EBV-Associated Gastric Carcinoma Displays Type I Viral Latency, Regardless of Host Inflammatory Response Pattern.
Jason R Pettus, Kyoung-Mee Kim, Jun-Hun Cho, Jason L Hornick, Amitabh Srivastava. Dartmouth-Hitchcock Medical Center, Lebanon, NH; Samsung Medical Center, Seoul, Korea; Brigham and Women's Hospital, Boston, MA
Background: EBV can be identified in approximately 10% of gastric carcinomas (GC). The patterns of host inflammatory response in EBV associated GC (EBV-GC) that resemble lymphoepithelioma-like carcinoma (LELC) and Crohn's disease-like lymphoid infiltrate (CLR) are associated with better prognosis than those which resemble conventional adenocarcinoma (CA). EBV associated neoplasms are associated with distinct types of viral latency, classified as type I, II, or III, based on expression patterns of EBV latency proteins. The aim of our study was to determine if the type of EBV latency correlates with the pattern of host inflammatory response in EBV-GC.
Design: A consecutive series of 1080 GC cases at a single large academic hospital in a region endemic for GC was subjected to screening by in-situ hybridization to detect EBER-positive GC. Cases with 100% EBER-positive tumor cells were classified as EBV-GC and were included in this study. Inclusion criteria for this study included pathologic stage 1B through IV (AJCC 2002), curative resection, and complete clinical follow-up data. All tumors were classified histologically into LELC, CLR, or CA based on host inflammatory response pattern using previously published criteria. A tissue microarray (TMA) was constructed using two representative cores from each case. Cores from normal gastric tissue were included as controls in the TMA. TMA slides were immunostained histochemically for EBV latency proteins LMP-1 and EBNA-2 in order to stratify each case into type I, II, or III latency.
Results: 106 EBV-GC (mean age: 56; range: 32-76 yrs; M/F ratio 76:30) formed the final study group. The pattern of inflammatory response in this group was classified as LELC in 39/106 (36.8%), CLR in 27/106 (25.5%), and CA in 40/106 (37.7%). Although EBER positivity was present in a 100% of cases, only 1/106 (0.9%) case was positive for EBNA-2, and this case showed a CLR pattern of inflammation. All cases were negative for LMP-1 expression by immunohistochemistry.
Conclusions: Our findings suggest that EBV-GC is almost universally associated with a type I EBV latency regardless of host inflammatory response pattern. Differences in inflammatory response and prognosis in EBV-GC are, therefore, more likely to be due to genotypic differences in the tumor cells rather than in type of EBV latency, which should be explored in future studies.
Tuesday, March 1, 2011 9:30 AM
Poster Session III # 161, Tuesday Morning