Inter-observer reproducibility of endometrial cytology by the Osaki Study Group method: utilising the Becton...

The purpose of the present study was to evaluate the reproducibility of the cytological diagnosis of endometrial lesions by the Osaki Study Group (OSG) method of new cytological diagnostic criteria using BD SurePath^™ (SP)-liquid-based cytology (LBC).

This cytological classification using the OSG method consists of six categories: (i) normal endometrium (NE), (ii) endometrial glandular and stromal breakdown (EGBD), (iii) atypical endometrial cells, cannot exclude atypical endometrial hyperplasia or more (ATEC-A), (iv) adenocarcinoma including atypical endometrial hyperplasia or malignant tumour (Malignancy), (v) endometrial hyperplasia without atypia (EH) and (vi) atypical endometrial cells of undetermined significance (ATEC-US). For this study, a total 244 endometrial samplings were classified by two academic cytopathologists as follows: 147 NE cases , 36 EGBD cases , 47 Malignant cases, eight ATEC-A cases, two EH cases and four ATEC-US cases. To confirm the reproducibility of the diagnosis and to study the inter- and intra-observer agreement further, a second review round followed at 3-month intervals, which included three additional cytopathologists.

The inter-observer agreement of NE classes improved progressively from 'good to fair' to 'excellent', with values increasing from 0.70 to 0.81. Both EGBD and Malignancy classes improved progressively from 'good to fair' to 'excellent', with values increasing from 0.62-0.63 to 0.84-0.95, respectively. The overall intra-observer agreement between the first and the second rounds was 'good to fair' to 'excellent', with values changing from 0.79 to 0.85. All kappa improvements were significant (P < 0.0001).

In this study, it seemed that the use of the OSG method as the new diagnostic criteria for SP-LBC preparation, may be a valid method to improve the precision (reproducibility) of endometrial cytology.