Objective: To evaluate the sensitivity and specificity of endometrial cytology obtained by intrauterine sample using a descriptive reporting format for endometrial cytological diagnosis.
Study design: 10,152 consecutive endometrial scrapings obtained in 13 different Japanese hospitals were analyzed. Cytological results were classified as 'negative for malignancy', 'atypical endometrial cells' (ATEC), 'endometrial hyperplasia', 'atypical endometrial hyperplasia' or 'malignant tumor'. ATEC was subclassified as 'ATEC, of undetermined significance' (ATEC-US) and 'ATEC, cannot exclude atypical endometrial hyperplasia or more' (ATEC-A). Cytological results were compared with the histological diagnosis as a gold standard. When the cytological result was 'negative for malignancy' and there was no subsequent histological examination, the case was considered a true negative when the endometrium was assessed as normal on transvaginal ultrasonography and there was no abnormal uterine bleeding.
Results: 1,083 cases in which histology was not performed, 557 cases of 'unsatisfactory specimen' and 76 cases of ATEC-US were excluded. In the remaining 8,436 cases, the sensitivity and specificity, positive predictive value and negative predictive value for detecting atypical endometrial hyperplasia or malignant tumors were 79.0 and 99.7, 92.9 and 98.9%, respectively.
Conclusion: The current diagnostic standards for endometrial cytology in Japan were established. Specificity is satisfactory for excluding cancer or precancerous diseases.
Copyright © 2012 S. Karger AG, Basel.