鈴木 加余子

OBJECTIVE: Eosinophilic gastrointestinal disorders (EGIDs) are pathologically diagnosed by manually counting the eosinophils in biopsy tissue under microscopy. However, the skill of the individual examiner is considered to influence the accuracy of the resulting eosinophil count (EC). This study aimed to examine the effects of different examiners and histopathological staining types on the EC results of pathological tissues from patients with EGIDs. METHODS: Infiltrating eosinophils in lesion tissues from 10 eosinophilic esophagitis and 28 eosinophilic gastroenteritis cases were counted by three pathologists and one cytotechnologist. The intra- and inter-observer variabilities in ECs related to hematoxylin-eosin (HE), May-Grünwald Giemsa (MG), and direct fast scarlet (DFS) staining were investigated. The effects of examiner expertise and staining method on ECs were analyzed using a linear mixed effects model. The difference in color value (ΔE) for each staining method was obtained using the Commission International de l'Eclairage luminance-a-b model (L*a*b*). RESULTS: There was no significant intra-observer variability in eosinophil counting. Regarding inter-observer agreement, the examiner with the most EGIDs experience reported higher ECs than the other examiners for all three staining types (P<0.001). ECs were significantly higher with MG and DFS staining than with HE staining, regardless of the examiner (both P<0.001). Additionally, the ΔE values with DFS were higher than those with MG and HE. CONCLUSIONS: DFS staining offered the most selective visualization of eosinophils. ECs may vary depending on both the skill of the examiner and the staining method.