Article Text

Download PDFPDF

Traditional serrated adenomas (TSAs) admixed with other serrated (so-called precursor) polyps and conventional adenomas: a frequent occurrence
  1. Runjan Chetty1,
  2. Sara Hafezi-Bakhtiari1,
  3. Stefano Serra1,
  4. Richard Colling2,
  5. Lai Mun Wang1
  1. 1Department of Pathology, Laboratory Medicine Program, University Health Network and University of Toronto, Toronto, Ontario, Canada
  2. 2Department of Cellular Pathology, Oxford University Hospitals, Oxford, UK
  1. Correspondence to Professor Runjan Chetty, Department of Pathology, Toronto General Hospital, 11th floor, Eaton wing, 200 Elizabeth Street, Toronto, Ontario, Canada M5G 2C4; runjan.chetty{at}


Background Traditional serrated adenoma (TSA) is a very characteristic type of serrated polyp that has a predilection for the left colon. Recent molecular advances have shown two molecular phenotypes of TSA: one associated with BRAF mutations and the other with KRAS mutations. The former is associated with hyperplastic polyps (HPs) and sessile serrated adenomas (SSAs), while the latter is associated with more conventional adenomatous dysplasia.

Aims The association of TSAs with so-called precursor lesions (HPs and SSAs) is not well recognised and the purpose of this study was to explore the coexistent presence of HPs, SSAs and adenomatous polyps within a large cohort of TSAs.

Methods In total 149 TSAs were examined for the presence of HP, SSA and adenomatous polyps.

Results Seen in 83 men and 65 women ranging in age from 32 to 89 years and 127 were left sided with 22 in the right colon. Seventy-eight of the 149 TSAs showed evidence of another polyp (52.34%): 32 were low-grade tubular/tubulovillous adenomas (TAs/TVAs; 41%), 28 were HPs (36%) and 18 were SSAs (23%). Eleven of the 22 right-sided TSAs were associated with a precursor lesion (1 HP and 7 SSA). In addition, five TSAs showed more than one polyp type: TSA with TA/TVA and HP (3); TSA with TA/TVA and SSA (2). The TAs/TVAs were adjacent to the TSA but occurred as a separate discrete polyp, while HPs and SSAs were intermingled with the TSA and present at the base and surface of the lesion.

Conclusions More than 50% of TSAs are associated with a precursor lesion or adjacent TA/TVA. Their recognition is important as this may have surveillance and management ramifications.


Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

It is now well accepted that three types of serrated polyps occur in the gastrointestinal tract: hyperplastic polyp (HP), sessile serrated polyp/adenoma (SSA) and traditional serrated adenoma (TSA). Much of the recent attention on these polyps has focused on their molecular pathogenesis and characterisation. However, some morphological issues that stem from a complex molecular pathogenesis still linger and need clarification. Most TSAs have a characteristic constellation of defining morphological features: an exophytic, villiform architecture, serrated or slit-like luminal profiles, intense cytoplasmic eosinophilia, the presence of ectopic crypt foci, and elongated, penicillate, slender nuclei with dispersed chromatin. Goblet cell-rich and flat variants have also been described, but generally the aforementioned criteria enable a diagnosis of TSA to be made. Detailed observations of TSAs have shown the coexistence of so-called precursor lesions (namely, HPs and SSAs) within a proportion of TSAs, but these have largely escaped recognition and emphasis in routine practice.1–7 Some of these studies have consisted of small numbers of cases or not described in detail the relationship of TSAs with precursor polyps, so named since they are putative harbingers of TSA. In terms of TSAs coexisting with other lesions, conventional adenomatous dysplasia and histological overlap with tubulovillous adenomas have also been described in TSA.8 ,9

The purpose of this study was to examine the incidence, extent and pattern of coexistence HP, SSA (‘precursor lesions’) and conventional tubular/tubulovillous adenoma areas within a large cohort of otherwise histologically typical TSAs.

Materials and methods

The computer records of the two pathology departments were searched for all cases diagnosed as TSA over a 4-year period (2010–2013 inclusive). All cases were reviewed to confirm the diagnosis of TSA and the relevant demographic and pathological data (age, gender, site) were also retrieved. To make the diagnosis of TSA, at least two of the following histological features were required: characteristic serrated or slit-like luminal profiles resembling small intestinal mucosa, intense cytoplasmic eosinophilia, the presence of ectopic crypt foci, and elongated, penicillate, slender nuclei with dispersed chromatin. A ‘flat’ variant of TSA was also recognised; hence, a villiform, exophytic architecture was not a prerequisite for the diagnosis. Once confirmed as TSAs, they were then examined for the coexistence of HP, SSA and conventional adenomas intimately admixed and/or immediately adjacent to predominantly TSAs. Adjacent to TSA was defined as being immediately next to the TSA without intervening normal mucosa. When TSA and non-TSA histological features were present within the same polyp, blending with each imperceptibly, but were still clearly recognisable as different epithelial types, this was regarded as intimately admixed. The presence of conventional adenomatous dysplastic tubules within a TSA was not deemed as a coexistent conventional adenoma but rather as dysplasia within a TSA, and polyps with these features were excluded. SSA was diagnosed using the one crypt architectural criteria tabled by the American College of Gastroenterology guidelines.10 The spatial relationship, in particular between TSA and other polyp types, was noted.


A total of 149 TSAs from 148 patients were retrieved: 83 men and 65 women ranging in age from 32 to 89 years (mean: 62.1 years); 127 were left sided (present at the splenic flexure and distally) in location, and 22 were right sided.

Of the 149 TSAs, a total of 78 showed evidence of another polyp (52.34%): 32 were low-grade tubular or tubulovillous adenomas (41%), 28 were HPs (36%) and 18 were SSAs (23%). In addition, five TSAs showed more than one polyp type: TSA with tubular/tubulovillous adenoma and HP (3); TSA with tubular/tubulovillous adenoma and SSA (2). In these five cases, the HP and SSA components were more prominent than the adenomatous polyps; hence, they were categorised as TSA with HP and SSA, respectively.

Interestingly, only 2 of the 78 were considered ‘flat’ TSAs; the remaining cases were all exophytic, protuberant villiform polyps.

The anatomical location of TSA with another coexistent precursor polyp was still predominantly left sided with only 11 occurring in the right colon. However, this represents 50% (11 of 22) of all right-sided TSAs, suggesting that right-sided TSAs are often accompanied by a precursor polyp. Only 1 TSA with an HP was located in the right colon (hepatic flexure); the remaining 27 were left sided. Of the 18 TSAs with SSA as a precursor lesion, 7 were located in the right colon.

With regard to the distribution of TSAs accompanied by tubular/tubulovillous adenomas, 27 of 32 were located in the left colon.

Pathology findings: TSA features formed the predominant histological pattern and features of other types of serrated or adenomatous polyps were either intimately admixed or immediately adjacent to the TSA. With regard to the presence of conventional adenomatous dysplasia, it is a moot point whether discrete, separate, tubular/tubulovillous adenomas are present or whether such lesions represent adenomatous dysplasia occurring at the periphery of a TSA. However, for the purposes of this study, an adenomatous area separate from, but adjacent to, a TSA were included and regarded as a coexistent tubular or tubulovillous adenoma (figure 1A, B). None of the tubular or tubulovillous adenomas contained high-grade dysplasia. The adenomas were discrete and clearly distinct from the TSA. The majority of HPs (figure 1C–F) and SSAs (figure 1G, H) with TSA were intimately admixed. Only one SSA and two HPs were adjacent to and separate from a TSA. Twenty-six of the 28 HPs were of the microvesicular type. Thirty of the 32 tubular/tubulovillous adenomas were adjacent to but separate from the TSA. The TSA areas were typical: villiform polyps composed of eosinophilic cells, deep serrations (resembling small intestinal villi) and the presence of ectopic crypt foci. The characteristic deep, slit-like serrations were the most consistent histological feature encountered in TSA. The two ‘flat’ TSAs did not have an exophytic protuberant growth.

Figure 1

(A and B) Traditional serrated adenoma (TSA; left) and an adjacent tubulovillous adenoma (TVA). The TVA has stratified penicillate, elongated nuclei with hyperchromatic chromatin. The tubules or glands within the TVA lack serrations and are lined by intestinal-type epithelium. (C–F) An example of a hyperplastic polyp (HP) intimately admixed with a TSA. There is clear intermingling of the two components and even extension of the HP to the surface of the TSA (E and F). (G and H) Sessile serrated adenoma (SSA) areas occurring with a TSA. As with HP, the SSA foci blended with the TSA areas, although a sharp transition between the two can be seen in (H).

The HP and SSA areas occurred within and adjacent to tubules showing typical TSA features. Although TSA was the striking and dominant pattern, the presence of HP and SSA tubules within the TSA suggested that these polyps could more appropriately be regarded as ‘hybrid’ polyps.

A common finding was the presence of single glands showing dual TSA and HP or SSA features with abrupt transitions between the two types of patterns (figure 1F). The SSA areas tended to be sited more towards the base of the TSA, with the TSA being luminally placed.


TSA is the least common type of serrated polyp, has a characteristic morphological appearance and its molecular correlates are being unravelled. 11 ,12 However, a close examination of TSA has revealed that a substantial proportion, more than 50% in the current study, are not ‘pure’ but show elements of HP, SSA and even tubular/tubulovillous adenomas. Rosty et al9 commented on SSA with cytological features of TSA, and the occurrence of tubulovillous adenomas with TSA areas has also been documented.8 Thus, TSA-like histological features have been documented in other polyps, and therefore it is not unexpected that the corollary be true as well (table 1).

Table 1

Precursor lesions and adenomatous polyps coexistent with TSA

The occurrence of TSAs mixed with other polyps was first observed by Iino et al1 who reported that 12 of 29 TSAs were ‘mixed’. The most common polyp accompanying TSA was HP, accounting for 9 of the 12 mixed polyps.1 Later, Lee et al2 demonstrated that HP was associated with 37% of the 35 TSAs in their series. Again, there is no documentation of the spatial arrangement of the HP with TSA. Kim et al3 identified what they dubbed a ‘precursor lesion’ in 35 of 112 TSAs. In this series, the most common precursor polyp associated with TSA was SSA, with 27 SSAs occurring in 35 TSAs.3 From the illustrations in this publication, it would appear that both HPs and SSAs were adjacent to the TSA.3

In an evaluation of SSA and TSA, Torlakovic et al4 describe only 1 coexistent SSA out of a total of 18 TSAs. In fact, they did not encounter any HP areas in their cohort of TSAs.4

In 2013, Kim et al5 observed that 56 (52%) of 107 TSAs contained precursor lesions. Forty-four of the 56 precursor lesions were HPs and 12 were SSAs. Additionally, 19 of 107 TSAs also had tubular adenoma-like areas. They commented on the spatial relationship between TSA and the accompanying precursor lesion.5 The HPs and SSAs were observed not only at the periphery and stalk of the TSA, “but also as intimately admixed forms with a typical TSA component”.5 Furthermore, these authors also found that TSAs associated with HP tended to be located in the distal colon or rectum, while those with SSA as a precursor lesion occurred in the right colon.5 Our study substantiates this lateralisation bias with 27 of 28 TSAs with HP as a precursor lesion being localised to the distal left colon. Furthermore, of the 18 TSAs with SSA, 11 were noted in the right colon. Thus, HP-associated TSAs have a marked predilection for the left colon while SSA-associated TSAs are more likely to occur in the right colon, as suggested by Kim et al.5 In their series, Kim et al5 also found that TSAs with tubular/tubulovillous adenomas had a left-sided predilection.

Wiland et al6 examined 55 left-sided TSAs and noted 13 cases (24%) to have precursor lesions. These consisted of six HPs that were intimately associated with TSA and three SSAs that were adjacent to the TSA. In addition, four precursor polyps were unclassified.6 The serrated precursor lesions in this study were mainly microvesicular HPs with only scattered goblet cells, and no pure goblet cell-rich lesions were encountered.

Bettington et al7 examined 200 TSAs and found that 38% of cases had a precursor lesion: 31% SSA and 7%. If the so-called advanced TSAs (those accompanied by dysplasia and/or carcinoma) were removed, then of the remaining 162 TSAs, 44% were associated with a precursor polyp (36% SSA and 8% HP).7 They required the precursor lesion to have “clear morphological distinction from the TSA component either at the edge or underlying the traditional serrated adenoma”.7 In this study, they noted that HPs tended to underlie the TSA, while SSAs were adjacent to the TSA.7 Our findings were similar in terms of the spatial arrangement of HP and SSA with TSA. In the series of cases reported by Bettington et al,7 all HPs were of the microvesicular variety.

In keeping with the studies by Wiland et al6 and Bettington et al,7 all HPs accompanying TSAs in our series were of the microvesicular type with only some containing occasional goblet cells.

There is enough evidence from this study and preceding studies to confirm that a significant proportion of TSAs contain a precursor lesion in the form of an HP or SSA. In most reported series, HP appears to be the more common precursor lesion associated with TSA. These lesions are often intimately admixed with the TSA, and in this study we have demonstrated glands with hybrid morphology of a precursor lesion and TSA. Adenomatous polyps, on the other hand, occur invariably as separate, discrete lesions in close proximity rather than within the TSA proper. When adenomatous dysplastic glands occur within a TSA, this most likely represents dysplasia occurring in a TSA, and these lesions should not be considered as precursor or coexistent polyps.

While the morphological appearances are what the practising pathologist will be first confronted with, it is important to be aware of the molecular events that underpin the morphological nuances. There is now convincing molecular evidence that two types of TSA exist: one associated with BRAF mutations and the other with KRAS mutations.11 ,12 These molecular events impact on the morphological appearance and associations of TSA. Those associated with BRAF mutations are usually associated with microvesicular HP and SSA (precursor lesions), have so-called serrated dysplasia and may occur in the right colon.11 ,12 The KRAS mutated TSAs are usually attended by conventional adenomatous dysplasia, are closely associated with adenomatous polyps and tend to be left sided.11 ,12

In conclusion, more than 50% of TSAs will be accompanied by another polyp type. These are usually microvesicular HPs and SSAs that are intimately intermingled with the TSA. This occurrence is a manifestation of a molecular event in the form of a BRAF mutation. When adenomatous polyps or indeed conventional adenomatous areas are noted within a TSA, a KRAS mutation is likely to be operative. Documentation of the occurrence of precursor lesions and adenomatous polyps are important for the practising pathologist since it may impact on future management protocols of these lesions.

Take home messages

  • At least 50% of traditional serrated adenomas (TSAs) are associated with areas of hyperplastic polyp (HP), sessile serrated adenoma (SSA) or tubular or tubulovillous adenoma (TA/TVA).

  • TSAs associated with HP and SSA are likely to have BRAF mutations.

  • TSAs with adenomas and conventional dysplasia are likely to have KRAS mutations.

  • HP and SSA are usually intermingled with the typical TSA glands.

  • TAs/TVAs are usually adjacent to TSA.

  • TA/TVA accompanying TSA should be separated from adenomatous dysplasia in a TSA.



  • Correction notice This article has been corrected since it was published Online First. The provenance and peer review statement has been corrected.

  • Handling editor Cheok Soon Lee

  • Contributors All the authors contributed to the planning, conduct and reporting of the work described in the manuscript. RC wrote the paper with contributions from all coauthors.

  • Competing interests None.

  • Provenance and peer review Not commissioned; internally peer reviewed.