Prediagnositic androgens and IGF-I and risk of ovarian cancer

A grant proposal will be submitted in May 2013 to a German funding agency (Deutsche Krebshilfe). The budget will include a lump sum for the OC3 coordinating center to cover costs for data pooling.

Epithelial ovarian cancer is a significant contributor to cancer mortality in women (Sieger R et al. 2012, Globocan 2008), but its etiology remains poorly understood. Evidence from experimental, clinical and epidemiological studies clearly indicates that ovarian cancer is a very heterogeneous disease (Shih & Kurman, 2004; Jarboe EA et al. 2008, Roh MH et al. 2010; Kurian AW et al., 2005, Gates MA et al., 2010). Along with differences by the major histological subtypes, new research findings suggest that the cell of origin (ovarian versus fallopian tube) could be also contribute to the observed heterogeneity (Lee Y. et al. 2007, Folkins AK et al. 2008). Nevertheless, because of the relatively low disease incidence and logistic difficulties in obtaining detailed tumor data, progress in characterizing these differences and using the knowledge for more efficient early detection and treatment has been slow. Understanding disease etiology may help to identify women at high risk of ovarian cancer and assist in the development of appropriate early detection and prevention strategies, which are not available at present._x000D__x000D_Reproductive history influences risk of ovarian cancer and it is believed that the observed associations with risk are mediated by exposure to endogenous hormones (Hunn & Rodriguez, 2012). So far, epidemiological studies relating prediagnostic concentrations of endogenous hormones to risk of epithelial ovarian cancer have yielded inconsistent results (Helzlsouer KJ et al, 1995; Lukanova et al, 2003; Rinaldi et al. 2007; Tworoger et al. 2008). One potential explanation could be that hormone-ovarian cancer associations differ by histological subtype / cell of origin and failing to account for such heterogeneity contributed to the overall null findings. Initial data from an on-going, large case-control study, nested in two Scandinavian Maternity Cohorts (1,052 cases), strongly supports such a possibility. A very consistent positive association of androgens (testosterone and androstenedione) with risk of most histological subtypes of ovarian tumors (both invasive and borderline), but not with invasive serous tumors, was observed._x000D__x000D_We propose to test the hypothesis that concentrations of endogenous hormones are associated with increased risk of non-serous epithelial ovarian cancer and with tumors with a dominant left or right ovarian mass (as a proxy for ovarian cell of origin). The lack of association observed in prospective studies so far could be because most of the included cases (particularly in postmenopausal women) had serous invasive ovarian cancer._x000D__x000D_An extension of the EPIC nested case-control study on endogenous hormones and epithelial ovarian cancer (Rinaldi et al., 2007) is in progress. The extended study includes 576 invasive cases (at least 291 serous, 41 mucinous and 90 endometrioid/clear cell tumors) and twice as many controls. We would like to invite other cohorts with available data to join with us in our effort to study in greater detail the role of endogenous hormones in ovarian cancer. The proposed pooled analyses are also a natural extension of an on-going research within the Ovarian Cancer Cohort Consortium (OC3). Through a project led by Dr. S. Tworoger, data on ovarian tumor dominance is being collected in all prospective studies that have evaluated the role of endogenous hormones in epithelial ovarian cancer. Dominant tumors will be considered those limited to one ovary (pelvic side) or with at least two-fold difference in tumor volume between ovaries, and will be used a surrogate marker of ovarian origin. Tumors more evenly sized between the ovaries will be classified as non-dominant and of likely tubal origin._x000D__x000D_Understanding ovarian cancer etiology and biology is only possible by combining the resources of most (if not all) existing studies, thereby maximizing access to cases with detailed tumor data, including histology and tumor dominance.

The main goal of this proposal is to understand the role of androgens and IGF-I in the etiology of epithelial ovarian cancer by exploring associations by the major histological subtypes of the tumors and by tumor dominance (as a proxy for cell of origin).

The primary aims of the study are:_x000D__x000D_1. To evaluate the associations of prediagnostic concentrations of androgens (testosterone, androstenedione and DHEAS) with risk of the major histological subtypes of epithelial ovarian cancer and by tumor dominance. _x000D__x000D_2. To evaluate the associations of prediagnostic concentrations of IGF-I with risk of the major histological subtypes of epithelial ovarian cancer and by tumor dominance.

We propose to pool available data from nested case control studies on prediagnostic concentrations of androgens and IGF-I and risk of epithelial ovarian cancer. Multivariable conditional logistic regression models will be used to assess the associations of endogenous hormones with ovarian cancer risk. Heterogeneity of the hormone-risk associations by histological subtype and tumor dominance will be assessed by log likelihood ratio tests. Subgroup analyses by age at diagnosis will also be conducted.

The role of endogenous hormones in ovarian cancer can be evaluated only in studies with prospectively collected blood samples. Despite being a leading cause of cancer mortality in women, ovarian cancer is relatively infrequent and the incidence of the major histological subtypes is even lower. Currently, no individual cohort has sufficiently large numbers to explore associations by histological subtype. Tumor dominance data will be available from an OC3 project.

15

Incident cases of invasive epithelial ovarian cancer (including fallopian tube and primary peritoneal cancers). Data on histological subtype, age at diagnosis, and, if available on tumor stage, grade and dominance.

Prediagnostic blood concentrations of any of the following hormones: androgens (testosterone, androstenedione, DHEA, DHEAS), SHBG and IGF-I, together with information on the type of blood aliquots analyzed and laboratory methods used for hormone quantification.

As available: at blood donation (age, menopausal status (pre, peri or post), fasting, and, for pre-menopausal women, phase of menstrual cycle), reproductive history (ages at menarche and menopause, parity, age at first and last full-term pregnancy, breast feeding, use of oral contraceptives and hormone replacement therapy (never, past, at blood donation, duration)), gynecological conditions (hysterectomies, unilateral ovariectomy, history of infertility), family history of ovarian cancer, family history of breast cancer, anthropometry, smoking, physical activity, education, and ethnicity.

No