Objective The rate of lymph node metastasis is extremely low in patients with low-risk endometrial cancer; lymphadenectomy may be unnecessary for these patients under an accurate preoperative diagnosis. The aim of this study was to evaluate the diagnostic accuracy of myometrial invasion (MI) on preoperative magnetic resonance imaging (MRI) and intraoperative frozen sections (FSs).

Materials and Methods Endometrial cancer was diagnosed in a total of 378 patients by preoperative endometrial curettage, preoperative magnetic resonance imaging MRI, and intraoperative FSs; the 378 patients underwent hysterectomy. The depth of MI was evaluated between the preoperative MRI, intraoperative FSs, and final paraffin sections (PSs). The histologic grade was also evaluated between preoperative endometrial curettage, intraoperative FSs, and final PSs.

Results The sensitivity, specificity, positive predictive value, and negative predictive value for deep MI (≥50%) on MRI were 57.8%, 92.0%, 69.3%, and 87.5%, respectively, with a kappa value of 0.53. These figures on FSs were 66.7%, 97.9%, 90.9%, and 90.4%, with a kappa value of 0.71. When grade 3 endometrioid adenocarcinoma, serous carcinoma, clear cell carcinoma, and carcinosarcoma were considered high-grade tumors, the grade evaluation at the time of FSs was 70.2%, 99.0%, 96.1%, and 89.7%, with a kappa value of 0.75. In the patients with low-grade tumors, including grade 1 or 2 endometrioid adenocarcinoma on preoperative endometrial curettage, the rate of unexpected lymph node metastasis did not differ significantly between the patients who had a diagnosis of MI and lymph node metastasis by MRI and those with diagnosis of MI and histological grade by FSs (4.0% vs 2.6%; P > 0.05).

Conclusions Frozen sections had a higher agreement rate for MI than MRI; however, MRI is still considered an acceptable modality to guide preoperative decisions regarding lymphadenectomy especially in grade 1 or 2 endometrioid adenocarcinoma.