and the relevant mutant cell lines were included as negative and positive controls with each batch of samples.

Statistical analysis

Analyses to detect association between categorical variables (sex, site, diagnosis, BRAF status) were performed using the χ² test or Fisher exact test, while the relationship between polyp type and age was determined by a Student t test. All reported probability (p) values were two-sided and a value of ⩽0.05 was considered significant. The SPSS v15.0 statistical package (SPSS, Chicago, Illinois, USA) was used for all statistical analyses.

RESULTS

There were 515 patients with a serrated polyp and 964 patients with a non-serrated (“conventional”) adenoma. The demographic details and the percentage distribution of each type of polyp are shown in table 1. There were no significant differences between individuals with serrated or non-serrated polyps in terms of age, although serrated polyps were more common in women (39.8% of polyps versus 31.1% in men; p<0.001), and were more likely to be left sided (28.6% of right-sided polyps versus 40.6% of left-sided lesions; p<0.001). There were insufficient data to meaningfully analyse the indications for endoscopy, since the submitted clinical information in the majority of cases was simply “polyp(s)”.

There were 49 SSAs, 11 TSAs and 10 MHAPs with tissue available for BRAF analysis. We also randomly selected 42 non-serrated adenomas and 36 hyperplastic polyps for analysis. When the latter were classified according to the criteria of Torlakovic et al,5 34 were microvesicular hyperplastic polyps (MVHPs) while two were goblet cell hyperplastic polyps (GCHPs). The results of the BRAF analysis are shown in table 2. Overall, serrated polyps were more likely to harbour BRAF mutations than non-serrated polyps (p<0.001). Within the serrated polyp group, SSA and MHAP were significantly more likely to have BRAF mutations than either TSA or HP (SSA versus TSA, p = 0.002; SSA versus HP, p<0.001; MHAP versus TSA, p = 0.02; MHAP versus HP, p = 0.001). Hyperplastic polyps and TSA showed no significant difference in their frequency of BRAF mutations.

DISCUSSION

In a calendar year, we found 515 patients with a serrated polyp; 85% were HPs, 11% SSAs, 2.1% TSAs and 1.9% MHAPs. Our BRAF analysis showed a strong association between a diagnosis of SSA and the presence of BRAF V600E mutation (table 2). We believe this finding validates our ability to use morphological criteria to identify this diagnostic entity.

All our MHAPs harboured a BRAF mutation, whereas only about half of our TSAs did so. These findings are in keeping with the results of others, and we believe it supports the concept that MHAPs generally represent progression along the MSI pathway, whereas only a proportion of TSAs do so.1 8 As expected, the great majority of non-serrated adenomas did not exhibit BRAF mutation.

Only two of our HPs subjected to BRAF analysis were of goblet cell type. We believe this low proportion is probably a consequence of over-representation of right-sided lesions in this group. As expected, neither GCHP contained a BRAF mutation. In contrast, about a third of the MVHPs showed BRAF mutation. This finding is in keeping with the hypothesis that some MVHPs can progress along the MSI pathway and develop into SSAs.1 3 However, the proportion of MVHPs with BRAF mutation in our study is lower than that reported by others.1 2 10 This could be because we were able to classify correctly a high proportion of serrated polyps with BRAF mutation as SSAs rather than MVHPs. We believe this is the result of recognising the diagnostic changes of SSA, even when present only focally in a lesion.

A previous prospective study that used routine chromoendoscopy found SSAs to represent 9% of all polyps, raising the possibility that chromoendoscopy increases the likelihood of identifying SSAs at colonoscopy.11 In contrast, the reported frequency of SSAs was only 2.2% of all polyps in another study.7 However, those authors required four distinct morphological criteria of SSA to be present before making the diagnosis; our findings suggest this requirement is overly restrictive. Other authors have recommended that the diagnosis of SSA be reserved for proximal polyps at least 1 cm in diameter;2 our findings suggest this restriction, too, is unnecessary.

In summary we have described the prevalence of colorectal polyps of different types in endoscopic biopsies from an unselected population. SSAs were common and comprised 3.9% of all polyps and 11% of serrated polyps. BRAF mutations were found at high frequency in SSAs and MHAPs, at intermediate rates in HPs and TSAs, and were rare in non-serrated polyps.