Breast cancer is one of the leading causes of cancer death in the United States, claiming about 46,000 lives eve U year. Marmnography has been proven effective in the diagnosis of breast cancer. However, the appearance of early breast cancer in mammograms is often subtle, necessitating the preservation of subtle features of mammograms when images are displayed to the radiologists for diagnostic interpretation. Digital mammography has the potential to hnprove the early detection of breast cancer. However, current medical soft-copy display devices have a number of limitations that may compromise the diagnostic effectiveness of digital mammography by failing to adequately represent the subtle features associated with early disease. Objective: The long-term goal of this project is to optimize the display quality"""""""" of digital mammograms, leading to improved early detection, diagnosis, and treatment of breast cancer. This proposal addresses one key aspect of display quality, resolution. This study aims 1) to substantiate the effects of the resolution characteristics of soft-copy display devices on the accuracy of mammographic interpretation in digital mammography, and 2) to define a resolution range for mammographic soft-copy display devices within which the diagnostic capability of digital mammograms to reveal the early signs of breast cancer is not compromised. Study Design: Initially, methods will be developed to simulate display presentations of soft-copy display devices and to simulate mammographic lesions (benign and malignant masses and microcalcifications). The quality of the simulations will be validated by objective quantitative methods. Mathematical perception models and the simulations in conjunction with real mammographic backgrounds will be used to assess the influence of the display resolution characteristics on the diagnostic interpretation of digital mammograms. The optimum display resolution setting and the resolution ranges within which diagnosis is not degraded more than specified levels will be determined. The validity of the model predictions will be examined in two observer performance experiments, one with mammographic backgrounds containing simulated lesions, and one with actual mammography cases. The results will be used to formulate recommendations to the breast imaging community concerning display resolution criteria for safe and optimized soft-copy display of digital mammograms.
