citokeratina en esofago de barret
TRANSCRIPT
-
8/12/2019 Citokeratina en Esofago de Barret
1/5
Cytokeratin Subsets for DistinguishingBarretts Esophagus From Intestinal Metaplasia
in the Cardia Using Endoscopic Biopsy SpecimensHala M. T. El-Zimaity, M.D., and David Y. Graham, M.D.
Gastrointestinal Mucosa Pathology Laboratory and Departments of Medicine and Pathology, Veterans
Affairs Medical Center and Baylor College of Medicine, Houston, Texas
OBJECTIVES: It has been suggested that Barretts epithelium
and intestinal metaplasia in the gastric cardia have different
cyotokeratin (CK) staining patterns and that Barretts epi-
thelium can be distinguished by CK staining pattern. The
aim of this study was to test the utility of CK staining fordistinguishing Barretts esophagus from gastric intestinal
metaplasia.
METHODS: Topographically mapped gastric biopsy speci-
mens were obtained from patients without Barretts esoph-
agus, and esophageal biopsies were obtained from patients
with long-segment Barretts esophagus (3 cm). Serial
sections were stained with Genta or El-Zimaity triple stain,
and biopsies with intestinal metaplasia were stained with
antibodies against CK 4, 13, 7, and 20.
RESULTS: Sections from 33 biopsies with Barretts esopha-
gus, 23 with intestinal metaplasia of the gastric cardia, 27with intestinal metaplasia of the gastric body, and 33 with
intestinal metaplasia of the antrum were examined. CK 4
and CK 13 stained squamous epithelium only. The proposed
diagnostic CK Barretts 7/20 pattern was found in only
39% of long-segment Barretts compared to 35%, 4%, and
24% in intestinal metaplasia from the gastric cardia, body,
and antrum, respectively. The criteria proposed had a sen-
sitivity of 45% and a specificity of 65%.
CONCLUSIONS: These results do not support keratin pheno-
typing as a tool for differentiating intestinal metaplasia
originating in the cardia from intestinal metaplasia of
Barretts. (Am J Gastroenterol 2001;96:13781382. 2001by Am. Coll. of Gastroenterology)
INTRODUCTION
Intestinal metaplasia is a common finding in biopsy speci-
mens taken from just below the gastroesophageal junction in
patients with current or past Helicobacter pylori infection
(15). This finding has prompted considerable discussion
regarding whether it is a variant of gastroesophageal reflux
disease (GERD) (6 9), (e.g., short-segment Barretts), a
consequence ofH. pyloriof the stomach, or both (15).
Barretts esophagus is considered a premalignant condi-
tion such that the distinction between H. pylorirelated
intestinal metaplasia of the cardia and short-segments Bar-
retts esophagus is important. Recently, Ormsby et al.(10),
using antibodies to cytokeratin (CK) 7 and CK 20, found
that the staining pattern of gastric intestinal metaplasia wasentirely different from that of Barretts epithelium. The
Barretts CK 7/20 pattern was defined as staining of the
superficial epithelium with CK 20 and staining of both the
superficial and deep metaplastic epithelium with CK 7. This
pattern was present in 97% of specimens with long-segment
Barretts and was not observed in gastric intestinal meta-
plasia (10). They hypothesized that the Barretts CK 7/20
pattern was specific for Barretts epithelium.
The current study was designed to confirm their hypoth-
esis and to extend the observation to biopsies taken at
endoscopy by investigating the utility of CK 7, 20, 4, and 13
in the histological distinction of Barretts esophagus fromintestinal metaplasia in the cardia.
MATERIALS AND METHODS
Patients and Histology
Mucosal biopsy specimens were obtained from patients who
had previous upper GI endoscopy with gastric mapping
which typically involved taking 14 biopsies from specified
sites (11). The anatomic cardia was defined as the mucosa
immediately below the site of the junction of the mucosa of
the tubular esophagus and the stomach (Z-line) and above
the beginning of the first gastric fold. Biopsies of the gastric
cardia (mean and median of two biopsies) were always
taken antegrade (not retrograde with retroflexion of the
endoscope). None of the patients had tongues of gastric type
epithelium. Each biopsy was placed in a separate bottle of
10% buffered formalin. Biopsies were embedded on edge,
sectioned at 5 m with six sections per slide, and stained
with the Genta stain (12) or El-Zimaity triple stain (13). The
median size of biopsy specimens (fixed tissue measured on
a glass slide) was 8 4 mm. Each specimen was reviewed
by one pathologist and scored using a visual analog scale
from 0 (absent/normal) to 5 (maximal intensity) for intes-
tinal metaplasia (14).
THE AMERICANJOURNAL OF GASTROENTEROLOGY Vol. 96, No. 5, 2001 2001 by Am. Coll. of Gastroenterology ISSN 0002-9270/01/$20.00Published by Elsevier Science Inc. PII S0002-9270(01)02355-3
-
8/12/2019 Citokeratina en Esofago de Barret
2/5
Cases were selected because of previously documented
intestinal metaplasia. Indications for upper GI endoscopy
included duodenal ulcer, gastric ulcer, and previous diag-nosis of intestinal metaplasia. Cases with previously docu-
mented long-segment Barretts metaplasia were obtained
from the files of the Veterans Affairs Medical Center.
High IronDiamine Staining
Biopsies with intestinal metaplasia were stained with high
irondiamine/Alcian blue (HID/AB) to identify neutral,
sialo-, and sulfomucins. Briefly, slides were immersed in
HID solution for more than 18 h, at 2325C. Slides were
then rinsed with deionized water and stained with 1% Alcian
blue, pH 2.5 for 2 min (15). Subtyping intestinal metaplasia
was done according to the system used by Jass and Filipe
(16, 17). Type I is classified as complete intestinal metapla-sia and type II and III are grouped incomplete metaplasia.
Immunohistochemical Studies
For immunophenotyping, 5-m thick sections were stained
using a modified streptavidin-biotin complex method with
antigen retrieval as required. Briefly, the following reagents
were used in sequential steps at 36C: inhibitor for endog-
enous peroxidase, primary antibody for 12 h, biotinylated
secondary antibody, avidin-biotin complex with horseradish
peroxidase, 3,3-diaminobenzidine tetrahydrochloride
(DAB). When indicated, slides were pretreated for antigen
retrieval by steam for 15 min in a Black and Decker(Schaumburg, IL) steamer in 10 mmol/L citrate buffer (pH
6.0), followed by cooling for 20 min. Slides were counter-
stained with hematoxylin. The antibody panel included CK
7, 20, 4, and 13 from Dako (Carpinteria, CA). After grading
slides for intensity of stain, intestinal metaplasia glands
were visually divided into thirds (upper, middle, and lower).
The staining pattern was considered superficial if the upper
third or two thirds was positive. Slides were then analyzed
for the different patterns as defined by Ormsby et al.(10),
e.g., superficial staining with CK 20 and strong CK 7 stain-
ing of both superficial and deep glands was defined as a
Barretts CK7/20 profile. The whole area of metaplasia was
also screened for patchy or diffuse staining.
Statistical Analyses
Scores were analyzed using Sigma Stat (Jandel Scientific
Software, San Rafael, CA). Fishers exact test or, whenappropriate, the 2 test (both two-tailed) were used for
comparison of proportions.
RESULTS
Intestinal Metaplasia
A total of 116 biopsies with intestinal metaplasia from 102
patients were examined (33 Barretts esophagus, 23 cardia,
27 body, and 33 antral). CK 4 and CK 13 stained squamous
epithelium only; areas with intestinal metaplasia remained
unstained. Intestinal metaplasia was incomplete in 79% and
65% of patients presenting with intestinal metaplasia in the
esophagus and cardia, respectively.
Intestinal Metaplasia in Barretts
A total of 33 biopsies from 29 patients with Barretts esoph-
agus were examined. The proposed diagnostic CK Barretts
7/20 as defined by Ormsbyet al.(10) was found in only 39%
of biopsies (Table 1) (Fig. 1).
Incomplete intestinal metaplasia was present in 27 biop-
sies from 23 patients; 11 biopsies had the esophageal pat-
tern, 12 biopsies had a gastric pattern, and four had other
patterns. Two patients had multiple biopsies. The first pa-
tient had two biopsies and both had the esophageal pattern.
The second patient had four biopsies, one had an esophagealpattern, and three had a gastric pattern.
Six biopsies from six patients had complete intestinal
metaplasia. Only three (50%) had the diagnostic Barrett
pattern.
Intestinal Metaplasia in the Stomach
A total of 83 gastric biopsies with intestinal metaplasia were
examined (23 cardia, 27 body, and 33 antral).
Intestinal Metaplasia in the Cardia
In all, 23 biopsies from 18 patients were examined. The
proposed diagnostic CK Barretts 7/20 pattern as defined
by Ormsby et al. (10) was found in eight of 23 biopsies
Table 1. Cytokeratin 7/20 With Barretts Esophagus and Gastric Intestinal Metaplasia, by Site
DiseaseN
(biopsies)Barretts CK7/20
PatternGastric CK7/20
PatternOther
Patterns
Barretts esophagus 33 13 (39%) 15 (15%) 5 (15%)6 complete 2 (33%) 3 (50%) 1 (17%)
27 incomplete 11 (41%) 12 (44%) 4 (15%)
Gastric intestinal metaplasiaCardia 23 8 (35%) 7 (30%) 8 (35%)
8 complete 2 (25%) 1 (13%) 5 (63%)15 incomplete 6 (40%) 6 (40%) 3 (20%)
Body 27 1 (4%) 7 (26%) 17 (63%)12 complete 0 5 (42%) 7 (58%)11 incomplete 1 (9%) 3 (27%) 7 (64%)
Antrum 33 8 (24%) 5 (15%) 20 (61%)14 complete 1 (7%) 2 (14%) 11 (79%)19 incomplete 7 (37%) 3 (16%) 9 (47%)
1379AJG May, 2001 Cytokeratin Subsets in Barretts Versus Stomach
-
8/12/2019 Citokeratina en Esofago de Barret
3/5
(35%), (Table 1). Incomplete intestinal metaplasia was
present in 15 biopsies from 11 patients. The esophageal
pattern was present in six (40%) (Fig. 2). Four patients had
two biopsies each. An esophageal pattern was seen in both
biopsies in one patient; the second patient had a gastric
pattern in both biopsies; and two patients had a mixed
pattern (i.e., one esophageal and one gastric in one patient,
and one gastric and one other pattern in the other patient).
Complete intestinal metaplasia was present in eight bi-
opsies from seven patients. The diagnostic pattern was
present in two of eight biopsies with complete intestinal
metaplasia (25%). One patient had two biopsies, one biopsy
had an esophageal pattern and the other biopsy had super-ficial staining of the glands with CK 7 and CK 20 (other
pattern).
Intestinal Metaplasia in the Corpus
The proposed diagnostic CK Barretts 7/20 pattern as de-
fined by Ormsby et al. (10) was found in only one of 27
biopsies (3%) (95% C.I. 0% to 20%) (p 0.0036 com-
pared to Barretts) (Table 1). The proposed diagnostic CK
Barretts 7/20 pattern was found in only one of 11 biopsies
(11 patients) with incomplete intestinal metaplasia. Al-
though CK 7 stained two cases with incomplete intestinal
metaplasia in the body, the characteristic Barretts stain-
ing of the entire length of the gland was observed in one
patient only. Similarly, complete intestinal metaplasia was
present in 12 biopsies from 11 patients; none had the Bar-
retts pattern.
Intestinal Metaplasia of the Antrum
The proposed diagnostic CK Barretts 7/20 pattern as de-fined by Ormsbyet al.(10) was found in nine of 33 biopsies
(27%) (Table 1). Incomplete intestinal metaplasia was
present in 19 biopsies from 16 patients. Seven biopsies
(37%) had the esophageal pattern. Three patients had two
biopsies each. One patient had an esophageal pattern and a
gastric pattern. The other patient had an esophageal pattern,
and superficial staining with CK 20 with no staining with
Figure 1. Representative section for cytokeratins 20 (A) and 7(B) in the esophagus. The proposed diagnostic pattern was observed in only39% of biopsies. Contrary to the proposed diagnostic pattern, CK 20 was expressed in the entire length of the gland in 17 biopsies (59%),and CK 7 was superficial or mixed, with areas of no staining in eight cases (28%).
Figure 2. Representative section for cytokeratins 20 (A)and 7 (B) in the cardia. The diagnostic cytokeratin Barretts 7/20 was found in35% of biopsies. Superficial staining with CK 20 was observed in 65% of biopsies, and CK 7 stained the entire length of the gland in 52%.
1380 El-Zimaity et al. AJG Vol. 96, No. 5, 2001
-
8/12/2019 Citokeratina en Esofago de Barret
4/5
-
8/12/2019 Citokeratina en Esofago de Barret
5/5
pattern of Helicobacter pylori gastritis in the gastric cardia.Am J Gastroenterol 1997;92:22204.
3. Goldblum JR, Vicari JJ, Falk GW, et al. Inflammation andintestinal metaplasia of the gastric cardia: The role of gastro-esophageal reflux and H. pylori infection. Gastroenterology1998;114:6339.
4. el-Serag HB, Sonnenberg A, Jamal MM, et al. Characteristics
of intestinal metaplasia in the gastric cardia. Am J Gastroen-terol 1999;94:6227.
5. Craanen ME, Blok P, Dekker W, Tytgat GN. Helicobacterpylori and early gastric cancer. Gut 1994;35:13724.
6. Voutilainen M, Farkkila M, Juhola M, et al. Specialized co-lumnar epitheilium of the esophagogastric junction: Preva-lence and associations. Am J Gastroenterol 1999;94:913 8.
7. Pereira AD, Suspiro A, Chaves P, et al. Short segments ofBarretts epithelium and intestinal metaplasia in normal ap-pearing oesophagogastric junctions: The same or two differententities? Gut 1998;42:65962.
8. Oberg S, Peters JH, DeMeester TR, et al. Inflammation andspecialized intestinal metaplasia of cardiac mucosa is a man-ifestation of gastroesophageal reflux disease. Ann Surg 1997;
226:52232.9. Spechler SJ, Zeroogian JM, Antonioli D, et al. Prevalence of
metaplasia at the gastro-oesophageal junction. Lancet 1994;344:15336.
10. Ormsby AH, Goldblum JR, Rice TW, et al. Cytokeratin sub-sets can reliably distinguish Barretts esophagus from intesti-nal metaplasia of the stomach. Hum Pathol 1999;30:28894.
11. El-Zimaity HM, Al-Assi MT, Genta RM, Graham DY. Con-firmation of successful therapy of Helicobacter pyloriinfection: Number and site of biopsies or a rapid urease test.Am J Gastroenterol 1995;90:19624.
12. Genta RM, Robason GO, Graham DY. Simultaneous visual-ization ofHelicobacter pyloriand gastric morphology: A newstain. Hum Pathol 1994;25:2216.
13. El-Zimaity HMT, Ota H, Scott S, et al. A new triple stain forHelicobacter pylori suitable for the autostainer. Arch PatholLab Med 1998;122:7326.
14. El-Zimaity HMT, Graham DY, Al-Assi MT, et al. Interob-server variation in the histopathological assessment of Heli-cobacter pylori gastritis. Hum Pathol 1996;27:35 41.
15. Spicer SS. Diamine methods for differentiating mucosub-
stances histochemically. J Histochem Cytochem 1965;13:21134.
16. Rokkas T, Filipe MI, Sladen GE. Detection of an increasedincidence of early gastric cancer in patients with intestinalmetaplasia type III who are closely followed up. Gut 1991;32:11103.
17. Silva S, Filipe MI, Pinho A. Variants of intestinal metaplasia
in the evolution of chronic atrophic gastritis and gastric ulcer.A follow up study. Gut 1990;31:1097104.
18. Moll R, Franke WW, Schiller DL, et al. The catalog of humancytokeratins: Patterns of expression in normal epithelia, tu-mors and cultured cells. Cell 1982;31:1124.
19. Petras RE, Sivak MV, Rice TW. Barretts esophagus: A re-view of the pathologists role in diagnosis and management.Pathol Annu 1991;26:132.
20. Krause WJ, lvey KJ, Baskin WN, et al. Morphological obser-vations on the normal human cardiac glands. Anat Rec 1978;192:5972.
21. Appelman HD, Kalish RJ, Clancy PE, et al. Distinguishingfeatures of adenocarcinoma in Barretts esophagus and in thegastric cardia. In: Spechler SJ, Goyal RK, eds. Barrettsesophagus: Pathophysiology, diagnosis, and management, 3rd
ed. New York: Elsevier, 1985:16787.22. Ellison E, Hassall E, Dimmick JE. Mucin histochemistry of
the developing gastroesophageal junction. Pediatr Pathol LabMed 1996;16:195206.
23. Antonioli DA, Goldman H. Changes in the location and typeof gastric adenocarcinoma. Cancer 1982;50:77581.
24. Blot WJ, Devesa SS, Kneller RW, et al. Rising incidence ofadenocarcinoma of the esophagus and gastric cardia. JAMA1991;265:12879.
25. Pera M, Cameron AJ, Trastek VF, et al. Increasing incidenceof adenocarcinoma of the esophagus and esophagogastricjunction. Gastroenterology 1993;104:5103.
26. Craanen ME, Dekker W, Blok P, et al. Time trends in gastriccarcinoma: Changing patterns of type and location. Am J
Gastroenterol 1992;87:5729.27. Bensley RR. The cardiac glands of mammals. Am J Anat1902;2:10556.
28. Miyagawa Y. The exact distribution of the gastric glands inman and in certain animals. J Anat 1921;55:5667.
29. Takubo K, Mafune K, Tanaka Y, et al. Pathology of the cardia.Nippon Geka Gakkai Zasshi 1998;9:54751.
1382 El-Zimaity et al. AJG Vol. 96, No. 5, 2001