Abstract 888: Personalized cutpoints of CA125 to improve early detection of ovarian cancer in postmenopausal women

Abstract

Abstract Background: Ovarian cancer is a deadly gynecologic cancer, accounting for approximately 14,000 deaths in the US annually. Lack of early detection modalities leads to advanced stage diagnosis with poorer prognosis. Cancer antigen 125 (CA125) is the most promising single biomarker for ovarian cancer screening. However, two large randomized screening trials comparing the combination of CA125 and transvaginal ultrasound to usual care did not show significant improvement in overall survival. Here, we examined whether using predicted personal CA125 cutpoints based on individual characteristics in postmenopausal women would improve discrimination of epithelial ovarian cancer cases from controls better than using a fixed cutpoint of 35 U/mL. Method: We first determined the effect estimates needed to calculate the predicted 90th and 95th percentile cutpoint of blood CA125 values for each individual using quantile regression model including validated predictors of CA125 (i.e. age, race, body-mass index, smoking status, parity, hysterectomy, age at last menstrual period, and duration of hormone therapy) using data on 28,998 women without ovarian cancer from Prostate Lung Colon and Ovarian Cancer Screening study (PLCO), European Prospective Investigation into Nutrition and Cancer (EPIC), Nurses' Health Studies (NHS/NHSII), and the New England Case Control Study. Then, we examined whether personalized cutpoints performed better than the fixed cutpoint of 35 U/mL in discriminating ovarian cancer cases from controls overall and by time between blood draw to clinical diagnosis using data on 581 incident ovarian cancer cases with CA125 measured up to 24 months prior to clinical diagnosis and 28,226 controls from PLCO, EPIC, and NHS/NHSII. Results: We observed that use of personalized CA125 cutpoints showed improved reclassification of ovarian cancer cases and controls compared to using the fixed cutpoint of 35 U/mL in blood CA125 measured 6 to 12 months prior to diagnosis with net reclassification improvement of 15% using the 90th percentile cutpoint and 12% using the 95th percentile cutpoint. The net reclassification improvement decreased as time between blood draw to clinical diagnosis increased. Conclusion: Our results suggest that redefining the upper limit of normal CA125 values using personalized cutpoints has the potential to modestly improve the performance of CA125 in ovarian cancer screening among average risk postmenopausal women. Citation Format: Naoko Sasamoto, Bernard A. Rosner, Renée T. Fortner, Britton Trabert, Allison F. Vitonis, Stephanie Nartey, Raina N. Fichorova, Linda J. Titus, Daniel W. Cramer, Nicolas Wentzensen, Shelley S. Tworoger, Rudolf Kaaks, Kathryn L. Terry. Personalized cutpoints of CA125 to improve early detection of ovarian cancer in postmenopausal women [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 888.