Objective: To assess the accuracy of serum CA125 at the level of more than 35 U/mL in predicting ovarian cancer, using histopathology as a gold standard.
Material and Method: Blood samples were obtained from 120 women with ovarian masses scheduled for elective surgery at Siriraj Hospital between October 1, 2003 and August 31, 2004 and sent for the assay of serum CA125 levels.
Results: Of the 120 women enrolled, ovarian cancer was found in 59 cases (49.2%) and benign ovarian mass in 61 cases (50.8%). The sensitivity, specificity, and accuracy of serum CA125 at the cutoff level of 35 U/mL for prediction of ovarian cancer were 83.1%, 39.3%, and 60.8%, respectively; with 57.0% positive predictive value, 70.6% negative predictive value, 60.7% false positive rate, and 16.9% false negative rate.
Conclusion: As stand-alone modality, serum CA125 of more than 35 U/mL in predicting ovarian cancer revealed modest diagnostic accuracy. There is a need to be careful for false positive in women at reproductive age group and false negative results in early-stage disease or ovarian cancer with low level of serum CA125.

Keywords: Serum CA125, Ovarian mass, Ovarian cancer