Gastric cancer in autoimmune gastritis: a molecular and histopathological study

Background: Gastric cancer (GC) is a the sixth most common cancer and the fourth cause of cancer related death worldwide. Most cases of GC originate in the preconditioned milieu of chronic gastritis and gastric atrophy, with H.pylori being the main risk factor for this tumor. Autoimmune gastritis (AIG) is a chronic, immune mediated disease that progressively leads to atrophy of the corpus/fundus mucosa of the stomach, which is responsible for the acidification of gastric content. Deriving hypochlorhydria leads to increased levels of gastrin that fuel ECL-cells hyperplasia and eventually to the development of type 1 neuroendocrine tumors (NETs). These tumors are mainly indolent and rarely metastasize. Whereas the increased risk for type 1 NETs is well-documented, no excess risk for GC seems to be associated to AIG. Oxyntic-restricted atrophy induced by AIG has specific characteristics, regarding both phenotype and distribution, that differentiates it from H. pylori-induced gastritis. Aim: The aim of this study was to describe molecular and histopathological features of the largest series of AIG-related GC, also in relation with predictive biomarkers that could guide therapeutical choices for these patients. Materials and methods: twenty-six biopsy specimens and 19 surgical specimens) tumors were evaluated in this retrospective multicentered study. All patients had histological and serological findings consistent with AIG. H. pylori infection was excluded through histological evaluation. All twenty-six samples were evaluated for HER2, PD-L1, CLDN18, MUC1, MUC2, MUC5AC, MUC6, CDX2, MMR proteins (MLH1, PMS2, MSH2, MSH6) and p53 IHC. EBV in situ hybridation was performed on all lesions. Next-generation sequencing analysis of 523 cancer-related genes was performed on seventeen cases, assessing all variant types including microsatellite instability (MSI) and Tumour Mutational Burden (TMB) (TrueSight Oncology 500, Illumina). Results: Fourteen patients (53.8%) were female, twelve were male (43.2%). Median age at diagnosis was 74,5 years. Most tumors (57.7%) were located in the corpus/fundus area. GC histotypes were distributed as follow: two (7.7%) adenocarcinomas within pyloric adenomas, sixteen (61.5%) conventional adenocarcinomas, two (7.7%) poorly differentiated adenocarcinomas with neuroendocrine differentiation, two (7.7%) adenocarcinomas with lymphoid stroma, three (11.5%) MiNeNs and one (3.8%) amphicrine carcinoma. Seven cases (26.9%) showed loss of MMR protein complex proteins, three (11.5%) were strongly positive (3+) for HER2, sixteen (61.5%) had PD-L1 CPS ≥ 5, six (23.1%) were CLDN18 positive, seven had aberrant expression of p53 (26.9%) and none was positive for EBER-ISH. Molecular analysis revealed five (29.4%) MSI cases, seven (41.1%) had high TMB. Most prevalent mutations were: TP53 (7/17, 41.1%), RNF43 (7/17, 41.1%), ERBB2 (5 mutations and 2 amplification cases), ARID1A (6/17, 35.3%), and PIK3CA (4/17, 23,5%). Conclusions: this study represents the largest series of GC in AIG analyzed in literature. Despite the fact that GC risk is not increased in these patients, tumors originating in the context of AIG can display a peculiar phenotype, especially in the neuroendocrine spectrum. Predictive biomarkers analysis revealed a high prevalence of microsatellite instability and a low frequency of CLDN18 positivity. Further studies should be performed to better understand the nature of this entity and confirm these findings on larger cohorts of patients.