Can preoperative magnetic resonance imaging replace intraoperative frozen sectioning in the evaluation of myometrial invasion for early-stage endometrial carcinoma?

To examine the performance of preoperative magnetic resonance imaging (MRI) and intraoperative frozen sectioning in the assessment of myometrial invasion during the early stages of endometrial cancer. This retrospective study employed data from patients with endometrial cancer who were operated on between January 2013 and November 2018. Patients who underwent preoperative MRI and were of FIGO 2009 stage I were included in the study. Radiological staging and intraoperative staging by frozen sectioning were carried out. The data were analyzed to assess agreement of the overall results concerning myometrial invasion. In total, 222 patients were enrolled. Their mean age was 58.3 ± 8.5 years. The accuracy of MRI for the detection of myometrial invasion was 88.7% and its sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 91.6%, 82.1%, 92.2%, and 80.9%, respectively, with a kappa coefficient of 0.734 (95% confidence interval [CI], 0.684-0.784; p < 0.001). The accuracy of intraoperative frozen sectioning was 94.4%, and its sensitivity, specificity, PPV, and NPV were 97.7%, 85.7%, 94.7%, and 93.4%, respectively, with a kappa coefficient of 0.856 (95% CI, 0.812-0.900; p < 0.001). No significant difference in accuracy was observed between MRI and frozen sectioning (p = 0.057). MRI and frozen sectioning were sensitive for the detection of myometrial invasion, according to receiver operating curve analyses (areas under the curve, 0.869 and 0.917, respectively; p < 0.001). The assessment of myometrial invasion by preoperative MRI and intraoperative frozen sectioning during the early stages of endometrial carcinoma was highly accurate.