|Home | About | Journals | Submit | Contact Us | Français|
Intestinal metaplasia (IM) of the gastric mucosa is a relatively frequent precancerous lesion (1). The inclusion of IM in a gastric biopsy pathology report often creates uncertainty for the gastroenterologist about the appropriate management. Although the risk of gastric cancer is increased in the presence of IM, the overall risk of gastric cancer in a patient with IM is extremely low compared with the risk of adenocarcinoma in a patient with Barrett’s esophagus (BE)(2). Although the incidence of gastric cancer is high in certain regions, such as Asia and Latin America, it is much lower in the United States and other Western countries, making it difficult to justify broad surveillance programs. The aims of this article are (i) to assist the clinician in identifying subgroups of patients with IM at increased risk for gastric cancer and (ii) to propose an algorithm for gastric IM management, considering the lack of universally accepted guidelines that can be applied to any population.
The gastric cardia is a poorly defined region whose anatomy and pathology are a subject of controversy. The cardiac mucosa, located immediately distal to the squamocolumnar junction (Z line) and proximal to the oxyntic mucosa, consists of columnar epithelium on the surface and mucous glands with morphology similar to that of the antrum. The diagnosis and management of IM of the gastric cardia are especially challenging in Western affluent societies because of the increasing incidence of adenocarcinoma of the cardia and lower esophagus (3,4) and its precursors, gastroesophageal reflux and BE. Carcinomas arising in the cardia or the lower esophagus tend to involve the gastroesophageal junction, making it often difficult to determine the site of origin of the tumor.
Esophageal adenocarcinoma has been causally related to gastroesophageal reflux. The risk factors for adenocarcinoma of the cardia are less well defined. A dual etiology, linking some tumors to Helicobacter pylori infection and others to reflux injury, is favored (5,6). IM of the cardia and BE differ in their risk for malignant transformation, and there are different implications for patient management (3). It is therefore important to distinguish between the two entities in biopsies from the gastroesophageal junction area. Several histopathology features have been found to be of special value (7). The presence of overlying squamous epithelium, esophageal gland ducts, and hybrid glands is seen exclusively in BE. IM of the incomplete type and multilayered epithelium also favor BE (7). However, in some cases the histology of BE and cardia IM is identical, and it is the endoscopic location of the biopsy that determines whether the patient is more likely to have BE rather than cardia IM.
Although the incidence of distal gastric cancer has decreased steadily in Westernized countries over the past century, this disease is still the second leading cause of cancer deaths worldwide (8). The International Agency for Research on Cancer categorized H. pylori infection as a type I carcinogen (9), and it is considered the primary cause of gastric cancer. The infection induces a chronic inflammatory process in the gastric mucosa. Over time, atrophy and IM may develop (10,11). IM is frequently identified in distal gastric biopsies, especially in populations at high risk for gastric cancer, such as those of eastern Asia, eastern Europe, and Andean Latin America. In the United States, the majority of the population is at low risk for gastric cancer, but there are several ethnic populations at high cancer risk, such as African Americans, Native Americans, and immigrants from Asia and Latin America (12,13). IM has been found to be more prevalent in those high-risk groups (14).
The accepted model for the development of gastric adenocarcinoma of the intestinal type consists of the following precancerous steps: non-atrophic gastritis, multifocal atrophic gastritis, IM, and dysplasia (1,10). Thus, IM represents a stage within a prolonged process. Identified risk factors for IM include H. pylori infection, high salt intake, smoking, alcohol consumption, and chronic bile reflux (15–21). The metaplastic foci tend to appear first at the antrum–corpus junction, especially at the incisura angularis. As the process advances, the foci enlarge and coalesce, extending to the neighboring mucosa in both the antrum and the corpus (22). Dysplastic foci may eventually appear within areas of IM; they are usually small and therefore subject to sampling error. The severity and tempo of progression of all the steps in the precancerous cascade may be influenced by the virulence determinants of the infecting H. pylori strain, as well as by environmental and host genetic factors (23–26).
In general, IM is easily recognized in histologic sections stained with hematoxylin and eosin (H&E). It has long been recognized that IM is heterogeneous, and several classifications have been proposed (11,27–30). Matsukura et al. based their classification of IM on the presence of small intestinal digestive enzymes (29). In complete-type IM, most or all of these enzymes are expressed. In incomplete-type IM, these enzymes are absent or only partially expressed. Discrimination of complete versus incomplete types of gastric IM has been widely adopted by pathologists and clinicians on the basis of the morphology on H&E-stained sections. Complete metaplasia is currently diagnosed when the epithelium resembles the small intestinal phenotype, with eosinophilic enterocytes displaying a well-defined brush border (representing absorptive microvilli) and well-formed goblet cells. Paneth cells may also be present. Incomplete metaplasia resembles a colonic epithelium phenotype with multiple, irregular mucin droplets of variable size in the cytoplasm and absence of a brush border (Figure 1).
Another classification, used in research but not in routine clinical practice, combines the morphologic characteristics described for complete and incomplete types with analysis of the types of mucins expressed. Histochemically, normal gastric mucins are pH neutral, and they stain magenta with periodic acid–Schiff (PAS). In IM, acid mucins replace the original gastric mucins and are stained blue with Alcian blue (AB) at pH 2.5. Therefore, combined PAS–AB staining discriminates between normal epithelium and IM. Acid mucins, in turn, can be sialic or sulfated; the latter stain brown with high-iron diamine (HID). Combined HID–AB staining can be used to assess for both types (Figure 2). Filipe et al. (11) studied mucin phenotypes using these discriminatory histochemical stains and proposed a classification system. IM type I (complete) expresses only sialomucins. Type III (incomplete) expresses sulfomucins. Type II (incomplete) is a hybrid form expressing a mixture of gastric and intestinal mucins. The question of whether these three types of IM follow a chrono-logic sequence is still under investigation (31–33). Accumulated experience with human specimens frequently shows the simultaneous expression of different types of mucins in the same metaplastic epithelial cell, suggesting that the different morphologic manifestations of the metaplastic process represent a gradual phenotypic change, with some types of mucins becoming less abundant, giving rise to new types of mucins (11,34). Neutral mucins present in normal mucosa gradually decrease during the initial development of IM, whereas sialomucins appear and become the predominant type of mucin. In more advanced stages of IM, sulfomucins appear and may become the predominant mucin. Figure 3 outlines the dynamics of the process and the types of mucins observed in each step.
Our proposed algorithm for the approach to the patient with gastric IM is shown in Figure 4. Subjects at high risk for gastric cancer who are undergoing upper endoscopy will benefit from mapping with biopsy samples from the antrum, corpus, incisura angularis, and any endoscopically visible lesion (35–38). Because of the multifocal nature of the metaplastic process, extensive mapping techniques have been proposed (16). Although a pathologic diagnosis of IM of the stomach is frequently reported without qualifiers, in our opinion, such a description is insufficient. The most important distinction should be between complete and incomplete IM, which can be made with routine H&E-stained sections. If the original report is not explicit about the type or extension of the IM, a dialogue with the pathologist is indicated. Several studies have demonstrated a significant relationship between incomplete IM and gastric carcinoma (11,27,39,40). A diagnosis of complete IM may not be an indication for prolonged surveillance and endoscopic monitoring. However, individual clinical judgment may dictate further evaluation. A review of the management of patients with IM suggests that for most patients in the United States, the risk of progression to cancer is very low, and surveillance is not clinically indicated in an “average-risk” patient (2). Likewise, guidelines by the American Society for Gastrointestinal Endoscopy state that endoscopic surveillance for gastric IM has not been extensively studied in the United States and therefore is not uniformly recommended (37). However, performance of endoscopic surveillance with topographic mapping of the entire stomach in subjects at increased risk of gastric cancer on the basis of ethnic background, immigration from a geographic location with high gastric cancer risk, or family history is recommended (2,37).
The aforementioned reports did not take into account the histologic type or extension of IM. Evidence indicates that both factors are of great importance in terms of gastric cancer risk. Therefore, subjects with incomplete IM or who have extensive IM should also be further evaluated to decide whether long-term endoscopic surveillance is justified. IM may be considered extensive when it involves at least two locations or when it is moderate or marked in more than one biopsy site (16).
A useful indicator of the extension of atrophic/metaplastic changes is the assessment of serum levels of pepsinogens, proposed initially by Samloff et al. as a potential “serologic biopsy” (41). Pepsinogen I (PGI, or PGA) is secreted by chief and mucous neck cells in the corpus and fundic glands, whereas PGII (or PGC) is produced by these cells and also by cells in the pyloric glands and Brunner’s glands (42). It is known that serum pepsinogen levels reflect the functional and morphologic status of the gastric mucosa (16,43). As mentioned, the precancerous process is usually initiated in the antrum–corpus junction. As atrophy and metaplastic changes advance, they tend to extend to the antrum and the corpus. A marked decrease in serum PGI levels is strongly associated with severe atrophy of the corpus mucosa. In Japanese studies, gastric atrophy is diagnosed when serum PGI levels are lower than 70 μg/l and the PGI/PGII ratio lower than 3.0; these criteria have shown high sensitivity (70.5%) and specificity (97%) for diagnosis of this condition (44). Using these criteria, Ohata et al. reported in a longitudinal cohort study that there is a stepwise increase in cancer risk driven by H. pylori infection in the initial steps and by gastric atrophy in the more advanced steps. The relative risk for cancer reached 61.85 when the pepsinogen levels were very low (PGI = 22.2 μg/l and PGI/PGII = 1.68) and the H. pylori infection had been lost (45). In some patients with gastric cancer, serum levels of pepsinogens may be within normal limits, because the atrophic/metaplastic changes are limited to the antral mucosa. In general, assessment of pepsinogen levels has proven to be very useful and suitable for cancer screening in populations at high risk of gastric cancer, because this indicator is low cost, minimally invasive, and easy to perform and provides quick results (46).
Although gastric dysplasia (also called noninvasive neoplasia or adenoma) is not the central subject of this article, its management is briefly mentioned. Patients with high-grade dysplasia (carcinoma in situ) confirmed by at least two gastrointestinal pathologists should undergo surgical or endoscopic resection because of the high probability of coexisting or metachronous invasive carcinoma (37,47). The management of low-grade dysplasia is less well defined. A review recommends annual endoscopic monitoring with rebiopsy (48). Another review suggests that endoscopic surveillance with a topographic mapping biopsy strategy should be performed every 3 months, for at least the first year; surveillance should be suspended when two consecutive endoscopies show negative results (49).
The strategy for gastric cancer prevention is to identify individuals at the highest risk, offer them intervention to decrease the risk whenever possible, and discuss with them the available options for surveillance. Since IM and the precancerous process in general may be driven by H. pylori infection, it is advisable to eradicate the bacterium in infected patients. Evidence suggests that curing the infection attenuates the precancerous process (19,50–52). Recently it has been reported in a very large cohort study from Taiwan that eradication of H. pylori within the first year of detection of peptic ulcers resulted in a significant reduction of risk for development of gastric carcinoma when compared with later eradication (53). Some controversy exists about the benefit for cancer prevention of eradicating the infection when IM has developed (54,55). Because it has been reported that IM scores did not regress in a 4-year follow-up after eradication in Japanese subjects (56), a surveillance protocol is still warranted in patients successfully treated for H. pylori. Similar data have been reported in an Italian cohort (57).
In the evaluation of the infection status, it is worthwhile to note that H. pylori colonization is typically absent in the IM foci of the complete type, whereas it may be present in the adjacent nonmetaplastic gastric mucosa and in IM of the incomplete type (58). In addition, decreased sensitivity of the rapid urease test is observed in subjects taking proton pump inhibitors. For these reasons, H. pylori infection should not be excluded on the basis of the negative results of tissue-based tests; other diagnostic tests, including serology, 13C-urea breath test, or stool antigen, should be considered. In particular, serologic testing has proven to be useful in detecting infection in subjects with IM without the demonstration of H. pylori in biopsy specimens using histology or rapid urease testing (59). Diagnostic testing and treatment for H. pylori are reviewed elsewhere (60). Other recommendations for subjects with IM can be derived from epidemiologic studies of gastric cancer causation: avoid smoking and excessive salt intake, and consume adequate amounts of fresh fruits and vegetables, which are abundant sources of antioxidant micronutrients.
A histopathologic diagnosis of complete-type IM, in and of itself, is not an indication for prolonged surveillance unless other indicators of risk for gastric cancer are present. H. pylori infection status must be determined by tissue-based tests and serology, and infected patients should be treated. A diagnosis of incomplete-type IM should be followed by endoscopic topographic mapping to evaluate its extension and rule out more advanced lesions such as dysplasia or early adenocarcinoma. Serum pepsinogen levels are an underutilized alternative for the assessment of the extension of the atrophic changes in the gastric mucosa. At the present time, there are no other recognized good markers of gastric dysplasia or cancer.
Financial support: Our research has been supported by Program Project Grant CA028842 of the National Cancer Institute, R01 DK053620, the National Institute of Diabetes & Digestive and Kidney Diseases and the Office of Medical Research, Department of Veterans Affairs.
CONFLICT OF INTEREST
Guarantor of the article: Pelayo Correa, MD.
Specific author contributions: All authors participated in the review of the literature and in the drafting and editing of the article for intellectual content and approved the final draft submitted.
Potential competing interests: None.