Interobserver variations in histopathological assessment of gastric pathology

Considering the fact that histology and its grading systems are accepted gold standards in diagnosis of diverse clinical disorders, assessing the accuracy and reliability of this method of diagnosis is of utmost importance. Thus, this study was performed to measure the agreement values between four independent histopathology readings for identical indices under one scoring guideline using three different approaches.

Four independent pathologists participated in this study and were blinded to the clinical diagnosis of patients. Various histological features were examined on gastric tissue specimens according to the updated Sydney system.

Statistical analysis revealed that our null hypothesis about the existing agreement between different histopathological observations is rejected for chronic gastritis, the presence of inflammatory activity, atrophy and Helicobacter pylori, whereas there were significant inter-observation agreements in regard to the presence of lymphoid follicles, intestinal metaplasia and dysplasia. Pairwise analysis showed that different grading scales resulted in different kappa values ranging from poor to excellent agreements. The best kappa values were observed for the evaluation of dysplasia between two independent pathologists.

This assessment has demonstrated that standardisation of less quantitative grading scales resulting in consistent readings is essential for affirmative clinical diagnosis and devising effective treatment strategies.