Premium
Visibility of simulated microcalcifications—A hardcopy‐based comparison of three mammographic systems a)
Author(s) -
Lai ChaoJen,
Shaw Chris C.,
Whitman Gary J.,
Johnston Dennis A.,
Yang Wei T.,
Selinko Veronica,
Arribas Elsa,
Dogan Basak,
Kappadath S. Cheenu
Publication year - 2005
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.1833011
Subject(s) - visibility , mammography , microcalcification , computer vision , artificial intelligence , medical imaging , computer science , medical physics , computer graphics (images) , medicine , optics , breast cancer , physics , cancer
Full‐field digital mammography systems are currently available for clinical use. These digital systems offer improved image quality, flexible image processing, display, storage, retrieval, and transmission. These systems employ a variety of different x‐ray detectors based on storage phosphors (in computed radiography), charge‐coupled devices (CCDs), or amorphous silicon flat panels (FPs). The objective of this study is to compare three different types of mammographic detectors: screen–film (SF) combination, aCsI ‐based FP detector, a CCD and x‐ray phosphor‐based detector for their performance in detection of simulated microcalcifications. Microcalcifications (MCs) were simulated with calcium carbonate grains of various sizes( 90 – 355 μ m ). They were overlapped with a slab of simulated 50% adipose/50% glandular breast tissue for a uniform background or an anthropomorphic breast phantom for a tissue structure background. Images of the phantoms, acquired with and without magnification, were reviewed by mammographers, physicists, and students. A five‐point confidence level rating was given for each MC reviewed. Average ratings from the mammographers were used to compare the performances of the three imaging systems, various MC size groups, and two magnification modes. The results indicate that with uniform background and no magnification, the FP system performed the best while the SF system did slightly better than the CCD system. With magnification added, all detection tasks were improved except for the smallest and largest one or two size groups. In particular, detection in the SF and CCD images was significantly improved over that in the FP images. With tissue structure background and no magnification, the FP system was outperformed by the SF and the CCD systems. With magnification added, the performance of the FP and the CCD systems was improved significantly. With this improvement, the SF and FP systems were outperformed by the CCD system.