[1157] Stage II Endometrial Carcinoma: Endocervical Gland Spread Is Not Reproducibly Distinguished from Endocervical Stromal Invasion.

Richard Zaino, Junjia Zhu, Catherine Abendroth, Anna Yemelyanova, Esther Oliva, Diana Lim, Ian S Hagemann, Kathleen Montone, Deborah Delair, Robert Soslow. Hershey Medical Center; Johns Hopkins Medical Center, Baltimore; Massachusetts General Hospital, Boston; U of Pennsylvania, Philadelphia; Memorial Sloan Kettering Cancer Center, NY

Background: FIGO stage is among the strongest predictors of survival for women with endometrial adenocarcinoma. For the past 20 years, cervical gland involvement and cervical stromal invasion have defined stage IIA and stage IIB disease respectively. In 2009, FIGO changed the criteria for stage II disease to include only those with cervical stromal invasion. Given the importance of this distinction, we wished to assess the ability of pathologists to distinguish various forms of cervical spread.
Design: A slide of endometrial adenocarcinoma paired with slides from the endocervix of 46 women with cervical spread of tumor were independently examined by 6 pathologists from 5 institutions, most of whom have a subspecialty interest in gynecologic pathology, with a range of 3 to 30 years of faculty experience. The pathologists were instructed to assess 5 features regarding the pattern of cervical spread according to their application of the FIGO rules in their routine practice. Fleiss' kappa statistic was used to examine the reproducibility of these assessments.
Results: The 6 pathologists agreed upon the various pattens of cervical spread statistically more often than by chance alone, but the frequency of disagreement was very high. While agreement on the presence of discontinuous spread was relatively good, agreement on all other patterns of spread varied from poor to fair.

Patterns of cervical spread - interobserver reproducibility

pattern of spread	range of pathologists' identification of feature	Kappa
gland involvement	63-91%	0.15
stromal invasion*	30-78%	0.28
vascular invasion only	0-9%	0.09
contiguous spread	56-80%	0.29
discontinuous spread	13-37%	0.49

*definition of FIGO stage II


Conclusions: The low level of diagnostic reproducibility among pathologists with an interest in gynecologic pathology to distinguish cervical gland spread from cervical stromal invasion of endometrial adenocarcinoma indicates that the current FIGO staging system needs modification. Specifically, depending upon the pathologist who examined the cases, from 30% of the women to 78% of the women would be diagnosed as having stage II endometrial adenocarcinoma. Either better criteria that distinguish stromal invasion from gland spread must be identified or the staging classification that relies on this distinction should be abandoned.
Category: Gynecologic & Obstetrics

Monday, February 28, 2011 1:45 PM

Platform Session: Section C, Monday Afternoon