Automatic Segmentation of Mammographic Density

Factors Affecting Sensitivity and Specificity of Screening Mammography and MRI in Women with an Inherited Risk for Breast Cancer

“…In our study, breast density was scored by different radiologists at different centers, so that inter-and intra-observer variation may play a role. Berg et al [38] found substantial inter-and intra-observer variability in mammographic interpretation, including breast density, while other studies reported that breast density assessed by different radiologists or by a radiologist and digital assessment was well correlated [39][40][41]. However, it is likely that this variation was random (and not systematic), leading to a dilution of the effect.…”

“…Additionally, the false positive rate was higher in women younger than 50 years than in older women. Using univariate ORs this finding was significant for women 40 …”

Increasing Mammographic Breast Density in Response to Hormone Placement Therapy and Breast Cancer Risk

“…Automated measurement of breast density is being developed (23)(24)(25)(26) and will greatly improve the ability to assess breast density changes over time in large populations though these methods will need to be validated and become more widely available. Breast density could then be assessed prospectively in a reproducible fashion on large populations.…”

Breast Density Segmentation: A Comparison of Clustering and Region Based Techniques

“…without taking the information of neighboring pixels into account. Examples of algorithms belonging to this category are the works of Boyd et al [4] and Sivaramakrishna et al [5] who used a gray-level thresholding technique to segment the breast into dense and fatty regions, or the proposals of Aylward et al [6] and Ferrari et al [7], who used Gaussian mixtures for segmenting the breast in five and four regions respectively. On the other hand, region based algorithms relies on classifying the pixels of the mammogram taking neighborhood information into account, i.e.…”