A Two-Stage Approach to Genetic Risk Prediction in Primary Care

• Activity Number: 351
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Philamer Atienza*+ and Swati Biswas and Jonathan Chipman and Christopher Amos and Kevin S. Hughes, M.D. and Banu K. Arun and Giovanni Parmigiani
• Companies: University of North Texas Health Science Center and University of North Texas Health Science Center and Dana-Farber Cancer Institute and MD Anderson Cancer Center and Harvard Medical School/Massachusetts General Hospital and MD Anderson Cancer Center and Harvard School of Public Health
• Keywords: BRCAPRO ; AUC ; specificity ; sensitivity ; predictive value ; breast cancer

Abstract

BRCAPRO estimates the probabilities of carrying mutations of the breast cancer genes BRCA1 and BRCA2 based on extensive family history information. Although it is widely used in genetic counseling, in primary care it is impractical to collect so much information. So we propose a two-stage approach: in the first stage, risk is assessed using limited family history information via a simplified version of BRCAPRO. If this risk is sufficiently high, the full BRCAPRO is used in the second stage. We explore several first stage tools including BRCAPROLYTE, which uses affection-only information on relatives up to the second-degree, and its variation BRCAPROLTYE-PLUS, which also utilizes the unaffected relatives with their ages imputed. We compute area under the ROC curve (AUC), sensitivity, specificity, and predictive values of the first stage and the overall two-stage approach by considering different combinations of the first and second stage cutoffs. We found that BRCAPROLYTE has high sensitivity. BRCAPROLYTE-PLUS has higher AUC and increased specificity for a given sensitivity leading to reduced counseling burden. We conclude that BRCAPRO can be adapted for primary care settings.