This study was a retrospective study conducted at the single tertiary center and involved lymphomas confirmed by endoscopic biopsy from June 2008 to February 2017. Forty-eight patients were enrolled in this study. The enrolled patients were divided into a primary gastric lymphoma group (primary group, n = 17) and a gastric involvement group (secondary group, n = 31) based on the presence of only gastric lesion (Fig. 1). Patients’ clinical characteristics, endoscopic findings and pathologic data were retrospectively reviewed using electronic medical records. The study was approved by our Institutional Review Board. All procedures were in accordance with the ethical standards of the Declaration of Helsinki. (IRB No.: KUGH 2019-10-021)

Primary gastric lymphoma was defined by the gastric location of the main lesion, with or without perigastric and/or abdominal lymph node involvement. 6 In the present study, the primary gastric lymphoma was defined based on the following Dawson’s criteria: 1) absence of peripheral lymphadenopathy at the time of presentation, 2) lack of enlarged mediastinal lymph nodes, 3) normal total and differential white blood cell counts, 4) predominance of gastric lesions at the time of laparotomy with only lymph nodes apparently affected in the immediate vicinity, 5) no lymphomatous involvement of liver and spleen.7 On the other hand, cases of secondary lymphoma were defined by distant lymph node involvement or involvement of other organs, deviating from the definition of primary lymphoma.

At time of initial diagnosis of gastric lymphoma, biopsy specimens from stomach, lymph node, or bone marrow were processed routinely with preparation of hematoxylin-eosin–stained sections. If initial hematoxylin-eosin stained slide showed suspicious lymphoma, fixed, paraffin-embedded tissue specimens were assessed immunohistochemically by pathologists using antibodies specific for CD3, CD5, CD10, CD20, cyclin D1, bcl-2, bcl-6, Cytokeratin cocktail, EMA and MUM-1. Some specimens also assessed for B-cell monoclonality using Immunoglobulin Heavy Chain (Ig H) gene rearrangement.

Endoscopic findings listed the most frequently mentioned items in the literature previously associated with gastric lymphoma. The most redundant and frequently mentioned findings were classified into 8 groups as follows: nodular, polypoid, ulcerofungating, ulceroinfiltrative, erosive, diffuse infiltrating, thickened fold-like, and mixed types.4,8–10 Representative cases under each classification are illustrated in Fig. 2.

The categorical variables were compared using Pearson’s chi-square, Fischer’s Exact test or linear by linear association, whereas the continuous variables were compared using Student’s t-test where appropriate. Statistical analysis was performed using IBM SPSS Statistics version 24.0 (IBM Co., Armonk, NY, USA) with a statistical significance level set at P < 0.05.

The mean age of the patients was 63.3 ± 13.1 years and 29 patients were female (60.4%). There was no clinical difference between patients with primary and secondary lymphomas, except for the involvement of the fundus (P = 0.035), when the invasion site of the gastric lesion was divided into the antrum, body, and fundus, respectively (Table 1). Only 41.7% (20/48) cases considered lymphoma in the first endoscopic diagnosis, including both primary and secondary lymphoma.

DLBLC was the most frequently diagnosed pathology (81.3%) based on all the histological findings in this study, similarly to previous studies. Approximately 56% of the invasion depth in the stomach was confined to the mucosa and submucosal layer, and there was no difference in primary and secondary lesions (Table 2). Involvement of lymph node in the upper gastrointestinal tract was significantly higher in secondary than in 29.4% of primary and 90.3% of secondary cases. Bone marrow test was performed in about 90% of cases, whereas the invasion was more frequent in secondary cases than in primary lymphoma. In 43% of cases tested for Helicobacter pylori, no difference in the pathology of primary and secondary lymphomas was detected.

No statistically significant differences were detected in the eight different types of primary and secondary gastric lymphoma based on endoscopic findings. However, the ulceroinfiltrative type, similar to a Borrmann type 3 stomach cancer was the most common, followed by the mixed type in both groups (Table 3).

The three representative cases included in this study are as follows. First, both cases were classified as polyoid lesions, with histological findings of diffuse large B-cell lymphoma. However, a case was classified as secondary lymphoma with ileal lymph node involvement. Both cases were difficult to differentiate because of similar endoscopic features (Fig. 3). Second, both cases showed ulcercofiltrative lesion, although primary gastric lymphoma was confirmed as mantle-cell lymphoma and secondary lymphoma showed diffuse large B-cell lymphoma histology. In secondary gastric lymphoma, positron emission tomography-computed tomography (PET-CT) showed invasion of multiple whole bones, bone marrow, left orbit, liver, both adrenal glands, and prostate as well as right subclavian lymph node (Fig. 4). In the last example, two cases were classified under a mixed type endoscopically. All cases were confirmed as diffuse large B-cell lymphoma. In the latter case, lymph node involvement of the whole body including cervical, left mediastinal, both axillary area, aortocaval and both inguinal areas was confirmed in PET-CT (Fig. 5).