ABSTRACT
PURPOSE: With improvements in breast cancer imaging, there has been a corresponding increase in false-positives and avoidable biopsies. There is a need to better differentiate when a breast biopsy is warranted and determine appropriate follow-up. This study describes the design and clinical performance of a combinatorial proteomic biomarker assay (CPBA), Videssa Breast, in women over age 50 years. EXPERIMENTAL DESIGN: A BI-RADS 3, 4, or 5 assessment was required for clinical trial enrollment. Serum was collected prior to breast biopsy and subjects were followed for 6-12 months and clinically relevant outcomes were recorded. Samples were split into training (70%) and validation (30%) cohorts with an approximate 1:4 case:control ratio in both arms. RESULTS: A CPBA that combines biomarker data with patient clinical data was developed using a training cohort (469 women, cancer incidence: 18.5%), resulting in 94% sensitivity and 97% negative predictive value (NPV). Independent validation of the final algorithm in 194 subjects (breast cancer incidence: 19.6%) demonstrated a sensitivity of 95% and a NPV of 97%. When combined with previously published data for women under age 50, Videssa Breast achieves a comprehensive 93% sensitivity and 98% NPV in a population of women ages 25-75. Had Videssa Breast results been incorporated into the clinical workflow, approximately 45% of biopsies might have been avoided. CONCLUSIONS: Videssa Breast combines serum biomarkers with clinical patient characteristics to provide clinicians with additional information for patients with indeterminate breast imaging results, potentially reducing false-positive breast biopsies.
