Histologic Changes of the Gastric Mucosa Associated With Primary Gastric Lymphoma in Endoscopic Biopsy Specimens Open Access

The histologic criteria useful for the diagnosis of primary gastric lymphoma (PGL) in endoscopic biopsies have been previously established.1,2 On the other hand, the histologic changes of the gastric mucosa adjacent to PGL have only recently been adequately described, in spite of the high frequency of intestinal metaplasia and atrophy and occasional dysplasia that are noted.3 Such epithelial disorders were initially described in gastrectomy specimens from patients with PGL and are similar to those found in the vicinity of gastric adenocarcinoma.4,5 To determine the frequency and type of histopathologic gastric disorders in mucosa adjacent to PGL in endoscopic specimens, 54 biopsies from patients with PGL were analyzed. We also discuss the clinical importance of recognizing these changes in endoscopic biopsies.

Biopsy results were obtained from the files of the Departments of Pathology at the Instituto Nacional de la Nutrición Salvador Zubirán and the Instituto Nacional de Cancerología in Mexico City. The findings for 54 endoscopic biopsies diagnosed as PGL were reviewed. The initial diagnosis of PGL was confirmed in 40 gastrectomies; in the remaining 14 cases, diagnosis was made by the histopathologic and immunohistochemical characteristics of primary gastric (mucosa-associated lymphoid tissue [MALT]-type) lymphoma described in endoscopic biopsy specimens.1,2 In each case, the patient's age and gender and the following histologic changes were recorded: intestinal metaplasia (complete or incomplete); atrophy; dysplasia (low grade and high grade); atypical regeneration of the glandular epithelium; and erosion of the gastric epithelium. Intestinal metaplasia and atrophy were evaluated primarily in tissue fragments without lymphomatous infiltration, as recommended by the International Workshop on Histopathology of Gastritis (updated Sydney System), Houston, 1994.6 Intestinal metaplasia was considered as slight when confined to few gastric glands in one high-power field (HPF) (40×); as moderate when it involved several HPFs; and as marked when the majority of the gastric mucosa was replaced by intestinal-type epithelium. Reactive gastric epithelium and gastric dysplasia (low grade and high grade) were classified according to Goldstein and Lewin.7 Immunohistochemical stains were performed with the avidin-biotin technique on areas of gastric dysplasia and glands with atypical regenerative changes using antibodies to p53 (dilution, 1:50; Dako A/S, Glostrup, Denmark) and bcl-2 (dilution, 1:40; Dako). The staining was classified as follows: (−), negative or equivocal staining; (+), weak expression (positive dysplastic cells were detected in 1–20% of the glands); and (++) strong expression (positive dysplastic cells detected in more than 20% of the glands).8 Nonepithelial changes, such as the grade of lymphoma as well as the presence of lymphoid follicles and Helicobacter pylori, were also recorded. For comparative purposes, 50 biopsies from adult patients diagnosed with H pylori–associated chronic gastritis and 50 biopsies from adult patients diagnosed with gastric adenocarcinoma (25 intestinal type and 25 diffuse type) were reviewed. For each case, the following histologic changes in the gastric mucosa were recorded: intestinal metaplasia, atrophy, lymphoid follicles, erosion of the superficial epithelium, atypical regeneration of the glandular epithelium, and epithelial dysplasia.

The patients included 24 men and 30 women. Their ages ranged from 29 to 71 years (mean, 58 years). There were 28 (52%) low-grade and 26 (48%) high-grade, primary gastric MALT-type lymphomas. The gastric epithelial changes in the endoscopic biopsy specimens from patients with PGL, carcinoma, and gastritis are summarized in the Table. Thirty-two (59%) out of 54 biopsies showed intestinal metaplasia; metaplasia was complete in 16 cases and incomplete in the remaining 16 (Figures 1 through 3). In 22 of the 32 cases, intestinal metaplasia was moderate or marked and was identified in several or most HPFs, whereas in the remaining 10 cases, metaplasia was slight and confined to isolated histologic fields. The existence of gastric lymphoma and intestinal metaplasia was observed in the same or different tissue fragments in most endoscopic biopsies. Lymphoepithelial lesions were not observed in the metaplastic glands. Twenty of 54 biopsies (37%) showed atrophy. In 17 of these biopsies, atrophy was associated with moderate or marked intestinal metaplasia (Figure 3); in 2 cases, intestinal metaplasia was slight, and in the remaining case, no metaplastic glands were identified. We found gastric dysplasia (one low-grade and one high-grade dysplasia) without lymphoepithelial lesions in 2 biopsies (Figure 4). In these cases, p53 nuclear staining was observed in approximately 20% and 40% of dysplastic cells, respectively, while bcl-2 staining was negative. In both cases, in addition to dysplastic glands, moderate intestinal metaplasia and atrophy were also observed. Erosion of the superficial gastric epithelium was found in 33 (61%) of the biopsies; in most cases, this change was present in tissue fragments with lymphomatous infiltration. Atypical regenerative glandular changes were noted in 10 biopsies, with erosion of the gastric epithelium (Figures 5 and 6). In all but 2 cases, atypical regenerative features resembled dysplastic glands; however, the presence of superficial epithelial erosion and inflammatory infiltrates, as well as the lack of immunoreactivity to p53, led to the recognition of these features as reactive changes.

Helicobacter pylori was identified in 31 of 54 cases (57%). Lymphoid follicles were present in 21 cases (39%); some of these follicles were small, whereas others were larger and exhibited irregular borders, no macrophages, and a poorly defined mantle zone.

The frequencies of histologic findings in biopsies from patients with H pylori–associated gastritis and gastric adenocarcinoma are shown in the Table. As a whole, the frequencies of intestinal metaplasia, gastric mucosal atrophy, superficial epithelial erosion, atypical regenerative changes, and lymphoid follicles were similar in biopsies from patients with PGL and patients with adenocarcinoma. Gastric epithelial dysplasia was more frequent in patients with adenocarcinoma. Atrophy or intestinal metaplasia were present in 10% of biopsies diagnosed as H pylori–associated chronic gastritis, although lymphoid follicles were present in most biopsies.

The concurrent existence of intestinal metaplasia, atrophy, and dysplasia of the gastric mucosa in PGL was first described in patients in whom gastrectomy was performed because of PGL. In 51 gastrectomies in patients with PGL, intestinal metaplasia was found in 69.5% of the cases and atrophy was found in 54% of the cases.3 Recently, we compared histopathologic changes of the gastric mucosa adjacent to PGL and to gastric adenocarcinomas in 50 gastrectomy specimens each from patients with PGL or gastric adenocarcinoma; these specimens were obtained during autopsies of 100 adult individuals who died of diseases other than gastric cancer (Arista-Nasr, unpublished observations, 2000). For gastric lymphomas and carcinomas, the frequency of intestinal metaplasia was 70% in the former group and 80% in the latter group, whereas atrophy was found in 54% and 68%, and glandular dysplasia was found in 12% and 26%, respectively. In contrast, intestinal metaplasia and atrophy, but no dysplasia, were demonstrated in 10% of the 100 autopsy specimens. In the current study, the frequency of epithelial changes was lower in the gastric samples obtained by endoscopy than that found in gastrectomy specimens (Table). This finding may be due to limited sampling of gastric mucosa. However, a significant percentage of the endoscopic biopsies from patients with PGL showed intestinal metaplasia and atrophy, although gastric dysplasia was identified in only 2 cases.

Lymphoid follicles are commonly found in mucosa affected by the gastritis associated with H pylori infection, as well as in gastrectomy specimens from patients with PGL.6,9 In this review, we found lymphoid follicles in 84% of endoscopic biopsies with H pylori–associated chronic gastritis (Table). However, only 39% and 40% of the PGL and gastric adenocarcinomas, respectively, diagnosed in endoscopic biopsies showed lymphoid follicles. The lower frequency of lymphoid follicles could be explained by virtue of the small quantity of tissue obtained during the endoscopic procedure and by the fact that biopsies are frequently obtained exclusively from the neoplasia. In addition, lymphoid follicles are eventually replaced by neoplastic cells.

The frequencies of intestinal metaplasia, atrophy, atypical regenerative changes, and lymphoid follicles in endoscopic biopsies of gastric lymphoma and adenocarcinoma was similar in this study (Table). These findings support the proposal that the chronic gastritis associated with H pylori infection forms the basis from which lymphoma and/or carcinoma might develop.3 The simultaneous presence of lymphoid follicles, atrophy, and intestinal metaplasia in an endoscopic biopsy could point to a higher risk for the development of a gastric neoplasia, and such disorders probably should be periodically watched.

The coexistent epithelial disorders and PGL in the gastric mucosa could be explained as long-term sequelae of H pylori infection. It is widely recognized that this microorganism results in chronic inflammation, atrophy, intestinal metaplasia, and dysplasia of the gastric mucosa.4,5,10 ,Helicobacter pylori also induces the development of lymphoid follicles, which eventually progress to gastric lymphomas.11–13 Moreover, the association of gastric adenocarcinoma and PGL occurs with a higher frequency than expected by chance, and development of both neoplasias has also been attributed to the carcinogenic capability of H pylori. The association of gastric lymphoma with adenocarcinoma can be synchronic or metachronic, or can be intermingled (collision tumor).14–16 Cases of gastric lymphoma associated with single or multiple early gastric carcinomas have also been reported,17,18 as well as coexistent malignant lymphoma and early carcinoma of the stomach.19 

In this study, H pylori was evident in 57% of the biopsies. This relatively low percentage may have been due to insufficient sampling of the gastric mucosa and also to the inability of H pylori to colonize metaplastic mucosa. As noted, H pylori is associated with the active chronic gastritis that initiates the chain of events, although it is no longer present once intestinal metaplasia of the gastric mucosa occurs, compromising acid production and bringing the pH of gastric juice close to neutrality.20 

As shown in the Table, endoscopic biopsies in many patients with PGL showed erosion of the superficial epithelium by the underlying neoplasm, which could be associated with atypical regenerative changes. In most biopsies, the differential diagnosis between atypical regenerative changes and dysplasia of the gastric mucosa can be established on routinely stained histologic sections. In some cases, however, such a distinction is not as easy as might be expected,7 because these conditions can share several architectural and cytologic features. In such cases, the presence of necrosis and inflammation around atypical glands and immunohistochemical studies employing p53 protein are useful for making the distinction. In several studies of gastric carcinogenesis involving both immunohistochemical and molecular genetic techniques, p53 mutations were demonstrated in gastric dysplasia, whereas these mutations were absent in regenerative changes of the gastric epithelium.21 In the present study, p53 protein was detected in the cases of gastric dysplasia, whereas bcl-2 protein was not detected. These results and the atypical architectural and cytologic findings in the gastric glands as a whole support the diagnosis of gastric dysplasia.7 

The presence of intestinal metaplasia, atrophy, and dysplasia of the gastric mucosa in patients with PGL should be recognized by endoscopists and pathologists. Endoscopists should carefully examine the mucosa adjacent to PGL to sample areas suggestive of damage. When biopsies are obtained exclusively from the neoplasia, recognition of other histopathologic changes may raise difficulties because neoplastic lymphocytes frequently replace gastric glands.1 In future years, the number of endoscopic procedures will likely increase due to the therapeutic response associated with H pylori eradication. This response is predominantly noted in patients with early low-grade MALT-type lymphomas.22,23 In these cases, periodic endoscopic studies will be useful not only for monitoring gastric lymphoma, but also for surveillance of epithelial changes of the gastric mucosa.