Objective The purpose of this study was to evaluate the predicting model for lymph node metastasis using preoperative tumor grade, transvaginal sonography (TVS), and serum cancer antigen 125 (CA-125) level in patients with endometrial cancer.

Materials and Methods Between January 2000 and February 2013, we identified 172 consecutive patients with surgically staged endometrial cancer. Transvaginal sonography was performed by an expert gynecologic radiologist in all patients. All patients had complete staging surgery including total hysterectomy with bilateral pelvic and para-aortic lymphadenectomy and were staged according to the 2009 International Federation of Gynecology and Obstetrics classification. Various clinicopathologic data were obtained from medical records and were retrospectively analyzed.

Results Of 172 patients, 138 patients presented with stage I (118 IA and 20 IB), 12 had stage II, 18 had stage III (2 IIIA, 1 IIIB, 8 IIIC1, and 7 IIIC2), and 2 had stage IV diseases. Most patients had endometrioid adenocarcinoma (88.4%), and others (12.6%) had nonendometrioid histology. Eighteen patients (10.5%) were found to have lymph node metastasis. Deep myometrial invasion on preoperative TVS (≥50%), high serum CA-125 level (≥ 35 IU/mL), preoperative grade 2 or 3 tumors were significant preoperative factors predicting lymph node metastasis. There was no significant association between preoperative histology and lymph node metastasis. We calculated the simple model predicting lymph node metastasis based on preoperative tumor grade, TVS findings, and CA-125 level using logistic regression analysis. The sensitivity and specificity of this model were 94% and 57%, respectively (area under the curve, 0.84; 95% confidence interval [CI], 0.74–0.93; P < 0.01).

Conclusions Preoperative tumor grade, myometrial invasion on preoperative TVS, and CA-125 can accurately predict lymph node metastasis in patients with endometrial cancer. The current study suggests the possibility that TVS could be positively used for preoperative evaluation strategy in the low-resource countries instead of expensive imaging modalities such as magnetic resonance imaging or positron emission tomography–computed tomography.