PC2164: Concordance Between Pathologists Grading Peak Eosinophil Count and Fibrosis in Eosinophilic Esophagitis

Background: Eosinophilic esophagitis (EoE) is a chronic immune mediated disorder characterized by esophageal inflammation. The condition is diagnosed when patients have symptoms of esophageal dysfunction and at least 15 eosinophils per high power field (eos/hpf) on esophageal biopsy. However, other tissue changes can occur with the disease process such as tissue fibrosis. Statistical concordance in reporting of eos/hpf and fibrosis between pathologists is not well studied.

Methods: Using a previously established data set of 70 pediatric patients with EoE (table 1), concordance between three pathologists in counting of eos/hpf and grading fibrosis. A total of 93 histologic cases from 70 patients with EoE were available, each case had 2 or 3 levels of esophageal biopsies collected from the distal, middle, and proximal esophagus. Each case was evaluated for peak eos/hpf and presence of lamina propria. If lamina propria was present hematoxylin and eosin stains were reviewed and fibrosis was graded on a 0-2 scale. 0 being thin, wispy collagen fibers and an absence of thick collagen bundles. 1 being the presence of few, non-coalescing thicker collagen bundles. 2 being the majority presence of coalescing thick collagen fibers, constituting an overtly fibrotic lamina propria. When possible, trichrome stains were also completed, and fibrosis was grading in the same manner. Concordance between the pathologists was assessed using Fleiss’ kappa with R studio.

Results: In total, 253 readings for peak eosinophil counts were completed by each of the 3 study pathologists. There was substantial agreement between pathologists in peak eos/hpf (k=0.77, p< 0.001). 73 of the 93 histologic cases (78%) had lamina propria identified on review. There was almost perfect agreement between pathologists in identification of lamina propria (k=0.95, p< 0.001). Identification of fibrosis showed higher concordance when trichrome staining was completed. With H&E staining, 117 slides were graded for fibrosis by each of the 3 study pathologists. There was moderate agreement between the pathologists (k=0.48, p< 0.001). In total 78 slides were stained with trichrome and graded for fibrosis by each of the 3 study pathologists, there was substantial agreement between pathologists (k=0.63, p< 0.001) (table 2).

Conclusion: We show that there is statistically substantial concordance between three pathologists in providing counts of eosinophils per high power field which is crucial to the diagnosis of EoE. We also show that when provided with samples including adequate lamina propria, pathologist are able to grade for fibrosis with substantial concordance as well. This is augmented by the use of trichrome staining. Use of consistent trichrome staining in evaluation of EoE may lead to better endotyping of the disease.