OBJECTIVE: The objective of our study was to compare the detection and distinguishability of microcalcifications on mammograms obtained with a digital direct flat-panel detector versus an analog system using an anthropomorphic breast phantom. MATERIALS AND METHODS: Studies were performed with a digital mammography system (Selenia) and an analog mammography system (Mammomat 3). Sixty-five transparent films were used as test specimens. Randomly distributed round and heterogeneous silicate particles (diameter, 100-1,400 microm) and an anthropomorphic scatter body were applied to the films. All radiographs were taken at identical settings and exposures. Six radiologists rated the films and monitor-displayed images independently of each other in random order on a standardized electronic questionnaire. RESULTS: Interpretations based on monitor reading produced superior results over those based on digital image reading and analog film reading. In 41.1% (95% CI, 38.7-43.5%) of all the monitor readings, 20.2% (18.2-22.2%) of all digital images, and 19.6% (17.6-21.6%) of all analog films, the number of detectable microcalcifications agreed with the gold standard method. The diameter of visible microcalcifications was interpreted correctly in 35.6% (33.2-38.0%) of monitor readings, 19.0% (17.1-21.0%) of digital images, and 21.0% (18.9-23.0%) of analog films; and microcalcification shape was interpreted correctly in 53.8% (51.4-56.3%) of monitor readings, 28.2% (26.0-30.4%) of digital images, and 28.3% (26.0-30.5%) of analog films. Microcalcification number and size were underestimated more frequently than overestimated. Regardless of display medium, accuracy increased proportionately with the diameter of the simulated microcalcifications for all evaluation variables. CONCLUSION: Digital flat-panel mammography is superior to the analog screen-film method for the detection and morphologic characterization of microcalcifications larger than 200 microm in diameter when the display medium is a monitor.
OBJECTIVE: The objective of our study was to compare the detection and distinguishability of microcalcifications on mammograms obtained with a digital direct flat-panel detector versus an analog system using an anthropomorphic breast phantom. MATERIALS AND METHODS: Studies were performed with a digital mammography system (Selenia) and an analog mammography system (Mammomat 3). Sixty-five transparent films were used as test specimens. Randomly distributed round and heterogeneous silicate particles (diameter, 100-1,400 microm) and an anthropomorphic scatter body were applied to the films. All radiographs were taken at identical settings and exposures. Six radiologists rated the films and monitor-displayed images independently of each other in random order on a standardized electronic questionnaire. RESULTS: Interpretations based on monitor reading produced superior results over those based on digital image reading and analog film reading. In 41.1% (95% CI, 38.7-43.5%) of all the monitor readings, 20.2% (18.2-22.2%) of all digital images, and 19.6% (17.6-21.6%) of all analog films, the number of detectable microcalcifications agreed with the gold standard method. The diameter of visible microcalcifications was interpreted correctly in 35.6% (33.2-38.0%) of monitor readings, 19.0% (17.1-21.0%) of digital images, and 21.0% (18.9-23.0%) of analog films; and microcalcification shape was interpreted correctly in 53.8% (51.4-56.3%) of monitor readings, 28.2% (26.0-30.4%) of digital images, and 28.3% (26.0-30.5%) of analog films. Microcalcification number and size were underestimated more frequently than overestimated. Regardless of display medium, accuracy increased proportionately with the diameter of the simulated microcalcifications for all evaluation variables. CONCLUSION: Digital flat-panel mammography is superior to the analog screen-film method for the detection and morphologic characterization of microcalcifications larger than 200 microm in diameter when the display medium is a monitor.
Authors: Kathrin Barbara Krug; Hartmut Stützer; Peter Frommolt; Julia Boecker; Henning Bovenschulte; Volker Sendler; Klaus Lackner Journal: Int J Breast Cancer Date: 2010-10-17