Reproducibility of histopathologic diagnosis of precursor lesions of gastric carcinoma in three Latin American countries

• Published in: Salud pública de México Vol. 52; no. 5; pp. 386 - 390
• Main Authors: Kasamatsu, Elena, Bravo, Luis Eduardo, Bravo, Juan Carlos, Aguirre-García, Jesús, Flores-Luna, Lourdes, Nunes-Velloso, María del Carmen, Hernández-Suárez, Gustavo
• Format: Journal Article
• Language: Spanish
• Published: Mexico 01.09.2010
• Subjects: Adult; Biopsy; Carcinoma - epidemiology; Carcinoma - prevention & control; Clinical Competence; Colombia - epidemiology; Consensus; Gastritis, Atrophic - diagnosis; Gastritis, Atrophic - pathology; Gastroscopy; Humans; Hyperplasia; Intestines - pathology; Metaplasia; Mexico - epidemiology; Observer Variation; Paraguay - epidemiology; Pathology, Clinical; Precancerous Conditions - diagnosis; Precancerous Conditions - pathology; Reproducibility of Results; Stomach - pathology; Stomach Neoplasms - epidemiology; Stomach Neoplasms - prevention & control; Paraguay; Colombia; Mexico
• ISSN: 1606-7916
• DOI: 10.1590/S0036-36342010000500005

The aim was to evaluate the concordance in the diagnosis of precursor lesions of intestinal-type gastric carcinoma among observers with different levels of experience. Gastric biopsies from 1 056 cases were studied: 341 from Colombia, 382 from Mexico, and 333 from Paraguay. Pathologists without experience (A) and with experience (B) in gastrointestinal pathology, as well as experts working in an international reference center (C) participated in the diagnosis of each case. The concordance (k) between pathologists with experience and those without was poor for the diagnosis of atrophic gastritis (k=0.04 to 0.12) and dysplasia (k=0.11 to 0.05), and good for the diagnosis of intestinal metaplasia (k=0.52 to 0.58). Supervision of pathologists without experience by those with experience remarkably improved the concordance in the diagnosis of atrophic gastritis (k=0.65) and intestinal metaplasia (k=0.91), and to a lesser degree, of dysplasia (k=0.28). The concordance among experts before and after the consensus meeting showed no variation in the diagnosis of atrophic gastritis (k=0.57); the concordance varied from good to excellent in the diagnosis of intestinal metaplasia (k=0.67 to 0.81) and from poor to good in that of dysplasia (k=0.18 to 0.66). The greatest differences arose in the diagnosis of chronic atrophic gastritis and dysplasia. The interobserver concordance depended on the experience of the observer and the consensus reading.