According to the World Health Organization (WHO), colorectal serrated lesions are a heterogeneous group of lesions characterized morphologically by a serrated architecture of the crypts, and classified histologically as hyperplastic polyps (HYPs), sessile serrated lesions/polyps (SSL/Ps) (with or without cytological dysplasia), or traditional serrated adenomas (TSAs). The features of serrated polyps of the colorectum have been discussed in comprehensive reviews, such as that by Rex
The distinction between HYP and SSL/P can be difficult histologically, particularly in the rectum, due to the range of normal histological appearances in the rectum including some bifurcated and dilated crypts and a higher ratio of goblet cells to absorptive cells compared with other areas of the colon [
Carcinomas of the serrated pathway are over-represented in studies of interval colorectal carcinomas [
With this in mind, our aim was to retrospectively review a series of serrated colorectal polyps from our institution, focusing on assessing the accuracy of pathological classification, and establishing the clinicopathological features of SSL/Ps and TSAs in our institution.
A search was performed using the laboratory information system (LIS) to identify all colorectal polyps assigned a “serrated adenoma” SNOMED code between January 1st 2004 and May 31st 2016. In our institution, the “serrated adenoma” SNOMED code is assigned to all SSL/Ps and TSAs (HYPs have a separate SNOMED code and were not included). Institutional ethical approval was granted.
All available and suitable haematoxylin and eosin- (H&E-) stained slides were reviewed by one pathologist (AMC), who was blinded to the original diagnosis and to the site in the colon of the polyp. All polyps were evaluated histologically and a diagnosis rendered as follows: HYP, SSL/P (with or without cytological dysplasia), TSA (with or without cytological dysplasia), or other.
In the United Kingdom, the Pathology sections of the British Society of Gastroenterology (BSG) and National Health Service (NHS) Bowel Cancer Screening Programme have approved the terminology developed by Bateman and Shepherd, namely, sessile serrated lesion (SSL), with or without dysplasia [
TSAs are characterized by a constellation of typical histological features, namely, striking granular eosinophilic cytoplasm, luminal serrations, the presence of ectopic crypt foci (ECF), and elongated, pencillate nuclei with evenly dispersed coarse chromatin and small inconspicuous nucleoli [
Histologically, SSL/P with dysplasia is identified by an abrupt transition from ordinary SSL/P to overt dysplasia. The 2010 WHO classification distinguishes two dysplasia patterns, namely, dysplasia resembling that of conventional adenomas and serrated dysplasia [
The main characteristics of conventional adenomatous dysplasia are the predominant location of the dysplastic component on the surface (i.e., “top–down” dysplasia), with preserved non-dysplastic SSL/P at the base of the lesion and complete similarity to the dysplasia of conventional adenomas. There is no serration, and the lesional dysplastic cells are columnar with at least focal goblet cell differentiation, elongated nuclei, and pseudostratification [
An eosinophilic appearance at low power with tightly packed crypts is characteristic of serrated dysplasia [
Similar to SSL/Ps, two forms of dysplasia are associated with TSAs: conventional dysplasia and serrated dysplasia (defined previously) [
As per the WHO, the grade of dysplasia is reported according to a 2-tier system, either low-grade dysplasia or high-grade dysplasia. Low-grade dysplasia is an unequivocal intraepithelial neoplastic condition that must be distinguished from inflammatory or regenerative changes. It is characterized by crowded crypts arranged in parallel, without complexity, back-to-back formation, cribriforming, or budding tubules. The nuclei retain basal orientation, being confined to the bottom half of the cells. Atypical mitoses, loss of polarity, or pleomorphism is not present. The morphological criteria for high-grade dysplasia can be divided into architectural and cytological atypia, with the diagnosis being based on architecture, supplemented by an appropriate cytology. The structural features of high-grade dysplasia are characterized by complex glandular crowding and irregularity, with back-to-back glands. The lesional cells in high-grade dysplasia display loss of cell polarity or nuclear stratification and have markedly enlarged nuclei, with vesicular chromatin and prominent nucleoli. Atypical mitoses are often seen, and prominent apoptosis is frequently present.
Cases that had an alternative diagnosis made following this review to that made by the reporting pathologist were categorised as “discordant cases.” All discordant cases, all SSL/Ps with dysplasia, and all TSAs were reviewed by a second specialized gastrointestinal pathologist (KS), who was blinded to the original diagnosis, to the site in the colon of the polyp and to the opinion of AMC.
Pathology reports were reviewed, and various demographics were extracted (e.g., patient age and gender, site, and microscopic size of polyps). Anatomic sites were based on the original specimen requisitions submitted by endoscopists, with right-sided colonic polyps being defined as those present in the caecum, ascending colon, hepatic flexure and transverse colon, and left-sided colonic polyps being regarded as those found in, and distal to, the splenic flexure (descending colon, sigmoid colon, and rectum). Specimens that were labelled by the clinician as “random colon” and specimens without any specific designation were all categorised as “colon NOS (not otherwise specified).”
Data was recorded and analysed using Microsoft Excel for Mac 2011 Version 14.6.4.
Over a 149-month period, 759 polyps were assigned a “serrated adenoma” SNOMED code. The H&E-stained slides of 664 of these polyps (from 523 patients) were available for review (endoscopic biopsy: 375; polypectomy: 255; endoscopic mucosal resection (EMR): 19; resection NOS: 11; piecemeal excision: 4). These polyps were from 267 male patients (51.1%) and 256 female patients (48.9%), with a median age of 64 years (range, 19-92 years) (Table
Demographics of patients included in this study (664 polyps from 523 patients), all of whom had at least one polyp originally reported as either SSL/P or TSA.
Gender |
Male | 267 (51.1%) |
Female | 256 (48.9%) | |
Age (years) | Median | 64 |
Range | 19–92 |
Abbreviations: SSL/P: sessile serrated lesion/polyp; TSA: traditional serrated adenoma.
All polyps had been reported by 9 general histopathologists; the workload of all of these pathologists comprised a large proportion of gastrointestinal biopsies.
All polyps included in this study were reviewed by one pathologist (AMC), and 41.1% of polyps (273/664) were reviewed by both pathologists (AMC & KS), with a consensus diagnosis assigned.
15.1% (100/664) of all polyps were reclassified, with the majority reclassified from SSL/P to HYP (66/664; 9.9%) (Table
Reclassification of serrated polyps (SSL/Ps and TSAs) following consensus review by two pathologists; (out of 664 reviewed).
Change in classification | Number (%) |
---|---|
Changed from SSL/P to HYP | 66 (9.9%) |
Changed from SSL/P to adenoma | 7 (1.1%) |
Changed from SSL/P to benign polyp NOS | 1 (0.2%) |
Changed from mixed serrated polyp to adenoma | 5 (0.8%) |
Changed from TSA to adenoma | 21 (3.2%) |
Abbreviations: SSL/P(s): sessile serrated lesion(s)/polyp(s); TSA(s): traditional serrated adenoma(s); HYP: hyperplastic polyp; NOS: not otherwise specified.
Hyperplastic polyp, with prolapse effect, characterized by dilated and congested submucosal blood vessels, thickening of the muscularis mucosae, and upward extension from the hypertrophic and splayed muscularis mucosae, with dilated crypts (a, b). Horizontal extension of crypt bases along the muscularis mucosae can be seen, mimicking the architecturally distorted, dilated, and/or horizontally branched crypts of SSL/Ps (c, d).
21 polyps (3.2%) were reclassified from TSA to conventional adenoma, and 7 polyps (1.1%) were ascribed a diagnosis of adenoma in lieu of SSL/P.
Following review of all 664 polyps by one or both pathologists, there were 520 SSL/Ps (78.3%, 520/664) (Figure
Sessile serrated lesion/polyp (SSL/P) without dysplasia (a). SSL/P with low-grade conventional adenomatous dysplasia (b). SSL/P with low-grade serrated dysplasia (c). Traditional serrated adenoma (TSA) without dysplasia (d). TSA with low-grade conventional adenomatous dysplasia (e). TSA with serrated dysplasia (f).
86.7% of SSL/Ps were located in the right side of the colon (Table
Characteristics of SSL/Ps identified over a 12-year period at our institution;
Parameter | All SSL/Ps | SSL/Ps without dysplasia | SSL/Ps with dysplasia (all LGD) |
---|---|---|---|
Number ( |
520 | 485 (93.3%) | 35 (6.7%) |
Age of patient (years) | |||
Median | 63 | 63 | 69 |
Range | 19-84 | 19-84 | 47-83 |
Site ( | |||
Right colon | 451 (86.7%) | 423 (87.2%) | 28 (80%) |
Left colon | 62 (11.9%) | 55 (11.3%) | 7 (20%) |
Colon NOS | 7 (1.4%) | 7 (1.4%) | — |
Size (mm) | |||
Median | 8 | 8 | 10 |
Range | 1-32 | 1-32 | 3-30 |
Size category ( | |||
< 1 cm | 167 (64.5%) | 162 (66.4%) | 5 (33.3%) |
≥1 cm | 92 (35.5%) | 82 (33.6%) | 10 (66.7%) |
Abbreviations: SSL/Ps: sessile serrated lesions/polyps; NOS: not otherwise specified; LGD: low-grade dysplasia.
Following consensus review, 13/520 (2.5%) SSL/Ps were downgraded from SSL/P with dysplasia to SSL/P without dysplasia. Overall, 6.7% of SSL/Ps exhibited dysplasia (35/520), all demonstrating low-grade dysplasia
The majority of SSL/Ps exhibited conventional adenomatous dysplasia (25/35; 71.4%) (Figure
Detection of SSL/Ps peaked in the most recent years reviewed (87.5% reported between 2013 and 2016, inclusive), coinciding with the introduction of “BowelScreen” (the Irish colorectal cancer screening programme,
The number of SSL/Ps detected between January 1st 2004 and May 31st 2016, with detection of SSL/Ps peaking in the most recent years included in this review (87.5% reported between 2013 & 2016, inclusive). This coincided with the introduction of “BowelScreen” (the Irish colorectal cancer screening programme).
85.7% of TSAs were located in the left side of the colon, with the majority being found in the rectum or sigmoid colon (32/40; 80%). 73.1% were large in size (≥1 cm), with a mean size of 18.6 mm (range, 2–60 mm) (Table
Characteristics of TSAs identified over a 12-year period at our institution;
Parameter | All TSAs | TSAs without dysplasia | TSAs with dysplasia |
---|---|---|---|
Number ( |
40 | 13 (32.5%) | 27 (67.5%) |
Age of patient (years) | |||
Median | 67 | 61 | 68 |
Range | 34-92 | 34-86 | 44-92 |
Site ( | |||
Right colon | 4 (10%) | 2 (15.4%) | 2 (7.4%) |
Left colon | 34 (85%) | 9 (69.2%) | 25 (92.6%) |
Colon NOS | 2 (5%) | 2 (15.4%) | — |
Size (mm) | |||
Median | 13 | 10 | 23.2 |
Range | 2-60 | 2-20 | 4-60 |
Size category ( | |||
<1 cm | 7 (26.9%) | 4 (44.4%) | 3 (17.6%) |
≥1 cm | 19 (73.1%) | 5 (55.6%) | 14 (82.4%) |
Dysplasia grade ( | |||
Low | 24 (60%) | N/A | 24 (88.9%) |
High | 3 (7.5%) | N/A | 3 (11.1%) |
Type of dysplasia ( | |||
Conventional | 26 (65%) | N/A | 26 (96.3%) |
Serrated | 1 (2.5%) | N/A | 1 (3.7%) |
Abbreviations: TSAs: traditional serrated adenomas; NOS: not otherwise specified; N/A: not applicable.
Table
Characteristics of SSL/Ps & TSAs identified over a 12-year period at our institution.
Parameter | All SSL/Ps | All TSAs |
---|---|---|
Number ( |
520 | 40 |
Age of patient (years) | ||
Median | 63 | 67 |
Range | 19-84 | 34-92 |
Site ( | ||
Right colon | 451 (86.7%) | 4 (10%) |
Left colon | 62 (11.9%) | 34 (85%) |
Colon NOS | 7 (1.4%) | 2 (5%) |
Size (mm) | ||
Median | 8 | 13 |
Range | 1-32 | 2-60 |
Size category ( | ||
<1 cm | 167 (64.5%) | 7 (26.9%) |
≥1 cm | 92 (35.5%) | 19 (73.1%) |
Dysplasia ( | ||
Low | 35 (6.7%) | 24 (60%) |
High | — | 3 (7.5%) |
Type of dysplasia ( | ||
Conventional | 25 (71.4%) | 26 (65%) |
Serrated | 3 (8.6%) | 1 (2.5%) |
Mixed | 7 (20%) | — |
Abbreviations: SSL/Ps: sessile serrated lesions/polyps; TSAs: traditional serrated adenomas; NOS: not otherwise specified.
We reviewed a large series of serrated polyps over a 12-year period, focusing on the histological diagnosis of SSL/Ps and TSAs, with and without dysplasia. Detection of SSL/Ps peaked in the most recent years included in this review (87.5% reported between 2013 & 2016, inclusive). This coincided with the introduction of “BowelScreen” (the FIT-based Irish colorectal cancer screening programme) and improved recognition of this entity by histopathologists. Similarly, Chetty and colleagues documented increasing awareness in their institution of SSL/Ps as an entity over a 4-year period [
Furthermore, this study clearly shows how challenging it can be to distinguish SSL/Ps from HYPs, as there are often only subtle differences, with 66/664 polyps in our cohort (9.9%) being reclassified from SSL/P without dysplasia to HYP following consensus review. In a related study, Gill and colleagues reviewed a large series of right-sided lesions originally diagnosed as HYPs and recategorised 30–64% of HYPs over a 4-year period to SSL/Ps, emphasising again how difficult the distinction between HYP and SSL/P can be [
Identification of architecturally distorted, dilated, and/or horizontally branched crypts (“L,” “boot,” or “anchor”-shaped crypts), in association with excessive/hyper-serration in the basal half of crypts, is required for a diagnosis of SSL/P (Table
Histological features of hyperplastic polyps with prolapse, sessile serrated lesions/polyps, and traditional serrated adenomas [
HYP with prolapse effect | HYP: |
SSL/P | (i) At least one unequivocal architecturally distorted, dilated, &/or horizontally branched crypt (“L,” “boot,” or “anchor”-shaped crypt) |
TSA | (i) Striking granular eosinophilic cytoplasm |
In contrast to SSL/Ps, HYPs are characterized by simple, elongated crypts with a serrated structure in the upper half of the crypts, with some proliferation in the basal (non-serrated) part of the crypts (Table
It is known that SSL/Ps with dysplasia are precursors for interval colorectal carcinomas, and that these lesions are rapidly progressive, difficult to detect endoscopically, commonly incompletely resected, and occasionally misdiagnosed histologically [
In a review of SSL/Ps from 2139 patients, Lash
The concept of dysplasia in TSAs is controversial. Many pathologists consider TSAs to be inherently dysplastic and routinely report low-grade dysplasia in TSAs mainly based on elongated, pencillate nuclei [
There are some limitations to our study. We did not retrieve and review all polyps that were classified as HYPs, to assess how many would be amended to SSL/P on review. We chose not to interrogate this area, as this has been previously studied and published by other authors [
As the malignant potential of SSL/Ps and TSAs has been clearly established, it is important that serrated polyps are identified and correctly classified histologically. It is therefore essential for all pathologists to strictly adhere to diagnostic criteria and to be aware of pitfalls in diagnosis. In particular, as has been established in the literature, failure to identify serrated polyps with dysplasia may result in inadequate surveillance and thus increases the risk of interval colorectal carcinoma.
All of the data used to support the findings of this study are included within the article.
The authors declare that they have no conflicts of interest.