Guidance

Bowel cancer screening: pathology guidance on reporting lesions

Updated 31 May 2021

This document supports consistent practice in pathology in the NHS Bowel Cancer Screening Programme (BCSP), as well as developing the evidence base for future recommendations regarding routine practice.

The NHS BCSP pathology clinical professional group (CPG) engages with pathologists across the UK and Ireland. The standards mentioned within this document refer specifically to the English programme. Different arrangements may apply elsewhere but most aspects of this guidance will apply to all.

These guidelines are consistent with the guidance and dataset produced by the Royal College of Pathologists (UK) (RCPath) for reporting colorectal cancer (CRC). They are also consistent with the quality assurance guidelines commissioned by the European Union in 2010.

The main updates to the previous version of this publication are:

• adding reference to the 2019 British Society of Gastroenterology, Public Health England and Association of Coloproctology of Great Britain and Ireland guidelines for post-polypectomy and post-cancer-resection surveillance
• adding further detail on standards which apply to pathologists reporting BCSP specimens
• adding further advice on assessment of lesion surgical margin – distinguishing margin involvement by dysplasia from margin involvement by adenocarcinoma
• updating advice on epithelial misplacement (EM) within adenomas versus adenocarcinoma
• updating the definition of sessile serrated lesions and serrated polyposis
• reinforcing the NICE recommendation for screening for Lynch Syndrome
• removing the section on villousness, now that this is not a criterion that affects the nature of post-polypectomy surveillance

This document helps pathologists report specimens which derive from BCSP practice. The vast majority of such specimens will be endoscopic in origin, either polypectomies or mucosal biopsies. A minority will be other forms of local excision specimen, such as endoscopic mucosal resection (EMR), or surgical resection specimens for colorectal neoplasms deemed unsuitable for local resection.

Most BCSP colonoscopy specimens are diagnosed as adenomatous or serrated polyps/lesions, with adenocarcinoma and inflammatory conditions each making up fewer than 10% of specimens in the BCSP setting. Within screen-detected CRCs, a higher proportion are early stage (stage pT1) compared to symptomatic CRCs, and frequently these can be treated by polypectomy or a form of local excision only.

This guidance places emphasis on reporting non-malignant lesions and on polypectomy or other local excision specimens which contain adenocarcinoma. It is beyond the scope of this document to discuss reporting of surgical resection specimens or colorectal inflammatory conditions in detail. Refer to the guidance and dataset produced by the Royal College of Pathologists (UK), major gastrointestinal pathology textbooks and others for more information on this.

Standardisation of pathology reporting is emphasised. This includes:

• which data items are reported
• the approach to data items
• the vocabulary used for data items

This guidance aims to standardise BCSP pathology reporting further, and to facilitate regional, national and international comparisons for audit and research purposes. We strongly encourage the use of standardised agreed datasets and electronic reporting systems, to aid collation of data by appropriate registries.

The British Society of Gastroenterology (BSG), Public Health England (PHE) and Association of Coloproctology of Great Britain and Ireland (ACGBI) recently published guidelines for post-polypectomy and post-cancer-resection surveillance and a figure summarising this new advice is reproduced below (see the polyp surveillance guidelines).

1. Standards

All people working within the BCSP need to be aware of the standards for the programme. The standard that relates to pathology turnaround is BCSP-S09: referral: pathology turnaround time.

1.1 Reporting requirements

A pathologist who wishes to start reporting BCSP cases should typically be in a substantive consultant post. However, the following individuals may also be considered for the role:

• consultant pathologists in long-term locum posts
• pathologists in non-consultant career grade posts
• pathologists in stage D (or equivalent) of training
• advanced practitioners holding the Advanced Specialist Diploma in Histopathology Reporting – Gastrointestinal Tract Pathology

The local cellular pathology department and the local lead BCSP pathologist decide whether to accept new members to the local BCSP pathology reporting team. The new pathologist must notify the clinical director and programme manager of the local screening programme and the regional screening quality assurance service (SQAS) team.

The pathologist or advanced practitioner must:

• register for participation in the national BCSP diagnostic external quality assurance (EQA) scheme. It is not required that pathologists or advanced practitioners new to BCSP reporting demonstrate satisfactory performance in this scheme before they begin reporting BCSP-derived specimens
• agree to attend at least one approved BCSP educational event during each 3 year period
• be aware of national guidance on BCSP pathology reporting
• be aware of local standard operating procedures (SOPs) related to the BCSP, such as managing difficult cases or changing a diagnosis after review
• double-report all cases of pT1 cancer with a second BCSP-accredited pathologist, who should be in a substantive consultant position

The local BCSP pathology lead will routinely audit the reporting of the whole team and should be careful to ensure that the reporting practice of new individuals falls within guidelines. They must be willing to mentor new members of the BSCP pathology reporting team if this is required.

Pathologists and advanced practitioners new to BCSP pathology reporting should:

• make sure they become known to other key members of the local BCSP team, within and beyond the pathology department such as the clinical director, programme manager and lead specialist screening practitioner (LSSP)
• be familiar with the local mechanisms for generating BCSP-related pathology reports, which for polyps and lesions will usually be in a proforma style
• become involved with audit projects related to BCSP activity, including audits of local BCSP pathology reporting
• lower the threshold for case discussion with a colleague while they gain experience in this area
• expect to designate around 5% of adenomas as showing high-grade dysplasia
• note there is currently no specified minimum number of BCSP-derived cases they must report before being accepted by the local lead pathologist to report cases independently

Double-reporting of cases of high-grade dysplasia is not currently required. However, as with other unusual and difficult cases, cases where the grade of dysplasia is uncertain should be discussed with a second BCSP-accredited pathologist, who should be in a substantive consultant post.

1.2 Standards relating to BCSP pathology

The standards which relate to BCSP are:

• to report BCSP cases, reporters must fulfil the criteria specified in the reporting requirements
• reporters must participate in local quality assurance (QA) visits and audit their own reporting practice alongside colleagues
• reporters must complete either the screening programme proforma or its computerised version for each BCSP case and each lesion reported. These should be returned to the screening centre who will enter the pathology data onto the central bowel cancer screening system (BCSS) (a list of the latest datasets to include in a proforma is available)
• reporters may also wish to provide a free text report directly to the clinician, in addition to the proforma
• turnaround time of pathology reports should allow patients who have had lesions removed at endoscopy to be managed appropriately and given a timely appointment to be seen at a follow-up clinic if required (current QA standards relating to turnaround times should be adhered to and ‘day zero’ should be counted as the date of receipt of sample in the laboratory and the end point should be counted as the day the report is authorised)
• interim reports are encouraged if cases are referred for second opinions
• around 5% of adenomas reported in the setting of a faecal occult blood test (FOBt) based screening programme are expected to be classified as having high grade dysplasia
• appropriate SNOMED codes should be applied to all cases
• cancers should be reported according to the latest Royal College of Pathologists (UK) and these BCSP guidelines

Double reporting of all pT1 cancers is required to minimise overdiagnosis of adenocarcinoma. Both reporters should be BCSP registered and named on the pathology report. If there is any doubt about a diagnosis of cancer, then referral is required for a further opinion. It is recommended that further opinions are sought:

• at local level first
• at regional level. This activity is not included in the role of the PHE pathology professional clinical advisor (PCA) and will require a funded referral to an experienced gastrointestinal pathologist within the region (the same happens in the symptomatic service)
• from the Expert Board, if the diagnosis is still in doubt. The opinions from the Expert Board will be funded by the referring Trust

These standards should be monitored by regular departmental audit, for all BCSP pathologists involved.

1.3 Standards relating to colorectal cancer pathology

These standards relate to the Royal College of Pathology dataset:

• multidisciplinary teams and/or pathology departments should audit their CRC resection pathology reports at regular intervals. Three standards are recommended and should be audited for each BCSP pathologist. These should be evaluated on a series of at least 50 resection specimens from symptomatic patients (not screen-detected cancers) who have not undergone pre-operative therapy
• the number of lymph nodes examined should be as high as possible with at least a median of 12 per case
• the frequency of serosal involvement should be at least 20% for colonic cancers
• the frequency of venous invasion, including intramural (submucosal and intramuscular) and extramural, should be at least 30%

1.4 Endoscopy requirements

It is essential that the histopathology request form accompanying any BCSP specimen includes all the relevant information to allow thorough and accurate pathology reporting. The responsible endoscopist must include:

• patient demographic details, including NHS and Health and Care number or another unique identifier
• date of endoscopy
• relevant clinical information
• endoscopic findings

The form should clearly indicate, with a suitable label or stamp, that the specimen has resulted from BCSP endoscopy and must therefore be reported by a BCSP pathologist and met the standards for turnaround in the programme.

A copy of the endoscopy report should be submitted with the request form, and if possible, accompanied by endoscopic photographs of any lesions identified. As a minimum, the request form should specify:

• the precise number and respective sites of any lesion submitted for histology
• any clinical suspicion of malignancy at endoscopy
• endoscopic lesion size(s)

Each individual lesion should, if possible, be submitted in separate specimen pots and appropriately labelled. The label should clearly state if each lesion has been endoscopically removed or only biopsied/partially removed, and the number of lesions represented in each specimen, to ensure accurate counting if multiple lesions are submitted in one specimen.

2. Specimen handling

BCSP specimens should be handled according to the same routine laboratory procedures as non-BCSP endoscopic or surgical practice. Most specimens received in the laboratory are in formalin. Local excision specimens may be pinned to cork and orientated with sutures. After adequate fixation (for at least 4 hours but typically overnight), endoscopic biopsy specimens are transferred to cassettes for processing, with the number of mucosal fragments in each specimen recorded.

2.1 Measurement

The largest dimension of polypectomy/lesion resection specimens should be macroscopically measured in millimetres (mm). Rounding up or down (terminal digit preference) should be avoided. These measurements should be recorded in a consistent way, for example ‘length x breadth x height’. Any stalk present should also be indicated, with its measurement. This conveys the maximum macroscopic information to the BCSP reporter.

Lesion size is clinically significant, notably at around the 5mm and 10mm diameter cut-off. It is important that lesions received intact are sectioned along their longer axis if possible. This means that the maximum dimension can be presented on the glass slide for microscopic measurement. If the largest axis within an intact polypectomy specimen is not presented on the glass slide, it is important to convey this information to the reporter by the dissector. This avoids potential underestimation of overall adenoma size and is particularly important for larger lesions.

2.2 Sectioning

Sectioning should be perpendicular to the lesion base excision margin if this is identifiable. Lesion bases do not need to be inked routinely as the excision margin is usually readily identifiable microscopically, often through the presence of diathermy artefact.

Lesions with a narrow stalk should be trimmed to keep the stalk intact and orientated to allow clear microscopic visualisation of the lesion base margin, through multiple levels if necessary.

Lesions with a broader stalk (‘semi-pedunculated’) or sessile lesions should be serially sectioned at 3mm intervals, perpendicular to the base margin if this is identifiable. All tissue should be processed for histological evaluation.

Good communication between the dissector and reporting pathologist is vital. It enables accurate assessment at microscopy of the maximum dimension of any adenomatous component of the lesion (see guidance on sizing classical adenomas). Adenoma measurements are used to inform individual risk stratification and recommend the surveillance interval for any follow-up colonoscopy.

2.3 Margins

If there is any suspicion of malignancy within a local excision specimen at endoscopic procurement or at dissection, any identifiable resection margins (mucosal and/or deep) should be painted to ensure subsequent microscopic identification. If any mucosal lesion has a surrounding mucosal margin of normal tissue that macroscopically measures less than 3mm, this margin should be examined perpendicularly by taking sections of the margin at right angles from a thicker slice. Macroscopic images are helpful to illustrate margin status and block sampling.

3. Reporting non-invasive lesions

3.1 Data

Unpublished histopathology data derived from 816,323 non-invasive lesions diagnosed at screening colonoscopy since the inception of the English programme revealed that:

• just over 70% were classical adenomas (57% tubular adenoma, 14% tubulovillous adenoma and <1% villous adenoma)
• 21% were reported as sessile serrated lesions, although this percentage is likely to grow as this entity was likely under-recognised in the early years of the programme
• all other diagnoses comprise <10% of all lesions (see frequencies of common histopathological diagnoses below)

3.2 Frequencies of common histopathological diagnoses from non-invasive lesions

This data relates to diagnoses from non-invasive lesions detected during screening colonoscopy since the inception of the English Bowel Cancer Screening Programme in April 2006 until March 2020.

Polyp type	Polyp subtype / architecture	Percent
Adenoma	Tubular adenoma	56.68
	Villous adenoma	0.70
	Tubulovillous adenoma	14.17
	No polyp architecture / sub type reported	0.24
Serrated lesion	Hyperplastic	18.17
	Mixed HP/adenoma	0.09
	Mixed polyp	0.14
	Serrated adenoma	0.52
	Sessile serrated lesion	1.43
	Sessile serrated lesion with dysplasia	0.09
	Traditional serrated adenoma	0.15
	No polyp architecture / sub type reported	0.01
Other polyp	N/A	2.12
Not polyp	N/A	5.50

3.3 Classical adenomas

The 2019 World Health Organisation (WHO) classification of premalignant epithelial lesions or adenomatous lesions is recommended. Adenomas must show dysplasia (intraepithelial neoplasia). Classical adenomas are divided into tubular, tubulovillous and villous types. Demarcation between them is based on the relative proportions of tubular and villous components.

The term ‘advanced adenoma’ (mainly employed in United States screening literature) most commonly describes any adenoma that is ≥10mm, has a villous component >25% (villous or tubulovillous) or has high grade dysplasia. This term should now be replaced by the phrase ‘advanced colorectal polyp’, as defined in the 2019 BSG/PHE/ACGBI guidelines. With respect to adenomas, those measuring ≥10 mm or with high grade dysplasia fulfil the new criteria for advanced colorectal polyp.

Sizing

Accurate lesion sizing requires careful macroscopic and microscopic correlation. The current available evidence indicates that the size of adenomas based on pathological assessment is more accurate and reliable than the endoscopy size. The pathology size of adenomas must be used for clinical decision making if both sizes are available (as described in research published in 2013a, 2013b and 2016).

Reporting pathologists should aim to provide the single maximum dimension of the adenomatous component of the adenoma lesion. This will often equate to the maximum macroscopic dimension of the formalin-fixed lesion, if all or almost all of the lesion is adenomatous.

Some large or unusually configured adenomas may preclude representation on the glass slide of the largest lesion axis. If microscopy demonstrates that such entire lesions are adenomatous, then the largest macroscopic dimension of the lesion, after formalin fixation, can be safely recorded as the maximum adenoma diameter. If the lesion includes a non-adenomatous component, record the maximum microscopic dimension.

Examples of measuring adenomas

This image show a bisected lesion resection specimen in which only part of the specimen is adenomatous. In such cases, the maximum microscopic dimension of the adenomatous component should be recorded as the adenoma size. In this case, the adenoma size (arrow) is 4mm, compared to 11mm maximum macroscopic dimension of the fixed lesion resection specimen.

Occasionally, the maximum microscopic dimension of an adenoma may be a vertical measurement. Correlation with macroscopic assessment of overall lesion dimensions is important and care should be taken not to include any non-adenomatous lesion stalk or non-adenomatous epithelial misplacement in the adenoma measurement. In this case, the adenoma size (arrow) is 6mm, compared to 10mm maximum (vertical) dimension of the entire lesion resection specimen.

If measuring on the glass slide, the BCSP graticule or an ISO accredited graticule should be used. Avoid terminal digit preference bias – measure in size to the nearest millimetre, not rounded up or down to the nearest 5 or 10mm.

Recording sizes of individual piecemeal fragments received is potentially confusing and not recommended. If a lesion is received piecemeal, or if a non-excision biopsy only is received, pathology size is not assessable and should be recorded as such and the endoscopic size should be used to determine follow-up.

Endoscopic sizing may take place in vivo, assisted by size comparison to open biopsy forceps, or after excision from within the specimen container. Auditing of the endoscopy size against the pathology size is an endoscopic quality measurement. Care should be taken not to record the endoscopy size under the pathology size.

Grading dysplasia

A 2-tier stratification of adenomatous dysplasia is now widely accepted. The terms used are ‘low grade’ and ‘high grade’. This provides a uniform system for integrating global histopathology grading data (particularly Western and Japanese). You should not use the terms mild, moderate or severe dysplasia.

The changes of high grade dysplasia should usually involve more than just one or 2 glands (except in tiny biopsies of lesions). This means there is enough to be identified at low power examination. Caution should be exercised in overinterpreting isolated surface changes that may be due to trauma, erosion or prolapse. Similarly, crush artefact should not be interpreted as glandular complexity.

High grade dysplasia is primarily diagnosed on architecture, supplemented by cytology. This means its presence is nearly always suspected by the appearance under low power of complex architectural abnormalities in structures whose epithelium looks thick, blue, disorganised and ‘dirty’. The architectural features of high grade dysplasia are:

• complex glandular crowding and irregularity
• a cribriform appearance and ‘back to back’ glands
• prominent intraluminal papillary tufting

Ectopic crypts are particularly characteristic of traditional serrated adenomas but may also occur in ‘classical’ adenomas. They impart a more complex appearance to the lesion but do not in themselves represent high grade dysplasia. Although many of these features often coexist in high grade dysplasia, individually they are neither necessary nor usually sufficient for a diagnosis of high-grade dysplasia. You should exercise caution in using just one feature to make a diagnosis.

These features may occasionally occur in lower grades of dysplasia, so it is also necessary to scrutinise the cytological features for signs of high grade dysplasia. The cytological features are:

• loss of cell polarity or nuclear stratification - to the extent that the nuclei are approximately equally, though haphazardly, distributed within all 3 thirds of the height of the epithelium
• markedly enlarged nuclei, often with a dispersed chromatin pattern and a prominent nucleolus
• atypical mitotic figures
• prominent apoptosis; giving the epithelium of the lesion a ‘dirty’ appearance

Again, these features usually coexist in high grade dysplasia, and caution must be exercised in using just one feature to make the diagnosis. To allow classification as high grade dysplasia, these cytological features should occur in a background of complex architectural abnormality.

Marked loss of polarity and nuclear stratification sometimes occur on the surface of small, architecturally regular, tubular adenomas that otherwise have features of low grade dysplasia (probably as a result of trauma or ulceration) and must not be used to classify a lesion as high grade.

The only potential exception to the rule is when the specimen consists of a small biopsy from the surface of a lesion – when there is insufficient tissue to assess the architecture properly. It is then permissible to regard marked cytological abnormalities alone as high grade dysplasia. This will usually lead to excision of the whole lesion so it will become possible to assess the whole lesion properly.

Examples of low and high grade dysplasia

Picture A shows an adenoma with predominantly villous architecture; even on low power magnification, the lack of any complex architectural features suggests low grade dysplasia throughout.

Higher magnification of the same adenoma in picture B confirms architectural and cytological features of low grade dysplasia.

Pictures C and D depict 2 separate adenomas, each demonstrating focal high grade dysplasia, with complex glandular crowding and irregularity, a cribriform appearance and luminal necrosis. Higher magnification to confirm the corresponding cytological features is rarely necessary given these low power appearances.

Assessment of lesion resection margin

The endoscopist should remove all adenomatous lesions completely, to prevent progression to adenocarcinoma. Ideally, removal should be in one piece to optimise assessment of the lesion. Larger lesions may require removal in multiple steps, resulting in a piecemeal lesion resection specimen.

The endoscopic impression of completeness of excision should always be conveyed on the pathology request form. The endoscopist’s impression is typically more valuable than that of the pathologist, who can comment only on any involvement of a diathermied margin by dysplasia (and specify high or low grade) within an intact lesion resection specimen. This does not equate to incomplete excision as diathermy may destroy a zone (up to several millimetres) of normal tissue, creating the impression of incomplete excision. Therefore, the phrase ‘involvement of diathermied margin by dysplasia’ is preferred when diathermy artefact is seen.

The vast majority of small specimens are not oriented and residual margin involvement by dysplasia is not assessable. Similarly, no useful comment can typically be made on margin involvement by dysplasia within a piecemeal lesion resection specimen unless a fragment specified as the margin has been clearly indicated.

Epithelial misplacement

Epithelial misplacement of adenomatous epithelium into the submucosa of a lesion is a well-recognised phenomenon, particularly common in large prolapsing lesions in the sigmoid colon. Distinction of epithelial misplacement, or so-called ‘pseudoinvasion’, from invasive adenocarcinoma is perhaps the single most difficult area of pathological diagnostic practice in BCSP pathology assessment. Large sigmoid colonic lesions are particularly prone to inflammation and ulceration, features which tend to enhance the dysplastic changes. When associated with epithelial misplacement, the potential for misdiagnosis of early adenocarcinoma (stage pT1) increases and the overall diagnostic difficulties become greater.

Double reporting by BCSP pathologists of all cases of endoscopically resected stage pT1 CRC is now mandatory in BCSP pathology practice. This is because misdiagnosis of adenocarcinoma may lead to an inappropriate major surgical intervention. Despite increased awareness of the problem and enhanced recognition of the classical features of epithelial misplacement versus adenocarcinoma (see guidance on differentiating epithelial misplacement from invasive adenocarcinoma), many cases are highly problematic as the features presented are overlapping.

Examples of epithelial misplacement versus early adenocarcinoma

Low power views of 4 lesions are presented to illustrate the important features evident at this magnification.

Picture A shows classical features of epithelial misplacement with a lobulated glandular arrangement, surrounding lamina propria rather than desmoplastic stroma, dysplastic change in continuity with surface adenoma and vascular congestion, haemorrhage and mucin extravasation resulting from gland rupture.

Picture B shows well differentiated adenocarcinoma featuring haphazardly infiltrating glands eliciting a desmoplastic stromal reaction.

Picture C shows a focal area of submucosal invasion within a tubulovillous adenoma. Note the thin band of intact muscularis mucosae towards the lateral aspects but central destruction of muscularis mucosae associated with small infiltrating glands.

Picture D shows a difficult case featuring glands with high grade dysplasia located within the submucosa of the lesion head. A lobular architecture is retained and there is probable gland rupture with adjacent abscess formation favouring benignity, but confident distinction from an early adenocarcinoma is not possible.

Lesions demonstrating the classical misplacement features of lobulated glands, surrounding lamina propria, haemosiderin deposition and prolapse-type muscular stroma usually provide no diagnostic problems. However, many cases show only some of these appearances, often accompanied by more concerning morphological features, such as apparent desmoplastic stroma, haphazard glandular ‘pseudoinfiltrative’ pattern, glandular angulation or single cells lying within stroma.

The most difficult cases often include misplaced glands showing high grade dysplasia, associated with gland rupture, mucin extravasation and secondary inflammatory changes. The correct diagnosis in such cases may be almost impossible to determine. Previous biopsy or partial lesion resection may also induce mucosal necrosis, ulceration with regenerative atypia and/or a desmoplastic stromal reaction, strongly mimicking malignancy. Any previous intervention should always be conveyed on the pathology specimen request form.

BCSP pathologists can access opinions from an experienced gastrointestinal pathologist within their region and, if required, an ‘Expert Board’ whereby these difficult diagnostic problems of epithelial misplacement versus polyp cancer are assessed by 3 specialist gastrointestinal pathologists. The current Board coordinator is Dr Adrian C. Bateman in Southampton (address given at the front of this document) and all such cases should be sent initially to him. The results for the first 5 years of these assessments have been published. These referrals are charged for, as per the usual NHS referral pathways.

Overall judgment must be based primarily on appropriate clinical management. If surgical intervention would not be warranted in a given case, regardless of a diagnosis of epithelial misplacement or adenocarcinoma (stage pT1 with no adverse features - poor differentiation, lymphovascular invasion and involvement of the surgical margin), it is considered prudent to reserve a diagnosis of malignancy for those cases with considerable certainty in the diagnosis.

Whilst a diagnosis of “suspicious of adenocarcinoma” can be entirely appropriate for an endoscopic biopsy specimen, this should be avoided as a final diagnosis when presented with a resection specimen. If there is diagnostic uncertainty, further opinion should be sought.

3.4 Differentiating epithelial misplacement from invasive adenocarcinoma

The pathological features used for differentiating between epithelial misplacement (EM) and adenocarcinoma include:

• epithelial ‘differentiation’ – in EM, it is usually similar to that of the surface adenomatous component, while in adenocarcinoma, it is variable but commonly different to the surface adenomatous component
• lamina propria accompaniment – it is very characteristic of EM but may be lacking when there is secondary inflammation and epithelial destruction, while in adenocarcinoma, it is usually absent
• accompaniment by non-adenomatous epithelium – in EM, it is characteristically seen when the EM is due to previous intervention, while in adenocarcinoma, it is absent
• haemosiderin deposition – in EM, it is characteristic and indicative of previous necrosis and/or haemorrhage, as may commonly occur due to previous torsion, while in adenocarcinoma, it is usually absent
• mucosal prolapse changes – in EM, it is often present, while in adenocarcinoma, it is usually absent
• mucus cysts – in EM, it is characteristic and commonly partly lined by adenomatous epithelium. Probably represents epithelial misplacement that has become ‘detached’ from the more superficial components, while in adenocarcinoma, it is usually only present in mucinous adenocarcinoma
• continuity with surface adenomatous component – in EM, it is characteristic but often only appreciated in multiple levels and/or 3D reconstruction studies, while in adenocarcinoma, it is usually absent but some cases do show continuity, even in 3D reconstruction studies
• involvement of muscularis propria – in EM, it is usually absent but can be seen very rarely, especially after previous intervention, while in adenocarcinoma, it is present if at least pT2
• budding – in EM, it is usually absent, but a similar phenomenon can be seen as a result of epithelial destruction and/or inflammation, while in adenocarcinoma, it is often present
• desmoplastic reaction to glands – in EM, it is usually absent but fibromuscular stromal proliferation can accompany EM, while in adenocarcinoma, it is usually present
• lymphatic and/or vascular invasion – in EM, it is absent, although exceptionally mimics can occur in the absence of adenocarcinoma, while in adenocarcinoma, it is usually diagnostic of cancer

3.5 Serrated lesions

The histopathological assessment of serrated colorectal lesions can be a problematic area. The terminology used to describe lesions within this spectrum is variable and the suggested minimum diagnostic criteria for some lesions differ between authorities. One of the most difficult areas is the nomenclature of, and diagnostic criteria for, sessile serrated lesions (SSLs; also referred to as sessile serrated polyps, sessile serrated adenomas or sessile serrated adenoma/polyps).

This is particularly important as these lesions, while bearing histological resemblance to hyperplastic polyps (HP), may be associated with the earlier development of epithelial dysplasia and adenocarcinoma. This topic is the subject of a recent review, an edited version of which forms the basis of the advice given within this document.

Lesions with serrated morphology should be given one of the following names according to their morphological features:

• hyperplastic polyp (HP)
• sessile serrated lesion (SSL)
• sessile serrated lesion with dysplasia
• traditional serrated adenoma (TSA)
• mixed polyp

Hyperplastic polyps (HPs)

Hyperplastic polyps (HPs) are small serrated lesions showing no features that would allow categorisation as SSL and no evidence of dysplasia. We use the term ‘dysplasia’ in this context to refer to the morphological appearances of epithelial neoplasia within the mucosa of the colon and rectum. For example, the epithelial changes recognised by histopathologists as characteristic of ‘classical’ adenomas.

HPs are usually small (less than 5mm diameter) and may occur anywhere within the colon and the rectum. However, they are particularly common in the distal colon and rectum and are often multiple. There are 2 common morphological forms of HPs.

Microvesicular HPs demonstrate vesicular mucin-containing epithelial cells and goblet cells are decreased in number compared with normal crypts. Microvesicular HPs tend to demonstrate BRAF mutations.

Goblet cell-rich HPs account for about a third of all hyperplastic polyps. They almost always occur in the left colon and rectum. Unsurprisingly, they show numerous goblet cells. KRAS mutations are predominant in the goblet cell-rich variant.

The cancer risk associated with small HPs is very low. Under the new BSG/PHE/ACPGBI guidelines, all hyperplastic polyps are considered premalignant polyps with the exception of those rectal hyperplastic polyps measuring ≤ 5 mm.

Sessile serrated lesions (SSLs)

SSLs are synonymous with sessile serrated adenomas/polyps and SSL is now the accepted terminology globally. When ‘pure’, these lesions show no evidence of adenomatous dysplasia but, in comparison to HPs, they contain one or more of the following histopathological features:

• irregular distribution of crypts
• dilatation of crypt bases
• serration present at crypt bases
• branched crypts
• horizontal extension of crypt bases
• herniation of crypts through the muscularis mucosae (see examples)

SSLs also show ‘dysmaturation’. This is a disorganised arrangement of proliferating cells and goblet cells within the lower half of the crypts, with subtle cytological abnormalities that are more pronounced than in hyperplastic polyps. Some pathologists believe that ‘dysmaturation’ represents a form of dysplasia but these changes are distinct from those that are recognised as dysplasia within ‘classical’ adenomas.

Opinion in the UK is that the term ‘adenoma’ is inappropriate for a lesion in which morphological dysplasia is not demonstrable and hence we would not use the term sessile serrated adenoma for such a lesion. Use of the 2019 WHO diagnostic criteria is recommended – the presence of a single characteristic crypt, as a minimum diagnostic requirement.

Examples of histology of sessile serrated lesions

These images contain examples of:

• A – branched crypts
• B – pronounced serration
• C – horizontal extension of crypt bases
• D – surface villousness
• E – crypt dilatation
• F – herniation of crypts through muscularis mucosae
• G and H – crypt bases showing ‘dysmaturation’, evidenced by nuclear enlargement, hyperchromasia, stratification and apposition of goblet and non-goblet cells

HPs and SSLs share many morphological features, and both are associated with mutations in the BRAF gene. It is therefore possible that they represent part of the same ‘spectrum’ of serrated lesions, with small HPs at one end and larger (10mm+) SSLs at the other. The condition originally termed ‘hyperplastic polyposis’ is now termed ‘serrated polyposis syndrome’ after the morphological features of SSLs were identified in this setting. In this model, it is unclear why tiny, often distal, HPs do not appear to be associated with a significant risk of CRC development, while larger lesions with the features of SSLs can be associated with the development of dysplasia and adenocarcinoma.

The WHO classification of 2019 proposes new criteria for diagnosing serrated polyposis syndrome, either ≥ 5 serrated lesions/polyps proximal to the rectum, all ≥ 5 mm in size, with two ≥ 10 mm in size (criterion 1) or > 20 serrated lesions/polyps of any size distributed throughout the colorectum, with ≥ 5proximal to the rectum (criterion 2). Importantly, these lesion counts can be cumulative over multiple endoscopies.

Sessile serrated lesions with dysplasia

SSLs may contain a focus of dysplasia as defined above. This dysplasia may be low or high grade in degree and is almost invariably present within a lesion that shows background features of a sessile serrated lesion without dysplasia. It has been suggested that these lesions may be associated with faster progression to adenocarcinoma than ‘classical’ adenomas. The term ‘mixed polyp’ has previously been used to describe this lesion (see below). Dysplasia arising in the context of an SSL commonly shows loss of immunohistochemical expression of the DNA mismatch repair enzyme hMLH-1, as part of a genetic signature that includes BRAF mutation and widespread DNA methylation (the ‘CpG island methylator phenotype’ – CIMP). Most so-called mixed polyps, especially in the right colon, are likely to represent various stages in the serrated neoplasia pathway, namely the presence of an SSL within which dysplasia has arisen.

Traditional serrated adenomas (TSAs)

TSAs are distinct from SSLs. They most commonly occur in the distal colon and rectum and may have a pronounced villiform or even filiform architecture. They are characterised by the presence of dysplasia (often subtle) together with a variable proportion of the lesion showing eosinophilic cytoplasm, pencillate nuclei and ectopic crypts. The serration in TSAs is imparted by a combination of undulations in the crypt epithelium and crypt budding. TSAs almost always comprise a mixture of foci showing the above characteristic features with areas showing a ‘classical’ adenoma growth pattern, in which obvious dysplasia is present.

The proportion of areas showing the characteristic TSA features and ‘classical’ adenoma features is variable and the minimum criteria for a diagnosis of TSA are not well defined. Molecular analysis has revealed that TSAs more commonly possess KRAS mutations and less commonly harbour BRAF mutations than SSLs. For these reasons, TSAs appear to be more closely related to ‘classical’ adenomas than SSLs.

TSAs are characterised by a disruption of the signalling pathways involved in stem cell control and cell fate determination. This results in the expansion of a progenitor cell population from the crypt base into the ectopic crypt foci or lateral buds that morphologically characterise this condition. These progenitor cells actively proliferate and accumulate somatic mutations with resultant dysplasia arising from outside the crypt base stem cell niche. This may be why TSAs seem to have a more rapid malignant potential compared to ‘classical’ adenomas, as the ectopic crypt foci/lateral buds act like extra crypt cell niches and are subject to additional mutations.

Mixed polyps

It is likely that the majority of ‘mixed’ polyps, especially in the right colon, represent SSLs with dysplasia. However, lesions may rarely be encountered, particularly in the left colon, that appear more likely to have arisen due to a true ‘collision’ event between a HP and a ‘classical’ adenoma. Furthermore, TSAs in which a significant component shows the features of a ‘classical’ adenoma are also seen. The minimum proportion of a TSA that is required to show features of a ‘classical’ adenoma in order for the polyp as a whole to be considered ‘mixed’ has yet to be defined. Occasionally polyps showing a combination of SSL and TSA-like features are encountered, with or without areas with a ‘classical’ adenoma appearance.

Another variant of the mixed polyp is the combination of HP changes and serrated low grade dysplasia with features of a TSA. These lesions are more unusual and are seen typically in the sigmoid colon and rectum. While a collision lesion is possible, we believe that the latter form of mixed polyp most likely represents different stages in the traditional serrated neoplasia sequence with serrated dysplasia deriving from a pre-existing HP.

Due to the existence of such lesions, it is sensible to retain the term ‘mixed polyp’ within the recommended terminology list, even if they may represent different serrated entities and different serrated neoplasia pathways.

Use of the term ‘mixed polyp’ should always be accompanied by a detailed description of the mixed features present and relative contributions to overall lesion composition. Where a lesion is showing a mixed type, the management should be based on the most clinically important component.

The principles of serrated lesion sizing and margin assessment are essentially the same as those for adenomatous lesions – as described in the guidance on classical adenoma sizing and grading dysplasia. Assessment of the proximity of hyperplastic polyps to the surgical margin is not required.

Advanced serrated polyps

Under the new BSG/PHE/ACPGBI guidelines, all serrated polyps (an umbrella term for hyperplastic polyps, sessile serrated lesions, traditional serrated adenomas and mixed polyps) which measure ≥10 mm or demonstrate dysplasia, are considered advanced colorectal polyps. This includes all traditional serrated adenomas, on the basis that all of these show dysplasia – although the evidence base for this inclusion is weak. Within the same guidelines, all serrated polyps apart from rectal hyperplastic polyps 5mm or less in size are considered ‘premalignant polyps’.

Serration in other situations

It is now recognised that serration may be seen as a complication of chronic inflammatory bowel disease. The significance of isolated epithelial serration in ulcerative colitis in particular is currently uncertain. Serration may also be seen in dysplasia arising in the context of chronic idiopathic inflammatory bowel disease.

It also appears that epithelial serration, in the colon and rectum, can occur in other settings. For instance, epithelial serration is associated with stromal lesions, particularly small stromal polyps. Previously, this has been documented in colonic neurofibromas, perineuriomas and so-called ‘benign fibroblastic polyp of the colon’. Further, particularly in the right colon, one may also see mucosal serration overlying submucosal lipomas.

Although this might represent the coincidence of a sessile serrated lesion and an underlying lipoma, there is recent unpublished literature to suggest that this may represent divergent differentiation in the same lesion. Evidence for the latter has certainly accrued for the combination of serrated pathology and perineuriomatous proliferations (as described in research published in 2010, 2011 and 2013). If in doubt, please consider carefully whether any serration found is a distinct lesion or secondary to an underlying pathology.

3.6 Inflammatory polyps

Inflammatory-type polyps are relatively common in BCSP practice and represent a heterogeneous group. Although they are most usually seen as a complication of chronic idiopathic inflammatory bowel disease (particularly chronic ulcerative colitis) they are also frequently encountered in association with diverticulosis and/or mucosal prolapse. Furthermore, sporadic single inflammatory-type polyps are well described in the colorectum. As the reporting pathologist may not know the clinical context of such polyps, specifically the full colonoscopic appearances, we recommend that all such polyps are classified as ‘inflammatory polyps’.

The morphological features may vary depending on the clinical context, but there is considerable overlap. Features may be sufficiently characteristic to allow diagnosis of a specific entity such as an inflammatory myoglandular polyp or colonic mucosubmucosal elongated polyp (CMSEP), entities reported to be distinct from other more common inflammatory and mesenchymal polyps (as described in research published in 1998 and 1992). For the purposes of data recording, such entities should also be included under the umbrella term ‘inflammatory polyp’.

4. Reporting invasive neoplasia

4.1 Definition of invasion

The WHO recommended definition of colorectal adenocarcinoma is the one in everyday use within the UK, namely: “invasion of neoplastic glandular epithelial cells through the muscularis mucosae into the submucosa of the bowel wall”. This definition does not allow comparison with Japanese series, in which a diagnosis of carcinoma can be made on purely cytological grounds or in cases with invasion into the lamina propria but not beneath the muscularis mucosae. However, it is consistent with United States and European literature and is also recommended in European guidelines for CRC pathology reporting.

This definition does not allow for the diagnosis of intramucosal carcinoma or stage pT in situ (pTis). Such terms are discouraged, to avoid overtreatment of lesions considered to have minimal or no risk of metastatic spread. The term high grade dysplasia should be used for such cases.

4.2 Reporting of diagnostic biopsy specimens

The definitive identification of adenocarcinoma in endoscopic biopsy specimens is one of the most difficult tasks faced by diagnostic pathologists reporting BCSP specimens. The diagnostic requirement to demonstrate submucosal invasion is problematic as submucosal tissue may not be represented in endoscopic biopsy material. Biopsies from endoscopically suspicious colorectal tumours therefore often fail to demonstrate clear-cut submucosal invasion.

The presence of a desmoplastic stromal response to neoplastic glands is considered acceptable for a diagnosis of adenocarcinoma in most clinical circumstances, with the notable caveat to exercise caution in the context of previous endoscopic biopsies or partial lesion resection from the same site. Also, juxtaposition of neoplastic glands to submucosal structures such as larger blood vessels, nerves and other neural structures may be sufficiently convincing to signify adenocarcinoma.

In the clinical context of a suspicious colonic mass on endoscopy and/or imaging which is locally unresectable, a histological diagnosis of primary glandular neoplasia – high grade dysplasia suspicious of adenocarcinoma or adenocarcinoma – is usually sufficient to proceed to surgery. More caution should be exercised with rectal lesions, given the greater number of therapeutic options for local excision and the higher morbidity of rectal surgery. Further, as rectal cancer is often treated with preoperative radiotherapy or chemoradiotherapy, it may be prudent to try to obtain a diagnostic sample should any form of molecular assay subsequently be requested. Such assays are best performed on treatment-naïve tissue specimens.

In summary, it is recommended to report the features that are evident microscopically, and to determine clinical management at a multidisciplinary team meeting discussion. A histological diagnosis of CRC requires, as a minimum, either definite histological evidence of submucosal invasion or a desmoplastic reaction to neoplastic glands in the setting of a clinically evident malignancy.

4.3 Stage pT1 adenocarcinoma

It is beyond the scope of this document to discuss in detail reporting of CRC surgical resection specimens as it is outside of the screening pathway. We recommend the guidance produced by the Royal College of Pathologists (UK) for reporting CRC (including local excision) specimens is recommended. However, within screen-detected CRCs, a high proportion (up to 20%) are of early stage (stage pT1 or ‘polyp cancers’) compared to symptomatic CRCs.

While the principles of reporting local excision specimens are the same as in reporting major surgical resections, a number of features require special attention in reporting local excisions of presumed early cancers. This is because they are used to evaluate the need for further surgical intervention. Areas of attention include:

• an assessment of margins to indicate completeness of excision
• measurement of parameters that are considered to predict the presence of lymph node metastatic disease, specifically:
• tumour size
• differentiation
• extent of submucosal invasion
• presence of lymphatic or venous invasion

The core data items recommended in the Royal College of Pathologists (UK) dataset for recording are:

• specimen type – polypectomy, endoscopic mucosal resection, endoscopic submucosal dissection, transanal endoscopic microsurgical excision, local surgical excision, major surgical excision stating the type of operation and specimen removed
• site of tumour
• overall specimen (usually polyp/lesion) size
• histological tumour type
• histological differentiation
• extent of local invasion
• lymphatic invasion
• venous invasion
• perineural invasion
• presence of a precursor lesion (or rarely other polyp/lesion type)
• margin involvement by carcinoma (deep/peripheral)
• minimum deep margin clearance of the invasive carcinoma (in millimetres)
• pT stage
• pN stage
• MMR/MSI tumour status, with an indication if the patient needs to undergo further testing for possible Lynch syndrome

4.4 Histological tumour type

Tumours should be typed using the WHO classification of 2019. The vast majority of malignant colorectal tumours are adenocarcinomas. Most are adenocarcinomas of no specific type, referred to as ‘not otherwise specified’ (NOS). Subtypes of adenocarcinoma recognised are:

• mucinous adenocarcinoma; adenocarcinoma with >50% composed of extracellular mucin
• signet ring cell adenocarcinoma; adenocarcinoma with >50% signet ring cell morphology
• medullary carcinoma
• serrated adenocarcinoma
• micropapillary adenocarcinoma; adenocarcinoma with >5% micropapillary morphology
• adenoma-like adenocarcinoma; adenocarcinoma with >50% adenoma-like morphology

Some of these histological variants are associated with characteristic biological and/or clinical features. Mucinous, signet ring and medullary carcinoma, when associated with mismatch repair (MMR) deficiency, have an excellent clinical prognosis. Micropapillary variants of adenocarcinoma tends to behave aggressively and, in the setting of early stage CRC, has been reported to have a significantly increased risk of regional lymph node metastatic disease. All other carcinoma types are rare in the colon and rectum.

4.5 Histological differentiation

Differentiation of CRCs is based primarily on architecture and specifically gland or tubule formation. Poorly differentiated tumours demonstrate either irregularly folded, ill-formed, small tubules or no tubule formation at all. Poor differentiation in early stage CRC is a significant risk factor for regional lymph node metastatic disease, and therefore a potential indicator for surgical intervention. However, publications assessing pT1 cancers and the risk of lymph node metastasis vary in whether poor differentiation is based on the predominant area or the worst area of differentiation within the tumour (as described in research published in 2015, 2004a, 2004b, 2014 and 2013).

As it is likely that most have used the worst area, in line with other guidance it is currently recommended that poor differentiation should be based on the presence of any area of definite poor differentiation until the situation is clarified by further research (as described in research published in 2018 and 2011). This approach helps minimise the risk of patient under-treatment. However, tumour ‘budding’ alone is not considered morphological evidence of poor differentiation (see guidance on tumour budding).

Poor differentiation and mucinous or signet ring cell morphology are among the morphological features suggesting involvement of the microsatellite instability (MSI) or mismatch repair (MMR)-deficient pathway to carcinogenesis. Other features include a medullary or solid architecture and prominent tumour-infiltrating lymphocytes. These features raise the possibility of underlying Lynch syndrome or, more likely in the age range of BCSP, sporadic MSI/MMR-deficient pathway CRC. There is considerable evidence that MMR-deficient CRCs have a better prognosis than MMR-proficient tumours (as described in research published in 2008 and 2011).

It is unclear if MMR status influences the significance of poor differentiation on lymph node metastatic risk in stage pT1 CRC. It is possible that poor differentiation is only an adverse feature in MMR-proficient tumours and should not be used as a factor triggering surgical intervention in MMR-deficient tumours. Further studies are required on this matter.

NICE recommends that MMR immunohistochemistry/MSI testing is undertaken in all new diagnoses of CRC. If MMR deficiency/MSI-high status is identified, further testing (with BRAF mutation analysis and hMLH1 methylation status assessment as appropriate) performed outside of the BCSP will be required to evaluate for possible Lynch syndrome.

4.6 Extent of local invasion

A variety of descriptive or quantitative methods have been proposed for stratifying stage pT1 CRCs, mainly with the purpose of determining the risk of regional lymph node metastatic disease when the cancer has been removed by local excision. In stage pT1 CRCs, the frequency of lymph node metastasis in sessile tumours that involve the thirds of the submucosa (so-called Kikuchi levels) has been reported to be:

• superficial (level sm1) – 2%
• middle (level sm2) – 8%
• deep (level sm3) – 23%

This is described in research published in 1995 and 2002.

In polypoid lesions, Haggitt and others identified the level of invasion into the stalk of the polyp as being important in predicting outcome and found that ‘level 4’ invasion was a factor. This is when a tumour extended beyond the stalk of the polyp into the submucosa but did not invade the muscularis propria. However, neither the Kikuchi nor Haggitt systems are easy to use in practice.

Haggitt level is particularly difficult in polypoid specimens lacking a clearly defined stalk (‘sub-pedunculated’) or if the specimen is poorly orientated. Kikuchi level cannot be evaluated accurately without representation of muscularis propria in the specimen, to allow division of the submucosa into thirds. Such representation is rare in local excision specimens, with the exception of some transanally derived specimens. Despite these limitations, which result in a limited clinical utility of Haggitt and Kikuchi levels in routine practice, they should still be recorded where possible in the absence of strong evidence to recommend alternative approaches.

Ueno and others have proposed that the depth of invasion of tumour beyond the muscularis mucosae and width of the invasive tumour provide more objective measures of potential for lymph node metastasis. Both of these quantitative measurements should also be recorded, in millimetres, in line with the recommendation of the Royal College of Pathologists (UK) dataset. Depth (or thickness) of invasive tumour should be measured from the muscularis mucosae where it is intact and identifiable. If the muscularis mucosae is obscured or destroyed by tumour, tumour thickness should be measured from the surface of intact mucosa or ulcer.

Forthcoming further evidence should indicate which of these methods of assessment of extent of local invasion will be most useful for clinical management decisions.

4.7 Lymphatic, venous and perineural invasion

Submucosal lymphovascular invasion is defined as tumour infiltration of endothelium-lined spaces in the submucosa is regarded as a significant risk factor in local excision specimens for lymph node or distant metastatic disease. Two separate meta-analyses examining studies of stage pT1 CRC concluded that lymphatic invasion and, to a much lesser extent, venous invasion, are powerful predictors of lymph node metastatic disease (as described in research published in 2004 and 2014). Therefore, it is now considered appropriate to attempt to evaluate lymphatic and venous invasion separately if possible.

It is important to distinguish lymphatic invasion from retraction artefact, and this may be assisted by application of D2-40 immunohistochemistry. This specifically identifies lymphatic channel endothelium and not venous channel endothelium (as described in research published in 2012 and 2009. Venous invasion is defined as tumour lying within an endothelium-lined space that is either surrounded by a rim of muscle or contains red blood cells. Elastic stains may highlight the rounded structure of a vein wall if tumour has obliterated the vein lumen. The greatest depth of lymphatic and venous invasion (intramural, comprising submucosal and intramuscular, or extramural) should be recorded, although, this will almost always be submucosal (intramural) within local excision specimens.

Lymphatic channels lack the muscular wall evident in veins and usually, though not always, contain no red blood cells. Confidently identifying thin-walled submucosal vessels as lymphatic channels or post-capillary venules may be extremely difficult, and application of D2-40 immunohistochemistry in selected cases is recommended given this potentially important distinction. Immunohistochemistry may also help distinguish lymphatic and/or venous invasion from retraction artefact or tumour ‘budding’. Lymphatic and/or venous invasion should only be recorded as positive if the features are considered definitive.

The significance of perineural invasion has only been demonstrated for CRC in surgical resection specimens not local resections. However, for consistency, the presence and deepest level (intramural or extramural) of perineural invasion should be reported for all CRC resection specimens.

4.8 Presence of a precursor lesion

Invasive carcinoma may destroy any precursor non-invasive lesion but, if any residual precursor is identified, the nature of this should be recorded. This will usually be a ‘classical’ adenoma but may on occasion be some other lesion such as a SSL or TSA.

4.9 Margin assessment

Both peripheral (mucosal) and deep margins need to be assessed. The peripheral margin of a local excision may be involved by invasive carcinoma, by non-invasive adenoma, or clear of both. The precise measurement of the closest proximity of the deep margin from invasive tumour should be recorded.

Most large lesions detected in BCSP practice are removed by diathermy snare or similar devices. Diathermy resection produces a zone of diathermy burn which can be up to several millimetres thick. Due to coagulation of tissue, and this introduces a number of secondary artefactual changes (see examples). These include:

• the diathermied plane of resection is drawn back into the stalk of the lesion and may be buried beneath the less affected mucosal rim
• coagulated blood vessels may stand proud of the rest of the retracted diathermy plane as they do not shrink to the same degree as the surrounding stroma
• possible marked clefting alongside the coagulated blood vessels because of the differential shrinkage of vessels and stroma – the coagulation zone is brittle and so may split or fragment during dissection
• the loose submucosal stroma may appear markedly disrupted beyond the zone of diathermy burn, probably due to a vaporisation effect

Care should be taken to take account of artefact, which could give rise to a false assessment of distance of tumour to the margin. Drawing a straight line to join the 2 edges of the retracted plane and using that as a putative plane will give an erroneous measurement if either the tumour is close to the margin and also retracts back into the lesion, or the lesion is sessile but develops a curved shape due to diathermy and fixation. For this reason, it is advised that the outer edge of the diathermy zone is used for assessment of the margin (examples B and C of diathermy artefact and margin assessment).

The process for this is:

1. Starting from the muscularis mucosae on one side, draw a smooth line following the outer edge of the diathermy burn to run to the muscularis mucosae of the opposite side. Include any indentations but ignore any artefactual splits and clefts.
2. Measure the distance of invasive carcinoma to the notional line. Distances should be recorded in millimetres to one decimal point.

Cancer in whatever context (for example in blood vessels or present as pools of mucinous carcinoma) should be considered when assessing tumour proximity to a margin. If there is coexisting epithelial misplacement care should be taken to ensure benign elements such as mucin lakes are not included.

If there is infiltration by malignant glands into the diathermy zone and this is associated with morphological distortion of tissue to the extent that it is not possible to confidently identify tumour clearance from the outer margin (see examples), then this should be regarded as margin involvement and a distance of 0 mm of clearance recorded. Cytokeratin or CDX2 immunohistochemistry may help assessment in this situation by identifying neoplastic glands within the diathermy zone and their relationship with the outer edge of the zone of diathermy.

4.10 Examples of diathermy artefact and margin assessment

Diathermy snare resection introduces a zone of secondary artefactual changes that makes margin assessment difficult.

Stromal vaporisation, clefting and withdrawal of the resection plane is most marked adjacent to thick submucosal blood vessels.

Measurement of tumour to resection margin should avoid such artefacts and a best attempt made to measure to the true resection margin (arrow). A distance of <1 mm from tumour to margin is considered margin involvement (R1 resection status).

In picture C, the zone of diathermy artefact can be several millimetres thick depending on the excision technique employed. The short arrow here is 2.1 mm long and the tumour is well clear of the outer (true) resection margin (the long arrow is 3.9 mm long). Measurement of clearance should not be to the inner aspect of the diathermy zone, in this event. However, any infiltration by malignant glands into the diathermy zone (short arrows) is regarded as margin involvement (0 mm distance recorded), as it is not possible to confidently determine the true extent of infiltration in this situation.

Inking the diathermy line is recommended, although the zone of diathermy burn should provide adequate evidence of the true margin. Problems in assessment may also arise because the axis of section does not include the diathermy line or because of cross cut elements in convoluted lesions. In general:

• when dissecting lesions ensure that the plane of section includes the diathermy line
• when assessing lesions be aware of their 3-dimensional configuration and their orientation
• sections cut at deeper levels may assist assessment, particularly if the lesion is convoluted or the artefact is marked
• in cases of doubt adopt a conservative approach and only measure to a margin about which you are confident

Involvement of a peripheral margin may indicate the need for repeat endoscopy and further local excision. Involvement of the deep resection margin by invasive tumour has traditionally been an indication for considering surgical intervention. There has been considerable discussion and controversy in the literature over the degree of clearance that might be regarded as acceptable in tumours, which extend close to the deep submucosal margin. Most existing guidance considers a clearance of <1 or 1 mm as needing consideration of further therapy (as described in research published in 2018, 2011 and 2013).

However, there is recent evidence to suggest that, for pT1 tumours, only tumour present at the true or estimated resection margin, or within the diathermy burn zone, should be considered for further treatment, which may be local re-excision if no other adverse pathology features are identified (as described in research published in 2016, 2015a and 2015b). However, it is considered that more substantive evidence is required before a change of protocol is made, and a margin of 1 mm or less remains the indicator for incomplete excision until this is forthcoming.

If tumour is close to the excision margin, multiple levels (at least 3 and preferably more) should be examined in the assessment of possible margin involvement, with adjunctive immunohistochemistry if required. Careful follow-up of the local excision site of carcinomas less than 1 mm from the margin is recommended if the MDT decide not to undertake a further excision or resection.

4.11 Tumour budding

There is considerable evidence emerging that identification of so-called ‘tumour budding’ in local excision specimens may predict outcome or indicate increased risk of nodal metastatic disease (as described in research published in 2015a, 2015b, 2004 and 2009). However, studies reporting the significance of tumour budding employ a wide variety of methods of assessment, some requiring cytokeratin immunohistochemistry, and there remain concerns over methodologies, cut-offs and reproducibility given the wide ranges of budding reported in CRC. For these reasons, routine reporting of tumour budding is currently not recommended. International standards have been agreed but these await validation in a screening context, as well as the relative value of budding versus other factors currently recommended to indicate the need for resection.

5. TNM 8 classification

Tumour, node and metastasis (TNM) staging is one of the systems used to stage bowel (colon and rectal) cancer and is what is used in the BCSP. A full list of TNM 8 classifications is available from the RCPath.

6. Bowel Cancer Screening System (BCSS) polyp dataset

The dataset to include in a proforma for collecting pathology data onto the central bowel cancer screening system (BCSS) is:

Polyp type: Adenoma

Subtype: Tubular adenoma / Tubulovillous adenoma / Villous adenoma

Polyp type: Serrated lesion

Subtype: Hyperplastic polyp / Sessile serrated lesion / Sessile serrated lesion with dysplasia / Traditional serrated adenoma / Mixed polyp

Polyp type: Inflammatory polyp / Peutz-Jeghers polyp / Juvenile polyp / other polyp

Subtype: Neuroendocrine tumour (carcinoid) / Lymphoid polyp / Lipoma / Stromal / other polyp

Polyp type: Normal mucosa

Polyp margin clear: Yes / No

Polyp size (mm):

Polyp dysplasia: Low grade / High grade / Not applicable

Polyp carcinoma: Yes / No