Cumulus Breast Density
Assessing breast cancer risk using mammography often includes evaluating breast density, but this has traditionally been studied using older film-based images. Today, most imaging is digital, and the images used in clinical care are processed for display. These processing steps can change how the image looks, raising questions about whether density measurements from these images are still reliable.
In this work, processed digital mammograms were used to measure breast density and evaluate its relationship with breast cancer risk. Density was assessed using Cumulus, a widely used tool for quantifying dense tissue on mammograms. The results showed that higher density remained strongly associated with increased risk, consistent with prior findings from film-based imaging. This supports the idea that routinely acquired clinical images can still provide meaningful and reproducible measurements.
Publication Link: PubMed
Figure A: Representative full-field digital mammography image prior to processing. This original image reflects the native acquisition resolution (70 micron pixel size) before any downsampling or filtering is applied.
Figure B: Processed full-field digital mammography image after downsampling (to 200 micron pixel size) and median filtering (radius of three pixels), with dense tissue segmentation generated using Cumulus (outlined in green). These preprocessing steps support more consistent density assessment.