Mammographic density assessed on paired raw and processed digital images and on paired screen-film and digital images across three mammography systems

Abstract

BackgroundInter-women and intra-women comparisons of mammographic density (MD) are needed in research, clinical and screening applications; however, MD measurements are influenced by mammography modality (screen film/digital) and digital image format (raw/processed). We aimed to examine differences in MD assessed on these image types.MethodsWe obtained 1294 pairs of images saved in both raw and processed formats from Hologic and General Electric (GE) direct digital systems and a Fuji computed radiography (CR) system, and 128 screen-film and processed CR-digital pairs from consecutive screening rounds. Four readers performed Cumulus-based MD measurements (n = 3441), with each image pair read by the same reader. Multi-level models of square-root percent MD were fitted, with a random intercept for woman, to estimate processed–raw MD differences.ResultsBreast area did not differ in processed images compared with that in raw images, but the percent MD was higher, due to a larger dense area (median 28.5 and 25.4 cm2 respectively, mean √dense area difference 0.44 cm (95% CI: 0.36, 0.52)). This difference in √dense area was significant for direct digital systems (Hologic 0.50 cm (95% CI: 0.39, 0.61), GE 0.56 cm (95% CI: 0.42, 0.69)) but not for Fuji CR (0.06 cm (95% CI: −0.10, 0.23)). Additionally, within each system, reader-specific differences varied in magnitude and direction (p < 0.001). Conversion equations revealed differences converged to zero with increasing dense area. MD differences between screen-film and processed digital on the subsequent screening round were consistent with expected time-related MD declines.ConclusionsMD was slightly higher when measured on processed than on raw direct digital mammograms. Comparisons of MD on these image formats should ideally control for this non-constant and reader-specific difference.Electronic supplementary materialThe online version of this article (doi:10.1186/s13058-016-0787-0) contains supplementary material, which is available to authorized users.

Automated mammographic breast density estimation using a fully convolutional network

“…Volumetric methods utilize the physics of x‐ray attenuations in breast tissue to estimate how much dense fibroglandular tissue exists in each given x‐ray path length. The area‐based PD is the most validated measure and researchers are actively using this measure, as indicated in recent studies . Among the above algorithms, Cumulus, a semi‐automatic area‐based method for estimating PD from the University of Toronto, is considered the gold standard for segmenting dense areas of the breast and estimating PD from given mammograms.…”

The distribution and determinants of mammographic density measures in Western Australian aboriginal women

“…unlikely to influence their intention to participate in screening). Finally, the processing applied to FFDM images to improve detection is known to alter the appearance of MD during measurement using the Cumulus software [35]. Despite reduced power to detect associations due to stratification by film/FFDM status, overall interpretation of the results was similar for both datasets; Aboriginal women have, on average, more NDA and lower amounts of MD than other Non-Aboriginal women.…”

Percent mammographic density prediction: development of a model in the nurses’ health studies

“…[22, 34] Further, studies which have assessed percent MD from digital mammograms and risk of breast cancer demonstrated that MD as assessed from digital mammography was valid and had similar association with breast cancer risk as was seen with film mammograms. [35] Additionally, recently published work from the International Pooling Project of Mammographic Density observed that in 128 paired images “MD differences between screen-film [mammograms] and processed digital [mammograms] on the subsequent screening round were consistent with expected time-related MD declines.”[36] Another limitation is that NHS and NHSII participants are predominantly Caucasian and are more similar to each other in various characteristics (e.g., education) than the general population, potentially limiting generalizability. Additional validation of the prediction model would be useful, especially in diverse populations.…”