Ovarian cancer risk factors by tumor aggressiveness: An analysis from the Ovarian Cancer Cohort Consortium
Renée T. Fortner,Elizabeth M. Poole,Nicolas A. Wentzensen,Britton Trabert,Emily White,Alan A. Arslan,Alpa V. Patel,V. Wendy Setiawan,Kala Visvanathan,Elisabete Weiderpass,Hans‐Olov Adami,Amanda Black,Leslie Bernstein,Louise A. Brinton,Julie Buring,Tess V. Clendenen,Agnès Fournier,Gary Fraser,Susan M. Gapstur,Mia M. Gaudet,Graham G. Giles,Inger T. Gram,Patricia Hartge,Judith Hoffman‐Bolton,Annika Idahl,Rudolf Kaaks,Victoria A. Kirsh,Synnove Knutsen,Woon‐Puay Koh,James V. Lacey,I‐Min Lee,Eva Lundin,Melissa A. Merritt,Roger L. Milne,N. Charlotte Onland‐Moret,Ulrike Peters,Jenny N. Poynter,Sabina Rinaldi,Kim Robien,Thomas Rohan,Maria‐José Sánchez,Catherine Schairer,Leo J. Schouten,Anne Tjonneland,Mary K. Townsend,Ruth C. Travis,Antonia Trichopoulou,Piet A. Brandt,Paolo Vineis,Lynne Wilkens,Alicja Wolk,Hannah P. Yang,Anne Zeleniuch‐Jacquotte,Shelley S. Tworoger
DOI: https://doi.org/10.1002/ijc.32075
2019-01-14
International Journal of Cancer
Abstract:Ovarian cancer risk factors differ by histotype; however, within subtype, there is substantial variability in outcomes. We hypothesized that risk factor profiles may influence tumor aggressiveness, defined by time between diagnosis and death, independent of histology. Among 1.3 million women from 21 prospective cohorts, 4,584 invasive epithelial ovarian cancers were identified and classified as highly aggressive (death in <1 year, n = 864), very aggressive (death in 1 to < 3 years, n = 1,390), moderately aggressive (death in 3 to < 5 years, n = 639), and less aggressive (lived 5+ years, n = 1,691). Using competing risks Cox proportional hazards regression, we assessed heterogeneity of associations by tumor aggressiveness for all cases and among serous and endometrioid/clear cell tumors. Associations between parity (phet = 0.01), family history of ovarian cancer (phet = 0.02), body mass index (BMI; phet ≤ 0.04) and smoking (phet < 0.01) and ovarian cancer risk differed by aggressiveness. A first/single pregnancy, relative to nulliparity, was inversely associated with highly aggressive disease (HR: 0.72; 95% CI [0.58–0.88]), no association was observed for subsequent pregnancies (per pregnancy, 0.97 [0.92–1.02]). In contrast, first and subsequent pregnancies were similarly associated with less aggressive disease (0.87 for both). Family history of ovarian cancer was only associated with risk of less aggressive disease (1.94 [1.47–2.55]). High BMI (≥35 vs. 20 to < 25 kg/m2, 1.93 [1.46–2.56] and current smoking (vs. never, 1.30 [1.07–1.57]) were associated with increased risk of highly aggressive disease. Results were similar within histotypes. Ovarian cancer risk factors may be directly associated with subtypes defined by tumor aggressiveness, rather than through differential effects on histology. Studies to assess biological pathways are warranted.