Abstract AS10: Ovarian cancer risk factor associations by tumor aggressiveness in the ovarian cancer cohort consortium (OC3)

Abstract

Abstract Introduction: Ovarian cancer is a heterogeneous disease with about half of patients dying within 3 years, but a third surviving at least 10 years, potentially reflecting differences not only for stage, but also in tumor biology. We evaluated if known and putative ovarian cancer risk factors were differentially associated with tumor aggressiveness, defined as rapidly fatal (death within 3 years) and less aggressive (all others). We assessed oral contraceptive (OC) use, parity, tubal ligation, height, body mass index (BMI), family history of breast or ovarian cancer, age at menarche and menopause, and smoking using pooled data from 18 prospective cohort studies. Methods: Ovarian cancer characteristics and diagnosis dates were abstracted from pathology reports or cancer registries; registries were used to ascertain vital status and date of death. We used competing risks Cox proportional hazards models to compute relative risks (RR) and 95% confidence intervals (CI) by tumor aggressiveness, stratified on study, age, and birth year, and adjusted for parity and OC use; p-heterogeneity (p-het) was evaluated by likelihood ratio tests. Results: Among 3,811 cases with known vital status and the potential for at least 3 years of post-diagnosis follow-up, 1,897 (49.8%; median survival=1yr) were rapidly fatal and 1,914 (50.2%; median survival=18yr) were less aggressive. 66% of rapidly fatal and 55% of less aggressive cases were of serous histology. Stronger inverse associations were observed for less aggressive than for rapidly fatal disease for tubal ligation (RR, yes vs. no=0.67 vs. 1.06, respectively; p-het=0.004), parity (RR, per child=0.87 vs. 0.93; p-het=0.001) pack years of smoking (RR, >30 pack-years vs. never=0.78 vs. 1.14; p-het=0.01), and suggestively for family history of breast cancer (RR, yes vs. no=1.15 vs. 0.99, p-het=0.18). Conversely, women with a BMI>30 vs. >22-25 kg/m2 were at higher risk of rapidly fatal disease (RR=1.40), but not less aggressive disease (RR=1.00; p-het=0.04). Associations for other risk factors did not differ by aggressiveness. Discussion: In this prospective analysis, we observed differential associations by tumor aggressiveness for some risk factors. The potentially stronger association of a family history of breast cancer with less aggressive disease is supported by reports of better survival in BRCA mutation carriers. Associations of parity, tubal ligation, and smoking with less aggressive disease may reflect differential associations by histology; additional analyses will account for histology and stage. The BMI association with rapidly fatal disease suggests that metabolic dysfunction may play a role in tumor aggressiveness. Our results support that pre-diagnostic factors may influence ovarian cancer development and aggressivness. Ultimately, understanding a woman's risk profile with respect to risk of rapidly fatal versus less aggressive disease at diagnosis may aid in determining the most optimal treatment strategy for long term survival. Citation Format: Shelley S. Tworoger, Elizabeth M. Poole, Alan A. Arslan, Lesley M. Butler, Victoria Kirsh, James V. Lacey, Jr., I-Min Lee, Alpa V. Patel, Kim Robien, Thomas Rohan, Dale P. Sandler, Leo J. Schouten, V. Wendy Setiawan, Kala Visvanathan, Elisabete Weiderpass, Emily White, Nicolas Wentzensen, for the OC3. Ovarian cancer risk factor associations by tumor aggressiveness in the ovarian cancer cohort consortium (OC3) [abstract]. In: Proceedings of the 10th Biennial Ovarian Cancer Research Symposium; Sep 8-9, 2014; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(16 Suppl):Abstract nr AS10.