Lower gastrointestinal endoscopy: guidance on indications for biopsy

Abstract

Lower gastrointestinal endoscopy is a commonly undertaken procedure and has assumed even greater prominence with the inception of the NHS Bowel Cancer Screening Programme (BCSP). Workloads are also constantly increasing within histopathology departments and this has led to a need for workload management by laboratories. Advanced endoscopic techniques now allow for targeted biopsies within settings such as inflammatory bowel disease surveillance and the BCSP. In this article, we provide guidance to the endoscopist for optimal biopsy protocols that are designed to maximise the chance of a subsequent histopathological examination providing definitive results and to reduce the number of unnecessary biopsies, in which histopathology is unlikely to deliver clinically useful information. The majority of the article focuses on biopsy taking within a defined range of clinical situations that are commonly encountered by endoscopists.

Introduction

This review provides guidance for the role of endoscopic biopsy in the diagnosis and management of a range of the most commonly encountered diseases during colonoscopy, including the terminal ileum. It includes advice as to when it is useful to perform a biopsy and when a biopsy is not likely to provide clinically relevant information or change management. Biopsy may be dangerous in certain situations—for example, perforation may result when biopsy is attempted in severe colitis or after biopsy of a mucosal diverticulum. In addition, the risk of bleeding is increased if vascular lesions undergo biopsy. Hot biopsy will tend to increase the risk of perforation, particularly within the right side of the colon. Furthermore, diathermy artefact associated with hot biopsy may inhibit accurate histological interpretation. Submucosal lesions are not infrequently biopsied but the diagnostic yield is low. A soft submucosal lesion is likely to be an entity such as a lipoma and does not usually warrant biopsy, whereas a firm lesion is likely to be a gastrointestinal stromal tumour or a neoplasm infiltrating the colon or rectum from outside. If histology is required, a drill biopsy technique is more likely to reveal adequate sampling; otherwise, other investigations may be appropriate—for example, an MRI scan.

Key information for the pathologist

Generally speaking, the quality of histological interpretation of an endoscopic biopsy is enhanced when appropriate clinical information is given to the histopathologist. ‘Pre-test likelihood’ can positively influence the opinion of the reporting histopathologist in the same way that a good clinical history is an essential component of a clinical consultation. For example, early inflammatory bowel disease (IBD) may show features almost indistinguishable from those of infective colitis and therefore features such as the length of the clinical history may facilitate histological interpretation. A history of radiotherapy may help the pathologist to interpret the early or late changes of radiation-induced injury (figure 1). When interpreting a colonic polyp biopsy, knowledge that the polyp has previously undergone partial polypectomy may reduce the risk of misinterpretation of misplaced adenomatous epithelium as adenocarcinoma. It is useful for the pathologist to know the site and type of polyp, as histology from a pedunculated sigmoid polyp that may have undergone torsion might be overinterpreted as cancer when it is actually benign. Similarly, the clinician needs to understand how to interpret ‘histopathological language’ when trying to translate histopathology into clinical practice—for example, the term ‘adenocarcinoma’ is commonly used, but does not necessarily mean that an early lesion is unresectable—and potentially curable—by polypectomy. Ideally, submission of the endoscopy report with copies of relevant pictures is encouraged.

Figure 1.

(A–D) Images taken at colonoscopy. (A) Superficial ulceration in the terminal ileum. This could represent Crohn's disease or another condition such as non-steroidal anti-inflammatory drug effect. Biopsies can help to confirm acute inflammation and/or ulceration and may reveal useful additional features such as granulomas, but may not show specific diagnostic features. (B) Sessile serrated lesion, present on a mucosal fold and highlighted with narrow band imaging. (C) Colitis in the sigmoid colon. This could represent inflammatory bowel disease or, in the appropriate clinical context, an alternative condition such as diverticular colitis. Biopsies can help to confirm the presence of colitis but may not be specific about the aetiology. (D) Radiation proctitis. A clinical history of radiotherapy is likely to help the pathologist arrive at the correct interpretation of biopsies from this area. (E–H) Histopathological images of gastrointestinal mucosal biopsies (all haematoxylin and eosin stain). (E) Part of a sessile serrated lesion showing pronounced serration and a branched crypt, deep within the lesion. Magnification×200. (F) Part of a traditional serrated adenoma showing eosinophilic cytoplasm and pencillate nuclei, particularly within the upper part of the image, together with crypt budding. Magnification×400. (G) Collagenous colitis showing a markedly thickened subepithelial collagen plate, together with a mild to moderate diffuse increase in lamina propria chronic inflammatory cell numbers. Early artefactual separation of the surface epithelium is also present. Magnification×200. (H) Segmental colitis secondary to diverticular disease, showing a branched and distorted crypt together with a mild diffuse increase in lamina propria chronic inflammatory cell numbers. Small numbers of neutrophils were also visible within the lamina propria and epithelium on high-power examination. The clinical history is essential here, since the histopathological features alone would otherwise be consistent with inflammatory bowel disease, for example. Magnification×200.

As well as the provision of appropriate clinical history, the anatomical site of origin of biopsies should always be clearly labelled. The normal appearance of proximal colonic mucosa differs from that of left-sided colonic mucosa and if the precise site of origin of biopsies is not stated, this might lead to an erroneous diagnosis of abnormality (eg, due to the presence of Paneth cells) within a biopsy that in fact has a normal appearance, but is derived from the right colon. Colonic series are best presented to the pathology laboratory either in multiwell cassettes or with the individual biopsies attached to a strip of Micropore filter paper that itself is clearly orientated to show the proximal and distal ends (eg, by ensuring that one end of the filter paper is pointed). Alternatively, it is possible to group biopsies into those derived from the right and left colons and to submit these to the histopathology laboratory in two containers. While this is quite a widely used technique, the multiwell or filter paper methods are preferable as they allow the histopathologist to provide a more precise assessment of disease distribution and reduce laboratory workload.

Encouraging best practice

Histopathologists can minimise the risk of diagnostic errors by working together in groups and by double reporting specimens in certain clinical situations, especially when there is diagnostic uncertainty. When a diagnosis of pT1 adenocarcinoma is being considered within a colorectal polypectomy specimen, a natural recommendation is that the case is seen by at least two pathologists with an interest in gastrointestinal pathology. This particular guidance is now mandatory within the UK Bowel Cancer Screening Programme. Full integration of histopathologists within the multidisciplinary team—including attendance at regular meetings—will help to ensure that histology reports are appropriate to the clinical context of each case. Histopathologists should also be members of appropriate external quality assurance schemes for diagnostic histopathology.

The remainder of this paper focuses on providing advice within a defined range of clinical situations that are particularly pertinent to the practice of clinical gastroenterology and endoscopic biopsy in colorectal disease (table 1).

Table 1.

Summary of guidance suggestions for endoscopic biopsy of the lower gastrointestinal tract

Clinical situation	Guidance for biopsy
Suspected microscopic colitis	At least 3–4 biopsies from each of the right and left colon. Biopsies from the distal colon and rectum may not always show the diagnostic features
Suspected active colitis, including IBD	As for suspected microscopic colitis; biopsies to be clearly arranged in order from proximal to distal origin. Biopsies from an inflamed sigmoid colon to be paired with biopsies from the rectum. Biopsies from an ulcer base may be useful for the identification of viral inclusions but otherwise biopsies of non-ulcerated mucosa are more likely to provide useful information
Surveillance of IBD	If monitoring for disease activity, as for suspected active IBD. If monitoring for neoplasia in longstanding disease, targeted biopsies should be used if the appropriate endoscopic techniques are available—for example, dye spraying. Suspected DALMs should undergo biopsy or ideally removal (eg, using EMR) and this should be accompanied by biopsies of the surrounding mucosa if the lesion is within an area affected by IBD
Proof that the ileocaecal valve has been reached during colonoscopy	This is not an indication for biopsy of the macroscopically normal terminal ileum. Photographic or video evidence should instead be used
Polypoid lesions	Small lesions can be removed by biopsy or polypectomy. Larger lesions may be removed by EMR if this is technically possible. This allows a potentially more definitive histopathological assessment and may well also be curative. Biopsy of lesions that are almost certain later to undergo EMR may lead to fibrosis and tethering of otherwise mobile lesions. The potential for sampling variability leading to failure to identify adenocarcinoma within a larger lesion that undergoes biopsy rather than formal excision should always be considered

DALM, dysplasia-associated lesion or mass; EMR, endoscopic mucosal resection; IBD, inflammatory bowel disease.

In what circumstances should colorectal biopsy specimens be taken when the mucosa appears normal at endoscopy?

Histopathological examination has long been the ‘gold standard’ for diagnosis when abnormal mucosa is seen during endoscopy. It is also recognised that mild/early disease or conditions that are only evident on microscopic examination of tissue may only be detected by biopsy and histopathological assessment. The commonest reason to biopsy endoscopically normal colorectal mucosa is to exclude microscopic colitis (especially lymphocytic and collagenous colitis) (figure 1).

A diagnosis of microscopic colitis should be considered in the appropriate clinical context. Both lymphocytic and collagenous colitis present with watery diarrhoea (ie, Bristol Stool Scale types 6 and 7), which may be chronic and which does not alternate with constipation, although patients may also present with this condition after a single episode of diarrhoea. Patients with collagenous colitis are most commonly middle-aged and elderly women while lymphocytic colitis is distributed more evenly between men and women. The latter form of colitis may be associated with coeliac disease, and both conditions may be associated with non-steroidal anti-inflammatory drug (NSAID) use. Both may also be associated with other symptoms—for example, arthropathy. Up to 20% of patients with chronic watery diarrhoea and in whom the colorectal mucosa appears normal at endoscopy, may have microscopic colitis.¹ Colorectal biopsies should generally only be performed in endoscopically normal colorectal mucosa when patients are middle aged or elderly and have persistent watery diarrhoea without bleeding. Interestingly, it has recently become recognised that subtle endoscopic abnormalities may be seen in microscopic colitis—for example, mucosal tears (figure 1).²

Colorectal biopsies in patients presenting with rectal bleeding and in whom the colorectal mucosa is endoscopically normal, are not indicated and do not help to identify the cause of the bleeding. In these patients, the most likely causes include haemorrhoids, anal fissures or diverticular disease.³ It is theoretically possible to identify very early IBD in patients where the colorectal mucosa is endoscopically normal. However, most series indicate that the detection rate of early IBD in this situation is extremely low.⁴ If there is clinical suspicion of IBD or other forms of colitis, then depending on the level of this suspicion, biopsies should be considered since the finding of normal mucosa might be helpful. For example, while colonic mucosa from patients with entirely quiescent IBD may appear histologically normal, biopsies from endoscopically normal mucosa might show features of healed/inactive chronic colitis. The term ‘focal active colitis’ has been used to describe the presence of colorectal mucosa with a normal appearance apart from an occasional focus of acute cryptitis or crypt abscess formation. This appearance has been linked to a subsequent diagnosis of Crohn's disease in a minority of younger patients, although more common alternative causes include resolving infection, NSAIDs and bowel preparation-associated changes.^5–7

When considering a diagnosis of microscopic colitis, how many biopsy specimens should be taken and from where?

When biopsies are being performed to exclude microscopic colitis, the histological features can be patchy and may also be more prominent within the proximal colonic mucosa, although sigmoidoscopy and rectal biopsy may provide a positive diagnosis in some cases.⁸ Therefore, definitive exclusion of a diagnosis of either lymphocytic or collagenous colitis requires full colonoscopy and multiple mapping biopsies, with at least three to four biopsies from the right colon (caecum—transverse colon) and the same number from the left colon (descending colon—rectum).

If the terminal ileum appears normal at endoscopy, should it be biopsied?

Colonoscopy allows the intubation of the terminal ileum in most patients. This can be useful when an abnormality is found—for example, ileal mucosal ulceration in a patient with localised Crohn's disease or chronic NSAID use (figure 1). With increasing calls for evidence to judge competency at colonoscopy, the practice of terminal ileal biopsy to demonstrate completeness of colonoscopy has become more common. However, ileal biopsy should not be used to demonstrate that colonoscopic examination has reached the ileum. Other methods of confirming ileal intubation are preferred, which are considerably cheaper— for example, photographic or video evidence. Nodularity of the ileal mucosa due to the presence of prominent lymphoid follicles is a common appearance and does not require biopsy.⁹ However, biopsies may be reasonable in this situation if there is clinical/endoscopic suspicion of lymphoma.

Ileal biopsies may provide useful clinical information if they are performed owing to the presence of an endoscopic lesion such as a polyp or ulcer. In Crohn's disease, granulomas are identified within ileal biopsies relatively rarely. However, the presence of other findings may contribute to the diagnosis. For example, ulcer-associated cell lineage (pseudopyloric metaplasia) can be a useful indicator of chronic inflammatory mucosal damage within the terminal ileum. The detection rate of features supporting a diagnosis of IBD within ileal biopsies is low unless an endoscopic abnormality is present.¹⁰ Also, most cases of isolated aphthous ulceration of the ileal mucosa do not progress to Crohn's disease and therefore histopathologists should take care not to overdiagnose Crohn's disease when superficial ulceration is the sole histopathological abnormality.

NSAIDs are associated with a variety of changes within the colon, as well as the ileum. These include non-specific colitis, ulceration, microscopic colitis and triggering a flare in IBD.¹¹ No particular biopsy strategies are recommended to detect these conditions, beyond those usually used to determine the extent and severity of the abnormalities present. A key factor to remember is that because NSAIDs are associated with a wide spectrum of mucosal changes, the histopathologist should be informed of any history of NSAID use. Diaphragm disease is classically considered to be a NSAID-related condition affecting the small intestine, but it may, rarely, occur within the large intestine.¹² Identification of diaphragm disease automatically implicates NSAIDs in the cause of the lesions, but this is essentially a macroscopic diagnosis made at endoscopy or within a surgical specimen, with biopsies disclosing non-specific features that require interpretation in the appropriate clinical context.

What is the optimal biopsy technique for patients with active ulcerative colitis?

The key reasons for performing biopsies in patients with suspected active ulcerative colitis are to help confirm the diagnosis and to determine the extent of the disease. Knowledge of the latter is important for determining the best management strategy. Biopsies from patients with active ulcerative colitis may show characteristic histopathological features, but these are rarely specific for the condition.¹³ For example, biopsies in early IBD—before the microscopic features of disease chronicity have had chance to develop—may show features that are essentially indistinguishable from those of self-limiting colitis (infective). Therefore, the clinical history and endoscopic appearance are important—the histopathologist may be willing to consider a diagnosis of early IBD in the differential diagnosis if the history is known to be short, when the absence of features of chronicity might otherwise suggest that IBD is unlikely. Opportunistic infections can be important to identify, whether they are the primary diagnosis or a secondary cause of relapse in IBD. Conditions such as tuberculosis and lymphogranuloma venereum are important differential diagnoses of IBD, especially Crohn's disease. Immunohistochemistry can be used to highlight and positively identify inclusion bodies (eg, for cytomegalovirus) and additional tests—for example, PCR-based tests—can be obtained where required (eg, for cytomegalovirus or Chlamydia trachomatis).

Biopsies from a colonic segment affected by diverticulosis may show overlying mucosal inflammation that can mimic either ulcerative colitis or Crohn's disease—the characteristics of chronic inflammation and/or other features commonly seen in Crohn's disease—for example, granulomas, may be present. Therefore, if a segmental colitis is encountered (figure 1), it is important to take biopsies from both the inflamed segment and a non-inflamed segment—for example, the rectum—to distinguish so-called diverticular colitis from chronic ulcerative colitis.¹⁴ Terminal ileal biopsy may provide useful additional information (eg, the presence of granulomas or ulcer-associated cell lineage) in patients where there is significant clinical suspicion of Crohn's disease.¹⁵

Ischaemic colitis can also be confused with IBD. The clinical picture of an elderly patient with known arterial disease, with hypotension or after major vascular surgery, increases the suspicion. The endoscopic changes can mimic IBD but a sharply demarcated disease distribution relating to the anatomy of the arterial supply may be a useful feature. Colonic biopsies in cases of subacute ischaemia may have a similar appearance to those of Crohn's disease, but features characteristic of subacute ischaemia—for example, variable degrees of lamina propria fibrosis (also seen in chronic ischaemia), haemosiderin deposition, capillary ectasia and microthrombi—may also be present.¹⁴

When mapping the extent of active ulcerative colitis, multiple colonic biopsies from the proximal colon to rectum are recommended and should be labelled according to their site of origin. Knowledge of the extent of disease is important for determining treatment, so a minimum of five to six biopsies covering different segments of the large intestine is required. Ulcerative colitis is described classically as always involving the rectum and extending for a variable distance into the proximal colon, but in some cases this distribution is not seen. Waxing and waning of the disease and the effect of partial treatment may result in sparing of the rectum and/or disease that shows a patchy distribution on endoscopic and/or histopathological examination.^{16 17} Furthermore, it is possible that the degree of inflammation assessed endoscopically may not completely correlate with the presence of inflammation within corresponding mucosal biopsies.¹⁶ In a patient in whom there is an otherwise confident diagnosis of ulcerative colitis, the presence of these features alone should not result in a change of diagnosis from ulcerative colitis to Crohn's disease. It is also entirely possible for mucosal biopsies from patients with quiescent IBD to have a histopathological appearance that is indistinguishable from normal.^{18 19} Therefore, the presence of microscopically normal colorectal mucosa does not preclude a diagnosis of previously active IBD.

What is the optimal biopsy technique for surveillance of chronic ulcerative colitis?

Patients with ulcerative colitis of greater than 8–10 years’ duration and/or those with pan-colitis, as opposed to more localised disease, are at the greatest risk of development of mucosal dysplasia and colorectal adenocarcinoma. Patients with both ulcerative colitis and Crohn's disease have an increased risk of colorectal cancer development of up to 20 times that of the normal population after this period of time.²⁰ The cumulative risk for development of colorectal cancer in patients with ulcerative colitis is 2%, 8% and 18% after 10, 20 and 30 years of colitis, respectively.²¹ The risk of cancer is further increased if coexistent primary sclerosing cholangitis is present, if there is a family history of colorectal cancer or if the inflammatory activity in the colitis is severe.²² Early detection of colorectal cancer or preinvasive mucosal lesions is vital to improve the outlook for these patients.

Traditionally, surveillance colonoscopy in longstanding ulcerative colitis will include obtaining random mucosal biopsies from the entire colon. This entails two to four random biopsies taken from every 10 cm segment of the colon, with further biopsies from suspicious areas.²³ The finding of low-grade flat dysplasia in random colorectal biopsies is associated with a >50% risk of the development of high-grade dysplasia or adenocarcinoma over 5 years.^{24 25} However, the detection rate of dysplasia is very low with this biopsy protocol. It has been suggested that 33 biopsies for each patient are required to produce a 90% confidence of excluding dysplasia,²⁶ which clearly becomes costly and time-consuming for endoscopy and histopathology departments. Furthermore, adherence to such biopsy protocols is variable and poor.

More recent data have indicated that advanced endoscopy methods—for example, dye spraying and narrow band imaging—increase the detection of dysplasia about threefold compared with conventional methods. Comparison of non-targeted biopsies with a targeted approach revealed no dysplasia within 2904 non-targeted biopsies whereas pan-colonic chromoendoscopy resulted in just 157 targeted biopsies but detected nine dysplastic lesions, seven of which were only visible after indigo carmine application.²⁷ Chromoendoscopy and confocal microscopy have been reported to increase the detection rate for dysplastic lesions 4.75-fold, but with 50% fewer biopsies. We believe that these targeted methods of biopsy must represent the future for the surveillance of patients with ulcerative colitis.^{28 29} However, it must be emphasised that a near-perfect bowel preparation is required and sufficient time must be available for the examination. Furthermore, the identification of mucosal lesions even with these advanced techniques can be challenging.

The current (2010) BSG guidelines for colorectal cancer screening and surveillance in moderate and high risk groups (updated from 2002) recommend pan-colonic dye spraying (usually with indigo carmine) to detect abnormal areas, with target biopsies for dysplasia or cancer.³⁰

Differentiating sporadic adenomas from colitis-associated lesions (ie, dysplasia-associated lesions or masses) can be problematic. A correlation has been identified between the site of colonic inflammation and the subsequent development of adenocarcinoma within an associated polypoid lesion.³¹ The development of endoscopic mucosal resection has made it possible to resect sessile polypoid lesions that previously would have required at least a segmental colectomy for removal and full evaluation. Conservative management does not appear to be associated with an increased risk of recurrence or further neoplasia.^{32 33} Therefore, it has now become possible to manage more patients with ulcerative colitis non-surgically when dysplastic lesions are identified. The following categories have previously been recommended for the management of dysplastic lesions arising in patients with ulcerative colitis³¹:

1. A polypoid dysplastic lesion not in an area of the colon affected by colitis

   These lesions can be removed by simple polypectomy.

2. A polypoid dysplastic lesion within an area of the colon affected by colitis

   These lesions should be removed by polypectomy and the surrounding mucosa should be biopsied to look for evidence of flat dysplasia.

3. A non-polypoid dysplastic lesion within an area of the colon affected by colitis

   These lesions should be removed locally if this is technically possible and the surrounding mucosa should be biopsied to look for flat dysplasia. Some of these patients may progress to colectomy if they are shown to possess widespread dysplasia within the colon.

4. Indeterminate lesions that do not fit within one of the above categories

   These lesions should be biopsied or removed locally if technically possible and the surrounding mucosa should be biopsied to look for flat dysplasia. The patient should then be followed up carefully.

What is the role of biopsy in the management of benign and malignant lesions?

The Bowel Cancer Screening Programme has recently advocated adoption of the Paris classification for reporting polyps identified at colonoscopy (table 2).³⁴ Closed biopsy forceps have been used to estimate the distinction between Is and IIa lesions. Usually, polyps should be excised and not biopsied. Diminutive polyps—for example, <2 mm—can be removed by biopsy. Larger polyps require snare excision (figure 1). The DISCARD study suggested that it was possible to diagnose polyp type by optical endoscopic assessment with a degree of accuracy that would allow biopsy samples to be discarded and deem histology unnecessary.³⁵ Further work is underway examining the feasibility of discarding tiny polyps on the basis that the yields of high-grade dysplasia and cancer are low in that subgroup. More recently, a community study in the USA showed that only 25% of gastroenterologists assessed polyps with ≥90% accuracy by this method, suggesting that this strategy is not yet widely applicable.³⁶ However, when multiple minute hyperplastic polyps are confined to the rectum, optical assessment and selection of which polyps to biopsy or resect is justifiable.

Table 2.

The Paris classification of superficial (‘type 0’) colorectal lesions

Basic subdivision	Paris category	Description
Polypoid lesions—that is, lesions protruding from the mucosal surface	0–Ip	Pedunculated
	0–Is	Sessile
Non-polypoid lesions	0–IIa	Slightly elevated
	0–IIb	Flat
	0–IIc	Slightly depressed
	0–III	Excavated (ulcerated)

If a polyp is endoscopically malignant, the overall management strategy is then based on the decision of whether to remove the lesion using an endoscopic method or via surgical resection. Biopsies will only provide a representative sample for assessment and, even if no invasive disease is identified within them, the possibility that adenocarcinoma exists elsewhere in such a lesion cannot be excluded entirely by biopsy alone.³⁷ This becomes important when assessing a large, mainly benign, polyp, but a potential small malignant area which should be the focus of biopsy is identified. Furthermore, there is a small risk of tissue ‘carry-over’ between specimens and possible diagnostic confusion if a biopsy is taken from a cancer before a benign polyp is biopsied or removed.

The development of advanced endoscopic mucosal resection techniques resulted in a major increase in endoscopy and away from surgery in the management of benign large polyps. This has led to the recognition that biopsies/partial resection to confirm benign adenomatous polyps can lead to submucosal fibrosis making endoscopic resection difficult, with increased risk, or impossible. Particularly when dealing with large flat laterally spreading polyps, biopsies from the flat portions of the polyp should be avoided and formal endoscopic mucosal resection performed instead.³⁸ It would be preferable when faced with a large polyp, to take photographs and/or a video with enhanced imaging (figure 1) or dye spray and then to discuss the case with an endoscopist who could remove the polyp confidently.