RESUMO
PURPOSE: To evaluate the concordance among the available histologic classifications for endometrial adenocarcinoma using interobserver and intraobserver agreement as well as the association of tumor histologic degree in the above mentioned classifications with cellular proliferation measured by Ki-67. METHODS: Seventy women who underwent surgical treatment of endometrial adenocarcinoma with histologic confirmation of endometrioid type were included in the study. Two experienced pathologists randomly analyzed the slides in 3 distinct timeframes with a maximum of 25 slides/timeframe. Tumor slides were classified according to the degree of differentiation using 4 different classifications: International Federation of Gynecology and Obstetrics (FIGO), modified FIGO, Lax, and Alkushi. RESULTS: Intraobserver agreement was reasonable for classification of FIGO (k 0.469 and 0.538), very good for modified FIGO (k 0.661 and 0.768), moderate for Lax classification (k 0.496 and 0.466), and moderate/good for Alkushi classification (k 0.528 and 0.736). Interobserver concordance was regular for FIGO classification (k = 0.271 and 0.343), good/moderate for modified FIGO classification (k = 0.661 and 0.522, respectively), regular/moderate for Lax classification (k = 0.258 and 0.465, respectively), and regular for Alkushi classification (k = 0.283 and 0.402). CONCLUSIONS: The prognostic value of histologic grading in endometrial carcinoma and its importance for a successful therapeutic plan have been documented repeatedly, but the best grading system, in terms of prognostication, reproducibility, ease of use, and universality (e.g., applicability to all tumor cell types), has not been unequivocally defined.