Diagnostic reproducibility determination of endometrial hyperplasia as well as well differentiated adenocarcinoma and comparison of morphologic with computerized morphometric data

Abstract

Introduction: Endometrial hyperplasia is common disease with difficulty in diagnosis and management in over years. According to WHOM classification low to moderate diagnostic reproducibility reported in different studies with need for new classification system. Presented new strategy based on objectively measured parameters such as VPS, SDSNA and OUTSD and calculated D-score. Morphometric analysis with higher accuracy and reproducibility will divided endometrial hyperplasia to EH and EIN subgroups.

Methods and materials: Archival paraffin-embedded biopsy and curettage specimens (n=100) of both endometrial hyperplasia and well differentiated Adenocarcinoma selected. Then inter & intra observer agreement assayed. 55 of 100 candidate specimens yielded suitable regions for morphometric assay. Analysis performed on delineated H&E stained sections. D-score was calculated, incorporating Volume percentage stroma (VPS), standard deviation of shortest nuclear axis (SDSNA) and gland outer surface density (OUTSD).

Results: Pathologists morphologic data showing overall moderate inter observer reproducibility (p<0.0001, K=0.5372) and excellent intra observer reproducibility (p<0.0000, K=o.8690) as calculated with statistical software and weighted kappa test. Higher diagnosis reproducibility were seen in SH & WDA subgroups and lower ones in CH & AH. Case by case comparison of computerized D-score and VPS with pathologists diagnosis shows a strong concordances with pathologist data no: 1 (p<0.05) with higher experience in gynecopathology. Discordances were seen between histopathology assessment of pathologists No: 2 & 3, (p>0.05) and D-score as calculated with post Hoc test and Sheffe exam.

Conclusion: Inter observer reproducibility is not sufficient  for treatment protocol and an alternative strategy such as morphometric analysis with higher accuracy and reproducibility shows more benefit in this regard especially in lower experience pathologists.