Application of ROC Curve in Differential Diagnosis of Ovarian Cancer and Benign Ovarian Tumor

Objective To study the difference of Ca125 values between ovarian cancer and benign ovarian tumor by receiver operating characteristic (ROC) curves, and to define the best cutoff point in the differential diagnosis of ovarian cancer and benign ovarian tumor. Methods Ca125 values of 54 ovarian cancer patients, 68 benign ovarian tumor patients and 104 normal individuals were detected by electrochemiluminescence immunoassay. We analyzed the Ca125 values using ROC curve of SPSS11.5, and calculated the sensitivity and specificity relative to different groups and different cutoff values. Results The area under the ROC curve (AUC) was 0.812 in the differential diagnosis of ovarian cancer and benign ovarian tumor using ROC analysis. Cutoff values were set to 60 U/ml, 100 U/ml and 150 U/ml, respectively. The sensitivity values were 0.80, 0.74 and 0.61 respectively. And the specificity values were 0.53, 0.79, and 0.93, respectively. Conclusions The best cutoff point of Ca125 obtained by our lab is 100 U/ml in the differential diagnosis of ovarian cancer and benign ovarian tumor. The diagnosis value of Ca125 is moderate.