Described 50 years ago by Norman Barrett, Barrett’s esophagus (BE) is currently defined as an endoscopically visible columnar mucosa in the distal esophagus, of any extension, proved to harbor intestinal metaplasia on biopsy, highlighted by the presence of goblet cells [
The incidence of EAC is rising in the western world. Its prevalence exceeds that of squamous cell carcinoma and EAC is the most common type of esophageal cancer in some populations [
Following evidence of dysplasia many patients will undergo excessive evaluations. However, some are only diagnosed with cancer at a late stage, in which there is already lymphatic spread; this results in poor outcomes [
Numerous studies have suggested that there is elevated Ki-67 expression in the metaplasia-dysplasia-adenocarcinoma sequence in BE [
The study population consisted of patients between the ages of 16 and 90 who were diagnosed with BE and EAC between August 2002 and December 2005. They were diagnosed and treated by the Surgery of the Esophagus, Stomach and Small Intestine Group at the Hospital de Clínicas de Porto Alegre (HCPA), Brazil.
We first reviewed anatomopathological records from the HCPA Pathology Service. The inclusion criteria were the following: (1) patients with dyspeptic symptoms and normal gastric mucosa on biopsy; (2) patients with esophageal columnar mucosa on endoscopy and intestinal-type metaplasia with goblet cells on biopsy; (3) patients with a diagnosis of EAC and esophagogastric junction tumor types I and II [
Taking all the clinical and histopathological data into account, the patients were divided into three groups: group 1 (controls), group 2 (BE), and group 3 (cancer). Sample size was calculated and it was found that at least 34 patients would be necessary: 12 in groups 1 and 2, respectively, and 10 for group 3. For nonparametric distribution, a sample size that is approximately 10% greater would be necessary (i.e., 38 patients).
The anatomopathological study was carried out separately by two experienced pathologists. Intestinal metaplasia was defined by the presence of goblet cells in the glandular mucosa. Dysplasia was defined as the presence of variation in nuclear size and shape, nuclear or nucleolar enlargement, increased nuclear to cytoplasmic ratio, hyperchromatism, and abnormal mitosis. Dysplasia was classified into negative, undefined, LGD, and HGD, as previously described [
Immunohistochemistry was performed at the Research Center of HCPA. Paraffin-embedded tissue sections fixed in formalin were used. Epitope retrieval was heat-induced in citrate buffer. Monoclonal antibody MIB-5 (DakoCytomation, Denmark) against the Ki-67 antigen was diluted 1 : 50. The avidin-biotin immunoperoxidase method was employed for Ki-67 staining, as described previously [
A Ki-67 index was determined for each patient, that is, the percentage of stained cells as a fraction of the total cells (at least 500) in an esophageal or gastric crypt (Figure
Ki-67 index: esophageal crypt scheme. Ki-67 index
Example of Ki-67 immunohistochemical staining of esophageal tissue in a patient with esophageal adenocarcinoma, in 400x field (stained cells are marked with an arrow).
The Ki-67 index had parametric distribution and the data are presented as the mean ± standard deviation. Comparisons between continuous variables of the three groups were assessed using analysis of variance (ANOVA). The Tukey test was used to localize differences, when they were present. The linear correlation between variables was analyzed with the Pearson correlation coefficient. Comparisons between categorical variables were made using the Chi-square test. Statistical significance was assumed at
This study was evaluated and approved by the Group of Research and Post-Graduation and Bioethics Committee of the HCPA, following all recommended ethical norms. The paraffin-embedded tissue specimens were obtained from the HCPA Pathology Service’s archives. Patients did not participate directly in the study, and their treatment protocols were not modified by the research. The clinical data, collected from the medical records, was used confidentially and anonymously.
Initially 80 patients were selected, of which 23 were excluded: 6 cases of CIM, 5 subcardial adenocarcinomas (Type III), and 12 which provided insufficient material. Of the remaining 57 patients, 19 had esophageal or esophagogastric adenocarcinoma, 21 had BE, and 17 were controls. The demographic data are presented in Table
Demographic data from patients whose tissue was used to investigate the relationship between expression of the antigen Ki-67 and stages within the metaplasia-adenocarcinoma sequence.
Group 1 (control) | Group 2 (BE) | Group 3 (cancer) | Total |
|
|
---|---|---|---|---|---|
|
|
|
|
||
Age (mean ± SD) |
|
|
|
|
|
Gender (%) | |||||
Men | 8 (47) | 9 (42.9) | 15 (78.9) | 32 (56.1) |
|
Women | 9 (52.9) | 12 (57.1) | 4 (21.1) | 25 (43.9) | |
Race | |||||
Caucasian | 16 (94.1) | 18 (85.7) | 19 (100) | 53 (93) |
|
Black | 1 (5.9) | 3 (14.3) | 0 (0) | 4 (7) |
*Patients with cancer were more likely to be men.
BE: Barrett’s esophagus.
The average overall Ki-67 index was
Example of immunohistochemical staining for Ki-67 antigen in Barrett’s esophagus under 200x microscopic magnification (stained cells are marked with an arrow).
Example of immunohistochemical staining for Ki-67 antigen in esophageal adenocarcinoma under 200x microscopic magnification (stained cells are marked with an arrow).
Ki-67 index variation between the three different groups (control, Barrett’s esophagus, and adenocarcinoma). There is increased expression of Ki-67 along the Barrett-adenocarcinoma sequence.
The correlation between Ki-67 antigen and the Barrett’s esophagus to adenocarcinoma sequence. Pearson coefficient = 0.6 (
The columnar epithelium extension in patients with BE was
Patients with EAC were classified according to stage. Stage 1 occurred in 5.9% of patients while stages 2, 3, and 4 corresponded to 31.6% of cases, respectively. No statistical difference in the Ki-67 index was observed between these stages. Eleven patients were resected, with a curative intent for five of these and a palliative intent for six. In eight patients, surgery was not carried out due to advanced disease (
As Barrett’s carcinogenesis is a multistep process that follows the typical metaplasia-dysplasia-adenocarcinoma sequence, prognostic markers for disease progression have been sought. These have included factors within the cell cycle, oncogenes, and tumor suppressor genes. Increased proliferative activity has been reported in different tumors [
Initial studies assessing Ki-67 antigen used flow cytometry. The results obtained using this method did not show significant differences in Ki-67 expression, when comparing BE patients with different degrees of dysplasia and adenocarcinoma [
The Ki-67 index (percentage of stained cells/total cells) has been used to evaluate the proliferative activity of tumors and requires immunohistochemistry in paraffin-embedded tissue. The Ki-67 index is now the method of choice for proliferation studies, due to its accuracy and ease of use. Although electronic counting is sometimes used for this method, we did not have access to the necessary equipment and used the more conventional manual counting method [
Hong et al. used the gastric epithelium as a control to evaluate Ki-67 in BE [
The crypt stratification, according to its depth, has been evaluated in some studies. These studies showed a difference in Ki-67 distribution, mainly between LGD and HGD. The proliferative activity moved from the deep compartment of crypt, in the BE with LGD, to the superficial compartment, in the BE with HGD [
Polkowski et al. analyzed the Ki-67 index in 25 esophagectomy-resected specimens, in different histological areas of BE [
Lauwers et al. analyzed Ki-67 expression in 20 esophagectomy specimens and reported 10% positivity in BE without dysplasia, 20% in LGD, and 50% in HGD [
Hong et al. evaluated the Ki-67 index in 43 patients with BE [
Rioux-Leclercq et al. assessed Ki-67 expression in 44 esophagectomy specimens, in different histological areas [
In a recent study, conducted by Feith et al., different histological areas were evaluated for Ki-67 in 24 esophagectomy specimens [
Another recent study by Szachnowicz et al. evaluated Ki-67 in 13 esophagectomy specimens and demonstrated “moderate” or “strong” proliferative activity in all cases of BE (
Bhargava et al. conducted a prospective study on the behavior of different markers in BE, using a rigorous esophageal biopsy protocol [
In summary, we found inconclusive and heterogeneous results in all of these previous studies, although there is agreement on a correlation between Ki-67 and disease evolution. In our study, the average Ki-67 expression was 10% in the normal gastric mucosa, in accordance with the literature. Patients with BE and EAC showed a Ki-67 positivity of 21% and 38%, respectively. These results differ from those previously reported, that is, up to 45% positivity in BE and 60% in cancer. These differences may be partly due to small and unrepresentative sample sizes, taken from studies based on different histological areas of esophagectomy-resected specimens. In such studies, only one patient may be analyzed using several histological sections. Our study sample is patient based and not specimen based, and each patient has only one diagnosis. Variations in immunohistochemical technique may also partly explain the variable results. These may include the types of antibodies, antigenic presentation, and assessment of the marker.
We demonstrated a significant correlation between the Ki-67 index, indicating proliferative activity, and the Barrett’s esophagus to adenocarcinoma progression. The results are concordant with the literature and confirm the progressive nature of this disease relative to the increasing prevalence of this marker.
In patients with EAC, we did not find an association between Ki-67 expression and either clinical staging, tumor penetration or nodal spread. These results suggest a limited role for Ki-67 as a prognostic marker in patients with this cancer; however, the small sample size used to carry out comparisons within the group must be considered. Other studies have similarly not found significant differences in Ki-67 expression relative to cancer staging [
In the last 10 years, more than 10 studies have noted that adequate gastroesophageal reflux control is associated with the histological regression of Barrett’s esophagus. Antireflux surgery was shown to be an important predictive factor for histological regression, occurring in 36% of patients undergoing surgery [
Regarding methodological aspects, we would like to point some fragilities of this study. First, as patients with cancer were more likely to be men, we could not rule out the impact of smoking or alcohol consumption. Also, the sample of patients is small; however we were able to reach statistical significance. At last, as this is a retrospective study, we could not obtain normal esophageal mucosa biopsy.
We found small absolute differences in Ki-67 expression between the three groups (controls, BE, and EAC), despite the fact that differences were statistically significant. This suggests a limited role for Ki-67 as a powerful marker of Barrett’s carcinogenesis. To better evaluate the prognostic value of this marker in the metaplasia-dysplasia-adenocarcinoma sequence, a prospective study with followup of patients at risk should ideally be conducted. Considering the variability of published results in this field, future studies require better standardization of the methods to allow improved comparisons between the outcomes.
The Ki-67 index was 10% in patients with normal gastric mucosa (control), 21% in patients with BE, and 38% in EAC patients. There was a significant difference between all the groups, with an increasing expression of Ki-67 relative to the progression of BE to adenocarcinoma.
There was linear correlation between Ki-67 expression and the metaplasia-adenocarcinoma progression in BE, demonstrating an increasing Ki-67 positivity relative to disease evolution.
The authors would like to thank the financial support from FIPE/HCPA.