Sacchi D, Cesaro S, Guzzinati S, Pizzi M, Zorzi M, Fassan M, Guzzardo V Mescoli C, Maddalo G, Salmaso R, Farinati F, Rugge M

8th Triennial Congress of Pathological Anatomy Siapec-iap 2019

Turin, 16-18 October  2019




Autoimmune Gastritis (AiG) is a chronic inflammatory disease mediated by autoantibodies against parietal cell (proton pump and/or intrinsic factor), and sensitized T-cells. In AiG, the inflammatory/atrophic lesions exclusively involve the oxyntic mucosal compartment. The reduced acid-production due to the oxyntic atrophy results in loss of intrinsic factor, vitamin-B12 deficiency and, eventually, into pernicious anemia1. Gastrin hypersecretion due to hypo-chlorhydria promotes the proliferation of oxyntic enterochromaffin-like (ECL) cells, thus creating the biological background where neuroendocrine tumors (NET) may develop. Based on the well-established correlation between gastric atrophy and gastric cancer (GC), the OLGA staging system consistently ranks the GC risk associated with the different score/topography of gastric atrophy. After the initial AiG diagnosis, the America Society for Gastrointestinal Endoscopy recommends a single endoscopic evaluation for neoplastic lesions 2. Among those AiG-patients showing linear and/or micronodular hyperplasia, a 3- to 5-years endoscopy surveillance has been suggested. Of note, however, the available guidelines do not specifically consider AiG patients being, more in general, applicable to any atrophic gastritis with active or eradicated H. pylori infection. Despite improvements on the current knowledge about AiG, no reliable markers are available for identifying patients at higher risk of gastric NET, and the precancerous meaning of AiG is largely debated.

Patients & Method

A consecutive series of 217 AiG patients with histologically-proved, serologically-confirmed AiG, and with no (previous or active) histologically proved H. pylori infection, was prospectively collected. At their enrollment, all patients underwent an initial endoscopy (Padua University Hospital); a follow-up endoscopy with biopsy sampling was always available (interval time ranging from 11 to 204 months; mean/median follow-up time= 84/72 months, respectively). Inclusion criteria were as follows: i) no epithelial neoplastic lesions; ii) availability of a gastric biopsy compliant with Sydney guidelines; iii) at least 2 biopsy sets collected per patient (initial-T0, and last-T1); iv) availability of immunohistochemical profiling, including chromogranin A (clone DAK-A3; DAKO). In total, 434 biopsy sets (as obtained from 217 patients) were considered. In all the available tissue samples, H. pylori status was further tested by genomic DNA extraction and RealTime-PCR for Hp detection; at the (post-recruitment) molecular testing 4/217 (1,8%) were weakly positive. All the original biopsy sets (Hematoxylin-Eosin and Giemsa stains) were histologically reconsidered ([SAD] blinded to patients’ endoscopic/clinical information) and the atrophy score was assessed (OLGA staging tutorial3). ECL-cell hyperplasia (as supported by the appropriate immunostaining) was assessed according to established criteria, distinguishing linear versus micro-nodular hyperplasia. Among micro-nodular ECL-hyperplasia cases, those showing more than 5 ECL aggregates per biopsy sample were defined as “adenomatous”. The strength of association between OLGA stage and the pathological and patients’ demographic features was obtained by applying Student’s t-test and the Wilcoxon nonparametric test for matched pairs, as appropriate. Significance was inferred at a P-value of less than 0.05. The statistical analysis was performed using the STATA 9.4 software (Stata Corporation, College Station, TX, USA).


In all, 434 biopsy sets obtained from 217 patients were considered. At the T0 endoscopy, the M:F ratio was 53:164 (mean age= 55,89 years; range= 15-91,5). Males mean age was significantly higher (M=59,21 years; range=28-80; F= 54,82 years; range=15-83; p=0,04). H. pylori-DNA was identified in 4/217 (1,8%) of cases. The gastritis OLGA stage significantly increased from T0 to T1 (no change= 172 cases [79,2%]; higher stage at T1= 30 [13,8%], lower stage at T1= 15 [6,9%]; p=0,04). The oxyntic atrophy showed non-significant progression from T0 to T1 (no change= 179 cases [82,5%]; higher stage at T1= 24 [11,1%]; lower stage at T1=14 [6,4%]; p= 0,07). Among the 4 H. pylori-positive cases, one progressed from stage II to stage III. At both T0 and T1, epithelial neoplastic lesions were never observed. In 18/217 patients (8,3%) an incidental NET was histologically documented (tumor grade: G1= 17; G2= 1). The mean age of NET-patients (M:F=5:13) was similar to that of non-NET patients (NET patients: mean age 58,1 years; range=31-76 years; non-NET patients: mean age 56,4; range=15-91,5; p=0,47). All NET-patients were H. pylori-negative and coexisted with ECL linear/micronodular hyperplasia and with oxyntic atrophy, involving more than 60% of the oxyntic biopsy specimens (score 3). Adenomatous ECL-hyperplasia was observed in 47/217 cases (21,6%); in particular, adenomatous ECL-hyperplasia was documented in 15/18 (83,3%) NET and 32/199 (14,7%) non-NET patients. Among these adenomatous patients, 45/47 (95,7%) showed score 3 oxyntic atrophy. The presence of adenomatous hyperplasia was an independent risk factor for gastric NET (risk ratio 22,3; 95%CI = 7,9-62,8).


In this large cohort of AiG-patients, the long-term endoscopy and histology follow-up did not documente any gastric epithelial precancerous/cancer lesion. In Western countries, the prevalence/incidence of both AiG and Gastric NETs are consistently increasing 4; moreover, fragmentary (even contradictory) information is available on the NET (AiG-associated) natural history. While AiG-related NETs account about 70-80% of all gastric NET, no reliable markers are available in order to identify patients at higher risk of gastric neuroendocrine tumors. This long-term follow-up study provides evidence that adenomatous ECL-hyperplasia is an independent risk factor for developing NET. Moreover, the prevalence of adenomatous ECL-hyperplasia was significantly higher in advanced corpus atrophy involving more than 60% of the biopsy oxyntic specimens.