The novel criteria for predicting pelvic lymph node metastasis in endometrioid adenocarcinoma of endometrium

Abstract

Objective: To identify clinicopathological risk factors for pelvic lymph node metastasis, and to evaluate the clinical validity of these factors in selecting patients who need pelvic lymph node dissection.

Methods: The data of 466 patients who had lymphadenectomy for endometrioid adenocarcinoma of the endometrium between January 2002 and December 2010 were reviewed retrospectively.

Results: All patients underwent pelvic lymphadenectomy and 192 (41.2%) patients also underwent paraaortic lymphadenectomy. The median number of pelvic lymph node was 16 (range: 2-46) and of paraaortic lymph node was 5 (range: 2-16). 10.1% (47/466) of all patients had pelvic lymph node involvement and 7.8% (15/192) of the patients had paraaortic lymph node involvement (LNI). Pelvic LNI was significantly more common in the presence of higher grades of tumor, LVSI, deep myometrial invasion, positive peritoneal cytology and cervical involvement. The logistic regression analysis revealed that LVSI, cervical glandular invasion and cervical stromal invasion remained to be the independent risk factors for LNI. When the LVSI and/or cervical involvement were considered as high risk for pelvic lymph node metastasis, NPV and specificity were found to be 96.3% and 68.4%, respectively. LNI was correctly estimated in 323 women (69%), overestimated in 132 women (28%) and underestimated in 11 women (2%).

Conclusion: LVSI, cervical glandular and stromal involvement were independent risk factors for pelvic LNI. These variables can be assessed pre- or intraoperatively with a high rate of accuracy, the model which uses these variables may be successfully used in the prediction of pelvic lymph node metastasis.