Brad M. Keller PhD, Presenter: Nothing to Disclose

Dania Daye BS, Abstract Co-Author: Nothing to Disclose

Jinbo Chen, Abstract Co-Author: Nothing to Disclose

Hao Jing PhD, Abstract Co-Author: Nothing to Disclose

Emily F. Conant MD, Abstract Co-Author: Consultant, Hologic, Inc

Despina Kontos PhD, Abstract Co-Author: Research Consultant, Hologic, Inc Support agreement, Hologic, Inc

Breast density, quantified as the projection percent area of mammographically dense tissue, is a strong risk factor for breast cancer. We investigate if volumetric breast density measures could improve breast cancer risk assessment over area-based density measures after adjusting for known demographic risk factors.

A case-control study was performed. The contralateral mediolateral oblique digital mammography images from 106 women with unilateral breast cancer and 212 age and side-matched controls were retrospectively analyzed. For each mammogram, both absolute and percentage estimates of breast density were computed. Area breast density was computed using a validated, custom-made computer algorithm, and volumetric breast density was assessed using FDA-cleared software (Quantra™, Hologic, Inc). For each woman in this study, established demographic risk factors regarding parity, first-degree relatives of breast cancer, ethnicity, and age at menarche were obtained via review of archived questionnaires. Multivariable logistic regression with stepwise feature selection was performed to determine which measures of breast density are most predictive of cancer status after adjustment for the standard demographic risk factors. Area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the predictive performance of the models.

A model based only on demographics has an AUC of 0.64 (p<0.001). Adding area percent breast density leads to a higher AUC of 0.70 (p<0.001). When both percent and absolute density measures are considered, only absolute volume (p=0.007) and absolute area (p=0.008) dense tissue estimates are retained as significant after adjustment for the demographic risk factors. This combined model achieves the highest AUC of 0.75 (p<0.001).

Volumetric breast density provides statistically significant, complementary information about breast cancer risk, even after adjusting for demographic risk factors and area breast density. Absolute measures of dense tissue could lead to further improvements in breast cancer risk assessment, compared to the commonly used percentage density measures, especially when volumetric measures are also considered.

Breast cancer risk assessment can be improved via simultaneous consideration of absolute measures of volumetric and area estimates of breast density, even after adjusting for demographic risk factors.

Keller, B, Daye, D, Chen, J, Jing, H, Conant, E, Kontos, D, The Role of Volumetric and Area Density Measures in Breast Cancer Risk Assessment.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12024631.html