Newly developed and validated eosinophilic esophagitis histology scoring system and evidence that it outperforms peak eosinophil count for disease diagnosis and monitoring.

Abstract

Eosinophilic esophagitis (EoE) is diagnosed by symptoms, and at least 15 intraepithelial eosinophils per high power field in an esophageal biopsy. Other pathologic features have not been emphasized. We developed a histology scoring system for esophageal biopsies that evaluates eight features: eosinophil density, basal zone hyperplasia, eosinophil abscesses, eosinophil surface layering, dilated intercellular spaces (DIS), surface epithelial alteration, dyskeratotic epithelial cells, and lamina propria fibrosis. Severity (grade) and extent (stage) of abnormalities were scored using a 4-point scale (0 normal; 3 maximum change). Reliability was demonstrated by strong to moderate agreement among three pathologists who scored biopsies independently (P ≤ 0.008). Several features were often abnormal in 201 biopsies (101 distal, 100 proximal) from 104 subjects (34 untreated, 167 treated). Median grade and stage scores were significantly higher in untreated compared with treated subjects (P ≤ 0.0062). Grade scores for features independent of eosinophil counts were significantly higher in biopsies from untreated compared with treated subjects (basal zone hyperplasia P ≤ 0.024 and DIS P ≤ 0.005), and were strongly correlated (R-square >0.67). Principal components analysis identified three principal components that explained 78.2% of the variation in the features. In logistic regression models, two principal components more closely associated with treatment status than log distal peak eosinophil count (PEC) (R-square 17, area under the curve (AUC) 77.8 vs. R-square 9, AUC 69.8). In summary, the EoE histology scoring system provides a method to objectively assess histologic changes in the esophagus beyond eosinophil number. Importantly, it discriminates treated from untreated patients, uses features commonly found in such biopsies, and is utilizable by pathologists after minimal training. These data provide rationales and a method to evaluate esophageal biopsies for features in addition to PEC.