Barrett Esophagus

Diagnosis

Barrett esophagus is diagnosed by endoscopic examination, and 2 criteria must be fulfilled. First, the endoscopist must ascertain that columnar epithelium lines 1 cm or greater of the distal esophagus. Second, biopsy specimens of that columnar epithelium must show evidence of metaplasia, which is a change from one adult tissue type to another. To ascertain that columnar epithelium lines the distal esophagus, the endoscopist first must locate the esophagogastric junction (EGJ), which is recognized as the most proximal extent of the gastric folds, and then determine that columnar epithelium extends 1 cm or greater above the EGJ into the esophagus ( Fig. 47.1 ). Endoscopically, columnar epithelium has a reddish color and velvet-like texture that can be distinguished readily from normal esophageal squamous epithelium, which is pale and glossy. There is disagreement among experts regarding the histologic type of epithelium required to confirm that there is metaplasia in the esophagus. Virtually all would agree that the finding of an intestinal-type epithelium with goblet cells (which has been called intestinal metaplasia, specialized intestinal metaplasia, or specialized columnar epithelium ) is clear evidence of metaplasia. Most published studies on Barrett esophagus have used intestinal metaplasia as a requisite diagnostic criterion. However, some authorities argue that cardiac-type mucosa, which is composed almost exclusively of mucus-secreting cells, also is metaplastic, has malignant predisposition, and can be considered diagnostic of Barrett esophagus. Although this debate remains unresolved, U.S. gastroenterology societies require the demonstration of intestinal metaplasia with goblet cells as a diagnostic criterion for Barrett esophagus.

Barrett esophagus can be further categorized as long segment (when the metaplastic epithelium extends at least 3 cm above the EGJ) or short segment (when <3 cm of metaplastic epithelium lines the distal esophagus). Short-segment Barrett esophagus, which is by far the most common form of the disease, was not widely recognized until it was described by Spechler et al. in 1994. Another proposed system for categorizing Barrett esophagus, the Prague C and M criteria, identifies the circumferential (C) and the maximum (M) extent of Barrett metaplasia. Data suggest that the cancer risk in Barrett esophagus varies with the extent of the metaplastic lining.

Epidemiology

Barrett esophagus typically is discovered during endoscopic examinations performed for the evaluation of GERD symptoms in middle-aged and older adults. The average age at the time of diagnosis is approximately 55 years. The condition is rare in children younger than age 10 and virtually nonexistent in children younger than age 5. White men predominate in most series, and, for unknown reasons, Barrett esophagus is uncommon in black and Asian populations, although the prevalence has increased recently in Asian countries. Cigarette smoking is also associated with Barrett esophagus. Among adult patients who have endoscopic examinations because of GERD symptoms, long-segment Barrett esophagus is found in 3% to 5%, whereas 10% to 20% have short-segment Barrett esophagus. In the general adult population of Western countries, the prevalence of Barrett esophagus (predominantly short segment) is between 1.6% and 6.8%.

Published estimates on the annual incidence of cancer in patients with long-segment Barrett esophagus have been as high as 2.9%, but a report published in 2000 showed that many of those estimates were exaggerated because they were based on older, small studies that suffered from publication bias. For the first decade of the new millennium, it was widely assumed that the cancer risk for patients with nondysplastic Barrett esophagus was approximately 0.5% per year. However, more recent studies suggested that the cancer risk for such patients is even lower, in the range of only 0.12% to 0.33% per year.

The epidemiology of esophageal adenocarcinoma is similar to that of Barrett esophagus. GERD is strongly associated with both conditions, and, like Barrett esophagus, esophageal adenocarcinoma affects white men predominantly. Obesity, especially with central adiposity, predisposes to both Barrett esophagus and esophageal adenocarcinoma, and the dramatic rise in the frequency of obesity in the USA has paralleled a similar rise in the prevalence of Barrett cancer. The mechanisms underlying these associations with obesity are not clear but may relate to the fact that central adiposity predisposes to GERD, perhaps by increasing intra-abdominal pressure (see also Chapters 46 and 48 ). Obesity also is associated with elevated serum levels of pro-proliferative hormones such as insulin-like growth factor I and leptin, and with decreased levels of the antiproliferative hormone adiponectin, factors that may contribute to carcinogenesis in Barrett esophagus.

It has been proposed that the declining frequency of infection with Hp in Western populations also may be contributing to the rising frequency of esophageal adenocarcinoma (see Chapter 48 ). A number of studies have suggested that Hp infection may protect against the development and neoplastic progression of Barrett esophagus, perhaps because, in a subset of patients, this infection may prevent GERD by decreasing gastric acid secretion. Other factors that appear to protect against the development of esophageal adenocarcinoma include the use of aspirin and other NSAIDs and the consumption of a diet high in fruits and vegetables. Although cigarette smoking and alcohol consumption are very strong risk factors for squamous cell carcinoma of the esophagus, cigarette smoking only modestly increases the risk for esophageal adenocarcinoma and alcohol does not appear to affect that risk at all. Finally, it has been proposed that the rising incidence of esophageal adenocarcinoma might be due to increased intake of dietary nitrate in green leafy vegetables that has resulted from the widespread use of nitrate-based fertilizers in Western countries following World War II.

Pathogenesis

Barrett pathogenesis begins with esophageal injury by GERD, and patients with long-segment Barrett esophagus often have severe GERD (see Chapter 46 ). Table 47.1 lists some physiologic abnormalities that have been reported in Barrett patients and suggests how those abnormalities might contribute to GERD severity. Individual patients may exhibit any, all, or none of those abnormalities, and their prevalence in Barrett esophagus is disputed. For example, some investigators have described normal gastric acid secretion in patients with long-segment Barrett esophagus. In addition, many patients with short-segment Barrett esophagus have no GERD symptoms and no endoscopic signs of esophagitis. Indeed, one large study has suggested that short-segment Barrett esophagus may affect approximately 5% of adults, irrespective of the presence of GERD symptoms. Studies have shown that, even in healthy volunteers, the very distal esophagus can be exposed to acid for more than 10% of the day. Such acid exposure can damage the esophagus directly and indirectly when nitrite (generated from dietary nitrate) reacts with acid to produce nitric oxide. High concentrations of nitric oxide in the distal esophagus have been observed in patients with GERD who have ingested nitrate.

In Barrett esophagus, reflux-damaged squamous epithelium is replaced by an intestinal-type epithelium that, presumably, is more resistant to GERD injury. Because intestinal-type cells are not found normally in the esophagus, this process must involve GERD-induced molecular reprogramming of key developmental transcription factors in the progenitor cellsgiving rise to the metaplasia. However, the identity of these progenitor cells is not clear, and there are several candidates. It has been suggested that Barrett metaplasia develops when GERD causes mature esophageal squamous cells to transdifferentiate into columnar cells or when GERD causes abnormal differentiation of immature progenitor cells. Potential esophageal sources for these progenitor cells include basal cells of the squamous epithelium, or cells lining the ducts of esophageal submucosal glands. Alternatively, as discussed in Chapter 46 , Barrett progenitor cells might migrate into the reflux-damaged esophagus from the gastric cardia, from a nest of residual embryonic-type cells that can be found at the gastroesophageal junction, or from a recently described population of transitional basal cells located at the squamocolumnar junction. Finally, in a rat model of reflux esophagitis, Barrett metaplasia was shown to develop from circulating bone marrow stem cells. Genes up-regulated by reflux esophagitis that have been proposed to contribute to the squamous-to-columnar metaplasia of Barrett esophagus include CDX genes, which are known to mediate the differentiation of intestinal epithelial cells, and the Hedgehog target genes BMP4 (bone morphogenetic protein-4) and FOXA2, which also are involved in columnar cell differentiation.

Barrett epithelial cells appear to be more capable of resisting reflux-induced esophageal injury than the native squamous epithelial cells. Unlike squamous cells, for example, Barrett cells secrete mucins and express the tight-junction protein claudin 18, features that render the epithelium more resistant to acid-peptic attack. Unfortunately, Barrett epithelium also is predisposed to neoplasia.

Molecular Biology of Neoplasia

During carcinogenesis, Barrett epithelial cells accumulate genetic and epigenetic alterations that endow the cells with the core physiologic attributes of malignancy proposed by Hanahan and Weinberg in 2000 (see Chapter 1 ). Those attributes include self-sufficiency in growth signals, insensitivity to antigrowth signals, evasion of apoptosis, limitless replicative potential, sustained angiogenesis, and the abilities to invade adjacent structures and to metastasize ( Fig. 47.2 ). More recently, Hanahan and Weinberg added 2 additional physiologic hallmarks of malignancy: the ability to reprogram energy metabolism to support continuous proliferation, and the ability to evade destruction by immune cells (T and B lymphocytes, macrophages, and natural killer cells).

Numerous genetic alterations have been described during the neoplastic progression of Barrett esophagus. Although a single such alteration may have multiple disparate effects, conceptually it can be useful to classify the alteration according to the major physiologic cancer attributes that it endows. For example, the expression of oncogenes (e.g., cyclin D1 , K-ras ), growth factors such as transforming growth factor-α and growth factor receptors such as EGF receptor enable Barrett cells to acquire self-sufficiency in growth signals. Insensitivity to antigrowth signals occurs primarily through the inactivation of tumor suppressor genes (e.g., TP53 , p16 ). Inactivation of TP53 also enables cells to evade apoptosis. Reactivation of the enzyme telomerase, which enables the cells to replace telomeres needed for cell division, can endow the cells with limitless replicative potential. Neoplasms can increase their vascular supply by secreting angiogenic factors such as vascular endothelial growth factor. Finally, for neoplastic cells to invade and metastasize, they must dissociate themselves from surrounding cells by disrupting cell adhesion proteins such as the cadherins and catenins, and by degrading the extracellular matrix through the secretion of enzymes such as matrix metalloproteases.

In Barrett esophagus, the acquisition of these core physiologic attributes of malignancy is facilitated by genomic instability and a tumor-promoting microenvironment. Barrett epithelial cells display genomic instability by gains or losses in segments of chromosomes that alter the cells’ DNA content. Aneuploidy is the condition in which there is abnormal cellular DNA content, and aneuploid cells are at increased risk for neoplastic progression. Aneuploidy can be detected by flow cytometry, by fluorescence in situ hybridization, and by automated image cytometry; the latter 2 techniques are more feasible for routine clinical practice. In some neoplasms, infiltration by immune cells with tumor-promoting effects also contributes to the acquisition of malignant attributes, but little is known about the tumor-promoting effects of immune cells in Barrett esophagus. Aneuploidy has been proposed as a biomarker for neoplastic progression in Barrett esophagus, as have a number of the genetic alterations discussed in the preceding paragraph. Although there have been some promising studies, especially those using panels of biomarkers, gastroenterology societies presently do not endorse the routine clinical use of biomarkers for clinical management of patients with Barrett esophagus.

Recent advances in genomic techniques such as whole-genome sequencing and whole-exome sequencing (which is limited to gene coding regions) have contributed greatly to knowledge of how Barrett cells become tumor cells. Whole-exome sequencing of DNA extracted from areas of Barrett metaplasia associated with Barrett adenocarcinomas has revealed a frequency of somatic mutation in the non-neoplastic metaplasia higher than that found in prostate or breast carcinomas and with a mutational pattern indicating genomic damage caused by oxidative stress (likely due to GERD) (see Fig. 47.2 ). Both the nondysplastic Barrett metaplasia and associated adenocarcinomas exhibit mutations in the p53 tumor suppressor gene, evidence that p53 mutations occur early in Barrett carcinogenesis. Oncogene activations appear to occur later, because they are found only in dysplastic and cancerous tissues.

Interestingly, the new genomic techniques have revealed 2 general pathways for Barrett carcinogenesis: the “traditional pathway” (which involves the step-wise accumulation of alterations in tumor suppressor genes, followed by oncogene activation, genomic instability, and malignant transformation) and the “genome-doubled pathway” in which p53 mutation is followed by whole genome doubling, leading to genomic instability, oncogene amplification, and malignant transformation) ( Fig. 47.3 ). Most tumors in Barrett metaplasia (62.5%) appear to develop through the genome-doubled pathway, which can be a far more rapid pathway to malignancy than the traditional pathway. Rapid cancer development through genome doubling might explain the frequent failure of endoscopic surveillance to detect cancer progression in Barrett esophagus.

Unlike the aforementioned studies in which somatic mutations are identified by genomic analyses of tissue specimens, genome-wide association studies (GWAS) identify germline alterations and typically are performed using DNA in whole-blood specimens. GWAS have shown that unrelated subjects with Barrett esophagus and esophageal adenocarcinoma have substantial overlap of single-nucleotide polymorphisms (SNPs) at specific locations in the genome, suggesting a shared genetic susceptibility for Barrett esophagus and its cancer. More recent GWAS analyzing selected SNPs within inflammatory pathways have found a significant association between germline variations in the cyclooxygenase pathway (specifically in the antioxidant microsomal glutathione S-transferase 1 gene) and the development of Barrett esophagus and esophageal adenocarcinoma. Another GWAS explored the association between 7 SNPs implicated as risk factors for Barrett esophagus with several well-known epidemiologic risk factors for Barrett esophagus (GERD, cigarette smoking, and BMI). Only an SNP in the FOXP1 gene (which regulates esophageal development) and the presence of at least weekly reflux symptoms were found to modify the risk of developing Barrett esophagus and esophageal adenocarcinoma. The identification of novel gene-environment interactions by these new “omics” approaches may well revolutionize our understanding of how Barrett esophagus develops and progresses to esophageal adenocarcinoma.

Dysplasia

Before neoplastic Barrett cells become malignant, some of the same genetic and epigenetic alterations that endow the physiologic attributes of malignancy also cause morphologic changes in the tissue that pathologists recognize as dysplasia ( Fig. 47.4 ). Dysplasia (also called intraepithelial neoplasia ) can be viewed as the histologic expression of genetic and epigenetic alterations that favor unregulated cell growth. Dysplasia is recognized by cytologic and architectural abnormalities in esophageal biopsy specimens that include (1) nuclear changes such as enlargement, pleomorphism, hyperchromatism, stratification, and atypical mitoses; (2) loss of cytoplasmic maturation; and (3) crowding of tubules and villiform surfaces. Dysplasia is categorized as low grade or high grade depending on the degree of histologic abnormalities, with more pronounced abnormalities assumed to reflect more severe genetic damage and greater potential for carcinogenesis. Pathologists have difficulty distinguishing low-grade dysplasia in Barrett esophagus from reactive changes caused by reflux esophagitis, and interobserver agreement for the diagnosis of low-grade dysplasia may be less than 50%. Interobserver agreement is better (approximately 85%) for high-grade dysplasia, but there is substantial disagreement among pathologists in distinguishing high-grade dysplasia from intramucosal carcinoma (see Chapter 48 ).

Dysplasia in Barrett esophagus often is associated with visible mucosal irregularities, but those irregularities frequently are subtle and easily missed, especially by nonexperts. Furthermore, dysplasia can be patchy in its extent and severity. These factors contribute to the substantial problem of biopsy sampling error in identifying dysplasia. To find dysplasia in Barrett esophagus, endoscopists traditionally have used the “Seattle biopsy protocol,” a random biopsy sampling system in which 4-quadrant biopsies are taken at 1- to 2-cm intervals throughout the length of Barrett metaplasia. However, a recent analysis of 2 radiofrequency ablation (RFA) trials found that dysplasia is most likely to be present in the proximal half of the Barrett metaplasia segment, suggesting that future biopsy protocols might be modified to biopsy that proximal segment preferentially. It is clear that the Seattle biopsy protocol often misses areas of dysplasia and even cancer. A recent systematic review and meta-analysis found that, among patients diagnosed with nondysplastic or low-grade dysplastic Barrett esophagus on an index endoscopy and who were followed for 3 years or more, 25% of those later diagnosed with esophageal adenocarcinoma had the cancer discovered within 1 year of the index endoscopy. This suggests that the cancers were present and missed at the time of the index procedure. In older series of patients who had esophagectomy performed because endoscopic biopsies revealed high-grade dysplasia in Barrett esophagus, invasive cancer was found in as many as 30% to 40% of the resected esophagi. However, a critical review of those older studies suggested that 13% was a more accurate estimate of the frequency of invasive cancer in this situation. With modern high-definition endoscopes and newer diagnostic techniques such as endoscopic mucosal resection (EMR, see later), the rate of missing invasive cancer in patients found to have dysplasia should be very low.

A number of advanced imaging techniques have been used to facilitate the detection of dysplasia and early cancer in Barrett esophagus, including chromoendoscopy, autofluorescence endoscopy, magnification endoscopy, narrow band imaging, optical coherence tomography, Raman detection methods, confocal laser endomicroscopy, and volumetric laser endomicroscopy (see Chapter 48 ). In the hands of endoscopists with proper training and expertise, these advanced techniques can identify neoplastic lesions that might not be apparent with routine, white-light endoscopy. However, gastroenterology societies do not mandate use of these advanced imaging techniques in routine clinical practice, and careful endoscopic examination with high-definition white-light endoscopy and random 4-quadrant biopsy remains the standard of care.

The overall incidence of cancer development in patients with nondysplastic Barrett esophagus is approximately 0.25% per year. One study suggests that patients who have non-neoplastic Barrett esophagus develop low-grade dysplasia at the rate of 4.3% per year and high-grade dysplasia at the rate of 0.9% per year. The natural history of low-grade dysplasia in Barrett esophagus remains disputed due to difficulties confirming the diagnosis and contradictory study results. In one study, 2 expert pathologists reviewed pathology slides in 147 patients in whom low-grade dysplasia had been diagnosed at community hospitals in the Netherlands; the experts confirmed the diagnosis in only 15% of cases. In those confirmed cases, however, the cumulative risk of neoplastic progression was 85% after 9 years. In contrast, an American study of 210 patients with low-grade dysplasia followed for a mean of 6.2 years found that the rate of progression of only 0.4% per year and diagnostic consensus among the pathologists was not associated with neoplastic progression. For patients with high-grade dysplasia, cancers develop at the rate of approximately 6% per year.