Grading of epithelial ovarian carcinomas has always been somewhat problematic, with different systems proposed for different histopathologic types [1]. Clear cell adenocarcinoma (CCC) has been the most problematic of all cell types, with almost all studies in the older literature failing to demonstrate a correlation between any grading system and survival (either disease-free or overall). This may have been, at least in part, because most published studies prior to the 1990s consisted of no more than 30 or 40 cases.

In 1998, Shimizu and colleagues [2, 3] proposed a “universal grading system” for ovarian epithelial carcinoma, which was meant to be applicable to all histopathologic types. In their own series, which consisted of over 400 cases overall, the system correlated well with survival for all cases, for early- and advanced-stage cases considered separately, and for all major cell types with the exception of CCC. The separation of survival of Grade 1, 2, and 3 cases according to the new system was statistically significant for all other cell types studied but only “suggestive” for CCC. Subsequent reports from other investigators have confirmed the applicability of this grading system to ovarian epithelial cancer in general but with variable results (usually similar to those of the Shimizu group) for CCC [4–6].

The most likely reason for the poorer results with CCC lies within the mechanics of the grading system itself and the usual histologic appearances of CCC. The system is based on assigning numerical scores for each of three microscopic features and then adding the scores for an individual tumor to determine the final grade. [1] Scores of 1–3 are assigned to each of the following features: (1) predominant architectural pattern (glandular/tubulocystic = 1, papillary = 2, solid = 3), (2) highest level of nuclear atypia (mild = 1, moderate = 2, severe = 3; these levels of atypia being defined more specifically in the Shimizu/Silverberg reports [1–3]), and (3) mitotic activity (0–9 mitotic figures counted per 10 high-power fields = 1, 10–25 MF/10 HPF = 2, 26 or more MF/10HPF = 3). When the three scores are added, a final tally of 3–5 is Grade 1, 6 or 7 is Grade 2, and 8 or 9 is Grade 3.

The problem with applying this system to CCC lies in the fact that CCC is always paucimitotic, so the score for mitotic activity will be 1 in the great majority of cases. Atypia is also not very variable, with most CCCs scoring 2 or 3 for nuclear atypia. Thus, the variability within a pool of CCC cases being graded is less than for most other epithelial ovarian carcinomas. We did not perform grading on all of the cases in the CCC series which form the basis for this Atlas, but at one point, we did grade a subset of these cases by the Shimizu system. Of 56 cases so graded, presumably representative of the entire series, 33 were Grade 1, 23 Grade 2, and none Grade 3. By individual scores, the mitotic score was 1 (fewer than 9 MF/10 HPF) in 52 of the 56 cases, with 3 cases scoring 2 and only one scoring 3 (and probably either the mitotic count or the diagnose of CCC is suspect in this case). Both nuclear atypia and architectural pattern peaked at scores of 2 (35 cases with this score for each of these determinations). A recent study of 129 graded CCC cases from Korea [7] reported similar findings but with a more uniform spread of architectural scores (38, 37, and 25 % scores 1, 2, and 3, respectively) and more skewing of nuclear atypia scores toward the severe level (57 % scoring 3 in this determination). As in our series, the majority of tumors (72.1 %) had a mitotic score of 1, with only 5.4 % scored as 3. Because of the higher estimation of nuclear atypia in this series, 52.7 % of cases received a universal grade of 2 or 3 (compared to 75.2 % FIGO Grade 2 or 3).