No evidence for increase in screening among women given report of moderately higher than average risk for breast cancer from personal genomics services: The PGen Study. S. W. Gray1, H. Q. Rana1,2, S. Gollust3, C. A. Chen4, S. Kalia2, J. Mountain5, T. Moreno6, J. S. Roberts7, R. C. Green2 for the PGen Study Group 1) Division of Medical Oncology and Population Sciences, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA; 2) Division of Genetics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA; 3) Division of Health Policy and Management, University of Minnesota School of Public Health, Minneapolis, MN; 4) Data Coordinating Center, Boston University School of Public Health, Boston, MA; 5) 23andMe, Inc., Mountain View, CA; 6) Pathway Genomics, San Diego, CA; 7) Department of Health Behavior and Health Education, University of Michigan School of Public Health, Ann Arbor, MI.

   Background: With the growing popularity of personal (direct-to-consumer) genomic testing, customers are receiving cancer risk assessments, but whether modestly increased risk influences screening behaviors is unclear. Responses from the Impact of Personal Genomics (PGen) Study were used to assess whether screening increased within 6 months of receiving modestly elevated breast cancer genetic risk information. Methods: The PGen Study is a web-based survey of consented personal genomics service customers from 23andMe and Pathway Genomics prior to and after receiving genetic risk information. We analyzed responses from women who did not have breast cancer and did not receive positive BRCA variant results (offered by one company). Risk results using common SNP variants were dichotomized as above average or average and below average as the predictor variable. Primary outcomes were self-reported use of clinical breast exam (CBE), mammography, and breast MRI within 6 months of return of results. Associations between predictor and outcome variables were evaluated using chi square testing and logistic regression. Results: Of 1050 women who took the baseline pre-test survey between March-July 2012, data from the 6 month survey was available for 55% at the time of analysis. Demographics of the 311 women over age 40 in our cohort were mean age 56 years (40-81); 90% White, 3% Black, 3% Hispanic; 74% college educated or above; 35% with family history of breast cancer; 96% insured; 11% above average risk (AAR). Women who received results from common variants indicating AAR did not report significantly more screening behaviors than those with average or below average risk results (ABR): CBE (AAR/ABR) 50%/49%, p=0.92; mammography (AAR/ABR): 41%/48%, p=0.43; breast MRI (AAR/ABR): 5.9%/4.0%, p=0.60. After controlling for socioeconomic characteristics, family history of breast cancer, health insurance, and testing company, there were no significant differences in the use of CBE (adjusted odds ratio (aOR) 1.41, 95% CI 0.65-3.04), mammography (aOR 0.93, CI 0.43-2.02) or breast MRI (aOR 3.74 CI 0.61-23.1) by risk group. Exploratory analysis for women under age 40 revealed similar findings. Conclusion: At 6 month follow-up, female customers of personal genomic testing who received common variant results of moderately above average breast cancer risk did not differ in breast cancer screening practices from customers who received average and below average risk results.

You may contact the first author (during and after the meeting) at