Ovarian cancer is the 5th leading cause of cancer death among US women; 62% of cases are diagnosed at advanced stages in which only 27% survive past 5 years. Yet, when ovarian cancer is detected at the localized stage (15% of cases), the 5 year survival rate is 94%. As survival is drastically improved if diagnosed early, methods to improve early detection through enhanced risk assessment, leading to better targeted primary prevention and screening are critical. The two validated risk assessment models from Rosner and Pfeiffer identify women at higher risk for ovarian cancer, but only have modest discriminatory power. Currently, no effective screening modality exists for ovarian cancer. Previous approaches for screening using single markers such as CA-125 in average risk populations have been unsuccessful, and led to many false positive results. Identifying novel panels of markers important to early carcinogenesis is key; epigenetic events (DNA methylation) are noted to occur early in carcinogenesis and reflect environmental insults and genetic vulnerability. Measurement of DNA methylation in cell free DNA has shown promise for early detection and screening for other cancers. Emerging evidence has shown that DNA methylation of select genes measured in tissue and plasma results in sensitivities >75% for detecting ovarian cancer. Yet, all evidence for ovarian cancer comes from small cross-sectional or retrospective clinical studies with no or limited epidemiologic data, raising concerns about temporality and confounding. We propose to replicate and build upon these promising findings by identifying a panel of DNA methylation markers that can detect ovarian cancer early. Using the Women's Health Initiative (WHI), one of the largest prospective studies in the US for women's health, we will verify differences in DNA methylation of 96 key genes (discovered by The Cancer Genome Atlas (TCGA)) measured in tumor vs. adjacent non tumor tissue using high throughput targeted resequencing. From these 96 TCGA genes, we will identify the top 10 genes that are differentially methylated by histology (npairs=200;
Aim 1). In a nested case control (ncases=610;
Aim 2), we will examine if DNA methylation of these top 10 genes measured in plasma are predictors of ovarian cancer risk years prior to cancer diagnosis.
In Aim 3, we will evaluate the overall performance of DNA methylation markers to enhance existing ovarian cancer risk models in terms of accuracy, discrimination and false positives (n=161,808 women). We will validate our findings in an independent prospective cohort, the Breast Cancer Family Registry ((BCFR); n=11,950 families; ncases=164;
Aim 4). The WHI and BCFR are ideal studies to conduct these aims; both have collected biospecimens and detailed epidemiologic data. We will employ a novel two-tiered approach for early detection of ovarian cancer that examines promising biomarkers (secondary prevention) across the spectrum of ovarian cancer risk (primary prevention). Our goal is through accurate identification of high risk individuals and reliable markers of early detection, we will reduce false positives, morbidity and mortality from ovarian cancer.
No routine screening method for ovarian cancer exists; yet early detection vastly improves survival. This project focuses on identification of markers for early detection in a large prospective study. By identifying relevant biomarkers, this research wil inform future screening tools and risk assessment models for this highly lethal disease.
