| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1996;62:1460-1465
© 1996 The Society of Thoracic Surgeons
Section of Thoracic Surgery, Department of Surgery, University of Michigan, Ann Arbor, Michigan
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentReferences |
|---|
Methods. Expression of SI and p53 were examined in 40 BAs and 25 cardia adenocarcinomas using immunohistochemical techniques.
Results. Sucrose isomaltase analysis revealed positive staining in 55% (22/40) of the BAs and 44% (11/25) of the cardia adenocarcinomas. Of 14 cardia adenocarcinomas that were SI negative, 100% (14/14) had no associated Barrett's mucosa. However, in 21 cardia adenocarcinomas with no associated Barrett's mucosa, 7/21 (33%) were SI positive. This suggests that SI-positive tumors may represent BA without the standard definition of Barrett's esophagus being met. P53 was present in 65% of BAs and 64% of cardia adenocarcinomas, demonstrating the importance and similarity of this gene alteration in both tumor types. Staining was positive for SI or p53 in 77% (50/65) of all tumors. Tumors of lower stage expressed SI more often than higher stage tumors.
Conclusions. These data suggest that a subset of cardia adenocarcinomas represent BAs. Surveillance endoscopy incorporating additional esophagogastric junction biopsies and assessment of SI or p53 may improve detection of intestinalized Barrett's mucosa and early dysplastic changes.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
The prevalence of adenocarcinoma of the esophagus is increasing at one of the most rapid rates of all visceral cancers in the United States [1]. This alarming trend represents a 100% increase in distal-third esophageal adenocarcinomas in the last two decades, with most patients receiving only limited palliation with operation or multimodality therapy [2]. Historically, adenocarcinoma of the esophagus was observed only in a limited number of patients with esophageal cancer. However, esophageal adenocarcinoma now accounts for up to 70% of esophageal cancers in the United States, surpassing esophageal squamous cell carcinoma [3]. Unlike squamous cell carcinoma, adenocarcinoma does not appear to be as related to the obvious risk factors of smoking and alcohol abuse. The development of adenocarcinoma of the esophagus is closely associated with chronic gastroesophageal reflux and the eventual replacement of the normal esophageal squamous cell epithelium by columnar-lined epithelium known as Barrett's metaplasia [4, 5]. Barrett's metaplasia is categorized into three histologic types: the gastric-fundic type, in which there are clusters of glands with parietal and chief cells similar to those found in the gastric body mucosa; the junctional type, which contains deep clusters of mucosa glands identical to those found at the gastric cardia; and the specialized or intestinal type, which is characterized by villi and tubules lined by intestinal-type goblet cells and other columnar epithelial cells. It is the specialized Barrett's metaplasia that has been most closely associated with adenocarcinoma of the esophagus [6]. By convention, Barrett's esophagus is diagnosed by the demonstration of biopsy-proven columnar epithelium at least 3 cm above the anatomic gastroesophageal junction. Routine surveillance endoscopy and biopsy have been recommended for patients with Barrett's esophagus because of the increased risk of adenocarcinoma [7]. Currently, esophagectomy is recommended for patients with Barrett's epithelium containing high-grade dysplasia, because of the presence of carcinoma in situ in as many as 50% of resected specimens [8].
Unlike cancer of the stomach, which has dramatically decreased in incidence, adenocarcinoma of the gastric cardia (CA) and Barrett's adenocarcinoma (BA) of the distal esophagus are occurring with increasing frequency [3, 9]. This suggests that the etiologic factors resulting in the development of adenocarcinoma in both of these areas may be similar. Consistent with this, recent reports have demonstrated the presence of Barrett's metaplasia in greater than 80% of patients with adenocarcinoma of the esophagus [10]. These observations have prompted investigators to suggest that use of the "3-cm rule" for defining Barrett's metaplasia may not identify all individuals with this condition. As a result, the finding of intestinal-type mucosa anywhere in the distal esophagus has been suggested to be indicative of Barrett's metaplasia [5, 11, 12].
The development of distal esophageal adenocarcinoma from intestinal-type metaplastic Barrett's epithelium is supported by the expression of intestinal specific enzymes such as sucrase isomaltase (SI), not only in Barrett's esophagus, but also in associated adenocarcinoma [13, 14]. Sucrase isomaltase is an intestinal disaccharidase that is abundantly expressed in more than 80% of patients with specialized Barrett's epithelium [13]. However, this protein is not expressed in normal esophageal mucosa, squamous cell carcinomas, or normal gastric mucosa. Thus the presence of this enzyme in intestinal-type Barrett's metaplasia, dysplasia, and the adenocarcinomas developing from this mucosa suggests a common cell of origin, although expression is often decreased in more advanced tumors [13].
Mutation of the p53 tumor suppressor gene clearly plays an important role in the development of esophageal adenocarcinoma. Mutations in p53 are observed in dysplastic Barrett's epithelium and in 60% to 70% of esophageal adenocarcinomas [1414–17]. The protein product of the mutated p53 gene is more stable than the wild-type p53 protein, causing it to accumulate in the nuclei of affected cells and allowing detection. In the absence of the wild-type protein, cells no longer undergo a G1-S phase DNA checkpoint arrest and thus may be predisposed to further DNA alterations, suggesting that p53 gene alterations are significant early events in the development of esophageal adenocarcinoma [18].
Previous studies have supported the conclusion that at least a subset of adenocarcinomas of the gastric cardia may actually represent BAs [6, 11]. The purpose of the present study was to examine the potential of SI expression and p53 nuclear accumulation to identify CAs that may have origin in Barrett's epithelium, perhaps explaining why BA and CA are clinically and histologically such similar tumors [10]. Twenty-five adenocarcinomas arising in the gastric cardia without preoperatively documented Barrett's epithelium and 40 adenocarcinomas arising in Barrett's epithelium were retrospectively identified and examined for these markers.
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
The barium swallow, computed tomography of the chest and abdomen, esophagogastroduodenoscopy, and original biopsy report confirming cancer were reviewed. There were 64 patients who underwent transhiatal esophagectomy and 1 who underwent transthoracic esophagectomy, all with cervical esophagogastric anastomoses. Biopsy specimens had been harvested in the operating room from the resected specimen and included samples from the proximal stomach, the tumor, and the normal distal esophagus immediately adjacent to the tumor. No patient in this study had received preoperative chemotherapy or radiation therapy.
Statistics
Results are reported as the mean ± standard error of the mean. Survival was calculated using Kaplan-Meier analysis for actuarial survival with confidence intervals at 95%. Comparisons between groups was performed with 
2 analysis. Follow-up was based on the most recent office visit or telephone contact for patients without recent follow-up data.
Immunohistochemical Analysis
All specimens were processed immediately after resection. Tissue was frozen in OCT compound with isopentane cooled to the temperature of liquid nitrogen. The following procedures were standardized to maintain uniform conditions for assessment of SI and nuclear p53 immunostaining as described previously [13, 14]. Five-micrometer cryostat tissue sections were mounted on poly-L-lysine–treated glass slides. For SI staining, the slides were fixed in acetone at 4°C for 10 minutes. For p53 staining, a 1:1 mixture of acetone and 100% methanol was used at the same time and temperature. Endogenous peroxidase activity was blocked by incubation of slides in 0.3% hydrogen peroxide in phosphate-buffered saline solution (PBS) for two changes of 15 minutes each. The slides were then rinsed twice in PBS for 5 minutes each followed by incubation with 0.1% bovine serum albumin in PBS for 10 minutes. The Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA) was used. Nonspecific binding was blocked using horse serum at the recommended dilution in PBS for 45 minutes. For SI staining, a monoclonal-antihuman SI primary antibody (a gift from Drs Ward Olsen and Mark Lloyd, University of Wisconsin, Madison, WI) was added at a 1:100 dilution in bovine serum albumin/PBS and incubated for 2 hours at room temperature. The primary antibody used for p53 protein detection was PAb 1801 (Oncogene Science Inc, Cambridge, MA) at a 1:100 dilution. This antibody recognizes residues 46–55 of human p53 and will react with both wild type and mutant p53. The tissues were washed twice for 5 minutes in PBS followed by biotinylated goat anti-mouse secondary antibody incubation at the recommended dilution for 45 minutes. The sections were again washed twice for 5 minutes each in PBS followed by incubation with avidin-biotin complex for 30 minutes. After two washes for 5 minutes each in PBS, the chromogen diaminobenzidine was applied for 5 minutes. The tissues being tested for SI were rinsed for 5 minutes in distilled water, lightly counterstained with hematoxylin (Sigma Chemical Co, St. Louis, MO), dehydrated, placed in toluene, and mounted with coverslips in Permount medium. Sections being tested for p53 were lightly counterstained with fast green (Sigma Chemical Co). All antibody incubation steps were performed in humidified chambers at room temperature. The positive control was a section of tissue previously determined to be positive for either SI or p53. The negative control was a section of each tissue that did not receive the primary antibody processed concurrently.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
All tumors were staged according to the TNM classification (Table 1
). In the BA group 31/40 (78%) were either stage II or III, and 19/25 (76%) of the CAs were stage II or III. Tumors were defined as stage IV based on celiac lymph node metastases in 8, liver metastases in 2, and metastasis to cervical lymph nodes in 1. Mean survival was 25 ± 3 months (range, 2 to 50 months) for all patients. All patients with stage I tumors are alive at 5 to 42 months after operation, and survival for stage II, III, and IV disease was 38 ± 4, 16 ± 3, and 12 ± 3 months, respectively. Actuarial survival curves based on tumor stage were calculated using Kaplan-Meier analysis for survival. Actuarial survival at 48 months was 100% for stage I, 59% for stage II, and 10% for stage III tumors. No patient with stage IV disease survived longer than 25 months (Fig 1). Mean survival was 22 ± 3 months for adenocarcinoma associated with Barrett's epithelium and 27 ± 4 months for cardia tumors. The overall actuarial survival curves for both groups are shown in Figure 2. There was no significant difference in survival between these groups or stages when appropriate statistical analysis could be carried out.
|
|
|
). The tumors that were of lower stage showed more extensive SI expression than higher stage tumors. Tumors were considered positive for SI even if expression was observed in a small percentage of the tumor cells. Of the CAs that were negative for SI, all (14/14) were histologically negative for Barrett's epithelium as well. There were 7/21 CAs (33%), however, that had no previously detectable Barrett's epithelium but stained positively for SI. Four of 25 patients with CA were found on histologic examination to have Barrett's epithelium at the level of the esophagogastric junction, but did not fulfill endoscopic criteria to be classified as Barrett's esophagus; all were positive for SI.
|
). There was no correlation between the intensity of p53 expression and the stage of disease for either tumor type, nor was there a significant survival advantage.
The expression of tumor markers fell into four separate categories: SI negative/p53 negative, SI negative/p53 positive, SI positive/p53 negative, and SI positive/p53 positive for the BA and CA groups (Table 2). Of 24 stage I and II tumors, there was SI expression in 15 stage III tumors (63%), as compared with 16/41 stage III tumors (39%) (p < 0.005). This difference in expression associated with stage was not appreciated with p53, in which 17/24 stage I and II (70%) tumors expressed p53 compared with 25/41 stage III and IV tumors (61%).
|
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
The current definition of what constitutes Barrett's mucosa clearly excludes some patients who have this condition and are at significant risk of the development of esophageal adenocarcinoma. Presently, the diagnosis of Barrett's esophagus is dependent on the endoscopic biopsy documentation of columnar epithelium at a minimum of 3 cm above the anatomic gastroesophageal junction [24, 25]. Patients so diagnosed then undergo more rigorous endoscopic surveillance. In our study, none of the patients with CA had Barrett's mucosa identified preoperatively, yet many of the CA group had similar genetic expression of SI (44%) and p53 (64%), markers also present in BA.
Two clinically significant differences in patients with BA and CA were identified: symptomatic reflux and dysphagia. Reflux symptoms were present in a significantly higher proportion of patients with previously documented Barrett's mucosa than in those with tumors of the gastric cardia. Symptomatic reflux is a relatively subjective finding, which may have selected patients to seek medical evaluation earlier. There was also a higher incidence of dysphagia in those patients with CA; these tumors tended to be of higher stage at the time of presentation. Although these clinical differences were significant, the long-term survival was no different for those patients with BA (22 ± 3 months) compared with those with CA (27 ± 4 months).
Sucrase isomaltase and other intestinal enzymes have previously been described as sensitive markers for Barrett's intestinalized epithelium [13, 14]. Sucrase isomaltase is expressed in the majority of adenocarcinomas associated with intestinal Barrett's epithelium; however, it is not present in esophageal squamous cell carcinoma. It has also been described in intestinal-type gastric adenocarcinoma [26] but is absent in normal gastric mucosa [13]. In the present study, all CAs that stained negatively for SI were also histologically negative for Barrett's mucosa. There were four tumors at the gastric cardia with small foci of Barrett's epithelium that was not appreciated preoperatively. These four tumors all stained positively for SI. Of the 21 tumors that had no pathologically reported evidence of Barrett's mucosa in the resected specimens, 7 of 21 (33%) were found to stain positively for SI. Our findings suggest that these tumors may have arisen from specialized Barrett's epithelium, and that during the growth of the tumor, "cannibalization" of any associated Barrett's mucosa may have occurred. It has been shown previously that Barrett's specialized intestinal epithelium may be found preoperatively in more than 80% of patients with esophageal adenocarcinoma, and that high-grade dysplasia is found in the majority of these [6, 10]. Clark and associates [6] have shown that the association of Barrett's specialized epithelium with cardia tumors is equally as high, although the frequency of high-grade dysplasia is somewhat lower. Our study demonstrates that some adenocarcinomas arising within 3 cm of the esophagogastric junction have genetic characteristics similar to traditional BAs, specifically the expression of SI in 44% and p53 in 64%. Expression of the intestinal enzyme SI is present in those patients with both metaplastic intestinal-type Barrett's epithelium and dysplastic changes of the esophageal epithelium [13]. During the events accompanying dysplasia in intestinalized Barrett's mucosa, this epithelium appears to become less differentiated, and the level of SI expression is often decreased [13]. This may account for the decreased level of SI expression in many of the higher stage tumors evaluated in our study. Sucrase isomaltase tended to be expressed in a greater percentage of tumors of lower stage: 63% of stage I and II tumors compared with 39% of stage III and IV tumors.
Nuclear accumulation of the p53 protein was almost equal in both BA and adenocarcinoma arising at the gastric cardia: 65% and 64%, respectively. This confirms the importance of this genetic alteration in tumors arising in Barrett's epithelium as well as cardia tumors. P53 mutations are frequently found in Barrett's mucosa with dysplasia and increase in frequency with the degree of dysplasia [16]. Alterations of this tumor suppressor gene may be an early marker for those patients at risk for the development of adenocarcinoma of the distal esophagus. A long-term evaluation has not yet been performed to determine what percentage of patients with p53 gene alterations in Barrett's mucosa without dysplasia will go on to have development of Barrett's adenocarcinoma or dysplasia.
More than 70% of gastric cardia adenocarcinomas express SI, p53, or both; therefore, these markers may be of use in the early detection of these neoplasms. Our results support previous suggestions that many CAs may actually represent a subset of BA [5, 6, 27]. This subset may not have been detected preoperatively because these patients failed to fulfill current endoscopic criteria for defining Barrett's esophagus. The presence of p53 nuclear protein accumulation and expression of SI in a patient with Barrett's metaplasia may be of value in predicting patients at risk for the development of esophageal adenocarcinoma. These markers may implicate the presence of higher risk specialized epithelium, even if the endoscopic criterion of a 3-cm length of Barrett's mucosa is not evident.
Our findings suggest that it may no longer be appropriate to diagnose Barrett's metaplasia of the distal esophagus strictly on the basis of an endoscopic measurement of the length of the abnormal mucosa. The risk of adenocarcinoma of either the gastric cardia or the distal esophagus may be equivalent in those patients with specialized columnar epithelium at any level. Endoscopic surveillance of Barrett's mucosa has allowed early identification of patients with low-grade malignancies and has improved survival with distal esophageal cancer. In patients without obvious Barrett's epithelium, immunohistochemical biopsy of the cardia and staining for p53 and SI may augment current surveillance techniques and have a positive impact on survival.
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Address reprint requests to Dr Iannettoni, Section of Thoracic Surgery, University of Michigan, 1500 E Medical Center Dr, 2120 Taubman Center, Box 0344, Ann Arbor, MI 48109-0344.
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  M. K. Ferguson and K. S. Naunheim RESECTION FOR BARRETT'S MUCOSA WITH HIGH-GRADE DYSPLASIA: IMPLICATIONS FOR PROPHYLACTIC PHOTODYNAMIC THERAPY J. Thorac. Cardiovasc. Surg., November 1, 1997; 114(5): 824 - 829. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
