Endoscopy Practice and Safety
Editor: Peter B. Cotton ed.
David N. Lewin
A gastroenterologist need not be an expert histopathologist or cytopathologist (hopefully they have one in their local pathology
department); however, they should understand the importance of gastroenterologistpathologist communication , specimen handling, and interpretation issues, and the relative efficacy of endoscopic sampling methods for a particular
clinical situation. Attention to quality of endoscopic biopsy plays a crucial role in maximizing the effectiveness of this
Histopathological and cytological examination of the gastrointestinal tract is reviewed. This chapter is not intended to be
a comprehensive review of all gastrointestinal pathology; however, a number of general concepts are examined and the major
gastrointestinal lesions are illustrated. The chapter has been divided into endoscopic biopsy specimens (histology), exfoliative
and fine-needle aspiration specimens (cytology), and an overview of the major conditions in each organ system. Additionally,
general comments pertaining to communication, special stains, and future trends are all presented.
Effective gastroenterologistpathologist communication drives the ability to make accurate diagnoses. Communication can take many forms, from a written
request to a verbal consultation. Communication clearly is a 'two-way street', with information communicated from the gastroenterologist to the pathologist and vice versa. There are two critical parts
of this communication:
1. Communicating adequate, concise information.
2. Using a common terminology.
Endoscopist communication responsibility
The pathologist needs the typical important patient information, including relevant clinical history, bowel preparation, and
current medications. Additionally, the endoscopist assumes the responsibility for two tasks normally completed by the pathologist
in autopsy and 'traditional' surgical pathology: doing the gross examination, via the endoscope, and taking adequate representative samples. Thus the
endoscopist needs to communicate the results of the gross (endoscopic) examination and where the biopsies were taken from.
More information about each organ system will be discussed later.
Pathologist communication responsibility
A pathology diagnosis and report is only helpful to the clinician if it is reproducible and useful. The pathologist should
provide feedback regarding the artifacts and adequacy of a specimen for interpretation. The pathologist must also resist the
urge to find an abnormality, no matter how subtle, in a biopsy specimen. A diagnosis with no subsequent clinical action plan
One method is to use a question-orientated approach. This will ensure that specific information is included in the pathology
report and that the endoscopist receives the answers to the specific question for which they are performing the endoscopy.
The clinician can provide simple focused questions for the pathologist to answer. For example, in a patient with long-standing
ulcerative colitis undergoing screening colonoscopy, the primary question is whether or not there is dysplasia present and
not to make a diagnosis of inflammatory bowel disease. The pathologist may agonize over trying to make a diagnosis of inflammatory
bowel disease, and not address the question of dysplasia if not asked.
Probably the most crucial aspect of communication is using a common terminology, the implications of which are understood
by all parties. While there have been a number of attempts to create standardized terminology for both endoscopy and pathology
, these are not universally used. Terminology may be different in various parts of the United States and is clearly different
between different countries . For the endoscopist this means describing the endoscopic exam using simple, descriptive terms. Jargon should be avoided.
For example, the term 'gastritis' can imply anything from colour changes to erosions. Endoscopic photographs of lesions for the pathologist can potentially
help mitigate this problem. For the pathologist this means using accepted pathological diagnoses. Non-standard diagnoses should
be avoided or at least explained with a comment. For example, in a study of inflammatory bowel disease  non-specific inflammation was used by various pathologists to mean the following:
1. Confident of IBD or irradiation colitis, but cannot discriminate.
2. Cannot discriminate between ulcerative colitis and Crohn's disease.
3. Probably Crohn's disease.
4. No histological evidence of inflammatory bowel disease.
Terms such as atypia need to be qualified, usually with a reassuring qualifier term such as 'reactive' or 'inflammatory', as to many clinicians these are often synonymous with dysplasia, which is not typically true in the United States, but is
true in Japan .
Endoscopic biopsy specimens
Endoscopic specimens comprise the majority of the pathology specimens received from gastroenterologists. Most are obtained
with pinch biopsy forceps and contain mucosa (epithelium and the lamina propria) down to the muscularis mucosa. Endoscopic
biopsy is a very good technique for diagnosing mucosal abnormalities; however, it is very poor for submucosal (and deeper)
lesions . Polypectomy and endoscopic mucosal resections will be discussed subsequently. There will be a brief overview of specimen
handling and interpretation issues, followed by major conditions in each organ site.
Specimen handling and interpretation issues
The handling of the specimens, both in the endoscopy suite and by the pathology laboratory, play a large role in the ability
of the pathologist to render an accurate diagnosis. Some of the major issues: orientation of the specimen, fixation, number
of biopsies per container, tissue processing, crush artifact, burn/cautery artifact, are discussed. Many of the interpretation discrepancies between pathologists occur secondary to poor biopsy
Orientation of mucosal biopsy is crucial for the pathologist. Ideally the pathologist would like to see the full thickness
of the mucosa (from surface epithelium to the muscularis mucosa below) (Fig. 1) . Tangential sections (Fig. 2) are much more difficult to interpret . This is especially the case in those instances of dysplasia in Barrett's esophagus, where one of the major criteria is nuclear
atypia that extend to the surface epithelium. Well-orientated specimens will do wonders to your pathologist's ability to make
Orientation of the biopsy can be accomplished in the endoscopy unit . This requires the endoscopy assistants to orientate and mount biopsies mucosal side up on some sort of mounting media. This
is done by relatively few centers and is probably not necessary. It can cause problems if done by untrained individuals. The
biopsies are very easy to crush and the surface epithelium can be dislodged easily. If this is something one is considering,
one needs good communication with the pathology laboratory to choose a mounting media and get continual quality control feedback.
Well-trained histotechnologists in the pathology laboratory can orientate most biopsies, assuming they are adequate (do not
put puny specimens into the fixative). Biopsies should be transferred from the forceps by using a needle (or toothpick) to
push the biopsy out of the forceps into the fixative. Shaking the biopsy off the forceps into the fixative can cause trauma
to the tissue. One can lose the surface epithelium very easily.
Fixation of the biopsy is very important. Biopsies should be transferred to the fixative immediately and not allowed to air
dry. This will prevent air drying artifacts, which can be very difficult to interpret. There are a number of different fixatives
used by pathology laboratories and each has its advantages and disadvantages. In general, your pathologist will have their
standard fixative and it is best to stick with that fixative. Every fixative has its own artifacts and changes the histology
slightly. The most common fixative used is formalin; however, many labs prefer to use Bouin's or Hollande's (a modification
of Bouin's solution). Bouin's creates less shrinkage artifact and gives better nuclear detail; however, it does leach out
granules of eosinophils and disposal is more difficult.
Number of biopsies per container
There is a great temptation to put numerous biopsies in a single specimen container, primarily because most pathology laboratory's
charge based on the number of containers. I have received ulcerative colitis surveillance cases with 20 biopsies in a single
container. There are a number of problems with this approach. (1) It is really not possible to line up more than four biopsies
in a tissue block (Fig. 3) and an attempt to put more biopsies in a tissue block will result in many of the biopsies being uninterpretable (Fig. 4). (2) An additional problem is that there is no way of identifying where the biopsy came from if it is not separately labelled.
If 1 of 20 biopsies has dysplasia, without the proper label there is no way to go back and take additional biopsies if needed.
One approach we have taken in our institution is to make multiple blocks from a single container if it contains more than
four biopsies. There is an increased technical charge from the lab, but not an increased professional charge.
Tissue processing is a component of the pathology laboratory and typically not under the purview of the endoscopist. This
also plays a crucial role in the ability of the pathologist to make an accurate diagnosis. There are a number of technical
flaws that are readily preventable. Problems that can be seen include misorientation of the biopsy specimen, sections that
are too thick, folding of the tissue (Fig. 5), and shattering of the tissue (Fig. 6) by either dull blades or overcooling of the tissue block. The sections should include multiple ribbons (Fig. 7) through the core of the tissue biopsy to allow adequate appraisal of the tissue.
A variety of different bowel preparation solutions have been reported to cause a number of different histological changes
in colonic mucosa. Oral sodium phosphate bowel preparation can cause aphthoid ulcers, focal active colitis [11,12], and rare reports of ischemic colitis . It also appears to induce colorectal crypt proliferation. Sodium phosphate enema has been reported to cause sloughing of
the surface epithelium. Magnesium citrate and 'X-prep' senna derivative appears to flatten the surface epithelial cells, cause mucin depletion, and increase lamina propria edema
. Anthraquinone glycoside  preparations appear to cause epithelial proliferation. Bisacodyl  has been reported to cause sloughing of the surface epithelium. Information given to the pathologist about the type of colon
preparation can be very helpful in interpreting the histological changes identified.
Endoscopic trauma can also induce artifacts seen on histology. Typically these are mild and difficult to differentiate from
prep-induced artifacts. They consist of edema and hemorrhage present in the lamina propria.
Every endoscopic biopsy has some crush artifact, typically at the periphery of the biopsy (Fig. 8) [18,19]). One would like to minimize that amount of tissue crush by minimizing the handling of the specimen (except in those institutions
that mount the tissue biopsy), not shaking the biopsy off the forceps or pushing the biopsy out of the cup from the top.
Endoscopic biopsies done with electrocautery current (hot biopsies and polypectomies) will cause cautery artifact (Fig. 9). This can be very useful in identifying the margin of resection for polypectomy specimen; however, depending on the amount
and type of cautery used it can make the tissue uninterpretable. Typically the cells and nuclei elongate (Fig. 10) with electrocautery and the differentiation between dysplastic adenomatous epithelium and hyperplastic epithelium becomes
Identification of the polyp stalk (Fig. 11) is an important aspect of interpreting these specimens. (Fig. 12) . After polypectomy and with fixation, the stalk retracts and can 'disappear' by the time it is received in the pathology laboratory. Ideally the specimen should be bisected or trisected by the pathologist
through the stalk; however, if they cannot identify the stalk, this may not be done. There are a couple ways of identifying
the stalk for the pathologist. One is to impale the polyp with a short 25-gauge needle, while a second is to put a drop of
india ink on the polyp stalk after removal. I prefer the second method, for reasons of safety in the pathology laboratory.
With the needle in the specimen there is always the risk of the pathologist blindly sticking themselves with the needle while
handling the specimen.
Endoscopic mucosal resection
Similar to polypectomy specimens, the margins of endoscopic mucosal resections (Fig. 12) need to be assessed . The pathologist needs to be informed that the specimen is an endoscopic mucosal resection and the margin needs to be assessed
(if a single intact specimen is procured). Typically the pathologist will use india ink to assess the margin microscopically
(Fig. 13). However, if the specimen is removed in a piecemeal fashion, the margins of the resection will be difficult, if not impossible
to assess. If these types of specimens are not submitted regularly to the pathologist, a discussion about how the specimen
is procured needs to occur with the pathologist prior to the first case.
Core biopsies are typically handled by surgical pathologists, although they are starting to be used as an adjunct to fine-needle
aspiration [24,25]. The indications are typically the same as for fine-needle aspiration. Where they appear to have the greatest utility is
in mesenchymal tumors (GI stromal tumors, leiomyomas, etc.), which are typically poorly sampled with traditional fine-needle
aspiration. They are handled by the pathology laboratory similar to endoscopic biopsies and should be fixed in the usual fixative
for endoscopic biopsies.
Most biopsy specimens, with the exception of cytology and fine-needle aspiration specimens, will be stained routinely with
a hematoxylin and eosin (H&E) stain (Fig. 14). The hematoxylin stains the nuclei of the cell blue while the eosin is a counter-stain that stains the cytoplasm of the
cell pink. This is the standard stain used by the pathologist for all tissues. Additional (special) histochemical and immunohistochemical
stains can also be performed and will be discussed later.
Exfoliative and fine-needle cytology
Cytology plays an important role in the evaluation of many gastrointestinal processes. Its most important use currently is
in the evaluation of the pancreato-biliary tree and as an adjunct to endoscopic ultrasound in the evaluation of mass lesions.
There are a number of different sampling techniques that are used. Aspiration of fluid, brushing a lesion or duct, and fine-needle
aspiration can all obtain cells for cytological analysis. Key to all these techniques is the use of imaging (visual, ultrasound,
fluoroscopic) to guide the needle or brush to the area of interest. If the needle is not in the mass, or the brush not on
the tumor, the entire procedure is completely worthless and may even cause harm. Sampling is one of the key components in
Once the specimen is obtained or reaches the laboratory there are a number of different preparation methods. Slides can be
made by direct application of the material to a slide. A brush can be smeared onto a slide (Fig. 15), or during a fine-needle aspiration a drop of material can be placed on a slide and then a second slide used to smear the
material onto the slide (Fig. 16). Alternatively, some newer techniques are available that put the material into a fixative (one example is ThinPrep) which
is transported to the lab and a monolayer smear of the material is made (Fig. 17). This has the advantage of avoiding some of the artifacts to be discussed subsequently (especially air drying, crush artifact,
and obscured smears).
Specimen handling; staining and fixation
Two stains are primarily used in cytology, Papanicolaou (Pap) stain and the Romanovsky stain. Each has advantages and provides
different information. Papanicolaou stain is the primary stain of cytology (used in evaluating pap smears of the cervix).
The cells are first fixed with alcohol, which is a coagulative fixative. This causes the cells to contract and appear more
rounded, almost spherical, like a boiled egg. The stain contains three cytoplasmic dyes: orange G, eosin Y (pink), light green,
and hematoxylin (dark blue) nuclear stain (Fig. 18). With the Romanovsky stain (Fig. 19) the slides are allowed to air dry, which tends to spread out and flatten the cells, similar to a fried egg. The stain is
a metachromatic stain (also used for blood smears) that is relatively rapid and often used for rapid assessment. The Papanicolaou
stain is a transparent type of stain and allows wonderful assessment of the nuclear detail. The Romanovsky stain, on the other
hand, provides nice staining of the cytoplasmic cellular features, background substances, mucin, and microbiological organisms.
The adage in cytology is that the nuclear features allow one to determine if a cell is benign or malignant, while the cytoplasmic
features tell one where the cell originated.
A number of artifacts can be induced (usually at the time of collection) that makes a cytological diagnosis impossible. Air
drying (Fig. 20) of a Papanicolaou-stained smear occurs when the smear is not placed into fixative quickly enough. The cell nuclei may enlarge
and take on the appearance of malignant cells. Cytoplasmic distortion also occurs. Crushing of the cells (Fig. 21) may occur when spreading the cells on the slides. Certain cell types (lymphocytes and small cell carcinoma) are especially
vulnerable to crush artifact. With crush, the cells lose their cytoplasm (naked nuclei) and the nuclei are elongated, precluding
cytological evaluation. Obscured smears occur when the material is not smeared adequately or in very bloody specimens. Ideally
a specimen should have tissue and very little blood. Excessive blood or cells on top of one another (Fig. 22) makes it very difficult to analyse the features of a single cell. Displaced epithelium (Fig. 23) refers to cells being picked up from another site. Virtually every fine-needle aspiration will pick up cells as one goes
through the gut wall, even with the stylet in place. It is for this reason that one does not biopsy lymph nodes through the
tumor. A positive node might be the result of displaced epithelium and not a true positive lymph node. We looked at a number
of test needles with a very tight gap, but were still unable to avoid displaced epithelium. Displaced epithelium can also
contaminate brush specimens.
The final process for the pathologist/cytologist is the diagnosis. One should remember that a diagnosis is just another name for an opinion. The diagnostic categories
typically used are unsatisfactory, benign, reactive, atypical, suspicious, and malignant regardless of site of origin. The
diagnostic categories can be viewed statistically, although it depends on the site biopsied and the likelihood of a positive
test. Also remember that sampling errors will come into play as well.
Unsatisfactory (Fig. 24): Insufficient tissue to make a diagnosis or the artifacts obscures the cells.
Benign (Fig. 25): The cells on the slide are not malignant.
Reactive (Fig. 26): The cells on the slide have some changes not seen in normal cells, but are unlikely to be malignant.
Atypical (Fig. 27): The cells have more unusual features than seen in reactive cells but are not definitely malignant. About 60% of these cases
will be malignant.
Suspicious (Fig. 28): The cells are worrisome for malignancy. Often this diagnosis is used when there are too few cells to feel comfortable with
a malignant diagnosis. About 90% of these cases will be malignant.
Malignant (Fig. 29): The cells have all the characteristics of malignancy; 99% of these cases will be malignant.
Figure 30 lists the typical features of malignancy by cytology.
This involves the use of a fine needle (22 gauge or smaller) to remove a sample of cells. There must be a target, 'mass', that can be identified by ultrasound. The needle typically has a stylet to diminish the amount of displaced tissue. Once
the needle is in the lesion, the stylet is removed, and suction is typically applied. A cutting motion is crucial with long
'piston-like' excursions and quick strokes (three strokes per second) . Aspiration by suction alone, with the needle at rest, is not sufficient to draw tissue into the needle; some aspirations
are performed with the cutting motion alone and no suction .
Major indications for FNA are mass lesions, hypertrophic folds, and lymph nodes. In general the diagnosis of metastatic tumor
in lymph nodes and primary epithelial, endocrine, and lymphoid malignancies is very good (with an accuracy of 8589% at our institution) . Where FNA is limited is in diagnosis of mesenchymal lesions. The cells in these lesions are difficult to aspirate (sticky)
and our diagnostic accuracy has been 38% . We have begun to experiment with EUS-guided core biopsy; however, it does not appear to be much better for these lesions
Organ system overview
Histological evaluation of mucosal biopsies of the esophagus (as elsewhere in the gastrointestinal tract) allows one to distinguish
between inflammatory, infectious, and malignant lesions. A few of the most common lesions are illustrated.
The information that the pathologist needs from the endoscopist is the location of the biopsy (typically given in centimeters
from the incisors); however, additional information is needed. The location of the gastroesophageal junction (or most proximal
location of the gastric folds) and location of the z-line (squamocolumnar junction) must also be provided in order for the
pathologist to accurately interpret the histopathological changes. Additionally, gross endoscopic findings should also be
conveyed to the pathologist.
Where and when to biopsy
Number of biopsies and where to biopsy depends on the type of lesion encountered and will be discussed with each specific
entity. A biopsy is typically performed to evaluate an endoscopic abnormality, patients with no endoscopic abnormality will
typically not have histological abnormalities either. Special care must be taken in evaluating erosions and ulcers, especially
in the setting of gastroesophageal reflux or Barrett's esophagus. Endoscopic biopsy can be useful to diagnose infectious lesions
such as cytomegalovirus (CMV) and herpes simplex. However, in the setting of Barrett's esophagus and GERD, I typically suggest
not to biopsy erosions or ulcers. Rather, treat the patient with proton pump inhibitors and bring the patient back for repeat
endoscopy. The reason for this is that the inflammatory changes are often extreme and can occasionally be interpreted by pathologists
as features of malignancy.
Gastroesophageal reflux disease
Endoscopic biopsy is probably not necessary for the diagnosis of GERD, with 24 h pH monitoring probably representing the gold
standard. The histological changes seen with reflux are erosions, acute inflammatory cells (neutrophils and/or eosinophils) in the surface epithelium, basal cell hyperplasia, and elongation of the rete papilla (Fig. 31). Additional features include capillary dilatation and numerous intraepithelial lymphocytes. Acute inflammatory cells are
the most specific feature; unfortunately they are seen in less than 40% of patients . Attention should be paid to the total number of eosinophils per high power field. Greater than 20 per high power field suggest
a diagnosis of eosinophilic esophagitis (Fig. 32) and not reflux . The basal cell hyperplasia and elongation of the rete papilla are sensitive for reflux (80% prevalence) (Fig. 33), however, not very specific (especially within 3 cm of the gastroesophageal junction) as they can be seen in asymptomatic patients .
The characteristic histological feature of Barrett's esophagus is the finding of goblet cells (intestinal metaplasia) within
the tubular esophagus (Fig. 34). These goblet cells are also found in association with cardia type mucosa and fundic type mucosa (typically seen in the
stomach). One can find intestinal metaplasia of the gastric cardia in association with Helicobacter pylori infection . Again, the site of the biopsy in relation to the gastroesophageal junction is crucial in the diagnosis. The goal of the
biopsy should be to establish a diagnosis of Barrett's esophagus, verify its length, and identify any dysplasia. Biopsies
should be taken from the following sites:
1. 2 cm below the gastroesophageal junction.
2. Gastroesophageal junction.
3. 2 cm intervals (four quadrant biopsies).
4. Proximal margin (z-line).
5. Target lesion, nodules, plaques.
Histologically the goblet cells have an acid mucin (as opposed to the neutral mucins of the stomach) that can be identified
using an Alcian blue stain at pH 2.5 (Fig. 35).
The primary infections identified histologically are candida, herpes simplex virus (HSV) and CMV. Each organism infects a
different cell or region within the mucosal biopsy and may require a biopsy from a different area within the esophagus.
Candida (Fig. 36) typically appears as raised white plaques. The organism is identified in the superficial squamous epithelium, often associated
with a neutrophilic inflammatory response. These areas may be lost if one attempts to orientate or mount these biopsies, so
these biopsies should just be placed into fixative without attempt at orientation. Histochemical stains (periodic acidSchiff or Gomori's methenamine silver stains) may be used to better identify the organisms.
Herpes simplex virus
Herpes simplex virus (Fig. 37) infects the squamous epithelial cells and will be best identified at the edge of an ulcerative lesion. The histological
features are multiple intranuclear inclusions with nuclear molding and margination of the chromatin. An immunohistochemical
stain can be used to highlight the virus.
CMV (Fig. 38) typically infects endothelial cells and thus can be best found in the center of ulcers . The diagnostic cell is an enlarged cell with a single 'owl's eye' eosinophilic intranuclear inclusion. Additionally, numerous small, basophilic, cytoplasmic inclusions may also be seen.
Again, an immunohistochemical stain can confirm the presence of CMV.
Adenocarcinoma and squamous cell carcinoma
Diagnosis of esophageal carcinoma using endoscopic biopsy is highly accurate. As few as two biopsies from the lesion will
give a diagnosis in 96% of the cases . The two major carcinoma types are squamous cell carcinoma (Fig. 39) and adenocarcinoma (Fig. 40). Squamous cell carcinoma will typically have intercellular bridges, ample eosinophilic cytoplasm, and evidence of keratinization,
while adenocarcinomas will form glands and have mucin production. Typically, in the well and moderately differentiated tumors
there is no difficulty differentiating the two entities. However, the poorly differentiated tumors (Fig. 41) can pose a diagnostic dilemma and there may be no way to differentiate the two types on the basis of the biopsy.
Where and when to biopsy
The important landmarks are the region of the stomach from which the biopsy came: antrum, body/fundus, cardia, greater curvature, lesser curvature, and angularis. This is especially important for the pathologist to determine
the degree of glandular atrophy as the thickness of the glands is different in the various areas of the stomach. It is at
its thickest in the mid portion of the body along the greater curvature. Additional useful information for the pathologist
is drug therapy that the patient has been on. Specifically, the use of antibiotics for H. pylori eradication (which will make the hunt for organisms much more difficult and may suggest to the pathologist to do one of the
many special stains for H. pylori), non-steroidal anti-inflammatory drugs (NSAIDs), and proton pump inhibitors. If a lesion is identified in the stomach and
a biopsy taken of the lesion it is also worthwhile taking a second (separately labelled) biopsy of the adjacent 'normal' mucosa to see the 'soil' in which the lesion is arising.
Inflammatory conditions; gastritis
Inflammatory and reactive changes of the gastric mucosa compose the bulk of gastric biopsy findings encountered in the surgical
pathology laboratory. The most common causes of gastric injuries are H. pylori and NSAIDs. There are numerous different nomenclatures. The most recent proposal is the Houston gastritis workshop , which expanded on the previous Sydney classification and takes into account the topographical distribution, histology, and
etiological factors. In general, gastric lesions can be divided into three broad categories: acute, chronic, and special forms.
Each of these categories is then subdivided by specific causes, distribution, and histological characterization. With this
classification system, clinical information and interaction of the pathologist and endoscopist is crucial. Lesions are also
characterized as either a gastritis, where the inflammatory infiltrate predominates the histological features, or gastropathy,
where the epithelial or vascular changes predominate.
The typical finding as a result of NSAID injury is reactive gastropathy (Fig. 42). Histologically one sees foveolar hyperplasia, cytologically reactive epithelial cells, edema, and congestion of the lamina
propria, and smooth muscle proliferation in the lamina propria. There may also be surface erosions. The inflammatory infiltrate
is typically minimal, although neutrophils may be present adjacent to an erosion.
H. pylori has a characteristic histological features (Fig. 43). The H. pylori organisms can be identified on the surface epithelium (either on the basic H&E or using special stains  such as Giemsa, Genta, modified WarthinStarry, and an immunohistochemical stain). There typically is a neutrophilic inflammatory infiltrate both in the lamina propria
and glandular epithelium (active inflammation), an increased mononuclear (lymphocytes and plasma cells) infiltrate in the
lamina propria, and lymphoid follicles with reactive germinal centers. Glandular atrophy and intestinal metaplasia may also
be present. In areas with extensive intestinal metaplasia it may be more difficult to identify the H. pylori organisms, as they do not live in the acid mucins of the intestinal metaplastic cells.
In patients with enlarged (hypertrophic) folds the differential diagnosis  includes infiltrating malignancy (carcinoma [Fig. 44] and lymphoma [Fig. 45]), infections, granulomatous inflammation (Fig. 46), and primary epithelial hyperplasia. Endoscopic biopsy is usually adequate for the evaluation of this lesion, although occasionally
endoscopic ultrasound fine-needle aspiration may help in the evaluation. If lymphoma is a concern, additional biopsies can
be obtained to use for flow cytometry, which will better evaluate the lymphoid process. In general, this is probably best
used in those patients who have a diagnosis of an atypical lymphoid infiltrate and a second endoscopy is scheduled to obtain
tissue for flow cytometry.
A number of neoplastic and non-neoplastic lesions can give rise to polyps within the stomach. These can additionally be classified
as epithelial or non-epithelial. Biopsy is relatively accurate, except in the submucosal mass type lesions. A biopsy of the
adjacent mucosa is often very helpful to identify whether the lesion is arising in a general pathological process or is focal.
It also serves as a control when the architectural abnormality of the polyp is relatively subtle. The most common polyp encountered
in the antrum is a hyperplastic polyp (Fig. 47), while the most common encountered in the gastric body is a fundic gland polyp (Fig. 48). Both of these lesions are non-neoplastic. Neoplastic polyps (adenomas) are relatively uncommon; however, they should be
removed in their entirety.
Tumors of the stomach are among the most common neoplasms in the world, and the majority are epithelial (adenocarcinomas)
(Fig. 49). In addition, neoplasms of endocrine (Fig. 50), lymphoid (Fig. 51), and mesenchymal (Fig. 52) tissue occur in the stomach less commonly. In general, endoscopic biopsy will diagnosis the majority of these lesions, the
exception being mesenchymal tumors. These lesions often arise in the submucosa or muscularis propria and are not sampled by
endoscopic biopsy. Fine-needle aspiration may be more helpful in these cases.
Biopsies of the small bowel are typically obtained from the duodenum or terminal ileum, which is the reach of standard upper
and lower endoscopy. In general these biopsies are done in the work-up of diarrhea and malabsorption. Orientation of the small
bowel biopsies in this case is very important as evaluation of the villus-to-crypt ratio plays a prominent role in the pathological
classification of small bowel lesions (Fig. 53). Duodenal biopsies should be taken distal to the duodenal bulb. Biopsies from the bulb may be difficult to interpret secondary
to artifactual blunting of the villi by the Brunner's glands. Most findings in the bulb are secondary to peptic-acid disease
and there is often a 'mild chronic duodenitis'. Usually 23 biopsies obtained with a regular-sized forcep from the second or third part of the duodenum should be sufficient for evaluation.
Small bowel mucosal lesions are categorized as specific (diagnostic) or non-specific and the severity of mucosal injury is
subjectively expressed as the degree of abnormality (or loss) of the villous architecture. Normal small bowel should have
a villus-to-crypt ratio of at least 3:1, fewer than 25 intraepithelial lymphocytes , and one should have at least 34 villi in a row to feel confident of the findings.
Celiac sprue is one of the more common lesions encountered in biopsies of patients with diarrhea and malabsorption. The histological
features are not specific; however, with the appropriate clinical information (improvement with gluten-free diet), they are
diagnostic. The typical histological lesion is absence of villi (total or subtotal villous atrophy), an abnormal surface epithelium,
increased numbers of intraepithelial lymphocytes, and an increased lymphoplasmacytic infiltrate in the lamina propria (Fig. 54). This typical flat mucosa represents one end of the spectrum of pathologic abnormalities in gluten-sensitive individuals.
The 'infiltrative lesion' of latent sprue  (more than 40 lymphocytes per 100 epithelial cells) characterizes the earliest recognizable pathological change.
Infectious lesions of the small bowel are typically categorized as specific (diagnostic) lesions with appropriate identification
of the organism. Common lesions identified include Whipple's disease, Mycobacterium aviumintracellulare, and Giardia.
Whipple's disease is a rare bacterial infection with Tropheryma whippelii that affects the small intestine in a diffuse fashion but can involve many other organ systems. The typical histological
feature  is massive infiltration of the lamina propria with foamy macrophages (Fig. 55). The macrophages contain the Whipple bacilli and will stain with a periodic acidSchiff (PAS) stain. Additionally, small collections of fat may be present in the lamina propria (the original name coined
by Whipple was lipodystrophy).
Mycobacterium aviumintracellulare have become more prominent in immunosuppressed patients . Histological evaluation reveals sheets of macrophages within the lamina propria. There is often thickening of the folds.
The macrophages are stuffed with organisms, which can be identified with acid fast stains (Fite) (Fig. 56).
Giardia is an extracellular parasite present on the surface epithelium, and easier to recognize in cytological smears. Histologically
the organisms appear slightly basophilic or grayish in the mucus adherent to villus tips or between villi (Fig. 57). They are approximately the size of an enterocyte nucleus and have a characteristic pear-shaped profile. The prominent paired
'owl's eye' central nuclei and indistinct flagella attached to the tapered end are usually visible only in smears. Trichrome or
Giemsa stains make the organism more prominent.
The majority of polypoid lesions in the duodenum consist of benign heterotopic or hamartomatous lesions. Heterotopic gastric
mucosa, including gastric surface metaplasia (Fig. 58) with or with out gastric gland, is the most common lesion . Brunner's gland nodules and hamartomas are also common in the first portion of the duodenum (Fig. 59). Neoplastic lesions (adenomas) do occur, predominately in the second portion of the duodenum. Many of these lesions are
found in patients with familial adenomatous coli . The criteria for the diagnosis of small intestine adenomas is the same as that for adenomas elsewhere in the gastrointestinal
tract, namely, the presence of dysplastic epithelium (Fig. 60). These may be virtually indistinguishable from adenomas elsewhere in the GI tract, but Paneth cells are often prominent.
Tumors of the small bowel are relatively uncommon, however; similar to the rest of the gastrointestinal tract the majority
are still epithelial (adenocarcinomas). Compared to the colon, the relative proportion of neoplasms of endocrine, lymphoid
, and mesenchymal tissue occur more commonly.
In general, endoscopic biopsy can diagnosis the majority of these lesions, the exception being mesenchymal tumors. These lesions
often arise in the submucosa or muscularis propria and are not sampled by endoscopic biopsy. Fine-needle aspiration may be
more helpful in these cases.
Colonoscopic endoscopic biopsy specimens are one of the most frequent specimens encountered in the pathology laboratory. Typically
they are obtained in the work-up of a patient with diarrhea or colon cancer screening, or the work-up of bleeding. A few clinical/endoscopic findings are typically found with these biopsies: a normal endoscopy in a patient with diarrhea, abnormal endoscopy
(redness, ulceration, friability), or polyp/mass lesion. Each will be addressed separately.
The diagnostic yield of lower endoscopy and biopsy is very good in the work-up of diarrhea. It is recommended that representative
biopsies be taken in all cases (both endoscopically normal and abnormal). Patel et al.  looked at the diagnostic outcome in 205 consecutive study patients with lower endoscopy for diarrhea. Of these, 152 (74%)
had a normal endoscopy and biopsy, 37 (18%) had a specific diagnosis based on endoscopy and biopsy (Fig. 61 reveals a list of the abnormal endoscopy and biopsy findings), and 16 (8%) had either an endoscopy or biopsy inconsistent
with the final clinical diagnosis. When the endoscopy diagnosis (typically colitis) was inconsistent with the final clinical
diagnosis the biopsies were normal (7 cases). The biopsy diagnoses that were inconsistent with the final clinical diagnosis
usually had a diagnosis of 'non-specific colitis' (7 of 9 cases).
Pathologists should not use the diagnosis of 'non-specific colitis'. Usually this is an overcall of normal colonic mucosa. The diagnosis of normal colonic mucosa or an unremarkable colonic
mucosal biopsy is an important one. It helps the endoscopist to rule out an inflammatory colonic lesion as the cause of the
diarrhea. There is often the feeling that the pathologist must make some sort of diagnosis, since the endoscopist took the
biopsy for 'some reason'. Unfortunately this gives the clinicians two options: treat everyone and over-treat the patients, or ignore the diagnosis
and miss those patients with true inflammatory disease. In a study  on the diagnosis of inflammatory bowel disease by a number of gastrointestinal pathologists, the diagnosis of non-specific
inflammation was used by different pathologists to mean 'confident of IBD or irradiation colitis, but cannot discriminate', 'cannot discriminate UC v CD', 'probably CD', and 'No histological evidence of IBD'. So one can see that the diagnosis can have many meanings to different pathologists and should not be used.
Biopsies of the normal colonic mucosa typically are comprised of parallel crypts lined by colonocytes, similar to test tubes
in a row if the biopsy is properly orientated. As in the small bowel there is a lymphoplasmacytic infiltrate in the lamina
propria and lymphoid aggregates (follicles) present . It is this normal baseline inflammatory infiltrate that is often over-interpreted as 'chronic non-specific colitis'. The inflammatory infiltrate is typically concentrated in the upper third of the mucosa (Fig. 62).
The majority of biopsy specimens taken in individuals with diarrhea with a normal endoscopy will be normal, however, approximately
110% will have a microscopic colitis . The term microscopic colitis is an unfortunate one as it typically encompasses the diagnosis of collagenous colitis (Fig. 63) and lymphocyte colitis (Fig. 64), but is occasionally used as a diagnosis on its own. In general the specific diagnoses should be used and not the ambiguous
term 'microscopic colitis'. Both lymphocytic and collagenous colitis can be seen in association with NSAIDs, celiac disease, and autoimmune diseases.
The etiology is unknown for both. The histological appearance for each is similar , with the differentiating feature being basement membrane thickening (greater than 10 µg) in collagenous colitis. This basement membrane thickening can be highlighted with a trichrome stain. The common histological
features are increased intraepithelial lymphocytes (IEL), increased numbers of eosinophils, flattening and detachment of the
surface epithelium, and an increased number of lymphocytes and plasma cells in the lamina propria. With collagenous colitis
the collagen thickening is patchy and typically best identified from biopsies in the transverse colon .
In patients with diarrhea and an abnormal endoscopy, the differential diagnosis includes the idiopathic inflammatory bowel
diseases (ulcerative colitis and Crohn's disease), ischemia, and infectious colitis. The endoscopic appearance of the bowel
and clinical history are very important in this differential diagnosis and need to be communicated to the pathologist.
Inflammatory bowel disease
The idiopathic inflammatory bowel diseases (IBDs) are chronic lesions and are characterized by chronic histological mucosal
findings. There is typically marked crypt architectural distortion (Fig. 65): irregularities of the crypts (loss of the typical parallel test tube architecture), crypt atrophy, lifting of the crypts
from the muscularis mucosa, and a basal plasmacytosis (plasma cells below the base of the crypts). These features are very
useful in differentiating IBD from acute self-limited colitis (ASLC) (Fig. 66), especially in adults. In children these features may be less prominent. There is also goblet cell mucin depletion, infiltration
of neutrophils and eosinophils in the lamina propria and glands, and reactive epithelial hyperplasia. However, these features
overlap with those of ASLC.
Differentiation of ulcerative colitis and Crohn's disease can be problematic on endoscopic mucosal biopsy. The characteristic
feature of Crohn's disease is the presence of epithelioid granulomas (Fig. 67). Unfortunately these are typically only seen in approximately 20% of biopsies from patients with Crohn's disease. Additionally,
microgranulomas and granulomas associated with a ruptured crypt can be seen in ulcerative colitis (and other lesions) (Fig. 68). The other useful differentiating histological feature is a microscopic patchiness to the inflammatory infiltrate (similar
to the endoscopic skip lesions) (Fig. 69). The best time to differentiate these two lesions is on initial endoscopy and biopsy with communication of the endoscopic
findings (diffuse vs. patchy, presence of rectal sparing) and representative biopsies of both normal (if present) and abnormal
areas. Like biopsies should be labelled and separately submitted. If a patient has patchy disease (suspicious for Crohn's
endoscopically), biopsies should be taken from the normal and abnormal areas and be put in separate containers. In biopsies
from a diffuse colitis (suspicious for ulcerative colitis), biopsies should be taken from the involved colon (no more than
four biopsies per container) and a separate container used with biopsies of the normal terminal ileum and proximal colon (if
present). Differentiating the two lesions after the initial biopsy becomes more difficult, as treatment of ulcerative colitis
can give patchiness, rectal sparing, and even histologicalal and endoscopic reversion to normal .
Pseudomembranous colitis (Fig. 70) can be seen associated with antibiotics (Clostridium difficile-associated colitis) and ischemia (Fig. 71). Pseudomembranes can be present with ischemia; however, in the majority of cases they are not present . Both lesions have a predominantly neutrophilic inflammatory infiltrate (as opposed to the prominent lymphoplasmacytic infiltrate
in the inflammatory bowel lesion) and a collection of neutrophils, fibrin, and necrotic cells making up the pseudomembrane
on the surface epithelium.
Ischemic colitis can be histologically very similar to pseudomembranous colitis, due to the presence of a pseudomembrane.
However, ischemia typically has a hyalinized lamina propria, is localized on endoscopy, and has atrophic crypts, often with
full-thickness mucosal necrosis. The hyalinization of the lamina propria (Fig. 71) is the most characteristic feature and best differentiates this lesion from pseudomembranous colitis.
Polyps are typically biopsied to identify patient risk for subsequent carcinoma and to remove the polyp to prevent it from
becoming a malignancy in the future . There are a variety of types of polyps; however, the focus here will be on the most commonly encountered: polyps adenoma
and hyperplastic polyps.
Adenomatous polyps are by definition neoplastic and dysplastic. All have a malignant potential; however, only a very few will
actually become a malignancy. A good demonstration of this is patients with familial adenomatous polyposis. These patients
have thousands of polyps, which usually begin to become apparent in the teens, and typically, by the age of 30 one or two
of the thousands of polyps will have become an invasive carcinoma. Thus with a given adenomatous polyp, the likelihood that
it will become malignant is low and it will take many years. All adenomatous polyps have dysplastic epithelium (Fig. 72) (characterized by nuclear enlargement and hyperchromasia) and thus appear basophilic (or blue) on the microscopic slide.
The lesion is confined by the basement membrane and there is no invasion into the adjacent tissue, so there is no possibility
of metastatic disease to the lymph nodes. Histologically the lesions are divided into tubular (Fig. 73), villous (Fig. 74), and mixed tubulovillous  based on the architecture of the polyp. Tubular lesions have crypt type architecture, while villous lesions mimic the finger-like
projections (villi) found in the small bowel mucosa. Traditional teaching is that there is a higher risk of carcinoma in villous
adenomas; however, this probably has more to do with the larger size of these lesions than the architecture seen. The degree
of dysplasia (low and high-grade) can be designated; however, for an individual polyp it has no prognostic value and is not
a feature upon which I typically comment . These lesions can be polypoid (with a stalk) or sessile. Recently, flat (Fig. 75) and depressed adenomas  have been described. These had been thought to perhaps have a higher incidence of carcinoma in smaller lesions; however,
this has not been confirmed by studies in the United States.
Hyperplastic polyps have typically been taken to be a benign lesion with no malignant potential. Histologically they have
epithelial hyperplasia with a serrated (sawtooth) appearance. On cross-section the glands look like a starfish (Fig. 76). By definition these lesions should not have any dysplasia. Recently a subset of these lesions has been described that appears
to have a malignant potential . These lesions are different from the typical adenomas from a histological and molecular standpoint. They appear to have
a defect in the DNA mismatch repair genes, the same defect found in patients with hereditary non-polyposis colon cancer (HNPCC)
and a minority of sporadic colon cancers . This is in contrast with the typical dysplasiacarcinoma sequence that begins with defects in the APC gene for most adenomas and carcinomas (and familial adenomatous polyposis).
The first lesion to be described was the serrated adenoma (Fig. 77). This lesion has the architecture of a hyperplastic polyp but with dysplastic nuclear changes. An additional lesion called
a sessile serrated adenoma (or sessile serrated polyp) (Fig. 78) has also been described without the dysplasia, but with more architectural changes than seen in the usual hyperplastic polyp.
Unfortunately these new lesions are still being described and the interobserver variability between pathologists is very high.
Long-term studies looking at these lesions have also not been performed, so the true meaning of these lesions is still being
The majority of mass lesions within the colon are adenocarcinomas (Fig. 79). One frustration for many endoscopists is that they will biopsy a mass lesion that they are certain is a carcinoma and the
pathology report will just refer to an adenoma or use words such as 'invasion cannot be excluded'. Many biopsies of these lesions will just obtain tissue from the surface of the tumor or from the pre-existing adenoma often
present at the edge of the lesion. The tissue obtained may not show the deeper invasive component of the cancer with its associated
desmoplastic reaction. The endoscopist should realize that this is not an indication for a second endoscopy to obtain 'diagnostic material'. Regardless of the diagnosis (adenoma vs. invasive carcinoma), in a mass lesion the patient is most likely going to have
the lesion resected. In truth the biopsy is really done to exclude a non-surgical etiology for the lesion, such as an infection
or a lymphoma. Thus a diagnosis of adenoma in a mass lesion should be referred for excision (endoscopic or surgical).
A number of special studies can be applied to tissue obtained for both histological and cytological evaluation. This includes
histochemical stains, immunohistochemical stains, in situ hybridization, flow cytometry, molecular pathology, cytogenetics, and electron microscopy.
These are the traditional stains used by pathologists for decades (H&E and Pap stains are examples). They essentially apply
chemistry and chemical reactions to a slide to highlight types of tissue of interest.
These are the application of antibodies to a specific protein with the use of some type of colorimetric (fluorescent or chromogen)
reporter to a slide.
In situ hybridization
This is similar to immunohistochemical stains; however, instead of using antibodies, complementary DNA or RNA is used to identify
a specific sequence on a slide.
In contrast to the techniques above, which are slide based, this involves measuring multiple characteristics of single cells
in a moving fluid stream. Cell size, internal structures, antigens, DNA ploidy, and cell cycle can all be analyzed. It requires
the specimen to be collected fresh or in a balanced media (RPMI) and run within 24 h of collection.
A technique that allows examination of intracellular architecture. It allows much higher magnification of individual cells
and can be helpful in determining the type of tumor. It is rarely done due to the availability of immunohistochemical stains.
These specimens should be fixed in glutaraldehyde fixative for optimum fixation.
In order to obtain a cytogenetic analysis one must grow the cells and get them to divide. Cytogenetics is essentially a gross
evaluation of the chromosomes of the cells. It allows one to look for major molecular abnormalities such as deletions, translocations,
and aneuploidy. Similar to flow cytometry, it requires fresh specimens. This technique is not typically used currently for
This includes PCR and extraction of DNA for various analyses. It is typically best done on fresh tissue, but can be performed
on formalin-fixed tissue. As the molecular bases of many diseases are becoming evident, this type of analysis is becoming
Outstanding issues and future trends
The future trends in pathology revolve around the further integration of the various 'omics' into the diagnostic and prognostic information generated based on tissue findings. The integration of proteinomics and genomics
will change the way we view pathology information in terms of risk stratification and management for individual patients.
Interestingly the more we learn about the molecular changes present in both inflammatory and malignant diseases, the more
they support the different phenotypic changes seen under the microscope by the pathologist.
The other major change is the improving imaging systems, toward the advent of the 'optical' biopsy . A number of different techniques are resulting in images approaching the quality and resolution of unstained histological
slides. The potential added information of the in vivo changes (as opposed to the static slide-based images a pathologist currently views) is very exciting. The challenge will
be the merger of the radiologistendoscopistpathologist into a single individual or multiple individuals working in concert. This has just begun, with the increased use
of endoscopic ultrasound and ERCP, and is sure to accelerate in the future.
1 Lewin, DN, Lewin, KJ & Weinstein, WM. Pathologistgastroenterologist interaction: the changing role of the pathologist. Am J Clin Pathol 1995; 103 (4) (Suppl. 1): S912. PubMed
2 Dixon, MF, Genta, RM & Yardley, JH et al. Classification and grading of gastritis: the updated Sydney System. International Workshop on the Histopathology of Gastritis,
Houston 1994. Am J Surg Pathol 1996; 20 (10): 116181. PubMed CrossRef
3 Lundell, LR, Dent, J & Bennett, JR et al. Endoscopic assessment of oesophagitis: clinical and functional correlates and further validation of the Los Angeles classification. Gut 1999; 45 (2): 17280. PubMed
4 Riddell, RH, Goldman, H & Ransahoff, DF et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983; 14 (11): 93168. PubMed
5 Lauwers, GY, Shimizu, M & Correa, P et al. Evaluation of gastric biopsies for neoplasia: differences between Japanese and Western pathologists. Am J Surg Pathol 1999; 23 (5): 51118. PubMed CrossRef
6 Theodossi, A, Spiegelhalter, DJ & Jass, J et al. Observer variation and discriminatory value of biopsy features in inflammatory bowel disease. Gut 1994; 35 (7): 9618. PubMed
7 Stamm, B, Heer, M & Buhler, H et al. Mucosal biopsy of vascular ectasia (angiodysplasia) of the large bowel detected during routine colonoscopic examination. Histopathology 1985; 9 (6): 63946. PubMed
8 Levine, DS & Haggitt, RC. Normal histology of the colon. Am J Surg Pathol 1989; 13 (11): 96684. PubMed
9 Lazenby, AJ, Yardley, JH & Giardiello, FM et al. Pitfalls in the diagnosis of collagenous colitis: experience with 75 cases from a registry of collagenous colitis at the Johns
Hopkins Hospital. Hum Pathol 1990; 21 (9): 90510. PubMed CrossRef
10 Lewin, KJ, Riddell, RH & Weinstein, WM (1992). Dialogue, handling of biopsies and resected specimens. Gastrointestinal pathology and its clinical implications, pp. 1518. Igaku-Shoin, New York.
11 Driman, DK & Preiksaitis, HG. Colorectal inflammation and increased cell proliferation associated with oral sodium phosphate bowel preparation solution. Hum Pathol 1998; 29: 9728. PubMed CrossRef
12 Zwas, FR, Cirillo, NW & El-Serag, HM et al. Colonic mucosal abnormalities associated with oral sodium phosphate solution. Gastrointest Endosc 1996; 43: 4636. PubMed
13 Oh, JK, Meiselman, M & Lataif, LE. Ischemic colitis caused by oral hyperosmotic saline laxatives. Gastrointest Endosc 1997: 31922.
14 Pockros, PJ & Foroozan, P. Golytely lavage versus a standard colonoscopy preparation: effect on normal colonic mucosal histology. Gastroenterology 1985; 88 (2): 5458. PubMed
15 Kliebeuker, JH, Cats, A & Zwart, N et al. Excessively high cell proliferation in sigmoid colon after an oral purge with anthraquinone glycosides. J Natl Cancer Inst 1995; 87: 4523. PubMed
16 Meisel, JL, Bergman, D & Graney, D et al. Human rectal mucosa: proctoscopic and morphological changes caused by laxatives. Gastroenterology 1977: 12749.
17 Ji, X, Shen, M & Jeng, Y et al. Diagnosis of mucosa-associated lymphoid tissue (MALT) lymphoma in gastroendoscopic biopsy specimens. J Environ Pathol Toxicol Oncol 1999; 18 (1): 613. PubMed
18 Woods, KL, Anand, BS & Cole, RA et al. Influence of endoscopic biopsy forceps characteristics on tissue specimens: results of a prospective randomized study. Gastrointest Endosc 1999; 49 (2): 17783. PubMed
19 Bernstein, DE, Barkin, JS & Reiner, DK et al. Standard biopsy forceps versus large-capacity forceps with and without needle. Gastrointest Endosc 1995; 41 (6): 5736. PubMed
20 Cooper, HS, Deppisch, LM & Gourley, WK et al. Endoscopically removed malignant colorectal polyps: clinicopathologic correlations. Gastroenterology 1995; 108 (6): 165765. PubMed
21 Coverlizza, S, Risio, M & Ferrari, A. et al. Colorectal adenomas containing invasive carcinoma: pathologic assessment of lymph node metastatic potential. Cancer 1989; 64 (9): 193747. PubMed
22 Masaki, T & Muto, T. Predictive value of histology at the invasive margin in the prognosis of early invasive colorectal carcinoma. J Gastroenterol 2000; 35 (3): 195200. PubMed CrossRef
23 Kondo, H, Gotoda, T & Ono, H et al. Early gastric cancer: endoscopic mucosal resection. Ann Ital Chir 2001; 72 (1): 2731. PubMed
24 Farmer, KD, Harries, SR & Fox, BM et al. Core biopsy of the bowel wall: efficacy and safety in the clinical setting. AJR Am J Roentgenol 2000; 175 (6): 162730. PubMed
25 Varadarajulu, S, Fraig, M & Schmulewitz, N et al. Comparison of EUS-guided 19-gauge Trucut needle biopsy with EUS-guided fine-needle aspiration. Endoscopy 2004; 36 (5): 397401. PubMed CrossRef
26 Smith, MJ, Kini, SR & Watson, E. Fine needle aspiration and endoscopic brush cytology: comparison of direct smears and rinsings. Acta Cytol 1980; 24 (5): 4569. PubMed
27 Young, NA, Mody, DR & Davey, DD. Misinterpretation of normal cellular elements in fine-needle aspiration biopsy specimens: observations from the College of
American Pathologists Interlaboratory Comparison Program in Non-Gynecologic Cytopathology. Arch Pathol Lab Med 2002; 126 (6): 6705. PubMed
28 DeMay, RM 1996 Fine needle aspiration biopsy. The art and science of cytopathology, pp. 46487. American Society for Clinical Pathology Press, Chicago.
29 Wallace, MB, Kennedy, T & Durkalski, T et al. Randomized controlled trial of EUS-guided fine needle aspiration techniques for the detection of malignant lymphadenopathy. Gastrointest Endosc 2001; 54 (4): 4417. PubMed CrossRef
30 Williams, DB, Sahai, AV & Aabakken, L et al. Endoscopic ultrasound guided fine needle aspiration biopsy: a large single centre experience. Gut 1999; 44 (5): 7206. PubMed
31 Ismail-Beigi, F, Horton, PF & Pope, CE. Histological consequences of gastroesophageal reflux in man. Gastroenterology 1970; 58 (2): 16374. PubMed
32 Winter, HS, Madara, JL & Stafford, RJ et al. Intraepithelial eosinophils: a new diagnostic criterion for reflux esophagitis. Gastroenterology 1982; 83 (4): 81823. PubMed
33 Ahmad, M, Soetikno, RM & Ahmed, A. The differential diagnosis of eosinophilic esophagitis. J Clin Gastroenterol 2000; 30 (3): 2424. PubMed CrossRef
34 Weinstein, WM, Bogoch, ER & Bowes, KL. The normal human esophageal mucosa: a histological reappraisal. Gastroenterology 1975; 68 (1): 404. PubMed
35 Goldblum, JR, Richter, JE & Vaezi, M et al. Helicobacter pylori infection, not gastroesophageal reflux, is the major cause of inflammation and intestinal metaplasia of
gastric cardiac mucosa. Am J Gastroenterol 2002; 97 (2): 30211. PubMed CrossRef
36 Goldblum, JR. The significance and etiology of intestinal metaplasia of the esophagogastric junction. Ann Diagn Pathol 2002; 6 (1): 6773. PubMed CrossRef
37 Gottfried, MR, McClave, SA & Boyce, HW. Incomplete intestinal metaplasia in the diagnosis of columnar lined esophagus (Barrett's esophagus). Am J Clin Pathol 1989; 92 (6): 7416. PubMed
38 Bonacini, M, Young, T & Laine, L. Histopathology of human immunodeficiency virus-associated esophageal disease. Am J Gastroenterol 1993; 88 (4): 54951. PubMed
39 Goodgame, RW, Genta, RM & Estrada, R et al. Frequency of positive tests for cytomegalovirus in AIDS patients: endoscopic lesions compared with normal mucosa. Am J Gastroenterol 1993; 88 (3): 33843. PubMed
40 Lal, N, Bhasin, DK & Malik, AK et al. Optimal number of biopsy specimens in the diagnosis of carcinoma of the oesophagus. Gut 1992; 33 (6): 7246. PubMed
41 El Zimaity, HM, Genta, RM & Graham, DY. Histological features do not define NSAID-induced gastritis. Hum Pathol 1996; 27 (12): 134854. PubMed CrossRef
42 Laine, L, Lewin, DN & Naritoku, W et al. Prospective comparison of H & E, Giemsa, and Genta stains for the diagnosis of Helicobacter pylori. Gastrointest Endosc 1997; 45 (6): 4637. PubMed
43 Chusid, EL, Hirsch, RL & Colcher, H. Hypertrophic gastropathy: a clinical review. Am J Gastroenterol 1965; 43: 30310. PubMed
44 Deppisch, LM & Rona, VT. Gastric epithelial polyps: a 10-year study. J Clin Gastroenterol 1989; 11 (1) (1989): 11015. PubMed
45 Kreuning, J, vd Wal, AM & Kuiper, G et al. Chronic nonspecific duodenitis: a multiple biopsy study of the duodenal bulb in health and disease. Scand J Gastroenterol Suppl 1989; 167: 1620. PubMed
46 Hayat, M, Cairns, A & Dixon, MF et al. Quantitation of intraepithelial lymphocytes in human duodenum: what is normal? J Clin Pathol 2002; 55 (5): 3934. PubMed
47 Weinstein, WM. Latent celiac sprue. Gastroenterology 1974; 66 (4): 48993. PubMed
48 Dobbins, WO III The diagnosis of Whipple's disease. N Engl J Med 1995; 332 (6): 3902. PubMed CrossRef
49 Roth, RI, Owen, RL & Keren, DF et al. Intestinal infection with Mycobacterium avium in acquired immune deficiency syndrome (AIDS): histological and clinical comparison
with Whipple's disease. Dig Dis Sci 1985; 30 (5): 497504. PubMed CrossRef
50 Brandborg, LL, Tankersley, CB & Gottieb, S et al. Histological demonstration of mucosal invasion by Giardia lamblia in man. Gastroenterology 1967; 52 (2): 14350. PubMed
51 Kreuning, J, Bosman, FT & Kuiper, G et al. Gastric and duodenal mucosa in 'healthy' individuals: an endoscopic and histopathological study of 50 volunteers. J Clin Pathol 1978; 31 (1): 6977. PubMed
52 Walk, L. Nodular hyperplasia of duodenal Brunner's glands: does it exist? Endoscopy 1982; 14 (5): 1625. PubMed
53 Jagelman, DG, DeCosse, JJ & Bussey, HJ. Upper gastrointestinal cancer in familial adenomatous polyposis. Lancet 1988; 1 (8595): 114951. PubMed CrossRef
54 Taggart, DP & Imrie, CW. A new pattern of histologic predominance and distribution of malignant diseases of the small intestine. Surg Gynecol Obstet 1987; 165 (6): 51518. PubMed
55 Patel, Y, Pettigrew, NM & Grahame, GR et al. The diagnostic yield of lower endoscopy plus biopsy in nonbloody diarrhea. Gastrointest Endosc 1997; 46 (4): 33843. PubMed
56 Goff, JS, Barnett, JL & Pelke, T et al. Collagenous colitis: histopathology and clinical course. Am J Gastroenterol 1997; 92 (1): 5760. PubMed
57 Lazenby, AJ, Yardley, JH & Giardiello, FM et al. Lymphocytic ('microscopic') colitis: a comparative histopathologic study with particular reference to collagenous colitis. Hum Pathol 1989; 20 (1): 1828. PubMed CrossRef
58 Offner, FA, Jao, FV & Lewin, KJ et al. Collagenous colitis: a study of the distribution of morphological abnormalities and their histological detection. Hum Pathol 1999; 30 (4): 4517. PubMed CrossRef
59 Nostrant, TT, Kumar, NB & Appelman, HD. Histopathology differentiates acute self-limited colitis from ulcerative colitis. Gastroenterology 1987; 92 (2): 31828. PubMed
60 Bernstein, CN, Shanahan, F & Weinstein, WM. Histological patchiness and sparing of the rectum in ulcerative colitis: refuting the dogma. J Clin Pathol 1997; 50 (4): 3545. PubMed
61 Kleer, CG & Appelman, HD. Ulcerative colitis: patterns of involvement in colorectal biopsies and changes with time. Am J Surg Pathol 1998; 22 (8): 9839. PubMed CrossRef
62 Levine, TS, Tzardi, M & Mitchell, S et al. Diagnostic difficulty arising from rectal recovery in ulcerative colitis. J Clin Pathol 1996; 49 (4): 31923. PubMed
63 Odze, R, Antonioli, D & Peppercorn, M et al. Effect of topical 5-aminosalicylic acid (5-ASA) therapy on rectal mucosal biopsy morphology in chronic ulcerative colitis. Am J Surg Pathol 1993; 17 (9): 86975. PubMed
64 Dignan, CR & Greenson, JK. Can ischemic colitis be differentiated from C difficile colitis in biopsy specimens? Am J Surg Pathol 1997; 21 (6): 70610. PubMed CrossRef
65 Winawer, SJ, Zauber, AG & Ho, MN et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993; 329 (27): 197781. PubMed CrossRef
66 Rubio, CA. Colorectal adenomas: time for reappraisal. Pathol Res Pract 2002; 198 (9): 61520. PubMed CrossRef
67 Terry, MB, Neugut, AI & Bostick, RM et al. Reliability in the classification of advanced colorectal adenomas. Cancer Epidemiol Biomarkers Prev 2002; 11 (7): 6603. PubMed
68 Yao, T, Tada, S & Tsuneyoshi, M. Colorectal counterpart of gastric depressed adenoma: a comparison with flat and polypoid adenomas with special reference to
the development of pericryptal fibroblasts. Am J Surg Pathol 1994; 18 (6): 55968. PubMed
69 Longacre, TA & Fenoglio-Preiser, CM. Mixed hyperplastic adenomatous polyps/serrated adenomas: a distinct form of colorectal neoplasia. Am J Surg Pathol 1990; 14 (6): 52437. PubMed
70 Torlakovic, E, Skovlund, E & Snover, DC et al. Morphologic reappraisal of serrated colorectal polyps. Am J Surg Pathol 2003; 27 (1): 6581. PubMed CrossRef
71 Nishioka, NS. Optical biopsy using tissue spectroscopy and optical coherence tomography. Can J Gastroenterol 2003; 17 (6): 37680. PubMed
Copyright © Blackwell Publishing, 2005