[1119] Frozen Section Does Not Reliably Predict the Likelihood of Lymph Node Involvement in Low Risk Endometrial Adenocarcinoma

Bojana Djordjevic, Soufiane El Hallani, Elizabeth D Euscher, Andres A Roma, Elizabeth J Moschiano, Rouba Ali-Fehmi, Elizabeth E Frauenhoffer, Delia P Montiel, Insun Kim, Sung R Hong, Denise Barbuto, Anais Malpica, Elvio G Silva. Multi-Institutional Study, University of Ottawa, The Ottawa Hospital, Ottawa, ON, Canada

Background: The role of lymphadenectomy in surgical treatment of low risk endometrial carcinoma is controversial. Intraoperative frozen section (FS) may be performed to exclude features that confer an increased risk of lymph node (LN) metastases, with the aim of identifying patients in whom lymphadenectomy may not be necessary. The reliability of FS in this setting, however, is not well established.
Design: In a multi-institutional study, 45 FIGO grade 1 and 2 endometrial endometrioid adenocarcinomas with LN involvement as the only site of extrauterine spread, and which were assessed by FS, were identified. Tumor grade, depth of myometrial invasion (MI) and presence of lymphatic vascular invasion (LVI) on FS and permanent section (PS) were recorded. The results were compared to a corresponding set of 137 cases without LN metastases.
Results: The results are summarized in the Table.

The rates of grade 1 and 2 carcinoma were comparable between the two groups. The rates of outer MI and LVI were significantly higher in the LN positive group on FS and PS. However, FS underestimated the number of cases with FIGO grade 2, outer MI and particularly LVI. At time of FS, LVI was either not reported or accurately perceived in a large proportion of cases.
Using a decision tree model, MI and LVI status at FS could discriminate between LN positive and LN negative cases with an overall performance of only 75%.

Conclusions: While outer MI invasion and LVI are more frequent in cases with positive LNs, inability to reliably recognize LVI in a large proportion of cases on FS renders FS a relatively inaccurate method for predicting LN involvement in low risk endometrial adenocarcinoma.
Category: Gynecologic & Obstetrics

Tuesday, March 20, 2012 9:30 AM

Poster Session III # 181, Tuesday Morning


Close Window