Evaluation of Digital Processing Methods for the Display of Digital Mammography

Purpose One of the advantages of digital mammography is the ability to acquire a mammography image with a larger contrast range. With this advantage comes the tradeoff of how to display this larger contrast range. Laser printed film, and video display both have smaller dynamic ranges than standard mammography film-screen systems. This work examines performance and preference studies for display processing methods for digital mammograms. Methods Processing techniques currently in practice for the display of digital mammography are being evaluated. Two sets of studies are ongoing at UNC as part of the International Digital Mammography Development Group (IDMDG). The first is a series of laboratory studies evaluating human observer detection rates for mammographic features. In these studies, we have conducted performance studies comparing default film screen (unenhanced), Fixed Intensity Windowing, Contrast Limited Adaptive Histogram Equalization, and Histogram-based Intensity Windowing. A second set of multi-center clinical studies is being performed by IDMDG are comparing filmscreen mammography to digital mammography, with the first step evaluating processing methods for the display of digital mammograms. This first clinical study is a preference study comparing analog film screen images taken at the same time as the digital images, with the digital images processed with different methods. The processing methods being evaluated are Manual Intensity Windowing, Histogram based Intensity Windowing, Mixture Modeling Intensity Windowing, Contrast Limited Adaptive Histogram Equalization, Unsharp Masking, MUSICA, Trex, and Peripheral Equalization. All of the processing methods are automatic, with fixed parameter choices, except Manual Intensity Windowing, which is done by an expert mammography technician Results and Conclusions The data analysis for this clinical preference study is not complete. Preliminary findings support earlier laboratory performance experimental results, and suggest that different algorithms will perform differently for the tasks of screening, mass characterization, and microcalcification characterization. This suggests the need for presenting different processed versions of the images to the radiologist.