Validation of the histologic grading for ovarian clear cell adenocarcinoma: a retrospective multi-institutional study...

Pathologic slides from 150 patients with clear cell adenocarcinoma from the collaborating institutions were reviewed independently by 2 pathologists, and each tumor was graded histologically using the Shimizu-Silverberg and International Federation of Gynecology and Obstetrics (FIGO) grading systems. For the Shimizu-Silverberg grading system, 3 parameters-architectural pattern, nuclear pleomorphism, and mitotic activity-were assessed and scored as 1 to 3. When the summed scores of these parameters were 3 to 5, 6 to 7, and 8 to 9, grades 1, 2, and 3 were assigned, respectively. The FIGO grade was based on the ratio of glandular/papillary growth versus solid growth: grade 1, less than 5% solid tumor; grade 2, 5% to 50% solid tumor; grade 3, greater than 50% solid tumor. Interobserver agreement levels for assignment of both gradings were fair (κ=0.32 and 0.24, respectively). After consensus had been acquired, 82 (55%), 56 (37%), and 12 (8%) tumors were classified as grades 1, 2, and 3 by the Shimizu-Silverberg grading system, and 88 (59%), 38 (25%), and 24 (16%) were classified as grades 1, 2, and 3 by the FIGO grading system, respectively. Survival analyses indicated that patients with grade 3 tumors, as defined by both the grading systems, tended to have a poor outcome, but any differences between them were not statistically significant. Multivariate analysis showed that only the presence of residual tumor after initial surgery was an independent prognostic factor for overall survival. These results suggest that the 2 tested grading systems have limited value for the prognostication of patients with clear cell adenocarcinoma, and that a more effective grading system for this tumor may be required.