Estudo inter e intra observadores da reprodutibilidade do diagnóstico histológico da gastrite crônica, de acordo com o sistema Sydney atualizado

Abstract

Objectives: To investigate the reproducibility of chronic gastritis histological diagnosis according to the updated Sydney system, in the working conditions of our environments pathology laboratories. Methods: We studied 168 dyspeptic patients from whom were obtained gastric endoscopic biopsies from the five locations recommended by the protocol above.The slides were stained by hematoxilin-eosin (HE), then were blindly and independently examined by three pathologists. They diagnosed and graduated the presence of infection by H. pylori, chronic inflammation, inflammatory activity, intestinal metaplasia and glandular atrophy. The subclassification proposed for chronic gastritis was also compared. The diagnoses agreements were investigated in intra and interobserver studies through kappa statistical method. In a second moment, the performance of the diagnosis of H. pylori infection and intestinal metaplasia was studied in comparison with the most used reference pattern.The applied methodology repeats the working conditions of our environments pathology laboratories. Results: In both intra and interobserver studies, in both pairwised studies and considering the set of observers, in both dichotomous analysis and comparison of graduation, a bad agreement for the histological diagnosis of the infection with H. pylori was observed, as well as inadequate indicators of sensitivity, specificity, predictive values and odds ratios. The agreement studies showed good levels of reproducibility in relation to the topics chronic inflammation, inflammatory activity, glandular atrophy and intestinal metaplasia, the last one being the one with the highest level of agreement between the observers. The search for intestinal metaplasia in the histological exam by HE showed, however, low diagnostic sensitivity. Both intra and interobserver agreement studies in relation to the subclassification of chronic gastritis showed disappointing results. Conclusions: It has been proved that in this study conditions, the histological exam of specimens obtained by endoscopic gastric biopsy constitutes a good diagnosis method to the several topics proposed by the updated Sydney system, showing good reproducibility among the observers, except for the research of H. pylori infection, suggesting the routine adoption of the reference pattern (carbolfucsin) to increase the reproducibility and the quality of the exam in this issue. The diagnosis of intestinal metaplasia is reliable when positive, failing the method, however, in the detection of a large number of cases (false negatives), suggesting the adoption of staining by the respective reference pattern (AB/PAS) in specific situations. The subclassification of chronic gastritis showed low reproducibility among the observers, either by the observerss unfamiliarity to the proposed nomenclature or difficulties inherent to the method.