Development of an Absolute Risk Model for Breast Cancer Subtypes

• Activity Number: 190
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Risk Analysis
• Author(s): Paige Maas*+ and Mitchell H Gail and Nilanjan Chatterjee
• Companies: and National Cancer Institute and National Cancer Institute
• Keywords: risk ; modeling ; heterogeneity ; missing data ; variance estimation ; cancer

Abstract

Absolute risk models for disease have many clinical applications, such as for determination of optimal age and frequency of screening and weighing risks and benefits of a preventive intervention. A model for absolute risk of breast cancer, widely known as the Gail model, is popularly used in clinical practice and various public health applications. Breast cancer is a heterogeneous disease, encompassing numerous subtypes based on tumor characteristics such as the presence of hormone receptors or growth factors. As different breast cancer subtypes lead to differential treatment and prognosis, it is important to develop a subtype-specific absolute risk models for breast cancers. We develop a methodology for building an absolute risk model for cancer subtypes that can integrate information from analytic cohort or case-control studies and from population-based registries such as the SEER database. Statistical challenges include handling of different types of missing information and variance estimation accounting for different sources of uncertainties. The methods will be applied to develop absolute risk models for subgroups of breast cancer defined by estrogen and HER2 receptor status.