Endometrial cancer: Can nodal status be predicted with curettage?

Objective. To determine whether histologic or molecular markers assessed in pretreatment curettage specimens predict nodal metastasis in endometrial cancer. Methods. Phenotypic and molecular variables (ploidy, proliferating cell nuclear antigen, MIB-1, p53, HER-2/neu, and bcl-2) were analyzed in preoperative specimens from 82 patients with endometrial cancer who had lymph nodes dissected. These 82 patients had been selected from a total population of 283 patients with endometrial cancer, using a case-cohort design. Weighted logistic regressions were then used to determine significant predictors of positive lymph nodes, and results were estimated for the total population of 283 patients. Results. Of the overall population, 12% of patients were estimated to have positive lymph nodes. Histologic subtype, p53, and bcl-2 each were significantly correlated (P < 0.05) with lymph node status. With application of stepwise logistic regression, p53 was the only independent predictor of lymph node status. In addition, a statistical model predictive of positive lymph nodes was generated which incorporated the risk factors p53, bcl-2, and histologic subtype. Conclusion. In pretreatment curettage specimens, the presence of unfavorable levels of p53 or bcl-2 or of nonendometrioid histologic features, or combinations of those, significantly predicted lymph node status, thus facilitating the preoperative identification of patients at risk of lymph node metastases.