barrett esophagus: diagnostic challenges

Available online at www.sciencedirect.com

www.elsevier.com/locate/semdp

S E M I N A R S I N D I A G N O S T I C P A T H O L O G Y 3 1 ( 2 0 1 4 ) 1 0 0 – 1 1 3

http://dx.doi.org/10.0740-2570/& 2014 El

nCorresponding auE-mail address:

Barrett esophagus: Diagnostic challenges

Catherine E. Hagen, MDa,b, Gregory Y. Lauwers, MDa,b,Mari Mino-Kenudson, MDa,b,n

aDepartment of Pathology, Massachusetts General Hospital, Boston, MassachusettsbHarvard Medical School, Boston, Massachusetts

a r t i c l e i n f o

Keywords:

Esophageal adenocarcinoma

Intestinal metaplasia

Crypt dysplasia

Foveolar dysplasia

Carditis

1053/j.semdp.2014.02.005sevier Inc. All rights rese

thor at: Department of Pamminokenudson@partn

a b s t r a c t

The incidence of esophageal adenocarcinoma and associated mortality has risen dramat-

ically over the past several decades, and, thus, it is increasingly important to understand

its pathogenesis and risk factors. Barrett esophagus is the established precursor to

esophageal adenocarcinoma that progresses through a metaplasia–dysplasia-carcinoma

sequence. Its risk of transforming to carcinoma is not as high as previously reported and

there appears to be a biological heterogeneity among patients with this disease. The overall

prevalence of Barrett esophagus in the United States ranges from 1% to 25% and is closer to

5% in patients with gastroesophageal reflux disease. Because of the frequency of Barrett

esophagus and associated implications, it is important for the practicing pathologist to

have a thorough understanding of this disease and its diagnostic pitfalls. In this review, we

will discuss issues associated with the diagnosis of Barrett esophagus, including the

definition of Barrett esophagus and its distinction from carditis with intestinal metaplasia.

We will also discuss challenges in the grading of dysplasia and new variants of dysplasia,

including crypt dysplasia and foveolar-type dysplasia. Finally, we will touch upon the

evaluation of dysplasia in endoscopic mucosal resection specimens.

& 2014 Elsevier Inc. All rights reserved.

Introduction

The incidence of esophageal adenocarcinoma and associatedmortality has risen dramatically over the past several deca-des, increasing approximately 6% per year, which is far morethan most other cancers in the United States.1 Consequently,it is increasingly important to understand the pathogenesisand risk factors for esophageal adenocarcinoma. Barrettesophagus is the established precursor lesion to esophagealadenocarcinoma, although its cancer risk is not as high aspreviously reported. The annual incidence of cancer amongpatients with Barrett esophagus is approximately 0.5% peryear, and it is much lower (0.12%) in populations that are notplagued by obesity and other risk factors.2,3 The risk of

rved.

thology, Massachusetts Gers.org (M. Mino-Kenudso

developing dysplasia is also limited. The annual incidenceof low-grade dysplasia is 4.3% and that of high-grade dyspla-sia is 0.9%.4 The risk of cancer development varies withextent of disease. Patients with long-segment Barrett esoph-agus, as defined by the presence of metaplastic glandularmucosa affecting an esophageal segment spanning at least3 cm, comprise the minority of Barrett esophagus patientsand are at higher risk for cancer development than thosewith less extensive disease.The definition of Barrett esophageal mucosa varies among

different geographic regions. In North America, Barrettesophagus is defined as replacement of squamous mucosain the distal esophagus with columnar metaplasia thatcontains goblet cells. However, this definition is not followed

eneral Hospital, 55 Fruit Street, Warren 122, Boston, MA.n).

http://dx.doi.org/10.1053/j.semdp.2014.02.005



http://crossmark.crossref.org/dialog/?doi=10.1053/j.semdp.2014.02.005&domain=pdf

mailto:[email protected]


S E M I N A R S I N D I A G N O S T I C P A T H O L O G Y 3 1 ( 2 0 1 4 ) 1 0 0 – 1 1 3 101

worldwide. A diagnosis of Barrett esophagus in the UnitedKingdom requires the presence of endoscopic features ofBarrett esophagus, but not detection of goblet cells in theaffected segment. Failure to detect goblet cells when there isan endoscopic evidence of disease can simply reflect inad-equate sampling. However, some evidence suggests that non-goblet columnar cells in the esophagus may be biologicallyintestinalized and at risk for cancer development, even in theabsence of goblet cells. The Japanese do not require thepresence of intestinal epithelium either, but do require thepresence of endoscopically evident palisading vessels in anapparent area of Barrett esophagus. The prevalence of Barrettesophagus depends on the population studied and presenceof associated risk factors. Estimates in the United Statesrange from 1% to 25% of the general population to 5% amongpatients with gastroesophageal reflux disease.5,6 Gastroeso-phageal reflux disease is a well-known risk factor for Barrettesophagus and is extremely common.7–9 Other well-recognized risk factors for Barrett esophagus include malesex, Caucasian race, advanced age, and increased body massindex. Cigarette smoking is a minor risk factor with a relative

Fig. 1 – Three types of columnar epithelium can be identified intype (A), cardiac type (B), and intestinal type (C).

risk of 2, although smoking in conjunction with other riskfactors is implicated in only 25–40% of cases.10–12

Definition of Barrett esophagus

Three types of columnar epithelium can be recognized in thesetting of endoscopically suspected Barrett esophagus: gas-tric–fundic type, cardiac type, and intestinal type, althoughthe former usually reflects the presence of a small hiatalhernia and warrants no further discussion (Fig. 1).13 The 2011American Gastroenterological Association (AGA) guidelinesdefine Barrett esophagus as specialized metaplastic columnarepithelium (i.e., intestinal type) that replaces the squamousepithelium of the distal esophagus and predisposes to cancerdevelopment.14 This definition is based on the long heldopinion that intestinal-type epithelium with goblet cells isthe only type of metaplastic columnar epithelium predis-posed to malignancy, and, thus, the presence of goblet cells isrequired for a diagnosis of Barrett esophagus in the UnitedStates.2,14–16

endoscopically determined Barrett esophagus: gastric–fundic


Detection of goblet cells is limited by sampling as well asthe length of Barrett esophagus. Barrett esophagus is classi-fied as long segment (43 cm), short segment (1–3 cm), andultrashort segment (o1 cm) based on endoscopic measure-ment.17 The longer the segment of Barrett esophagus, themore likely that goblet cells are identified on biopsy.18,19

Harrison et al.20 demonstrated that at least eight biopsiesare required to identify goblet cells in 67.9% of patients withendoscopic evidence of columnar epithelium in the distalesophagus. Obtaining only four biopsies detects goblet cells inonly 34.7% of patients.Multilayered epithelium represents a composite epithelium

of abortive acid mucin-containing glandular cells and imma-ture, basally located squamous cells (Fig. 2). This type ofepithelium is not uncommon among patients with chronicgastroesophageal reflux disease and has been strongly asso-ciated with intestinal metaplasia, suggesting that it is aprecursor epithelium to goblet cells.21–23 However, there areno established guidelines for reporting or surveillance ofmultilayered epithelium, and it has not been clearly associ-ated with any cancer risk. Thus, it should not be consideredto represent Barrett esophagus for clinical purposes untilprospective data regarding its biologic risk become available.There have been conflicting data as to whether goblet cells

are necessary for a diagnosis of Barrett epithelium. Severalretrospective studies have shown immunohistochemicalexpression of intestinal markers, such as CDX2, DAS-1, villin,and MUC2 in non-goblet mucinous epithelial cells in Barrettesophagus,24,25 and some molecular abnormalities are sharedby cardiac-type epithelium and specialized columnar epithe-lium.26 Thus, it is clear that cardiac-type metaplastic epithe-lium without goblet cells is abnormal, although thesefindings have not been linked to definite dysplasia or cancerrisk in the absence of goblet cells, and thus, their biologicrelevance in cardiac-type epithelium is yet unknown.Some investigators have suggested that intestinal meta-

plasia is not a prerequisite to cancer development near the

Fig. 2 – Multilayered epithelium. (A) At low power, multilayeredepithelium. (B) At higher power, multilayered epithelium is comlayer of mucin-containing cells.

gastroesophageal junction, but rather, cardiac-type epithe-lium alone is a cancer precursor. Kelty et al.27 retrospectivelyevaluated 712 patients with columnar-lined esophagus andfound similar rates of malignant transformation amongpatients with intestinal metaplasia (4.5%) and those withnon-specialized columnar epithelium (3.6%). Unfortunately,this study suffers methodologic flaws that detract from theirfindings. Several cases were classified as lacking intestinalmetaplasia based on as few as two mucosal biopsies, andpatients with serial biopsies were eliminated from the anal-ysis. Takubo et al.31 evaluated the histologic features of smallesophageal adenocarcinoma (mean diameter 0.67 cm; 98.5%confined to the mucosa) from 141 patients who underwentendoscopic mucosal resection. They found that 71% of caseswere associated with cardiac-type epithelium alone at theperiphery of the cancers, and 57% had no intestinal meta-plasia anywhere in the mucosal resection specimen.Subsequent large population-based studies have shown

that cancer risk is significantly higher in patients withspecialized columnar epithelium compared to those withnon-specialized columnar epithelium (0.38% per year vs.0.07% per year, respectively).28,29 Other studies have shownthat esophageal adenocarcinomas are uniformly associatedwith intestinal metaplasia when the lesions are small(o1 cm) or confined to the mucosa.30 The likelihood ofdetecting intestinal metaplasia decreases with increasingtumor size and, thus, the absence of intestinal metaplasiain some cancer cases may reflect overgrowth of this epithe-lium by tumor or a failure to sample intestinal metaplasia inlimited material.The preponderance of currently available data support the

notion that intestinal metaplasia is a prerequisite to adeno-carcinoma at the gastroesophageal junction, and thus, thepresence of specialized columnar epithelium with goblet cellsremains a requirement to establish a diagnosis of Barrettesophagus in the United States. However, as more data isaccumulated on the carcinoma risk of patients with non-

epithelium (asterix) is seen adjacent to mature squamousposed of basal squamous cells (arrows) overlaid by a single


specialized columnar metaplasia, the definition of Barrettesophagus may continue to evolve.32

Utilization of ancillary stains in the diagnosis ofBarrett esophagus

Goblet cells are typically dispersed on a background of non-goblet neutral mucin-containing cells similar to those of thegastric mucosa. They are readily detected upon evaluation ofhematoxylin-and-eosin-stained slides. Goblet cells contain acidmucin that imparts blue discoloration to the mucin vacuole,which compresses the nucleus and laterally displaces themembranes of adjacent cells. Some pathologists routinely useancillary stains to detect goblet cells, although this practice isnot recommended, nor is it supported by data in the literature.The alcian blue/periodic acid Schiff stain is most commonlyused to facilitate identification of goblet cells. This stain colorsthe neutral mucin of gastric foveolar epithelium red, while theacidic mucin of goblet cells is blue (Fig. 3). However, a variety ofgoblet cell mimics may also show alcian blue positivity, therebyrepresenting diagnostic pitfalls when these stains are utilized.Injured, hyperplastic foveolar-type epithelial cells assume agloboid shape and contain mostly neutral mucin with someacid mucins. These pseudo-goblet cells are often positive foralcian blue stains. Other columnar cells contain blue-tingedmucin (“columnar blues”), but lack a goblet shape, and alsostain with alcian blue. Other alcian blue-positive goblet-cellmimics include columnar cells lining ducts that drain mucosaland submucosal glands, submucosal mucinous glands, andmultilayered epithelium. None of these goblet cell mimics havebeen shown to confer a risk of malignant progression andshould not be considered to represent specialized epithelium.33

Some immunohistochemical stains have been used todetect goblet cells. Although many are highly sensitive withrespect to detecting goblet cells, they lack sufficient specific-ity to be clinically useful. Phillips et al.34 found that 100% ofBarrett esophagus cases showed CDX2 positivity, compared

Fig. 3 – Barrett esophagus with intestinal metaplasia (A) as highlfoveolar epithelium and cardiac-type glands stains red (lower righleft).

to only 32% of cases composed of columnar epitheliumwithout intestinal metaplasia (Fig. 4). Immunohistochemicalstains for MUC2, an intestinal-type mucin, show betterperformance characteristics. McIntire et al. found MUC2staining to be a highly specific (96%) marker of Barrettesophagus, although they noted staining of non-goblet cellcolumnar epithelium in 78% of patients with goblet cellselsewhere (Fig. 4). These results may reflect an underlyingalteration of columnar epithelium that precedes histologicevidence of intestinal differentiation, although larger, pro-spective studies are necessary.35

Barrett esophagus versus intestinal metaplasia ofthe cardia

The squamocolumnar junction (Z line) normally corresponds tothe gastroesophageal junction and appears as an abrupt tran-sition between squamous and glandular mucosae. However,the squamocolumnar junction may be nodular, or irregular, inpatients with chronic gastroesophageal reflux disease, in whichcase it may be difficult for the gastroenterologist to knowwhether samples of an abnormal area are obtained from themost distal part of the esophagus or the proximal stomach(cardia). Cardiac-type mucosa is frequently encountered inbiopsies from this region and may reflect metaplasia of thedistal esophagus (i.e., columnar-lined esophagus) or atrophy offundic mucosa in the proximal stomach. The latter may reflectphysiologic gastroesophageal reflux and begins to develop inchildhood with progression over time. Alternatively, cardiac-type mucosa on the gastric side of the gastroesophagealjunction may result from chronic infection with Helicobacterpylori. Intestinal metaplasia may develop in mucosa of theproximal stomach or the distal esophagus and is quite com-mon, being present in 6–25% of patients with an endoscopicallynormal gastroesophageal junction and approximately 30% ofpatients with an abnormal squamocolumnar junction. Somedata suggest that distinguishing intestinal metaplasia of the

ighted by AB/PAS stain (B). The neutral mucin of the gastrict), while the acidic mucin of the goblet cells stains blue (upper

Fig. 4 – Barrett esophagus with intestinal metaplasia (A). An alcian blue stain highlights goblet cell mucin (B). Goblet cells andsome non-goblet columnar cells exhibit nuclear expression of CDX2 (C), and cytoplasmic expression of MUC2 (D) consistentwith an intestinal phenotype.


proximal stomach from that of the distal esophagus is clinicallyimportant, as the cancer risk of intestinal metaplasia in thestomach appears to be less than that of the esophagus.36,37

Srivastava et al.38 described several morphologic featuresthat can be used to distinguish intestinal metaplasia of theesophagus from that of the stomach. Features specific toBarrett esophagus include squamous epithelium overlyingcrypts with intestinal metaplasia, hybrid glands containingsuperficially located goblet cells in mucinous cardiac-typeglands, and esophageal glands or ducts associated withintestinal metaplasia (Fig. 5A–C). Less specific features ofBarrett esophagus include multilayered epithelium (Fig. 2),gland disarray and atrophy (Fig. 5D), incomplete intestinalmetaplasia, and diffuse intestinal metaplasia affecting 450%of the biopsy. Unfortunately, many cases lack clear-cutevidence of intestinal metaplasia in the esophagus and, thus,are probably best classified as showing “cardiac-type” epithe-lium with intestinal metaplasia and correlated with endo-scopic findings before making a diagnosis of Barrettesophagus.

Most immunohistochemical stains studied to date are oflimited practical value in distinguishing Barrett esophagusfrom intestinal metaplasia of the cardia. Several markershave been studied, including CK7, CK20, MUC antigens,HepPar-1, and CD10.39–42 Most studies have shown utility ofthese markers in distinguishing intestinal metaplasia of theesophagus from that of the distal stomach (antrum), but notthe gastric cardia, so they are of limited clinical utility.Ormsby et al.42 originally described a unique staining patternof CK7/CK20 in Barrett esophagus that they claimed to bediagnostically useful. The “Barrett pattern” of staining con-sisted of strong CK7 staining of superficial and deep glands incombination with weak CK20 surface staining. Unfortunately,most others have been unable to replicate these results.41,43

Dysplasia arising in Barrett esophagus

Dysplasia is defined as a proliferation of neoplastic epi-thelium confined to the basement membrane of the gland

Fig. 5 – Features useful for distinguishing Barrett esophagus from carditis with intestinal metaplasia. Buried metaplasia(intestinal metaplasia overlaid by esophageal squamous epithelium) (A), presence of esophageal ducts (B, arrows), and hybrid glands(glands with both intestinal and cardiac type epithelium) (C) are specific features. Presence of gland atrophy (D) is also helpful.


within which it arose. Although several investigators havedescribed ancillary stains and molecular techniques tofacilitate a diagnosis of dysplasia, the histologic diagnosisof dysplasia remains the gold standard and basis forclinical decision making. Biopsies from patients withBarrett esophagus are classified as negative for dysplasia,low-grade dysplasia (encompassing mild-to-moderatedysplasia), high-grade dysplasia (encompassing severedysplasia and carcinoma in situ), indefinite for dysplasia,and intramucosal carcinoma. Although this classificationscheme distributes cases in defined groups, one shouldbe aware that negative for dysplasia, low-grade dysplasia,high-grade dysplasia, and intramucosal carcinomarepresent points on a spectrum, and some cases thatmay be considered negative for dysplasia by some maybe classified as low-grade dysplasia by others. It is alsoimportant to remember that “indefinite for dysplasia” isnot a point on this spectrum. Rather it is a provisionalcategory to be used in specific situations, as describedsubsequently.

Negative for dysplasia

Criteria that should be considered when evaluating biopsiesfrom patients with Barrett esophagus include maturation of thesurface epithelium compared to deep glands, gland architec-ture, cytologic features, and the presence or absence of inflam-mation. Biopsies that are negative for dysplasia generally showsurface maturation characterized by a greater amount ofcytoplasm in the surface epithelial cells compared to the deepglands. Glands are evenly dispersed in the mucosa and sur-rounded by abundant lamina propria. Cytologic features arebland. Mitotic figures are confined to the glands and show cellpolarity with basally located nuclei, small nucleoli, and evenlydispersed chromatin. Inflammation is variable and, if present,generally located in the superficial mucosa (Fig. 6A).

Low-grade dysplasia

Low-grade dysplasia shows a lack of surface maturation. Thesame atypical cells that populate the deep gland region are


present at the luminal surface. There are mild architecturalabnormalities in the form of glandular crowding, but complexglands are lacking. Cells show mild cytologic abnormalities,including nucleomegaly, hyperchromasia, and increasedmitotic activity, although there is a minimal loss of polarity.Cytologic abnormalities of dysplasia usually show an abrupttransition from adjacent non-dysplastic epithelium, whichcan be a helpful clue to the diagnosis (Fig. 6B). Because

Fig. 6 – Spectrum of dysplasia grading. (A) Negative for dysplasilimited to the normal proliferative zone with evidence of surfacehyperchromatic nuclei with pseudostratification. Crypt architectprominent mitotic activity, enlarged nuclei, and loss of nuclear pshows cytoarchitectural atypia, equivalent to that of low-grade dsurface maturation and the stromal changes suggestive of a repa

inflammation-related cytologic atypia shows substantialoverlap with the features of low-grade dysplasia, it is gen-erally unwise to make a diagnosis of low-grade dysplasia inthe presence of epithelial inflammation.The estimated risk of esophageal adenocarcinoma varies

widely among patients with low-grade dysplasia, reflectinginconsistent diagnostic criteria and interobserver variabilitywith respect to this diagnosis.4,44–47 Extent of low-grade

a. Although some nuclear hyperchromasia is noted, it ismaturation. (B) Low-grade dysplasia shows basally oriented,ure is preserved. (C) High-grade dysplasia demonstratesolarity with evidence of glandular crowding. (D) This lesionysplasia, in the deep glands. However, given the presence ofrative process, it is best classified as indefinite for dysplasia.


dysplasia is a significant risk factor for esophageal adenocar-cinoma, so quantifying the number of biopsies involved bydysplasia may be helpful in assessing risk of dysplasiaprogression.48 Patients with low-grade dysplasia are typicallytreated with aggressive anti-acid therapy for several months,followed by repeat endoscopy and biopsy within 6 months torule out a higher-grade lesion.44 If a high-grade lesion is notfound, the patient should undergo yearly endoscopy until nodysplasia is found for 2 consecutive years.49

High-grade dysplasia

Cytoarchitectural abnormalities of high-grade dysplasia aremore pronounced than those of low-grade dysplasia, and thisdiagnosis is rarely confused with reactive atypia. At lowmagnification, lesional areas lack surface maturation. Archi-tectural abnormalities, including gland crowding, fusedglands, and micropapillae, may be striking. Nuclear polarityis lost and full thickness nuclear stratification is commonlyobserved. Hyperchromasia, nuclear membrane irregularities,

Fig. 7 – Crypt dysplasia showing unequivocal dysplasia in the cryof the crypt base demonstrating features of dysplasia characterizcrowding (B). Immunohistochemical stains for p53 (C) and Ki-67 (

and prominent nucleoli are generally present. Numerousmitotic figures, including atypical ones, are a frequent finding(Fig. 6C).44,50 Of note, foci of high-grade dysplasia are fre-quently inflamed, but as the degree of atypia exceeds thatwhich can be explained by reactive changes, a diagnosis ofhigh-grade dysplasia can still be rendered.A diagnosis of high-grade dysplasia confers substantial risk

of esophageal adenocarcinoma progression and representsthe point at which therapeutic intervention will be recom-mended for most patients. Flat (i.e., endoscopically inappar-ent) dysplasia is evaluated with repeat endoscopy and biopsywithin 3 months following acid suppression, whereas raisedlesions are removed by endoscopic mucosal resection.49

Other endoscopic therapies, such as photodynamic therapy,cryotherapy and, radiofrequency ablation (RFA), representalternatives to resection in some patients. Of these, RFA isassociated with a high rate of complete eradication of bothdysplasia and intestinal metaplasia with a reduced risk ofdisease progression and has supplanted photodynamictherapy.51

pt base and evidence of surface maturation (A). Higher powered by an increased N/C ratio, prominent nucleoli, and glandD) show strong expression within the dysplastic gland bases.


Indefinite for dysplasia

Biopsies may be considered to be indefinite for dysplasia intwo main situations. Cases with cytologic atypia in thesurface epithelium that meets criteria for low-grade dysplasiamay be considered indefinite for dysplasia in the presence ofneutrophilic inflammation. This provisional diagnosis mayalso be used when cytologic atypia in deep glands overlapswith that of low-grade dysplasia, but the overlying surfaceepithelium is mature. There are no criteria for a diagnosis ofindefinite for dysplasia. The surface epithelium is usuallymore mature than that of the deep glands, which may show aslight degree of crowding. Cytologic features include nuclearhyperchromasia and irregularities and more numerousmitotic figures with maintained cell polarity. Inflammationis usually present (Fig. 6D). The differential diagnosis isgenerally limited to low-grade dysplasia due to the presenceof similar histologic features, whereas high-grade cytologicatypia that exceeds what can be explained by reactive orregenerative atypia should always be considered to representdysplasia.52 In practice, patients with diagnoses of low-gradedysplasia and indefinite for dysplasia show comparableprogression-free survival, and these two groups are managedsimilarly for surveillance purposes.46

Variants of dysplasia

Several previously unrecognized morphologic variants ofBarrett dysplasia have been recently described. These sub-types have not been fully investigated and most data arelimited to small, retrospective studies.

Crypt dysplasiaCrypt dysplasia (to which we prefer the term basal glanddysplasia) is defined as cytologic atypia that meets criteria foreither low- or high-grade dysplasia and is limited to the deepglands (Fig. 7A and B). Lomo et al.53 first described thisfinding in association with dysplasia and/or invasive

Fig. 8 – Non-adenomatous dysplasia (A) and foveolar dysplasiawith rounded nuclei and lacking the pseudostratified nuclei of ccontaining cytoplasm, reminiscent of foveolar epithelium, is see

adenocarcinoma among patients with Barrett esophagusand noted that it frequently displayed abnormal p53 and Ki-67 immunolabeling (Fig. 7C and D). Later studies demon-strated progressively increasing aneuploidy in cases of non-dysplastic Barrett esophagus, crypt dysplasia, low-grade dys-plasia, high-grade dysplasia, and esophageal adenocarci-noma, suggesting that crypt dysplasia may represent aprecursor lesion to low-grade dysplasia.54 However, it shouldbe emphasized that data regarding crypt dysplasia are limitedby small numbers of cases and the retrospective nature ofavailable studies. In the sentinel paper, for example, 87% ofthe patients with crypt dysplasia had prior or concurrentconventional dysplasia or esophageal adenocarcinoma,which likely had some bearing on patient outcome in thatstudy. Given that most cases of crypt dysplasia seem to be, atmost, low-grade lesions and the predictive value of thisfinding is largely unknown, it is probably prudent to classifycases with low-grade abnormalities limited to the deepglands to be indefinite for dysplasia. However, high-gradecytologic abnormalities cannot be explained by regenerativeatypia in deep glands and should always be considered torepresent dysplasia, as they are at risk for progression tocancer.52

Non-adenomatous dysplasiaNon-adenomatous dysplasia is characterized by a back-to-back gland pattern. Lesional cells are cuboidal with a highnuclear-to-cytoplasmic ratio, nuclear pleomorphism, roundto ovoid nuclei, and, in many instances, prominent nucleoli(Fig. 8A). The cells may have clear, mucin-containing cyto-plasm, leading some to classify it as foveolar-type dysplasia(Fig. 8B). In contrast to adenomatous dysplasia, elongatedpencil-shaped nuclei with clumped chromatin and nuclearpseudostratification are not typical. Non-adenomatous dys-plasia accounts for 6–8% of all dysplasia cases, particularlyhigh-grade dysplasia, and shows a high rate of DNA abnor-malities.55 Patients with non-adenomatous dysplasia developcarcinoma at rates similar to those of patients with high-

(B) showing back-to-back glands composed of cuboidal cellsonventional intestinal-type dysplasia. In addition, mucin-n in the foveolar dysplasia (B).


grade adenomatous dysplasia. Of note, the recognition offoveolar-type dysplasia has been used to support the hypoth-esis that two pathways of malignant transformation inBarrett esophagus exist: (1) an intestinal pathway withintestinal metaplasia and type 1 (intestinal/classical) dyspla-sia and (2) progression via a pathway similar to that of gastriccarcinogenesis with development of type 2 (foveolar) dyspla-sia.56–58

Adenocarcinoma in Barrett esophagus

Superficial biopsies of invasive adenocarcinoma, especiallyintramucosal carcinoma, can show overlapping features withthose of high-grade dysplasia, yet the management of thesesituations may be different. Carcinomas frequently show alack of surface maturation and substantial architecturalabnormalities, including crowded glands, fused or syncytialgrowth of tumor cells, small angulated “abortive glands”invading the lamina propria, dilated glands with intraluminaldebris, and foci of single cell infiltration into the lamina

Fig. 9 – Features of intramucosal carcinoma include dilated glandlamina propria (B, arrows), and fused tumor glands (C). High-graspray (arrows) should not be overinterpreted as invasive carcino

propria (Fig. 9A–C). The latter two features appear to be highlypredictive of adenocarcinoma in the subsequent resectionspecimens.59 Importantly, invasive carcinomas do not elicit adesmoplastic stromal response when they invade themucosa, so this feature is not reliable in the detection ofcarcinoma. Invasive carcinomas usually show high-gradecytologic features with numerous mitotic figures and prom-inent nucleoli. Inflammation is variably present. Pagetoidspread of tumor cells, although rare (5%), is always a mani-festation of invasive adenocarcinoma and is never seen incases of high-grade dysplasia.60

It is worth emphasizing that the musculo-fibrous anoma-lies, such as duplication of the muscularis mucosae, prolapsechanges, and/or gland entrapment, are frequently seen inBarrett esophagus with or without neoplasia, presumablyreflecting chronic injury.61 It is important to remember thatthe most superficial layer of a duplicated muscularis mucosa,as well as the “submucosa-like” connective tissue directlysubjacent to it, are biologically equivalent to the laminapropria. Only lesions that penetrate the deeper muscle layer

s with intraluminal debris (A), abortive glands infiltrating thede dysplasia with glandular entrapment in the muscularma (D).


of the muscularis mucosae should be considered to demon-strate submucosal invasion. High-grade dysplasia mayinvolve “misplaced” glands underneath the superficial layerof duplicated muscularis mucosae, resulting in the misinter-pretation of submucosal invasive carcinoma (Fig. 9D). In thissetting, the presence of (aforementioned) unequivocalcytoarchitectural atypia of adenocarcinoma needs to beidentified before making a diagnosis of adenocarcinoma.

The role of immunohistochemistry in assisting thediagnosis of dysplasia

Multiple studies have evaluated the role of biomarkers in thedetection and classification of dysplasia. Of these, p53 hasbeen the most extensively studied. The p53 gene is rarelymutated in non-dysplastic gastrointestinal epithelium, butthe frequency of mutations increases dramatically in high-grade dysplasia and adenocarcinoma in Barrett esophagus.Immunohistochemical stains for p53 represent a surrogatemarker of underlying p53 mutations, although they should beinterpreted with caution. Some p53 mutations lead to aprotein product that is not readily degraded, thereby accu-mulating in the nucleus and producing a positive stainingresult. Others result in an unstable product that is rapidlydegraded. In this situation, an immunostain will fail to shownuclear p53 positivity. Younes et al.62 evaluated 53 cases ofBarrett esophagus and found that p53 staining occurred in 0%of cases considered negative, 9% of those considered low-grade, 55% of high-grade dysplasia, and 87% of esophagealadenocarcinomas. Unfortunately, most other studies haveshown p53 staining to be an unreliable marker of dysplasia,being present in up to 10% of cases considered morphologi-cally negative for dysplasia.63–65 However, p53 staining incases of low-grade dysplasia may represent a predictivemarker of progression to high-grade dysplasia or esophagealadenocarcinoma.66,67 Alpha-Methylacyl-CoA racemase(AMACR) is a mitochondrial and peroxisomal enzyme that

Fig. 10 – Low power view of EMR specimen (A). An immunohistduplication of the muscularis mucosa, consisting of the newly f(arrow heads) (B).

acts as a catalyst for racemization of a-methyl branchedcarboxylic coenzyme-A thioesters and plays a role in thebeta-oxidation of branched-chain fatty acids and fatty acidderivatives. Strong AMACR staining is a specific marker ofhigh-grade dysplasia and carcinoma (96%), but it is notparticularly sensitive (10%). On the other hand, any degreeof staining may be seen in biopsies that are either negativefor dysplasia or show high-grade dysplasia. Strong staining oflow-grade dysplasia may predict disease progression but doesnot perform better than routine histology.68 Thus, the clinicalvalue of these markers beyond histologic evaluation isnot clear.

Changing paradigms in management of Barrett-related neoplasia

Given the advancement of endoscopic therapeutic modalitiesand the risk for mortality and morbidity of esophagectomy,the American Gastroenterology Association recently releaseda medical position statement regarding management ofBarrett esophagus. It recommends endoscopic ablative ther-apy, especially radiofrequency ablation, and/or endoscopicmucosal resection (EMR) for patients with confirmed high-grade dysplasia in Barrett esophagus.14 Localized intramu-cosal carcinoma may be treated endoscopically.69 Recom-mendations further support endoscopic mucosal resectionfor patients who have dysplasia in Barrett esophagus asso-ciated with a visible mucosal irregularity, because thesepatients are more likely to harbor carcinoma in the lesion.14

Indeed, patients who would have been treated by esophagec-tomy for early Barrett esophagus-related neoplasia in thepast are increasingly managed with less invasive techniques.For these reasons, endoscopic resection specimens areincreasingly submitted for pathology evaluation.Endoscopic mucosal resection was developed as a diagnos-

tic modality (strip-off biopsy) in the early 1980s, but soonbecame a therapeutic tool to remove superficial neoplasms of

ochemical stain for smooth muscle action (SMA) highlightsormed superficial layer (arrows) and the original deep layer


the esophagus and stomach. Its efficacy is limited whenpositive circumferential and/or deep resection margins areobtained, but it is a valuable therapeutic modality for high-grade dysplasia and intramucosal carcinoma when per-formed by experienced gastroenterologists and surgeons.Several studies have shown that pathologists are better ableto establish accurate diagnoses and stage of superficiallesions when evaluating these relatively large specimenscompared to biopsy samples.70–72 The degree of interobserveragreement regarding classification of Barrett esophagus-related neoplasia also improves with review of endoscopicmucosal resection specimens compared to biopsy samples.This may be in part because the larger tissue samplesincrease the ability to detect important mucosal landmarks,such as duplicated muscularis mucosae (Fig. 10).73

Endoscopic mucosal resection should be considered assurgical resection by both clinicians and pathologists whoshould be aware of the tissue orientation for purposes ofmargin assessment. Some clinicians mount the excisionalspecimen on a wax block and stretch it gently prior to fixationfor at least 12 h. Taking a photograph of the specimen helpscorrelate the histologic with the endoscopic findings. Tissuesections are obtained at 2-mm intervals, which optimizeevaluation of important histologic features. If the specimenis wide enough, radial sections of the ends should besubmitted for evaluation of the circumferential margin, butthey may submitted en face when specimens are small.Fragmented excisional specimens are difficult to evaluatefor tumor stage and margin adequacy, and this is a relativelyfrequent complication when samples are obtained usingband ligation techniques. Pathologic findings that should bereported for these specimens include the status of themargins, the grade and stage of the neoplasm (low- or high-grade dysplasia, intramucosal or submucosal carcinoma), andpresence or absence of vascular invasion in invasive tumors.The status of the tissue margins is particularly important, asit predicts a relatively high risk of recurrence. Technicalartifacts, such as hemorrhage and electrodiathermic effects,may be present in endoscopically obtained specimens andmay limit histologic interpretation.74

Summary

Current guidelines require the presence of specialized col-umnar epithelium to establish the diagnosis of Barrett esoph-agus in the United States, although some data suggest thatany type of glandular metaplasia in the distal esophagusrepresents a potential cancer risk factor. Goblet cells, thehallmark of intestinal-type epithelium, are readily detectedupon evaluation of hematoxylin-and-eosin-stained slidesand routine use of ancillary stains to detect goblet cells aregenerally discouraged by most authorities. Indeed, ancillarystains, such as the Alcian blue/periodic acid Schiff stain andCDX2 and MUC2 immunostains, may enhance the appear-ance of goblet cells, but they also stain many non-goblet celltypes of epithelia that do not definitively confer a risk ofmalignant transformation. Given the potential psychologicaland financial impact of a diagnosis of Barrett esophagus onpatients, pathologists should not render the diagnosis lightly.

Diagnosis and grading of neoplasia is an important aspect ofpathologic evaluation of Barrett esophagus. It can be chal-lenging and complicated by interobserver variability, sincenegative for dysplasia, low-grade dysplasia, high-grade dys-plasia, and intramucosal carcinoma represent points on aspectrum. However, the use of certain histologic criteria,recognition of dysplasia variants, and understanding ofpotential pitfalls inherent in endoscopically obtained materi-als can facilitate the correct diagnosis.

r e f e r e n c e s

1. Hur C, Miller M, Kong CY, et al. Trends in esophageal adenocar-cinoma incidence and mortality. Cancer. 2013;119(6):1149–1158.

2. Spechler SJ, Sharma P, Souza RF, Inadomi JM, Shaheen NJ,American Gastroenterological AssociationAmerican Gastro-enterological Association technical review on the manage-ment of Barrett's esophagus. Gastroenterology. 2011;140(3):e18–e52[quiz e13].

3. Hvid-Jensen F, Pedersen L, Drewes AM, Sorensen HT, Funch-Jensen P. Incidence of adenocarcinoma among patients withBarrett's esophagus. N Engl J Med. 2011;365(15):1375–1383.

4. Sharma P, Falk GW, Weston AP, Reker D, Johnston M,Sampliner RE. Dysplasia and cancer in a large multicentercohort of patients with Barrett's esophagus. Clin GastroenterolHepatol. 2006;4(5):566–572.

5. Fan X, Snyder N. Prevalence of Barrett’s esophagus in patientswith or without GERD symptoms: role of race, age, andgender. Dig Dis Sci. 2009;54(3):572–577.

6. Ronkainen J, Aro P, Storskrubb T, et al. Prevalence of Barrett'sesophagus in the general population: an endoscopic study.Gastroenterology. 2005;129(6):1825–1831.

7. Zagari RM, Fuccio L, Wallander MA, et al. Gastro-oesophagealreflux symptoms, oesophagitis and Barrett's oesophagus inthe general population: the loiano-monghidoro study. Gut.2008;57(10):1354–1359.

8. Westhoff B, Brotze S, Weston A, et al. The frequency ofBarrett's esophagus in high-risk patients with chronic GERD.Gastrointest Endosc. 2005;61(2):226–231.

9. Veldhuyzen van Zanten SJ, Thomson AB, Barkun AN, et al. Theprevalence of Barrett's oesophagus in a cohort of 1040 Canadianprimary care patients with uninvestigated dyspepsia undergoingprompt endoscopy. Aliment Pharmacol Ther. 2006;23(5):595–599.

10. Thrift AP, Kramer JR, Richardson PA, El-Serag HB. No signifi-cant effects of smoking or alcohol consumption on risk ofBarrett's esophagus. Dig Dis Sci. 2014;59(1):108–116.

11. Cook MB, Shaheen NJ, Anderson LA, et al. Cigarette smokingincreases risk of Barrett's esophagus: an analysis of theBarrett's and esophageal adenocarcinoma consortium. Gas-troenterology. 2012;142(4):744–753.

12. Anderson LA, Watson RG, Murphy SJ, et al. Risk factors forBarrett's oesophagus and oesophageal adenocarcinoma:results from the FINBAR study. World J Gastroenterol. 2007;13(10):1585–1594.

13. Paull A, Trier JS, Dalton MD, Camp RC, Loeb P, Goyal RK. Thehistologic spectrum of Barrett's esophagus. N Engl J Med.1976;295(9):476–480.

14. American Gastroenterological AssociationSpechler SJ,Sharma P, Souza RF, Inadomi JM, Shaheen NJAmericanGastroenterological Association medical position statementon the management of Barrett's esophagus. Gastroenterology.2011;140(3):1084–1091.

15. Skinner DB, Walther BC, Riddell RH, Schmidt H, Iascone C,DeMeester TR. Barrett's esophagus. Comparison of benignand malignant cases. Ann Surg. 1983;198(4):554–565.

http://refhub.elsevier.com/S0740-2570(14)00012-4/sbref1






















































16. Smith RR, Hamilton SR, Boitnott JK, Rogers EL. The spectrumof carcinoma arising in Barrett's esophagus. A clinicopatho-logic study of 26 patients. Am J Surg Pathol. 1984;8(8):563–573.

17. Sharma P, Morales TG, Sampliner RE. Short segment Barrett'sesophagus—the need for standardization of the definition andof endoscopic criteria. Am J Gastroenterol. 1998;93(7):1033–1036.

18. Oberg S, Johansson J, Wenner J, et al. Endoscopic surveillanceof columnar-lined esophagus: frequency of intestinal meta-plasia detection and impact of antireflux surgery. Ann Surg.2001;234(5):619–626.

19. Chandrasoma PT, Der R, Ma Y, Peters J, Demeester T.Histologic classification of patients based on mapping biop-sies of the gastroesophageal junction. Am J Surg Pathol.2003;27(7):929–936.

20. Harrison R, Perry I, Haddadin W, et al. Detection of intestinalmetaplasia in Barrett's esophagus: an observational compa-rator study suggests the need for a minimum of eightbiopsies. Am J Gastroenterol. 2007;102(6):1154–1161.

21. Shields HM, Rosenberg SJ, Zwas FR, Ransil BJ, Lembo AJ, OdzeR. Prospective evaluation of multilayered epithelium in Bar-rett's esophagus. Am J Gastroenterol. 2001;96(12):3268–3273.

22. Glickman JN, Spechler SJ, Souza RF, Lunsford T, Lee E, OdzeRD. Multilayered epithelium in mucosal biopsy specimensfrom the gastroesophageal junction region is a histologicmarker of gastroesophageal reflux disease. Am J Surg Pathol.2009;33(6):818–825.

23. Shields HM, Zwas F, Antonioli DA, Doos WG, Kim S, SpechlerSJ. Detection by scanning electron microscopy of a distinctiveesophageal surface cell at the junction of squamous andBarrett's epithelium. Dig Dis Sci. 1993;38(1):97–108.

24. Hahn HP, Blount PL, Ayub K, et al. Intestinal differentiation inmetaplastic, nongoblet columnar epithelium in the esopha-gus. Am J Surg Pathol. 2009;33(7):1006–1015.

25. DeMeester SR, Wickramasinghe KS, Lord RV, et al. Cytoker-atin and DAS-1 immunostaining reveal similarities amongcardiac mucosa, CIM, and Barrett's esophagus. Am J Gastro-enterol. 2002;97(10):2514–2523.

26. Liu W, Hahn H, Odze RD, Goyal RK. Metaplastic esophagealcolumnar epithelium without goblet cells shows DNA contentabnormalities similar to goblet cell-containing epithelium.Am J Gastroenterol. 2009;104(4):816–824.

27. Kelty CJ, Gough MD, Van Wyk Q, Stephenson TJ, Ackroyd R.Barrett's oesophagus: intestinal metaplasia is not essentialfor cancer risk. Scand J Gastroenterol. 2007;42(11):1271–1274.

28. Bhat S, Coleman HG, Yousef F, et al. Risk of malignantprogression in Barrett's esophagus patients: results from alarge population-based study. J Natl Cancer Inst. 2011;103(13):1049–1057.

29. Westerhoff M, Hovan L, Lee C, Hart J. Effects of dropping therequirement for goblet cells from the diagnosis of Barrett’sesophagus. Clin Gastroenterol Hepatol. 2012;10(11):1232–1236.

30. Chandrasoma P, Wickramasinghe K, Ma Y, DeMeester T. Isintestinal metaplasia a necessary precursor lesion for adeno-carcinomas of the distal esophagus, gastroesophageal junc-tion and gastric cardia? Dis Esophagus. 2007;20(1):36–41.

31. Takubo K, Aida J, Naomoto Y, et al. Cardiac rather thanintestinal-type background in endoscopic resection speci-mens of minute Barrett adenocarcinoma. Hum Pathol.2009;40(1):65–74.

32. Riddell RH, Odze RD. Definition of Barrett's esophagus: timefor a rethink—is intestinal metaplasia dead? Am J Gastro-enterol. 2009;104(10):2588–2594.

33. Younes M, Ertan A, Ergun G, et al. Goblet cell mimickers inesophageal biopsies are not associated with an increased riskfor dysplasia. Arch Pathol Lab Med. 2007;131(4):571–575.

34. Phillips RW, Frierson HF Jr, Moskaluk CA. Cdx2 as a marker ofepithelial intestinal differentiation in the esophagus. Am JSurg Pathol. 2003;27(11):1442–1447.

35. McIntire MG, Soucy G, Vaughan TL, Shahsafaei A, Odze RD.MUC2 is a highly specific marker of goblet cell metaplasia inthe distal esophagus and gastroesophageal junction. Am JSurg Pathol. 2011;35(7):1007–1013.

36. Spechler SJ. Intestinal metaplasia at the gastroesophagealjunction. Gastroenterology. 2004;126(2):567–575.

37. Sharma P, Weston AP, Morales T, Topalovski M, Mayo MS,Sampliner RE. Relative risk of dysplasia for patients withintestinal metaplasia in the distal oesophagus and in thegastric cardia. Gut. 2000;46(1):9–13.

38. Srivastava A, Odze RD, Lauwers GY, Redston M, Antonioli DA,Glickman JN. Morphologic features are useful in distinguish-ing Barrett esophagus from carditis with intestinal metapla-sia. Am J Surg Pathol. 2007;31(11):1733–1741.

39. Glickman JN, Shahsafaei A, Odze RD. Mucin core peptideexpression can help differentiate Barrett's esophagus fromintestinal metaplasia of the stomach. Am J Surg Pathol. 2003;27(10):1357–1365.

40. Glickman JN, Wang H, Das KM, et al. Phenotype of Barrett'sesophagus and intestinal metaplasia of the distal esophagusand gastroesophageal junction: an immunohistochemicalstudy of cytokeratins 7 and 20, das-1 and 45 MI. Am J SurgPathol. 2001;25(1):87–94.

41. Mohammed IA, Streutker CJ, Riddell RH. Utilization of cyto-keratins 7 and 20 does not differentiate between Barrett’sesophagus and gastric cardiac intestinal metaplasia. ModPathol. 2002;15(6):611–616.

42. Ormsby AH, Goldblum JR, Rice TW, et al. Cytokeratin subsetscan reliably distinguish Barrett's esophagus from intestinalmetaplasia of the stomach. Hum Pathol. 1999;30(3):288–294.

43. Glickman JN, Ormsby AH, Gramlich TL, Goldblum JR, Odze RD.Interinstitutional variability and effect of tissue fixative onthe interpretation of a Barrett cytokeratin 7/20 immunoreac-tivity pattern in Barrett esophagus. Hum Pathol. 2005;36(1):58–65.

44. Montgomery E, Bronner MP, Goldblum JR, et al. Reproduci-bility of the diagnosis of dysplasia in Barrett esophagus: areaffirmation. Hum Pathol. 2001;32(4):368–378.

45. Reid BJ, Levine DS, Longton G, Blount PL, Rabinovitch PS.Predictors of progression to cancer in Barrett's esopha-gus: baseline histology and flow cytometry identify low-and high-risk patient subsets. Am J Gastroenterol. 2000;95(7):1669–1676.

46. Montgomery E, Goldblum JR, Greenson JK, et al. Dysplasia as apredictive marker for invasive carcinoma in Barrett esoph-agus: a follow-up study based on 138 cases from a diagnosticvariability study. Hum Pathol. 2001;32(4):379–388.

47. Skacel M, Petras RE, Gramlich TL, Sigel JE, Richter JE, Gold-blum JR. The diagnosis of low-grade dysplasia in Barrett'sesophagus and its implications for disease progression. Am JGastroenterol. 2000;95(12):3383–3387.

48. Srivastava A, Hornick JL, Li X, et al. Extent of low-gradedysplasia is a risk factor for the development of esophagealadenocarcinoma in Barrett's esophagus. Am J Gastroenterol.2007;102(3):483–493 [quiz 694].

49. Wang KK, Sampliner RE. Practice parameters committee ofthe American College of Gastroenterology. Updated guide-lines 2008 for the diagnosis, surveillance and therapy ofBarrett's esophagus. Am J Gastroenterol. 2008;103(3):788–797.

50. Reid BJ, Haggitt RC, Rubin CE, et al. Observer variation in thediagnosis of dysplasia in Barrett's esophagus. Hum Pathol.1988;19(2):166–178.

51. Shaheen NJ, Sharma P, Overholt BF, et al. Radiofrequencyablation in Barrett's esophagus with dysplasia. N Engl J Med.2009;360(22):2277–2288.

52. Younes M, Lauwers GY, Ertan A, et al. The significance ofindefinite for dysplasia grading in Barrett metaplasia. ArchPathol Lab Med. 2011;135(4):430–432.










































































































































53. Lomo LC, Blount PL, Sanchez CA, et al. Crypt dysplasia withsurface maturation: a clinical, pathologic, and molecularstudy of a Barrett's esophagus cohort. Am J Surg Pathol.2006;30(4):423–435.

54. Zhang X, Huang Q, Goyal RK, Odze RD. DNA ploidy abnor-malities in basal and superficial regions of the crypts inBarrett's esophagus and associated neoplastic lesions. Am JSurg Pathol. 2008;32(9):1327–1335.

55. Rucker-Schmidt RL, Sanchez CA, Blount PL, et al. Nonade-nomatous dysplasia in Barrett esophagus: a clinical, patho-logic, and DNA content flow cytometric study. Am J SurgPathol. 2009;33(6):886–893.

56. Brown IS, Whiteman DC, Lauwers GY. Foveolar type dysplasiain Barrett esophagus. Mod Pathol. 2010;23(6):834–843.

57. Khor TS, Alfaro EE, Ooi EM, et al. Divergent expression ofMUC5AC, MUC6, MUC2, CD10, and CDX-2 in dysplasia andintramucosal adenocarcinomas with intestinal and foveolarmorphology: is this evidence of distinct gastric and intestinalpathways to carcinogenesis in Barrett esophagus? Am J SurgPathol. 2012;36(3):331–342.

58. Demicco EG, Farris AB 3rd, Baba Y, et al. The dichotomy incarcinogenesis of the distal esophagus and esophagogastricjunction: intestinal-type vs cardiac-type mucosa-associatedadenocarcinoma. Mod Pathol. 2011;24(9):1177–1190.

59. Patil DT, Goldblum JR, Rybicki L, et al. Prediction of adeno-carcinoma in esophagectomy specimens based upon analysisof preresection biopsies of Barrett esophagus with at leasthigh-grade dysplasia: a comparison of 2 systems. Am J SurgPathol. 2012;36(1):134–141.

60. Abraham SC, Wang H, Wang KK, Wu TT. Paget cells in theesophagus: assessment of their histopathologic features andnear-universal association with underlying esophagealadenocarcinoma. Am J Surg Pathol. 2008;32(7):1068–1074.

61. Lewis JT, Wang KK, Abraham SC. Muscularis mucosae duplica-tion and the musculo-fibrous anomaly in endoscopic mucosalresections for Barrett esophagus: implications for staging ofadenocarcinoma. Am J Surg Pathol. 2008;32(4):566–571.

62. Younes M, Lebovitz RM, Lechago LV, Lechago J. p53 proteinaccumulation in Barrett's metaplasia, dysplasia, and carci-noma: a follow-up study. Gastroenterology. 1993;105(6):1637–1642.

63. Ramel S, Reid BJ, Sanchez CA, et al. Evaluation of p53 proteinexpression in Barrett's esophagus by two-parameter flowcytometry. Gastroenterology. 1992;102(4 Pt 1):1220–1228.

64. Bhargava P, Eisen GM, Holterman DA, et al. Endoscopicmapping and surrogate markers for better surveillance inBarrett esophagus. A study of 700 biopsy specimens. Am J ClinPathol. 2000;114(4):552–563.

65. Jones DR, Davidson AG, Summers CL, Murray GF, Quinlan DC.Potential application of p53 as an intermediate biomarker inBarrett's esophagus. Ann Thorac Surg. 1994;57(3):598–603.

66. Kastelein F, Biermann K, Steyerberg EW, et al. Aberrant p53protein expression is associated with an increased risk ofneoplastic progression in patients with Barrett's oesophagus.Gut. 2013;62(12):1676–1683.

67. Skacel M, Petras RE, Rybicki LA, et al. p53 expression in lowgrade dysplasia in Barrett's esophagus: correlation withinterobserver agreement and disease progression. Am J Gas-troenterol. 2002;97(10):2508–2513.

68. Kastelein F, Biermann K, Steyerberg EW, et al. Value of alpha-methylacyl-CoA racemase immunochemistry for predictingneoplastic progression in Barrett's oesophagus. Histopathol-ogy. 2013;63(5):630–639.

69. Luna RA, Gilbert E, Hunter JG. High-grade dysplasia andintramucosal adenocarcinoma in Barrett's esophagus: therole of esophagectomy in the era of endoscopic eradicationtherapy. Curr Opin Gastroenterol. 2012;28(4):362–369.

70. Mino-Kenudson M, Brugge WR, Puricelli WP, et al. Manage-ment of superficial Barrett's epithelium-related neoplasms byendoscopic mucosal resection: clinicopathologic analysis of27 cases. Am J Surg Pathol. 2005;29(5):680–686.

71. Mandal RV, Forcione DG, Brugge WR, Nishioka NS, Mino-Kenudson M, Lauwers GY. Effect of tumor characteristics andduplication of the muscularis mucosae on the endoscopicstaging of superficial Barrett esophagus-related neoplasia. AmJ Surg Pathol. 2009;33(4):620–625.

72. Wani S, Abrams J, Edmundowicz SA, et al. Endoscopicmucosal resection results in change of histologic diagnosisin Barrett's esophagus patients with visible and flat neo-plasia: a multicenter cohort study. Dig Dis Sci. 2013;58(6):1703–1709.

73. Mino-Kenudson M, Hull MJ, Brown I, et al. EMR for Barrett'sesophagus-related superficial neoplasms offers better diag-nostic reproducibility than mucosal biopsy. GastrointestEndosc. 2007;66(4):660–666 [quiz 767, 769].

74. Odze RD, Lauwers GY. Histopathology of Barrett's esophagusafter ablation and endoscopic mucosal resection therapy.Endoscopy. 2008;40(12):1008–1015.

























































































barrett esophagus: diagnostic challenges

Documents