Quantifying Breast Thickness for Density Measurement

The largest source of error in the estimation of volumetric breast density typically arises from inaccuracies in the measurement of breast thickness. We present a method which accounts for paddle tilt and show that the variation within the compressed breast region may be as much as 21.2mm in the chest wall to nipple direction. The system-indicated value of breast thickness results in an average underestimation of compressed breast volume of 10.5% which has implications for density measurement. Paddle deformation in the lateral direction is shown to be insignificant in comparison. A method of modelling the thickness at the breast periphery is presented with examples of the thickness maps generated.