Breast Cancer Risk Calculator

About the Model

About 1 in 8 U.S. women develop invasive breast cancer over their lifetime. Mammograms allow for early detection of breast cancer and help determine which lifestyle changes can lower one’s risk. In addition, a woman’s risk of breast cancer is an important factor in determining what preventive measures to take. The purpose of this breast cancer risk calculator is to predict a woman’s risk of breast cancer within one year.

This calculator’s predictions are based on a logistic regression model developed by Barlow et al. (JNCI 98:1204-1214, 2006) using data from the Breast Cancer Surveillance Consortium. The model requires information about age group, breast density (from the screening mammography), previous breast procedures, first-degree family history of breast cancer, and – for post-menopausal women only – race/ethnicity, body mass index, age at birth of first child, current use of hormone therapy, surgical menopause, and previous mammographic outcome. There are separate prediction models for premenopausal and postmenopausal women.

The Barlow model has been evaluated using the receiver operating characteristic (ROC) curve. The area under the ROC curve (AUC) is a commonly used statistical measure for evaluating and comparing prediction models. A useless prediction model (i.e., a random guess) would have an AUC of 0.5. The AUC of the Barlow model is estimated to be approximately .622 (for both pre- and post-menopausal women), which is significantly higher than 0.5.