Building a Better Model: A Personalized Breast Cancer Risk Model Incorporating Breast Density to Stratify Risk and Improve Application of Resources.

摘要

Purpose: Development and validation of a personalized breast cancer risk assessment model that includes automated measurement of breast density. Scope: Assemble a cohort of women with known breast cancer risk factors and digital mammogram files for women diagnosed with breast cancer using existing data sources and match them to controls (Harvey/Knaus). Validate and refine automated breast density software (Yaffe/Harvey). Build and validate the initial comprehensive model (Knaus/Yaffe/Harvey). Major Findings: During this second year, we have recruited over 3200 women (622 cases). Breast density measurement has been evaluated for accuracy using a second test set showing very good correlation with 2D methods, for precision using repeated measures with excellent correlation, and variation between vendors. A survey instrument was developed, vetted using focus groups, and administered (nearing completion) to evaluate screening knowledge, practices, and willingness to change practice.