Defining Cancer Risk in Barrett’s Esophagus: A Pathologist’s Perspective

Barrett’s esophagus is an acquired metaplastic change of the distal esophagus that occurs secondary to chronic gastroesophageal reflux. The normal squamous epithelium is replaced by a columnar epithelium containing a distinctive specialized intestinal epithelium, defined pathologically by the presence of goblet cells. The significance of establishing a diagnosis of Barrett’s esophagus lies in the fact that this condition carries with it a risk for the development of esophageal adenocarcinoma. Carcinomas that arise in the setting of Barrett’s esophagus are thought to develop as a part of a dysplasia-adenocarcinoma sequence similar to that seen in patients with inflammatory bowel disease.

Many chronic inflammatory conditions are associated with an increased risk of neoplastic transformation. For example, in the gastrointestinal tract, cancer risk is increased in patients with chronic inflammatory bowel disease and in those with long-standing Helicobacter pylori infection. When dysplasia develops in the setting of a chronic inflammatory condition, it is usually multifocal and widespread, suggesting that neoplastic transformation may result from abnormalities occurring over large regions of the affected mucosa. These alterations are likely attributable to an increased rate of genetic mutation facilitated by chronic mucosal inflammation and repetitive cycles of epithelial regeneration. Concurrent effects of increased proliferation resulting from the repair of the damaged mucosa also contribute to genetic damage. Rapidly dividing cells are likely at increased risk for undergoing mutation when compared to quiescent cells. Furthermore, cell proliferation is necessary for the permanent establishment of genetic mutations within the cell population.

The cancer risk associated with Barrett’s esophagus progressively increases as the epithelium undergoes changes from Barrett’s metaplasia to low-grade dysplasia, high-grade dysplasia, and ultimately invasive carcinoma. As a result, all Barrett’s esophagus patients undergo routine endoscopic surveillance so that precancerous mucosal alterations might be detected prior to the development of invasive cancer. Several difficulties are encountered, however, in determining which patients will ultimately progress to cancer.

First, diagnostic difficulties may exist in establishing the diagnosis of dysplasia in the first place. The problems a pathologist may face in establishing a diagnosis of dysplasia include difficulties relating to sampling error, distinguishing reactive changes from changes due to dysplasia, differences in observer interpretation of the diagnosis of dysplasia, and difficulties in differentiating high-grade dysplasia from invasive carcinoma.

Second, even if the pathologist can confidently make a diagnosis of dysplasia, it is not certain that any given patient will continue to progress through the metaplasia-dysplasia-carcinoma sequence. Certainly, many patients with Barrett’s esophagus will never develop dysplasia or invasive carcinoma.

Because of the difficulties in distinguishing between regenerative changes and dysplasia, and because not all persons with Barrett’s esophagus develop dysplasia or carcinoma, attempts have been made to apply numerous adjunctive tools to establish a diagnosis of dysplasia and to identify those patients who are likely to subsequently develop a carcinoma. A huge number of studies have examined potential markers of neoplastic risk in Barrett’s esophagus patients, but only a few markers appear to have potential utility in this regard. These include DNA ploidy studies, expression of proliferation markers, expression of α-methylacyl-coenzyme A racemase—or AMACR—and expression of the tumor suppressor proteins p53 and p16.

“Several difficulties are encountered in determining which patients will ultimately progress to cancer.”

DNA PLOIDY STUDIES

In general, abnormalities in DNA ploidy correlate well with conventional histologic diagnoses of dysplasia and carcinoma, and several studies suggest that this marker might represent a valuable adjunctive tool in the evaluation of patients with Barrett’s esophagus. ^1,2 Careful mapping studies demonstrate that early carcinomas arise within a single aneuploid cell population.¹ Aneuploidy is detected with increased frequency as the histologic grade of the epithelium increases during neoplastic progression.¹ As a result, DNA analyses may be valuable in differentiating reactive atypia from dysplasia in a small biopsy, because high levels of aneuploidy generally occur only in patients with high degrees of dysplasia or adenocarcinoma. In addition, some specimens that are histologically negative or indefinite for dysplasia may contain aneuploid cells. In some patients, these aneuploid cell populations extend over large segments of the Barrett’s mucosa, suggesting that a single abnormal cellular clone may spread to involve large mucosal areas. Identification of such aneuploid areas in patients without histologic evidence of dysplasia might prompt rebiopsy or more frequent endoscopic surveillance. In contrast, a normal DNA ploidy pattern in a biopsy indefinite for dysplasia might provide reassurance that the lesion may not progress onward.

It is important to note, however, that such comparative histologic and ploidy analyses must generally be carried out using Feulgen staining and image analysis since flow cytometric DNA ploidy analysis and histologic analysis cannot be performed on the exact same tissue specimen. Such techniques are expensive and cumbersome, and are therefore difficult or impossible to implement as a part of routine daily practice.

“Certainly, many patients with Barrett’s esophagus will never develop dysplasia or invasive carcinoma.”

PROLIFERATION MARKERS

Immunohistochemical staining for MIB-1, the Ki-67 proliferation antigen, shows gradually increasing expression in the Barrett’s esophagus-dysplasia-adenocarcinoma sequence.^3,4 As would be expected, MIB-1 expression is seen in a significantly higher proportion of dysplastic or adenocarcinoma cells than in cells of nondysplastic Barrett’s epithelium. In addition, the pattern of MIB-1 expression is altered in dysplastic vs. nondysplastic mucosa. In Barrett’s esophagus without dysplasia the cells expressing Ki-67 are limited to the bases of the crypts, whereas in dysplasia, the proliferating cells extend upward into the upper portion of the crypt and onto the mucosal surface. A recent study suggests that the combined use of MIB-1 and p53 staining may be of value in decreasing interobserver variation in the diagnosis of Barrett’s associated dysplasia.⁵

p53 ALTERATIONS

The p53 tumor suppressor gene is the most frequently mutated gene in human cancers, and is implicated in the regulation of cell proliferation and differentiation, DNA repair and synthesis, and programmed cell death. p53 affects cell-cycle arrest in the G₁ phase in response to DNA damage, presumably allowing injured cells time to effect DNA repair before entering S phase. Loss of this checkpoint control potentially leads to replication of damaged DNA, and resultant genomic instability in affected cells.

Most p53 mutations are point mutations that alter the half-life of the p53 protein. This stabilized p53 protein is detectable by immunohistochemistry, while the normal wild-type protein, which is rapidly degraded, is not. p53 immunostaining does not identify all mutations, since those resulting in a truncated or absent protein may not be detected. In addition, p53 immunostains may also detect simple overexpression of the wild-type protein as occurs, for example, in the setting of repair of radiation-induced injury. However, p53 immunostaining is simple to perform, and is relatively fast and inexpensive in comparison with mutation analysis.

Alterations in p53 expression are common in esophageal adenocarcinomas and Barrett’s associated high-grade dysplasia.⁵ p53 abnormalities also occur in low-grade dysplasia and metaplastic Barrett’s epithelium without dysplasia, albeit at a lower frequency. In addition, p53 alterations reportedly are seen with greater frequency among those patients who will ultimately progress to high-grade dysplasia or carcinoma. This finding has prompted some to suggest that p53 immunohistochemistry be used as a complementary test with histologic evaluation in the diagnosis of dysplasia in patients with Barrett’s esophagus. However, it is important to note that there remain many patients without evidence of p53 abnormalities who progress to develop cancer, and there are some with p53 mutations who do not. Because of these difficulties, it is not possible to justify the routine use of p53 immunohistochemistry in the evaluation of biopsies from patients with Barrett’s esophagus.

p16 INACTIVATION

Inactivation of the tumor suppressor p16, located on chromosome 9, is one of the most common abnormalities identified in human tumors. p16 encodes a cell-cycle regulatory protein that inhibits cyclin-dependent kinases 4 and 6, preventing phosphorylation of the retinoblastoma gene product, Rb. This results in a block of cell-cycle progression in the G₁-S phase. Inactivation of p16 may occur as a result of mutation, homozygous deletion, or methylation of the promoter of the gene. Allelic loss of p16 is common in esophageal adenocarcinomas, and appears to be an early event in the Barrett’s esophagus-dysplasia-adenocarcinoma sequence that provides affected cells with a survival advantage.^7,8 This finding suggests that p16 loss may be a marker for patients at risk for later development of dysplasia or carcinoma. Loss of p16 expression may be identified using immunohistochemistry, but the practical utility of this stain in evaluating Barrett’s-associated dysplasia is unclear.

AMACR EXPRESSION

AMACR overexpression occurs in a variety of carcinomas, most notably prostatic adenocarcinoma. AMACR is also expressed in three quarters of colonic adenomas and most adenocarcinomas, but is rarely expressed in normal colonic epithelium.⁹ Recent studies reported no evidence of AMACR expression in non-neoplastic Barrett’s mucosa, but positive expression in 11%–38% of low-grade Barrett’s dysplasias, 64%–81% of high-grade dysplasias, and 72%–75% of esophageal adenocarcinomas.^10,11 Based on these findings, some authors suggest that immunohistochemical detection of AMACR expression is useful in establishing the diagnosis of dysplasia. However, other studies suggest that the sensitivity of this technique is low,¹² particularly in the situation that causes the pathologist the most difficulty, distinguishing low-grade dysplasia from reactive atypia. The relationship of AMACR expression to risk for progression of dysplasia is unknown at the present time.

“The situation that causes the pathologist the most difficulty is distinguishing low-grade dysplasia from reactive atypia.”

CONCLUSIONS

From all of these studies, it is clear that there is no single molecular marker that will suffice to allow us to predict who will or will not develop cancer in the setting of Barrett’s esophagus. Carcinogenesis is a multistep process that occurs as a result of alterations in many different genes. Therefore, it is likely that we will be required to develop panels of markers, which may, in differing combinations, allow us to predict neoplastic risk in individual patients. We have clearly not yet achieved this level of sophistication in our understanding of Barrett’s-associated neoplasia. With additional large, long-term follow-up studies, however, we may someday reach this goal.