The diagnosis of mucinous lesions in endometrial samplings by gynaecological pathologists: an analysis of diagnostic reproducibility

The purpose of this study is to assess the reproducibility among gynaecological pathologists in their diagnosis of mucinous alterations in endometrial sampling specimens. Twenty-six cases were independently reviewed by four experienced gynaecological pathologists from four academic medical centres. Pathologists were asked to classify each case into one of four groups, including three World Health Organization (WHO)-recognised categories: (1) mucinous metaplasia; (2) atypical mucinous glandular proliferation; (3) carcinoma; and (4) 'other' (absence of a true mucinous alteration and/or an alteration of non-endometrial origin). The overall reproducibility was 'fair' (κ = 0.39). In an analytical scenario that established three clinically significant groups ('benign/non-neoplastic', 'atypical', and 'carcinoma') by redistributing all group 4 responses, the resultant kappa improved to 0.51 (moderate reproducibility). In another analysis with only two categories-'benign/non-neoplastic' versus 'atypical/carcinoma'-reproducibility was similarly moderate (κ = 0.46). However, with one exception, all cases that were ultimately diagnosed as carcinoma in a follow-up hysterectomy specimen, were classified as atypical or carcinoma in the preceding sampling. For 11 cases that were classified as either 'carcinoma' or 'atypical' by all observers, there was moderate reproducibility (κ = 0.53) in making that distinction, and none of a wide array of morphological features were found to significantly distinguish between these two categories. For five cases that all observers classified as either mucinous metaplasia or benign endocervix, reproducibility was substantial (κ = 0.67). In summary, gynaecological pathologists show moderate reproducibility in categorising mucinous alterations in endometrial sampling specimens as benign, atypical, or carcinomatous. They accurately classify as at least 'atypical' those cases that are ultimately diagnosed as carcinoma in the subsequent resection. Our findings suggest that there are indeed some mucinous alterations which have features that do not allow for reproducible assignment by pathologists into the WHO-recognised categories. In this subset of cases, there may be a need for better-defined diagnostic criteria and/or extra-morphological diagnostic tools.