Background: The aim of this paper was to assess the accuracy of frozen sections histological examination and preoperative CA-125 to select patients with high risk endometrial cancer.
Methods: We reviewed women with type I endometrial cancer treated from January 2011 through January 2013 at the same university hospital. Preoperative CA-125 and intraoperative frozen sections were analyzed to select patients at high risk for metastases, according to Mayo Clinic algorithm. All patients underwent hysterectomy with bilateral adnexectomy. High risk patients underwent complete surgical staging. Respectively, we compared the accuracy of CA-125, frozen sections, and an algorithm combining Ca-125 plus frozen sections, with permanent sections histology as positive control. χ2 test, Landis and Koch kappa statistics (k), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy were determined for each variable.
Results: One hundred seventy-two women were included. CA-125 levels, using 8.3 U/ml as cut-off value, showed 63.4% sensitivity, 51.6% specificity, 84.7% PPV, 25.0% NPV, 61.1% accuracy, and a low kappa statistics (k=0.106, P<0.125). Frozen sections demonstrated 97.3% sensitivity, 100% specificity, 100% PPV, 90.0% NPV, 97.8% accuracy and an optimal kappa statistics (k=0.934, P<0.001). The algorithm combining CA-125 with frozen sections showed 99.1% sensitivity, 48.1% specificity, 88.8% PPV, 92.9% NPV, 89.2% accuracy, and a satisfactory kappa statistics (k=0.578, P<0.001).
Conclusions: We proved the utility of Mayo algorithm even in a different institution. Combining CA-125 plus frozen sections doesn't look like advantageous compared to frozen sections alone.