Back to Authors

Garg, Shipra

0000-0002-4967-9056

Mayo Clinic

Department of Pathology and Laboratory Medicine, Phoenix Children's Hospital, Phoenix, Arizona.

Citations 1

Diagnosis of Pediatric Non-Esophageal Eosinophilic Gastrointestinal Disorders by Eosinophil Peroxidase Immunohistochemistry
Hasan, Shaina, Garg, Shipra , Doyle, Alfred D. , Bauer, Cindy S. , Wright, Benjamin L. , Schroeder, Shauna , Hasan, Shaina H, Garg, Shipra , Wright, Benjamin L , Taylor, Steve, Buras, Matthew

Pediatric and Developmental Pathology | June 28, 2021 | Volume: 24 | Issue: 6 | Pages: 513-522

10.1177/10935266211024552 Get It

Background Diagnosis of non-esophageal eosinophilic gastrointestinal disorders requires quantification of tissue eosinophils. Our objective was to evaluate eosinophil peroxidase (EPX) immunohistochemistry (IHC) as a method for histologic diagnosis of eosinophilic gastritis (EG) and eosinophilic duodenitis (EoD). Methods We performed a retrospective analysis of biopsies from pediatric EG/EoD cases and controls. Subjects with EG or EoD had ≥30 eosinophils per high power field (eos/hpf) in ≥5 hpf in the stomach and/or ≥3 hpf in the duodenum, respectively. Controls had no histopathologic diagnosis recorded. Tissue eosinophil counts were assessed by hematoxylin & eosin stains. EPX stains were assessed using a unique histopathologic scoring system. Slides were digitized and EPX+ staining area/mm 2 was quantified by image analysis. Results Twenty-six EG/EoD cases and 40 controls were analyzed. EPX scores and EPX/mm 2 levels were markedly elevated in EG/EoD (p ≤ 0.0001). Eosinophil density (eos/mm 2 ) correlated strongly with EPX scores and EPX/mm 2 levels in the stomach (r ≥ 0.77) and moderately with EPX scores and EPX/mm 2 levels in the duodenum (r ≥ 0.52); (p < 0.0001). EPX quantification identified EG/EoD subjects with high diagnostic accuracy (EPX score: AUC = 1 for EG and EoD; EPX/mm 2 : AUC = 0.98 (95%CI 0.96-1) for EG, AUC = 0.91 (95%CI 0.81-1) for EoD). Conclusion EPX-based assessment of eosinophilic inflammation may facilitate automated histologic diagnosis.