Primary Motility  Disorders of the  Esophagus
 The Esophageal
 Esophagogastric  Junction

  Browse by Author
  Browse by Movies
Volume: The Esophageal Mucosa
Chapter: The dysplastic mucosa

Movie:  Dysplasia-Carcinoma sequence (Commentaries Dr.Validire)

Is there a great variation in histologic grading of dysplastic mucosa among highly-experienced gastrointestinal pathologists?

K. Geboes (Leuven)

Dysplasia is a term used in various subspecialties of medicine including gynecology, urology, radiology and gastroenterology for the description of macroscopic and microscopic lesions. It is originally an ancient Greek word which means "malformation", or "a morphologic structure which is different from what normally is expected". As such it can indicate either a congenital or an acquired malformation, and it is still used with this meaning for macroscopic and radiological lesions. For microscopic and pathological lesions "dysplasia" has various significations. It can indicate a "malformation" or a lesion having a relation to malignancy. An example of the former is the term "angiodysplasia" which was used for vascular lesions, especially of the colon. A more appropriate term for this condition is "vascular ectasia" [1].

From these considerations it is clear that at present no general definition of dysplasia is available. The lack of such a definition for a term which is so commonly used is responsible for many misunderstandings and a great deal of confusion. The purpose of the present review is to explain the term "dysplasia" as it is used in gastrointestinal pathology and to indicate some differences, rather than to present a new classification or to indicate one single meaning for the word. We will, however, indicate the most appropriate definition as it is now currently used, while other

definitions are still available and are not necessarily wrong. It is therefore extremely important that the pathologist and the clinicians working closely together are aware of the meaning of the term, and when using it do so in complete understanding of each other.

In gastrointestinal pathology the use of the term "dysplasia" is mainly limited to microscopic epithelial and nonepithelial lesions related to malignancy. The various epithelia of the gastrointestinal tract in which dysplastic lesions are described have different characteristics and a different microenvironment. The esophagus is covered by a multilayered squamous epithelium, which may be replaced by a metaplastic columnar epithelium. The colon is lined by a columnar epithelium. In the stomach the glandular compartment of the epithelium is much more developed than in the other segments. However, because of these differences the criteria for the diagnosis of dysplasia and, more important, the consequences of the finding of dysplasia differ according to organs.

Definition of dysplasia

The term "dysplasia" has been introduced in gastrointestinal pathology in studies on gastrointestinal cancer. It was defined as "a process of disordered cell growth" [2], "epithelial atypia or alteration" or "excessive abnormal proliferation of the gastric epithelium". The lesions are subdivided into several categories and usually a distinction is made between mild, moderate and severe or grade I, grade II and grade III [3,4]. Clinical and experimental studies have shown that mild and moderate degrees or grade I and II dysplasia, as used in the classifications of Ming [2] and Oehlert [3], are probably clinically insignificant and largely reactive or regenerative in nature [5,6].

The fact that regenerative alterations of columnar epithelia may be mistaken for dysplasia is not strange because all cells in regenerative epithelia show signs of cell growth and proliferation, including larger nuclei, basophilia of the cytoplasm and architectural changes.

Therefore, because of the obvious differences in behavior between regenerative and degenerative epithelial alterations, there was a need for a more precise definition of dysplasia. This was introduced in 1983, when an international group studying dysplasia in ulcerative colitis defined dysplasia as: "epithelial changes, lesions or alterations that are unequivocally neoplastic", thus excluding all reactive (regenerative) changes and equivocal changes [7]. A similar definition has been introduced for dysplasia in Barrett's esophagus [8,9]. The evaluation of the neoplastic nature of the lesions is based on light microscopic observations (cytological and architectural alterations). More sophisticated methods such as immunohistochemistry using specific markers (Ki67, recognizing a nuclear antigen expressed in all phases of the cell cycle, PCNA - proliferating cell nuclear antigens - and others), or flow cytometry, may help in the future to identify the neoplastic nature more precisely. Furthermore, for the esophagus it must be remembered that Barrett's mucosa is a metaplastic mucosa that is quite different from the colonic mucosa from which colitic carcinomas arise.

Dysplastic lesions have been identified in the whole gastrointestinal tract. Yet the classifications that have been proposed and that are used in the various segments of the gastrointestinal tract are not entirely similar. Roughly, we can state that for simple colorectal adenomas and for the stomach, the lesions are subdivided into mild, moderate and severe [10], and that in the earlier studies concerning the stomach, dysplasia is not always unequivocally neoplastic [5].

In Barrett's esophagus the lesions are classified into negative, indefinite or positive for dysplasia. The latter is subdivided into low grade and high grade [8,9]. (Figs. 1 and 2). Yet other classifications have been used and have appeared in the literature. Hamilton classified dysplasia in columnar cell lined esophagus into three grades: low, intermediate and high, similar to an international classification of gastric dysplasia, although the intermediate grade can be included in the high-grade category [8,11]. Schmidt and coworkers used two types of dysplasia: low grade and high grade, based upon the appearance and position (basal or apical) of the nuclei. For each grade of dysplasia these authors also describe two types [12]. In general most authors also include carcinoma in situ or intraepithelial carcinoma in the high-grade dysplasia category [9].

If we compare these classifications and regard dysplasia as unequivocally


Figure I. .High-grade dysplasia in Barrett's mucosa: on the left squamous epithelium (1). The columnar epithelium is diffusely dysplastic (with areas of minimal invasion). A dilated tubular duct of an esophageal gland is present in the submucosa (2) (H.E. x 50).


Figure 2. .Higher magnification of Fig. 1 showing the dysplastic nature of the surface and glandular epithelium characterized by enlarged nuclei with loss of polarity and striking cytological abnormalities (A: H.E. x 125; B: H.E. x 200).

neoplastic, it appears that low-grade dysplasia in Barrett is equivalent to forms of mild dysplasia in the stomach, while high-grade dysplasia equals moderate and severe dysplasia. The moderate or intermediate grade is not classically used for Barrett's esophagus.

The fact that different classifications are available, the incorrect use of words and the lack of a proper definition, rather than differences in opinion, are often the cause of confusion. Misunderstandings can thus be avoided by using proper terminology in a specific setting and by defining exactly the meaning of terms.

Interobserver and intraobserver variation in the classification of dysplasia

The microscopic diagnosis of dysplasia is based upon the identification of cytological and architectural criteria. According to most authors cytological criteria are superior to disorganized architecture and more important for the classification of the lesions as low grade or high grade.

Although not clearly established, various studies show that, in specimens from patients operated for high-grade dysplasia, a carcinoma is frequently already present. Therefore, it seems that the criteria used to distinguish low grade from high grade are adequate. On the other hand, some studies have shown a discordance between histologic dysplasia and the finding of aneuploidy on flow cytometry. Flow cytometry might identify other subgroups at risk, and the microscopic evaluation of the presence of dysplasia might not be the best or the only good technique for the early identification of cancer [13].

Another problem is to see whether pathologists of equal experience assess dysplasia reliably. We must realize that as such the alterations characteristic for dysplasia may present interpretative problems for the pathologists. Clearly, this is a very important issue, because if clinicians will decide esophagectomies on the basis of pathology reports they need to be confident of the ability of the pathologists to diagnose dysplasia.

This problem has been studied extensively and it appears that: - indefinite and low-grade dysplasia are difficult to distinguish and there may be

classification problems. An interobserver agreement of 70% can be reached


- the interobserver and intraobserver (the same pathologist seeing the same biopsy at a different moment) agreement is good for high-grade dysplasia. An interobserver agreement of 87% was reached for high-grade dysplasia [14,15].

- similar results are obtained by experienced pathologists and by pathologists of different specific or general interest.

- the terminology used in reporting the lesions may be different (low grade or mild, high grade or severe, three categories or two) but, if a good agreement is made on terminology, no major difference of opinion exists among pathologists for the classification, especially not for the high-grade lesions.


1. Boley S, Brandt U. Vascular ectasias of the colon. Dig Dis Sci 1986:31: 26S-42S.

2. Ming SCH, Bajtai A, Correa P. Gastric dysplasia: significance and pathologic criteria. Cancer 1984:54:1794-1801.

3. Oehlert W, Keller P, Henke M, Strauch M. Die dysplasien der Magenschleimhaut. Dtsch Med Wochenschr 1975;100: 1950-1956.

4. Sipponen P. Gastric dysplasia. In: Williams GT (ed) Gastrointestinal Pathology. Berlin: Springer Verlag, 1990;61-76.

5. Farini R, Farinati F, Leandro G. Gastric epithelial dysplasia in relapsing and nonrelapsing gastric ulcer. Am J Gastroenterol 1982;77:844-853.

6. Farinati F, Cardin F, Di Mario F. Follow-up in gastric dysplasia patients. Am J Surg Pathol 1989;13:173-174.

7. Riddell RH, Goldman H, Ransohoff D, Appelman HD, Fenoglio CM, Haggitt RC, Ahren C, Correa P, Hamilton SR, Morson BC, Sommers SC, Yardley JH. Dysplasia in inflammatory bowel disease. Hum Pathol 1983:14:931-967.

8. Hamilton SR Reflux esophagitis and Barrett's esophagus. In: Goldman H, Appelman HD, Kaufman N (eds) Gastrointestinal Pathology. Baltimore: Williams & Wilkins, 1990;11-68

9. Potet F, Duchatelle V. Barrett's Oesophagus In: Williams GT (ed) Current Topics in Gastrointestinal Pathology. Berlin: Springer Verlag, 1990;43-60.

10. Pascal RR. Consistency in the terminology of colorectal dysplasia. Hum Pathol 1988:19:1249-1250.

11. Hamilton SR, Smith RRL. The relationship between columnar epithelial dysplasia and invasive carcinoma arising in Barrett's esophagus. Am J Clin Pathol 1987:87:301-312.

12. Schmidt HG, Riddell RH, Walther BC, Skinner DB, Riemann JF. Dysplasia in Barrett's esophagus. J Cancer Res Clin Oncol 1985;110:145-152.

13. Fennerty MB, Sampliner RE, Way D, Riddell RH, Steinbronn K, Garewal HS. Discordance between flow cytometric abnormalities and dysplasia in Barrett's esophagus. Gastroenterology 1989;97:815-820.

14. Reid BJ, Haggitt RC, Rubin CE, Roth G, Surawicz CM, Van Belle G, Lewin K, Weinstein WM, Antonioli DA, Goldman H, MacDonald W, Owen D. Criteria for dysplasia in Barrett's esophagus: a cooperative consensus study. Gastroenterology 1987;88:1552.

15. Reid BJ, Haggitt RC, Rubin CE, Roth G, Surawicz CM. Observer variation in the diagnosis of dysplasia in Barrett's esophagus. Hum Pathol 1988;19:166-178.

S.R. Hamilton (Baltimore)

Adenocarcinoma arising in columnar-lined mucosa is a well-known complication of Barrett's esophagus (columnar epithelial replacement of previously squamous-lined esophageal mucosa, usually due to chronic gastroesophageal reflux). Mortality in patients presenting with symptomatic Barrett's adenocarcinoma is high. Therefore, the strategy of surveying patients with known Barrett's esophagus for preinvasive neoplasia in the form of columnar epithelial dysplasia (intraepithelial neoplasia) and for early adenocarcinoma is currently recommended in patients who are suitable surgical candidates for esophagectomy. In the absence of proven efficacious endoscopic ablation procedures or pharmacologic agents effective against dysplasia, prophylactic esophagectomy is the currently available treatment. Esophagectomy is recommended for Barrett patients with dysplasia severe enough to warrant major concern about synchronous or metachronous adenocarcinoma, although the natural history of the dysplasia-carcinoma sequence is not yet well defined in large numbers of patients. Esophagectomy carries the risk of serious morbidity and mortality even

in the hands of the expert esophageal surgical teams who should be entrusted with the prophylactic procedure. Therefore, accurate histopathologic interpretation of dysplasia in endoscopic biopsy specimens and brush cytology specimens of Barrett's mucosa is of great concern, because of the implications for patient management.

Dysplasia is identified histopathologically on the basis of abnormalities in mucosal architecture, epithelial morphology and cytology. The histopathology of columnar epithelial dysplasia in Barrett's mucosa is impressively heterogeneous. In addition, dysplasia can be difficult to distinguish from reactive columnar epithelial changes due to the effects of ongoing gastroesophageal reflux. Intraobserver and interobserver variation in the interpretation of the spectrum of dysplasia is therefore expected.

The most useful classification schemes for communication among gastroenterologists, esophageal surgeons and pathologists regarding dysplasia in Barrett's mucosa employ modifications of the scheme developed for dysplasia in idiopathic inflammatory bowel disease. As shown in Table 1, the findings are classified into negative for dysplasia, indefinite for dysplasia and positive for dysplasia.

The negative category indicates no concern for dysplasia in the specimens. The indefinite category is used for abnormalities which are of concern, but not interpreted as diagnostic of dysplasia, usually due to inflammation or artifact complicating interpretation, but also due to uncertainty about the interpretation of the findings. This category can be subdivided into probably negative, of unknown significance, and probably positive for the pathologist to express the level of concern [1]. In the positive for dysplasia category, three grades were used by the group (low-, intermediate- and high-grade dysplasia) to classify the severity of the dysplasia. This approach is based on the study of features of dysplasia associated with invasive adenocarcinoma, in order to establish criteria [1 ]. Some investigators use three grades analogous to gastric dysplasia [2]. Other groups use two grades (low- and high-grade) with criteria adapted from colonic dysplasia in idiopathic inflammatory bowel disease [3] (Table 1). The two-grade system permits greater observer agreement, but cases at the low end of the high-grade dysplasia spectrum in Barrett's mucosa, using colonic criteria for idiopathic inflammatory bowel disease, have a relatively infrequent association with synchronous adenocarcinoma.

In the clinical application of dysplasia classification, two decision points are

Table 1. .Classification schemes for dysplasia in Barrett's mucosa

Three grades of dysplasia [1]

Two grades of dysplasia [2]

Negative for dysplasia

Negative for dysplnsia

Indefinite for dysplasia

Indefinite for dysplasia

Probably negative

Of unknown significance

Probably positive

Positive for dysplasia

Positive for dysplasia








important: the interface between indefinite for dysplasia vs. negative for dysplasia, which may influence frequency of follow-up, and the interface between high-grade dysplasia vs. low-grade/intermediate-grade dysplasia, which will influence consideration of prophylactic esophagectomy. One study addressed the agreement among eight experienced gastrointestinal pathologists in the diagnosis of dysplasia, with two grades in the positive for dysplasia category [4]. A slide set of 71 sections, including biopsy and resection specimens, was chosen to include some especially difficult cases. Interobserver agreement of 85 and 87% was found in two successive rounds of evaluation comparing the combined category of high-grade dysplasia and inter-mucosal carcinoma against the combined category of low-grade dysplasia, indefinite for dysplasia and negative for dysplasia. At the other end of the spectrum, interobserver agreement was 71 and 72% in the two rounds when negative for dysplasia was compared against the combined category of indefinite for dysplasia, positive for dysplasia and intermucosal adenocarcinoma.

All diagnostic tests have inherent variability. In ideal circumstances, variation is reduced to a minimum. The study described above demonstrates that there is acceptable reproducibility of classification at the two clinically important decision points in the hands of experienced gastrointestinal pathologists. The level of agreement is better than that for preinvasive breast disease evaluated by experienced breast pathologists [5].

No marker which improves on dysplasia as an indicator of risk of adenocarcinoma is yet available. Such markers may never become available, because the diverse molecular, genetic and biochemical pathways which provide potential markers are all associated with the same morphologic endpoint (i.e., dysplasia and ultimately adenocarcinoma). Furthermore, dysplasia appears to be related to accumulated molecular abnormalities of various types, as evidenced by the heterogeneity of results in studies published thus far. For example, abnormalities in p53 occur in a subset of Barrett's adenocarcinomas, but certainly not all cases [6,7]. Thus, none of the currently identified genetic abnormalities are absolutely associated with neoplasia, and furthermore neoplasia is not uniformly associated with any of the genetic abnormalities. Although important for investigational studies of the pathogenesis of dysplasia and carcinoma, the individual genes and biochemical pathways seem unlikely to provide markers with clinical utility exceeding dysplasia itself.

In conclusion, variability in the grading of dysplasia in Barrett's mucosa by gastrointestinal pathologists occurs at an acceptable level for satisfactory clinical management. Given the rarity of patients with dysplasia in Barrett's esophagus and the high stakes for the patient being considered for prophylactic esophagectomy, surgical management in centers with a highly experienced esophageal surgery team and gastrointestinal pathologist, who has reviewed the biopsy specimens, is prudent.


1. Hamilton SR, Smith RRL. The relationship between columnar epithelial dysplasia and invasive adenocarcinoma in Barrett': esophagus. Am J Clin Pathol 1987:87:301-312.

2. Morson BC, Sobin LH, Grundmann E, Johansen A, Nagayo T, Serck-Hanssen A. Precancerous conditions and epithelial dysplasia in the stomach. J Clin Pathol 1980;33:711-721.

3. Riddell RH, Goldman H, Rensohoff DF et al. Dysplasia in inflammatory bowel disease. Standardized classification with provisional clinical applications. Hum Pathol 1983; 14:931-968.

4. Reid BJ, Haggitt RC, Rubin CE et al. Observer variation in the diagnosis of dysplasia in Barrett's esophagus. Hum Pathol 1988;19:166-178.

5. Schnitt SJ, Connolly JL, Tavassoli FA et al. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria. Am J Surg Pathol 1992;16:1133-1143.

6. Flejou JF, Potet F, Muzeau F et al. Overexpression of p53 protein in Barrett's syndrome with malignant transformation J Clin Pathol 1993;46:330-333.

7. Blount PL, Ramel S, Raskind WH et al. 17p allelic deletions and p53 protein overexpression in Barrett's adenocarcinoma. Cancer Res 1991 ;51;5482-5486.

Publication date: May 1994 OESO©2015