World J. Surg. 16, 308-312, 1992
World Journal of Surgery O 1992 by the Soci6t~ Internationale de Chimrgic
The Neoplastic Potential of Columnar-Lined (Barrett's) Esophagus G . N . J . Tytgat, M.D. and W. H a m e e t e m a n , M.D. Department of Hepatogastroenterology, University of Amsterdam, Amsterdam, The Netherlands Reflux of gastric contents is the most important factor contributing to the development of a columnar-lined esophagus (CLE). CLE should be considered a "mosaic" of cells, glands, and architectural types showing variable degrees of maturation towards intestinal and gastric epithelium. Dysplasia refers to an unequivocally neoplastic alteration in the epithelium. High-grade (severe) dysplasia consists of neoplastic cellular and/or architectural changes within the columnar epithelium, without invasion through the basement membrane. The prevalence of dysplasia is usually 5% to 10%. Dysplasia has been found in the vast majority in the specialized columnar-type or intestinal-type epithelium. Adenocarcinoma in CLE is most often seen in males; At the time of diagnosis transmural tumor infiltration is found in 60% to 88% of patients, and lymph node invasion in 55% to 74% of patients. The prognosis is poor with a low 5 year survival. Estimates of incidence of cancer vary from 1 case per 52 patient years of follow-up to 1 case per 441 patient years of follow-up.
D e v e l o p m e n t and Characteristics o f C o l u m n a r - L i n e d Esophagus
Reflux of gastric contents, especially of acid, is probably the most important factor contributing to the development of a columnar-lined esophagus (CLE). Also delayed clearance of the refluxed material, secondary to impaired esophageal motility, may contribute to prolonged exposure of the esophagus to acid. Endoscopic studies have shown progressive upward migration of columnar mucosa in the presence of persistent gastroesophageal reflux, with erosions and ulceration of the squamous epitheIium [I, 2]. CLE should be considered a "mosaic" of cells, glands, and architectural types showing variable degrees of maturation towards intestinal and gastric epithelium [3]. Architectural forms vary from foveolar, to ridged, to true villous forms. Besides mucus secreting columnar cells, goblet cells, neuroendocrine cells, and Paneth cells, chief and parietal cells may be found. Absorptive cells with a basally located nucleus, a uniform cytoplasm, and a prominent terminal webb comparable to small intestinal absorptive enterocytes may also be found. There are variable degrees of superimposed inflammation and atrophy [3, 4]. We found an incomplete variant of intestinal metaplasia, called the variant or pseudoabsorptive cell, with apical cytoplasmic mucin and an incompletely develReprint requests: Prof. Dr. G.N. Tytgat, Maag-, Darm-, Leverziekenten, Academisch medisch Centrum, Meibergdreef 9, NL-1105 AZ, Amsterdam, The Netherlands.
oped or absent brushborder and microviili, in up to 80% of our patients [5, 6]. In several studies a zonation has been suggested with fundictype epithelium located in the more distal esophagus and specialized columnar-type epithelium located in the proximal columnar segment [7]. An increase in the frequency of intestinal-type epithelium with increasing length in the columnar-lined segment has been reported. Dysplasia
Dysplasia refers to an unequivocally neoplastic alteration in the epithelium. Dysplasia may vary from mild or low-grade to severe or high-grade. In the European literature, Morson's criteria of dysp|asia are used [8]; in the American studies, criteria derived from dysplasia in inflammatory bowel disease are used [9, 10]. High-grade (severe) dysplasia consists of neoplastic cellular and/or architectural changes within the columnar epithelium without invasion through the basement membrane. Low-grade dysplasia consists of less striking changes that are in(ermediate between high-grade dysplasia and ordinary Barrett's metaplasia. In the absence of overt carcinoma, the prevalence of dysplasia is usually 5% to 10% [11, 12]. Several arguments support the view that high-grade dysplasia is to be considered as an unequivocal premalignant lesion. Studies of patients with Barrett's adenocarcinoma have demonstrated dysplasia in approximately 90% of cases [3, 13-15]. Esophagectomy for high-grade dysplasia can reveal previously unrecognized adenocarcinoma in the specimens [11, 16-18]. In endoscopic follow-up studies, detection of a carcinoma was preceded by the detection of increasing severity of dysplasia [5, 19, 20]. Though dysplasia has been found in all epithelial types, in the vast majority it has been seen in the specialized columnar- or intestinal-type epithelium [5]. While the presence of high-grade dysplasia is an ominous sign, the time course for development of invasive carcinoma from high-grade dysplasia varies considerably. High-grade dysplasia can be present for more than 3-4 years without evidence of invasive carcinoma, as shown by endoscopic ultrasonography and/or surgery [5, 17, 20, 21]. Even less information is available concerning the natural history of low-grade dysplasia. It is not known if adenocarcinoma can arise directly from low-grade dysplasia.
G.N.J. Tytgat and W. Hameeteman: Neoplasia and Barrett's Esophagus
309
Table 1. Screening for adenocarcinoma in patients with columnar-lined esophagus.
Study (yr.) (reference)
Screening method
No. of pts.
Mean follow-up (yr.)
No. of carcinomas
Incidence per patient year follow-up
Spechler (1984) (33) Sprung (1984) (34) Cameron (1988) (32) Sampliner (1986) (35) Robertson (1988) (19) v.d. Veen (1989) (36) Hameeteman (1989) (5) Total Endoscopically screened
Endoscopy Endoscopy Post" Endoscopy Endoscopy Post~ Endoscopy
105 41 122 25 56 155 50 554 276
3.3 4.0 8.5 2.1 3.0 4.4 5.2 4.9 3.6
2 2 2 I 4 4 5 20 14
1/175 1/81 1/441 1/92 1/56 1/170 1/52 1/135 1/71
"Post: follow-up by postal inquiries of patients, relatives or general practitioner.
Adenocarcinoma
Many reports have stressed the association between CLE and esophageal adenocarcinoma. Adenocarcinoma in CLE is most often seen in males with an average age of 55-60 years. The male-female ratio is approximately 5.5:i. Dysphagia is a prominent presenting symptom reported in 70% to 95% of patients. Weight loss is present in 42% to 63% of patients. Bleeding from the malignancy has been reported in 11% to 36% of patients. Carcinoma in CLE may be present multifocally, nearly always surrounded by dysplasia in the adjacent columnar epithelium. At the time of diagnosis the tumor is usually far advanced; transmural tumor infiltration is found in 60% to 88% of patients, and lymph node invasion in 55% to 74% of patients [14, 15, 22-24]. The results of esophagectomy for Barrett's carcinoma are comparable to the results obtained for other esophageal carcinomas. The extent of wall penetration and spread of tumor to lymph nodes influence the prognosis. With the majority of patients presenting with an advanced tumor, the prognosis is poor with a 5 year survival of 7% [15], 16% [25], or 22% [23]. With tumor limited to the mucosa or submucosa and without lymph node invasion, prognosis is considerably better and cure can be obtained [15, 17, 26-29]. While the association between CLE and adenocarcinoma has been demonstrated beyond doubt, there is still debate on the magnitude of the risk. The reported prevalence of adenocarcinoma in CLE (the coexistence of the two lesions at a single time) ranges from zero to 46.5% [30, 31]. Interpretation of these figures is difficult as the total population from which these patients were obtained is hard to define. Patients without symptoms will not be included as they obviously will not visit their physician. Cameron and coworkers [32] suggested from comparison of clinical incidence and autopsy figures that only 5% to 10% of patients with CLE might come to medical attention. Patient selection thus is likely to influence prevalence figures considerably. To assess the incidence of carcinoma in CLE (the frequency of the occurrence of cancer developing over time), a long follow-up is necessary of patients who had no detectable neoplasm at the initial examination. Estimates of incidence differ considerably from 1 case per 52 patient years of follow-up to 1 case per 441 patient years of follow-up (Table 1). Such variation is to be expected as none of the series rePorted more than 5 cases of carcinoma, and in 4 studies only 1 or 2 cases were collected. Variation in incidence may also be related to the
method of follow-up. The more thorough the follow-up, the more carcinomas might be revealed [37]. Surveys in which regular endoscopy is used give better and often higher incidence rates. Considering the endoscopic studies only, the average incidence of adenocarcinoma is 1 case in 71 patient years of follow-up. Compared with the risk in the general population, patients with a CLE therefore have a 30-fold to 350-fold increased risk of developing an esophageal carcinoma [5, 19, 31, 33]. Dysplasia-Carcinoma Sequence
In recent prospective endoscopic follow-up studies [5, 19, 20], the detection of carcinoma was preceded by increasing severity of dysplasia. In Robertson's [19] endoscopic surveillance study of 56 patients with CLE, 4 patients developed high-grade dysplasia and 3 patients developed carcinoma. Two of the adenocarcinomas were preceded by progressively more severe dysplastic changes. In our own Amsterdam study [5] in which patients with a CLE were seen for endoscopy and histologic investigation at regular intervals, 50 patients were followed for a mean period of 5.2 years, At study entrance 7 patients were found to have dysplasia, 6 scored as low-grade, 1 as high-grade. At the end of the study period dysplasia was found in 13 patients, 10 low-grade and 3 high-grade. Moreover, adenocarcinoma had developed in 5 patients. In the years preceding the detection of the carcinoma, biopsy specimens had shown increasing degrees of severity of dysplasia (Fig. 1). Similar observations were recently presented by Schnell and coworkers [20], emphasizing that patients with high-grade dysplasia may progress to adenocarcinoma. In general there is wide variation in the time course of the development of dysplasia and carcinoma. With CLE usually being a consequence of longlasting gastro-esophageal reflux, one might question whether the development of dysplasia is also a consequence of persisting reflux. There are no clear answers to this question. Development of carcinoma in CLE has been described after anti-reflux surgery, but there is no information on the efficacy of the various procedures in the abolishment of reflux. It has been stated that an effective anti-reflux operation might prevent an increase in severity of dysplasia [29]. True regression of CLE is a rare event [38, 39]. Recently regression was described in a few patients after extensive anti-reflux surgery [40] and after long term medical acid suppressive treatment with 60 mg of omeprazole [41]. In
310
World J. Surg. Vol. 16, No. 2, Mar./Apr. 1992
DYSPLASIA ,
.3
f¢
/
.2
/
/
/
/
I:
/'
•
H
./" i/
f"
/ // i
L /"
\
[
x .,"
/
/
"%..
!/
N 0
z
4
/ "~ . .
6
,.~°"
i¢
i
1
8
t0
YEARS
earlier surgical studies in which mainly a Nissen fundoplication was performed, regression was seen only rarely and the methods by which this regression had been established have frequently been criticized.
Fig. 1. Development and progression of dysplasia in 5 patients with columnar-lined esophagus who developed carcinoma. N: negative for dysplasia; L: low-grade dysplasia; H: high-grade dysplasia; C: carcinoma.
DETECTION OF DYSPLASIA
/
Have diagnosis conflrmed.~by another pathologist Intensive medical therapy 8 - 12 weeks
Practical Recommendations
CLE is a premalignant condition with an increased risk of esophageal carcinoma. Development of adenocarcinoma is high.~rade low/moderate no~dysplasla / dysRlasla ~ likely to be preceded by an increase in severity of the dysplastic changes. Dysplasia develops predominantly in intestinal-type En~osono Surveillance Surveillance nography eptihelium. It is essential to take multiple biopsy specimens Repeat endoscopy every year_ . every 2years from the entire columnar segment to minimize the risk of every 3 - 6 months till age 70(?) till age 70(?) or missing intestinal-type epithelium and/or dysplasia. Two to 4 biopsy specimens should be taken, in a circular fashion, from esophageal resection every 2-3 cm of CLE, with additional biopsies from any mucosal abnormalities. High-grade dysplasia may be present in Fig. 2. A flow diagram for a screening program for patients with columnar-lined esophagus. patients with CLE for as much as 4 years without signs of progression to carcinoma, as seen by endoscopy with biopsies and endoscopic ultrasonography. The risk of cancer for CLE R6sum~ seems to warrant an effective surveillance program [42, 43]. A Le reflux du contenu gastrique est le facteur le plus important flow diagram for such a screening program is shown in Figure 2. qui contribue au d6veloppement de l'endobrachyoesophage. The efficacy of a surveillance program in reducing the morbidity L'endobrachyoesophage pourrait 6tre consid6r6 comme une and mortality of esophageal cancer remains to be demonstrated "mosa~que" de cellules, de glandes et types d'architectures, [44]. Obviously surveillance is only indicated for those patients pr6sentant des degr6s variables de maturation vers l'6pith61ium in whom esophagectomy is considered a therapeutic option if intestinal et gastrique. La dysplasie consiste en une alt6ration early malignancy is discovered. Recently investigators at the n6oplasique indiscutable de l'6pithElium. Une dysplasie s6v6re Cleveland Clinic have estimated the direct costs of annual (de grade 61ev6) comporte des celiules n6oplasique et/ou des endoscopic surveillance for CLE at $62,000 for each carcinoma modifications architecturales de l'6pith61ium cylindrique sans detected [45]. A program of surveillance every other year would franchissement de la membrane basale. La pr6valence de la cost substantially less. dysplasie est habituellement de l'ordre de 5 ~ 10%. La dysplasie Dysplasia is the only clinically useful parameter important in est, dans la majorit6 des cas, mise en 6vidence dans l'6pith61ium follow-up studies. Despite intensive investigation, all other cylindrique ou intestinal. L'ad6nocarcinome au sein de l'endotechniques proposed to identify epithelia with high malignant brachyoesophage est retrouv6 principalement chez l'homme. potential are of doubtful clinical usefulness at present. Recently Au moment du diagnostic, une infiltration tumorale transpariresearchers in the United States of America have developed a 6tale est retrouv6e dans 60 ~i 88% des cas, et un envahissement technique for culturing endoscopic Barrett biopsy specimens. ganglionnaire dans 55 h 74% des cas. Le pronostic est mauvais Cytogenetic analysis of the cultured tissue revealed chromo- avec un taux faible de survie h 5 ans. L'estimation de l'incisome 7 abnormalities associated with enhanced expression of dence de cancer est variable, allant de 1 cas sur 52 gt I cas sur epidermal growth factor receptors [46]. 441 patients par ann6e de suivi.
G.N.J. Tytgat and W, Hameeteman: Neoplasia and Barrett's Esophagus
311
Resumen E1 reflujo del contenido g~tstrico es el factor m~is importante en el desarrollo del es6fago de Barrett. E1 es6fago de Barrett debe ser considerado como un " m o s a i c o " de crlulas, gl~indulas y tipos arquitect6nicos, que exhiben grados variables de maduraci6n hacia el epitelio intestinal y g~istrico. El termino displasia se refiere a una inequivoca alteracirn neopl~sica del epitelio. El alto grado (severo) de displasia consistente en cambios celulares neopl~isicos y/o arquitectrnicos del epitelio columnar, sin invasi6n de la membrana basal. L a incidencia de la displasia es Usualmente del orden de 5-10%. La displasia ha sido hallada principalmente en el epitetio del tipo columnar especializado o intestinal. El adenocarcinoma en el es6fago de Barrett se observa principalmente en hombres. En el momento del diagn6stico la infiltraci6n transmural del tumor se encuentra con una incidencia de 60-88%, y la invasi6n de n6dulos linf~iticos con incidencia de 55-74% de los casos. El pron6stico es pobre, on baja posibilidad de sobrevida a 5 afios. La incidencia estimada de c~ncer varia desde 1 caso por 52 a 1 caso por 441 pacientes afio de seguimiento.
14.
15. 16. 17.
18. 19. 20.
21. References I. Endo, M., Kobayashi, S., Kozu, T., Takemoto, T., Nakayama, K,: A case of Barrett epithelization followed up for five years. Endoscopy 18(suppl 3): 11, 1974 2. Savary, M., Monnier, P.: Diagnosis, pathophysiology and adenocarcinogenesis of Barrett's esophagus. In Esophageal Disorders: Pathophysiology and Therapy, T.R. DeMeester, D.B. Skinner, editors, New York, Raven, pp. 101-8, 1985 3. Thompson, J.J., Zinsser, K.R., Enterline, H.T.: Barrett's metaplasia and adenocarcinoma of the esophagus and gastroesophageal junction. Hum. Pathol. 14:43, 1983 4. Levine, D.S., Rubin, C.E., Reid, B.J., Haggitt, R.C.: Specialized metaplastic columnar epithelium in Barrett's esophagus: A comparative transmission electron microscopy study, Lab. Invest. 60:418, 1989 5. Hameeteman, W., Tytgat, G.N.J,, Houthoff, H.J., van den Tweel, J.G.: Barrett's esophagus, development of dysplasia and adenocarcinoma. Gastroenterology 96:1249, 1989 6. Hameeteman, W.: Columnar-lined (Barrett's) esophagus. Thesis, University of Amsterdam 1989; Drukkery All Zn B.V., Katwijk, The Netherlands 7. Wesdorp, I.C.E., Bartelsman, J., Schipper, M.E.I., Offerhaus, J., Tytgat, G.J.: Malignancy and premalignancy in Barrett's esophagus: A clinical, endoscopical, and histological study of 100 patients. Gastroenterology 11:317, 198t 8. Morson, B.C., Sobin, L.H., Grundmann, E., Johansen, A., Nagayo, T., Serck-Hansen, A.: Precancerous conditions and epithelial dysplasia in the stomach. J. Clin. Pathol. 33:711, 1980 9. Riddell, R.H.: Dysplasia and regression in Barrett's epithelium. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 143-52, 1985 10. Reid, R.J., Haggitt, R.C., Rubin, C.E., Rabinovitch, P.S.: Barrett's esophagus: Correlation between flow cytometry and histology in detection of patients at risk for adenocarcinoma. Gastroenterology 93:I, 1987 tl. Schmidt, H.G., Riddell, R.H., Walther, B., Skinner, D.B., Riemann, J.F: Dysplasia in Barret.t's esophagus. J. Cancer Res. Clin.. Oncol. 110:145, 1989 12. Kerlin, P., D'Mellow, G., van Deth, A.: Barrett's esophagus: Clinical, endoscopic, and histologic spectrum in fifty patients. Aust. N.Z.J. Med. 16:198, 1986 13. Hamilton, S.R., Smith, R.R.L., Cameron, J.L.: Prevalence and characteristics of Barrett's esophagus in patients with adenocarci-
22. 23. 24. 25. 26. 27. 28. 29.
30.
31. 32.
33. 34. 35. 36.
noma of the esophagus or esophagogastric junction. Hum. Pathol. 19:942, 1988 Haggitt, R.C., Dean, P.J.: Adenocarcinoma in Barrett's epithelium. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 153-66, 1985 Monnier, P.H., Frbntolliet, C., Savary, M., Ollyo, J.B.: Barrett's oesophagus or columnar epithelium of the lower oesophagus. Bailliere's Clin. Gastroenterol. •:769, 1987 Skinner, D.B., Walther, B.C., RiddeU, R.H., Schmidt, H., lascone, C., DeMeester, T,R.: Barrett's esophagus: Comparison of benign and malignant cases. Ann. Surg. 198:554, 1983 Reid, R.J., Weistein, W.M., Lewin, K.J.: Endoscopic biopsy detect early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic lesions. Gastroenterology 94:81, 1988 Palley,S.L., Sampliner, R., Garewal, H.S.: Editorial: Management of high-grade dysplasia in Barrett's esophagus. J. Clin. Gastroenterol. 11:369, 1989 Robertson, C.S., Mayberry, J.F., Nicholson, D.A., James, P.D., Atkinson, M.: Value of endoscopic surveillance in the detection of neoplastic change in Barrett's oesophagus. Br. J. Surg. 75:760, 1988 Schnell, T., Sontag, S., Chejfec, G., Chintam, R., O'Connell, S., Kurucar, C.: High grade dysplasia in Barrett's esophagus: Experience with 36 patients (abstract). American College of Gastroenterology, October 23, 1989 Hamilton, S.R., Smith, R.R.L.: The relationship between columnar epithelial dysptasia and invasive adenocarcinoma arising in the Barrett's esophagus. Am. J. Clin. Pathol. 87:301, 1987 Haggitt, R.C., Tryzelaar, J., Ellis, F.H., Colcher, H.: Adenocarcinoma complicating columnar epithelium-lined (Barrett's) esophagus. Am. J. Clin. Pathol. 70:1, 1978 Skinner, D.B.: En bloc resection for neoplasms of the esophagus and cardia. J. Thorac. Cardiovasc. Surg. 85:59, 1983 Steiger, Z., Wilson, R.E., Leichman, L., Busuito, M.J., Rosenberg, J.C.: Primary adenocarcinoma of the esophagus. J. Surg. Oncol. 36:68, 1987 Sanfey, H., Hamilton, S.R., Smith, R.R.L.: Carcinoma arising in Barrett's esophagus. Surg. Gynecol. Obstet. 161:570, 1985 Witt, T.R., Bains, M.S., Zaman, M.B., Martini, N.: Adenocarcinoma in Barrett's esophagus. J. Thorac. Cardiovasc. Surg. 85:337, 1983 Smith, R.R.L., Hamilton, S.R., Boinott, J.K., Rogers, E.L.: The spectrum of carcinoma arising in Barrett's esophagus. Am. J. Surg. Pathol. 8:563, 1984 Rosenberg, J.C., Budev, H., Edwards, R.C.: Analysis of adenocarcinoma in Barrett's esophagus utilizing a staging system. Cancer 55:1353, 1985 Skinner, D.B., Walther, B.C., Little, A.G.: Surgical treatment of Barrett's esophagus. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 211-21, t985 Bremner, C.G., Hamilton, D.G.: Barrett's esophagus: Controversial aspects. In Esophageal Disorders: Pathophysiology and Therapy, T.R. DeMeester, D.B. Skinner, editors, New York, Raven, pp. 233-9, 1985 Cameron, A.J., Ott, B.J., Payne, W.S.: The incidence of adenocarcinoma in columnar-lined (Barrett's) esophagus. N. Engl. J. Med. 313:857, 1985 Cameron, A.J., Zinsmeister, A.R., Ballard, D.J., Carney, J.A.: A population-based study of Barrett's esophagus: Comparison of clinically diagnosed prevalence versus autopsy determined prevalence. Gastroenterology 94:A57, 1988 Spechler, S.J., Schimmel, E.M., Dalton, J.W., Doos, W., Tries, J.S.: Barrett's epithelium complicating lye ingestion with sparing of the distal esophagus. Gastroenterology 81:580, t981 Sprung, D.J., Ellis, F.H., Gibb, S.P.: Incidence of adenocarcinoma in Barrett's esophagus. Am. J. Gastroenterol. 79:817A, 1984 Sampliner, R.E.: Editorial. Does Barrett's esophagus need to be detected? Arch. Intern. Med. 96:438A, 1986 van der Veen, A.H., Dees, J., Blankensteijn, J.D,, van Blankenstein, M.: Adenocarcinoma in Barrett's oesophagus: An overrated risk. Gut 30:14, 1989
312 37. Atkinson, M.: Barrett's oesophagus: To screen or not to screen? Gut 30:2, 1989 38. Wesdorp, I.C.E., Bartetsman, J., Schipper, M.E.I., Tytgat, G.N.: Effect of long-term treatment with cimetidine and antiacids in Barrett's oesophagus. Gut 22:724, 1981 39. Sampliner, R.E., Garewal, H.S., Fennerty, M.B., Aickin, M.: Lack of impact of therapy on extent of Barrett's esophagus in 67 patients. Dig. Dis. Sci. 35:93, 1990 40. Perniceni, T., Leymarios, J., Molas, G., Fekete, F.: L'endobrachyoesophage regresse-t-il apres diversion duodenale totale? Gastroenterol. Clln, Biol. •2:709, 1988 41. DeViere, J., Buset, M., Dumonceau, J.M., Rickaert, F., Cremer, M.: Regression of Barrett's epithelium with omeprazole. N. Engl. J. Med. 320:1497, 1989
World J. Surg. Voi. 16, No. 2, Mar./Apr. 1992 42. Spechler, S.J.: Barrett's esophagus: What's new and what to do. Am. J. Gastroenterol. 84:220, 1989 43. Dent, J.: Approaches to oesophageal columnar metaplasia (Barrett's oesophagus), Scand. J. Gastroenterol. 24(suppl 168):60, 1989 44. Spechler, S.J.: Endoscopic surveillance for patients with Barretts' esophagus: Does the cancer risk justify the practice? Ann. Intern. Med. 106:902, 1987 45. Achkar, E., Carey, W.C.: The cost of surveillance for adenocarcinoma complicating Barrett's esophagus. Am. J. Gastroenterol. 83:291, 1988 46. Garewal, H., Leibovitz, A., Sampliner, R., Prabhata, R., Trent, J., Korc, M., Sloan, D.: Tissue culture and characterization of epithelial cells devised from Barrett's esophagus, a premalignant lesion. Gastroenterology 94:A143, 1988
World Journal of Surgery O 1992 by the Soci6t~ Internationale de Chimrgic
The Neoplastic Potential of Columnar-Lined (Barrett's) Esophagus G . N . J . Tytgat, M.D. and W. H a m e e t e m a n , M.D. Department of Hepatogastroenterology, University of Amsterdam, Amsterdam, The Netherlands Reflux of gastric contents is the most important factor contributing to the development of a columnar-lined esophagus (CLE). CLE should be considered a "mosaic" of cells, glands, and architectural types showing variable degrees of maturation towards intestinal and gastric epithelium. Dysplasia refers to an unequivocally neoplastic alteration in the epithelium. High-grade (severe) dysplasia consists of neoplastic cellular and/or architectural changes within the columnar epithelium, without invasion through the basement membrane. The prevalence of dysplasia is usually 5% to 10%. Dysplasia has been found in the vast majority in the specialized columnar-type or intestinal-type epithelium. Adenocarcinoma in CLE is most often seen in males; At the time of diagnosis transmural tumor infiltration is found in 60% to 88% of patients, and lymph node invasion in 55% to 74% of patients. The prognosis is poor with a low 5 year survival. Estimates of incidence of cancer vary from 1 case per 52 patient years of follow-up to 1 case per 441 patient years of follow-up.
D e v e l o p m e n t and Characteristics o f C o l u m n a r - L i n e d Esophagus
Reflux of gastric contents, especially of acid, is probably the most important factor contributing to the development of a columnar-lined esophagus (CLE). Also delayed clearance of the refluxed material, secondary to impaired esophageal motility, may contribute to prolonged exposure of the esophagus to acid. Endoscopic studies have shown progressive upward migration of columnar mucosa in the presence of persistent gastroesophageal reflux, with erosions and ulceration of the squamous epitheIium [I, 2]. CLE should be considered a "mosaic" of cells, glands, and architectural types showing variable degrees of maturation towards intestinal and gastric epithelium [3]. Architectural forms vary from foveolar, to ridged, to true villous forms. Besides mucus secreting columnar cells, goblet cells, neuroendocrine cells, and Paneth cells, chief and parietal cells may be found. Absorptive cells with a basally located nucleus, a uniform cytoplasm, and a prominent terminal webb comparable to small intestinal absorptive enterocytes may also be found. There are variable degrees of superimposed inflammation and atrophy [3, 4]. We found an incomplete variant of intestinal metaplasia, called the variant or pseudoabsorptive cell, with apical cytoplasmic mucin and an incompletely develReprint requests: Prof. Dr. G.N. Tytgat, Maag-, Darm-, Leverziekenten, Academisch medisch Centrum, Meibergdreef 9, NL-1105 AZ, Amsterdam, The Netherlands.
oped or absent brushborder and microviili, in up to 80% of our patients [5, 6]. In several studies a zonation has been suggested with fundictype epithelium located in the more distal esophagus and specialized columnar-type epithelium located in the proximal columnar segment [7]. An increase in the frequency of intestinal-type epithelium with increasing length in the columnar-lined segment has been reported. Dysplasia
Dysplasia refers to an unequivocally neoplastic alteration in the epithelium. Dysplasia may vary from mild or low-grade to severe or high-grade. In the European literature, Morson's criteria of dysp|asia are used [8]; in the American studies, criteria derived from dysplasia in inflammatory bowel disease are used [9, 10]. High-grade (severe) dysplasia consists of neoplastic cellular and/or architectural changes within the columnar epithelium without invasion through the basement membrane. Low-grade dysplasia consists of less striking changes that are in(ermediate between high-grade dysplasia and ordinary Barrett's metaplasia. In the absence of overt carcinoma, the prevalence of dysplasia is usually 5% to 10% [11, 12]. Several arguments support the view that high-grade dysplasia is to be considered as an unequivocal premalignant lesion. Studies of patients with Barrett's adenocarcinoma have demonstrated dysplasia in approximately 90% of cases [3, 13-15]. Esophagectomy for high-grade dysplasia can reveal previously unrecognized adenocarcinoma in the specimens [11, 16-18]. In endoscopic follow-up studies, detection of a carcinoma was preceded by the detection of increasing severity of dysplasia [5, 19, 20]. Though dysplasia has been found in all epithelial types, in the vast majority it has been seen in the specialized columnar- or intestinal-type epithelium [5]. While the presence of high-grade dysplasia is an ominous sign, the time course for development of invasive carcinoma from high-grade dysplasia varies considerably. High-grade dysplasia can be present for more than 3-4 years without evidence of invasive carcinoma, as shown by endoscopic ultrasonography and/or surgery [5, 17, 20, 21]. Even less information is available concerning the natural history of low-grade dysplasia. It is not known if adenocarcinoma can arise directly from low-grade dysplasia.
G.N.J. Tytgat and W. Hameeteman: Neoplasia and Barrett's Esophagus
309
Table 1. Screening for adenocarcinoma in patients with columnar-lined esophagus.
Study (yr.) (reference)
Screening method
No. of pts.
Mean follow-up (yr.)
No. of carcinomas
Incidence per patient year follow-up
Spechler (1984) (33) Sprung (1984) (34) Cameron (1988) (32) Sampliner (1986) (35) Robertson (1988) (19) v.d. Veen (1989) (36) Hameeteman (1989) (5) Total Endoscopically screened
Endoscopy Endoscopy Post" Endoscopy Endoscopy Post~ Endoscopy
105 41 122 25 56 155 50 554 276
3.3 4.0 8.5 2.1 3.0 4.4 5.2 4.9 3.6
2 2 2 I 4 4 5 20 14
1/175 1/81 1/441 1/92 1/56 1/170 1/52 1/135 1/71
"Post: follow-up by postal inquiries of patients, relatives or general practitioner.
Adenocarcinoma
Many reports have stressed the association between CLE and esophageal adenocarcinoma. Adenocarcinoma in CLE is most often seen in males with an average age of 55-60 years. The male-female ratio is approximately 5.5:i. Dysphagia is a prominent presenting symptom reported in 70% to 95% of patients. Weight loss is present in 42% to 63% of patients. Bleeding from the malignancy has been reported in 11% to 36% of patients. Carcinoma in CLE may be present multifocally, nearly always surrounded by dysplasia in the adjacent columnar epithelium. At the time of diagnosis the tumor is usually far advanced; transmural tumor infiltration is found in 60% to 88% of patients, and lymph node invasion in 55% to 74% of patients [14, 15, 22-24]. The results of esophagectomy for Barrett's carcinoma are comparable to the results obtained for other esophageal carcinomas. The extent of wall penetration and spread of tumor to lymph nodes influence the prognosis. With the majority of patients presenting with an advanced tumor, the prognosis is poor with a 5 year survival of 7% [15], 16% [25], or 22% [23]. With tumor limited to the mucosa or submucosa and without lymph node invasion, prognosis is considerably better and cure can be obtained [15, 17, 26-29]. While the association between CLE and adenocarcinoma has been demonstrated beyond doubt, there is still debate on the magnitude of the risk. The reported prevalence of adenocarcinoma in CLE (the coexistence of the two lesions at a single time) ranges from zero to 46.5% [30, 31]. Interpretation of these figures is difficult as the total population from which these patients were obtained is hard to define. Patients without symptoms will not be included as they obviously will not visit their physician. Cameron and coworkers [32] suggested from comparison of clinical incidence and autopsy figures that only 5% to 10% of patients with CLE might come to medical attention. Patient selection thus is likely to influence prevalence figures considerably. To assess the incidence of carcinoma in CLE (the frequency of the occurrence of cancer developing over time), a long follow-up is necessary of patients who had no detectable neoplasm at the initial examination. Estimates of incidence differ considerably from 1 case per 52 patient years of follow-up to 1 case per 441 patient years of follow-up (Table 1). Such variation is to be expected as none of the series rePorted more than 5 cases of carcinoma, and in 4 studies only 1 or 2 cases were collected. Variation in incidence may also be related to the
method of follow-up. The more thorough the follow-up, the more carcinomas might be revealed [37]. Surveys in which regular endoscopy is used give better and often higher incidence rates. Considering the endoscopic studies only, the average incidence of adenocarcinoma is 1 case in 71 patient years of follow-up. Compared with the risk in the general population, patients with a CLE therefore have a 30-fold to 350-fold increased risk of developing an esophageal carcinoma [5, 19, 31, 33]. Dysplasia-Carcinoma Sequence
In recent prospective endoscopic follow-up studies [5, 19, 20], the detection of carcinoma was preceded by increasing severity of dysplasia. In Robertson's [19] endoscopic surveillance study of 56 patients with CLE, 4 patients developed high-grade dysplasia and 3 patients developed carcinoma. Two of the adenocarcinomas were preceded by progressively more severe dysplastic changes. In our own Amsterdam study [5] in which patients with a CLE were seen for endoscopy and histologic investigation at regular intervals, 50 patients were followed for a mean period of 5.2 years, At study entrance 7 patients were found to have dysplasia, 6 scored as low-grade, 1 as high-grade. At the end of the study period dysplasia was found in 13 patients, 10 low-grade and 3 high-grade. Moreover, adenocarcinoma had developed in 5 patients. In the years preceding the detection of the carcinoma, biopsy specimens had shown increasing degrees of severity of dysplasia (Fig. 1). Similar observations were recently presented by Schnell and coworkers [20], emphasizing that patients with high-grade dysplasia may progress to adenocarcinoma. In general there is wide variation in the time course of the development of dysplasia and carcinoma. With CLE usually being a consequence of longlasting gastro-esophageal reflux, one might question whether the development of dysplasia is also a consequence of persisting reflux. There are no clear answers to this question. Development of carcinoma in CLE has been described after anti-reflux surgery, but there is no information on the efficacy of the various procedures in the abolishment of reflux. It has been stated that an effective anti-reflux operation might prevent an increase in severity of dysplasia [29]. True regression of CLE is a rare event [38, 39]. Recently regression was described in a few patients after extensive anti-reflux surgery [40] and after long term medical acid suppressive treatment with 60 mg of omeprazole [41]. In
310
World J. Surg. Vol. 16, No. 2, Mar./Apr. 1992
DYSPLASIA ,
.3
f¢
/
.2
/
/
/
/
I:
/'
•
H
./" i/
f"
/ // i
L /"
\
[
x .,"
/
/
"%..
!/
N 0
z
4
/ "~ . .
6
,.~°"
i¢
i
1
8
t0
YEARS
earlier surgical studies in which mainly a Nissen fundoplication was performed, regression was seen only rarely and the methods by which this regression had been established have frequently been criticized.
Fig. 1. Development and progression of dysplasia in 5 patients with columnar-lined esophagus who developed carcinoma. N: negative for dysplasia; L: low-grade dysplasia; H: high-grade dysplasia; C: carcinoma.
DETECTION OF DYSPLASIA
/
Have diagnosis conflrmed.~by another pathologist Intensive medical therapy 8 - 12 weeks
Practical Recommendations
CLE is a premalignant condition with an increased risk of esophageal carcinoma. Development of adenocarcinoma is high.~rade low/moderate no~dysplasla / dysRlasla ~ likely to be preceded by an increase in severity of the dysplastic changes. Dysplasia develops predominantly in intestinal-type En~osono Surveillance Surveillance nography eptihelium. It is essential to take multiple biopsy specimens Repeat endoscopy every year_ . every 2years from the entire columnar segment to minimize the risk of every 3 - 6 months till age 70(?) till age 70(?) or missing intestinal-type epithelium and/or dysplasia. Two to 4 biopsy specimens should be taken, in a circular fashion, from esophageal resection every 2-3 cm of CLE, with additional biopsies from any mucosal abnormalities. High-grade dysplasia may be present in Fig. 2. A flow diagram for a screening program for patients with columnar-lined esophagus. patients with CLE for as much as 4 years without signs of progression to carcinoma, as seen by endoscopy with biopsies and endoscopic ultrasonography. The risk of cancer for CLE R6sum~ seems to warrant an effective surveillance program [42, 43]. A Le reflux du contenu gastrique est le facteur le plus important flow diagram for such a screening program is shown in Figure 2. qui contribue au d6veloppement de l'endobrachyoesophage. The efficacy of a surveillance program in reducing the morbidity L'endobrachyoesophage pourrait 6tre consid6r6 comme une and mortality of esophageal cancer remains to be demonstrated "mosa~que" de cellules, de glandes et types d'architectures, [44]. Obviously surveillance is only indicated for those patients pr6sentant des degr6s variables de maturation vers l'6pith61ium in whom esophagectomy is considered a therapeutic option if intestinal et gastrique. La dysplasie consiste en une alt6ration early malignancy is discovered. Recently investigators at the n6oplasique indiscutable de l'6pithElium. Une dysplasie s6v6re Cleveland Clinic have estimated the direct costs of annual (de grade 61ev6) comporte des celiules n6oplasique et/ou des endoscopic surveillance for CLE at $62,000 for each carcinoma modifications architecturales de l'6pith61ium cylindrique sans detected [45]. A program of surveillance every other year would franchissement de la membrane basale. La pr6valence de la cost substantially less. dysplasie est habituellement de l'ordre de 5 ~ 10%. La dysplasie Dysplasia is the only clinically useful parameter important in est, dans la majorit6 des cas, mise en 6vidence dans l'6pith61ium follow-up studies. Despite intensive investigation, all other cylindrique ou intestinal. L'ad6nocarcinome au sein de l'endotechniques proposed to identify epithelia with high malignant brachyoesophage est retrouv6 principalement chez l'homme. potential are of doubtful clinical usefulness at present. Recently Au moment du diagnostic, une infiltration tumorale transpariresearchers in the United States of America have developed a 6tale est retrouv6e dans 60 ~i 88% des cas, et un envahissement technique for culturing endoscopic Barrett biopsy specimens. ganglionnaire dans 55 h 74% des cas. Le pronostic est mauvais Cytogenetic analysis of the cultured tissue revealed chromo- avec un taux faible de survie h 5 ans. L'estimation de l'incisome 7 abnormalities associated with enhanced expression of dence de cancer est variable, allant de 1 cas sur 52 gt I cas sur epidermal growth factor receptors [46]. 441 patients par ann6e de suivi.
G.N.J. Tytgat and W, Hameeteman: Neoplasia and Barrett's Esophagus
311
Resumen E1 reflujo del contenido g~tstrico es el factor m~is importante en el desarrollo del es6fago de Barrett. E1 es6fago de Barrett debe ser considerado como un " m o s a i c o " de crlulas, gl~indulas y tipos arquitect6nicos, que exhiben grados variables de maduraci6n hacia el epitelio intestinal y g~istrico. El termino displasia se refiere a una inequivoca alteracirn neopl~sica del epitelio. El alto grado (severo) de displasia consistente en cambios celulares neopl~isicos y/o arquitectrnicos del epitelio columnar, sin invasi6n de la membrana basal. L a incidencia de la displasia es Usualmente del orden de 5-10%. La displasia ha sido hallada principalmente en el epitetio del tipo columnar especializado o intestinal. El adenocarcinoma en el es6fago de Barrett se observa principalmente en hombres. En el momento del diagn6stico la infiltraci6n transmural del tumor se encuentra con una incidencia de 60-88%, y la invasi6n de n6dulos linf~iticos con incidencia de 55-74% de los casos. El pron6stico es pobre, on baja posibilidad de sobrevida a 5 afios. La incidencia estimada de c~ncer varia desde 1 caso por 52 a 1 caso por 441 pacientes afio de seguimiento.
14.
15. 16. 17.
18. 19. 20.
21. References I. Endo, M., Kobayashi, S., Kozu, T., Takemoto, T., Nakayama, K,: A case of Barrett epithelization followed up for five years. Endoscopy 18(suppl 3): 11, 1974 2. Savary, M., Monnier, P.: Diagnosis, pathophysiology and adenocarcinogenesis of Barrett's esophagus. In Esophageal Disorders: Pathophysiology and Therapy, T.R. DeMeester, D.B. Skinner, editors, New York, Raven, pp. 101-8, 1985 3. Thompson, J.J., Zinsser, K.R., Enterline, H.T.: Barrett's metaplasia and adenocarcinoma of the esophagus and gastroesophageal junction. Hum. Pathol. 14:43, 1983 4. Levine, D.S., Rubin, C.E., Reid, B.J., Haggitt, R.C.: Specialized metaplastic columnar epithelium in Barrett's esophagus: A comparative transmission electron microscopy study, Lab. Invest. 60:418, 1989 5. Hameeteman, W., Tytgat, G.N.J,, Houthoff, H.J., van den Tweel, J.G.: Barrett's esophagus, development of dysplasia and adenocarcinoma. Gastroenterology 96:1249, 1989 6. Hameeteman, W.: Columnar-lined (Barrett's) esophagus. Thesis, University of Amsterdam 1989; Drukkery All Zn B.V., Katwijk, The Netherlands 7. Wesdorp, I.C.E., Bartelsman, J., Schipper, M.E.I., Offerhaus, J., Tytgat, G.J.: Malignancy and premalignancy in Barrett's esophagus: A clinical, endoscopical, and histological study of 100 patients. Gastroenterology 11:317, 198t 8. Morson, B.C., Sobin, L.H., Grundmann, E., Johansen, A., Nagayo, T., Serck-Hansen, A.: Precancerous conditions and epithelial dysplasia in the stomach. J. Clin. Pathol. 33:711, 1980 9. Riddell, R.H.: Dysplasia and regression in Barrett's epithelium. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 143-52, 1985 10. Reid, R.J., Haggitt, R.C., Rubin, C.E., Rabinovitch, P.S.: Barrett's esophagus: Correlation between flow cytometry and histology in detection of patients at risk for adenocarcinoma. Gastroenterology 93:I, 1987 tl. Schmidt, H.G., Riddell, R.H., Walther, B., Skinner, D.B., Riemann, J.F: Dysplasia in Barret.t's esophagus. J. Cancer Res. Clin.. Oncol. 110:145, 1989 12. Kerlin, P., D'Mellow, G., van Deth, A.: Barrett's esophagus: Clinical, endoscopic, and histologic spectrum in fifty patients. Aust. N.Z.J. Med. 16:198, 1986 13. Hamilton, S.R., Smith, R.R.L., Cameron, J.L.: Prevalence and characteristics of Barrett's esophagus in patients with adenocarci-
22. 23. 24. 25. 26. 27. 28. 29.
30.
31. 32.
33. 34. 35. 36.
noma of the esophagus or esophagogastric junction. Hum. Pathol. 19:942, 1988 Haggitt, R.C., Dean, P.J.: Adenocarcinoma in Barrett's epithelium. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 153-66, 1985 Monnier, P.H., Frbntolliet, C., Savary, M., Ollyo, J.B.: Barrett's oesophagus or columnar epithelium of the lower oesophagus. Bailliere's Clin. Gastroenterol. •:769, 1987 Skinner, D.B., Walther, B.C., RiddeU, R.H., Schmidt, H., lascone, C., DeMeester, T,R.: Barrett's esophagus: Comparison of benign and malignant cases. Ann. Surg. 198:554, 1983 Reid, R.J., Weistein, W.M., Lewin, K.J.: Endoscopic biopsy detect early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic lesions. Gastroenterology 94:81, 1988 Palley,S.L., Sampliner, R., Garewal, H.S.: Editorial: Management of high-grade dysplasia in Barrett's esophagus. J. Clin. Gastroenterol. 11:369, 1989 Robertson, C.S., Mayberry, J.F., Nicholson, D.A., James, P.D., Atkinson, M.: Value of endoscopic surveillance in the detection of neoplastic change in Barrett's oesophagus. Br. J. Surg. 75:760, 1988 Schnell, T., Sontag, S., Chejfec, G., Chintam, R., O'Connell, S., Kurucar, C.: High grade dysplasia in Barrett's esophagus: Experience with 36 patients (abstract). American College of Gastroenterology, October 23, 1989 Hamilton, S.R., Smith, R.R.L.: The relationship between columnar epithelial dysptasia and invasive adenocarcinoma arising in the Barrett's esophagus. Am. J. Clin. Pathol. 87:301, 1987 Haggitt, R.C., Tryzelaar, J., Ellis, F.H., Colcher, H.: Adenocarcinoma complicating columnar epithelium-lined (Barrett's) esophagus. Am. J. Clin. Pathol. 70:1, 1978 Skinner, D.B.: En bloc resection for neoplasms of the esophagus and cardia. J. Thorac. Cardiovasc. Surg. 85:59, 1983 Steiger, Z., Wilson, R.E., Leichman, L., Busuito, M.J., Rosenberg, J.C.: Primary adenocarcinoma of the esophagus. J. Surg. Oncol. 36:68, 1987 Sanfey, H., Hamilton, S.R., Smith, R.R.L.: Carcinoma arising in Barrett's esophagus. Surg. Gynecol. Obstet. 161:570, 1985 Witt, T.R., Bains, M.S., Zaman, M.B., Martini, N.: Adenocarcinoma in Barrett's esophagus. J. Thorac. Cardiovasc. Surg. 85:337, 1983 Smith, R.R.L., Hamilton, S.R., Boinott, J.K., Rogers, E.L.: The spectrum of carcinoma arising in Barrett's esophagus. Am. J. Surg. Pathol. 8:563, 1984 Rosenberg, J.C., Budev, H., Edwards, R.C.: Analysis of adenocarcinoma in Barrett's esophagus utilizing a staging system. Cancer 55:1353, 1985 Skinner, D.B., Walther, B.C., Little, A.G.: Surgical treatment of Barrett's esophagus. In Barrett's Esophagus: Pathophysiology, Diagnosis, and Management, S.J. Spechler, R.K. Goyal, editors, New York, Elsevier, pp. 211-21, t985 Bremner, C.G., Hamilton, D.G.: Barrett's esophagus: Controversial aspects. In Esophageal Disorders: Pathophysiology and Therapy, T.R. DeMeester, D.B. Skinner, editors, New York, Raven, pp. 233-9, 1985 Cameron, A.J., Ott, B.J., Payne, W.S.: The incidence of adenocarcinoma in columnar-lined (Barrett's) esophagus. N. Engl. J. Med. 313:857, 1985 Cameron, A.J., Zinsmeister, A.R., Ballard, D.J., Carney, J.A.: A population-based study of Barrett's esophagus: Comparison of clinically diagnosed prevalence versus autopsy determined prevalence. Gastroenterology 94:A57, 1988 Spechler, S.J., Schimmel, E.M., Dalton, J.W., Doos, W., Tries, J.S.: Barrett's epithelium complicating lye ingestion with sparing of the distal esophagus. Gastroenterology 81:580, t981 Sprung, D.J., Ellis, F.H., Gibb, S.P.: Incidence of adenocarcinoma in Barrett's esophagus. Am. J. Gastroenterol. 79:817A, 1984 Sampliner, R.E.: Editorial. Does Barrett's esophagus need to be detected? Arch. Intern. Med. 96:438A, 1986 van der Veen, A.H., Dees, J., Blankensteijn, J.D,, van Blankenstein, M.: Adenocarcinoma in Barrett's oesophagus: An overrated risk. Gut 30:14, 1989
312 37. Atkinson, M.: Barrett's oesophagus: To screen or not to screen? Gut 30:2, 1989 38. Wesdorp, I.C.E., Bartetsman, J., Schipper, M.E.I., Tytgat, G.N.: Effect of long-term treatment with cimetidine and antiacids in Barrett's oesophagus. Gut 22:724, 1981 39. Sampliner, R.E., Garewal, H.S., Fennerty, M.B., Aickin, M.: Lack of impact of therapy on extent of Barrett's esophagus in 67 patients. Dig. Dis. Sci. 35:93, 1990 40. Perniceni, T., Leymarios, J., Molas, G., Fekete, F.: L'endobrachyoesophage regresse-t-il apres diversion duodenale totale? Gastroenterol. Clln, Biol. •2:709, 1988 41. DeViere, J., Buset, M., Dumonceau, J.M., Rickaert, F., Cremer, M.: Regression of Barrett's epithelium with omeprazole. N. Engl. J. Med. 320:1497, 1989
World J. Surg. Voi. 16, No. 2, Mar./Apr. 1992 42. Spechler, S.J.: Barrett's esophagus: What's new and what to do. Am. J. Gastroenterol. 84:220, 1989 43. Dent, J.: Approaches to oesophageal columnar metaplasia (Barrett's oesophagus), Scand. J. Gastroenterol. 24(suppl 168):60, 1989 44. Spechler, S.J.: Endoscopic surveillance for patients with Barretts' esophagus: Does the cancer risk justify the practice? Ann. Intern. Med. 106:902, 1987 45. Achkar, E., Carey, W.C.: The cost of surveillance for adenocarcinoma complicating Barrett's esophagus. Am. J. Gastroenterol. 83:291, 1988 46. Garewal, H., Leibovitz, A., Sampliner, R., Prabhata, R., Trent, J., Korc, M., Sloan, D.: Tissue culture and characterization of epithelial cells devised from Barrett's esophagus, a premalignant lesion. Gastroenterology 94:A143, 1988
