Agreement and intraobserver variability in detection of helicobacter pylori infection using routine histology, H. pylori immunohistochemical stain, and warthin-starry special stain on gastric tissue biopsy samples

Document Type

Conference Proceeding

Publication Date

10-28-2021

Publication Title

Am J Clin Pathol

Abstract

Introduction/Objective: Helicobacter pylori (HP) is a prevalent cause of gastritis in the US and worldwide. Accurate detection of the organism is important for appropriate treatment. Diagnosis is made using immunohistochemistry and special stains including Warthin-Starry stain (WS) which is low cost, technically easy to perform on tissue sections, and can be automated. We aimed to assess interobserver variability, reproducibility and validity of hematoxylin and eosin stain (H&E), WS and anti-Helicobacter pylori immunohistochemical staining (HP-IHC) for the histopathological identification and evaluation of HP organisms on gastric mucosa biopsies. Methods/Case Report: We prospectively evaluated gastric biopsies from ten adult patients. These patients underwent upper gastrointestinal endoscopy with subsequent biopsy for various abdominal complaints. Seven of the ten biopsies (7/10) were HP infection positive, and three cases (3/10) were HP negative by H&E stain and HP-IHC (determined by GI Pathologist). The presence or absence and density of HP, were assessed on H&E, HP-IHC and WS in blinded fashion by five general surgical (GS) pathologists. The GS pathologists assessed the cases for presence and density of HP using the semi-quantitative modified Sydney classification (none, mild, moderate, and severe). Percentage agreement and interobserver variablilty using Cohen-Kappas statistics (KS) were calculated. Results (if a Case Study enter NA): The percentage agreement for presence or absence of HP in the biopsies with H&E, HP-IHC and WS stains were 91%, 98% and 99%, respectively. The interobserver agreement for evaluation of presence of HP was excellent with WS staining method (Overall KS = 0.737 95%, Confidence Interval (CI) = 0.501- 0.973) and HP-IHC (Overall KS = 0.783, 95% CI = 0.585-0.980). Agreement for H&E was moderate (Overall KS = 0.532, 95% CI= 0.392-0.672). HP density agreement was excellent using WS and HP-IHC and H&E was in the moderate range. Conclusion: This study found excellent interobserver agreement using IHC and WS. While IHC is the most specific stain, WS is more sensitive in identifying the shape and morphology of HP organisms and is an efficient and low-cost alternative with excellent morphology in general surgical practice.

PubMed ID

Not assigned.

Volume

156

Issue

SUPPL 1

First Page

S61

Find It @ Sladen

Share

COinS