Evaluation of automatic exposure control options in digital mammography.

Abstract

To quantify the trade-offs of dose and image quality among pre-loaded automatic exposure control (AEC) options in digital mammography, two AEC tables from the Hologic Selenia digital mammography system were compared: the default AEC "table 0" and AEC "table 1". Realistically-shaped phantoms consisting of tissue-equivalent material of various thicknesses (4.5 cm-7 cm) were imaged to obtain a figure of merit (FOM), the squared contrast-to-noise ratio per mean glandular dose. To relate the results to pathological findings and to evaluate the overall performance, the measured contrast-to-noise ratios were applied to simulated lesions on the anthropomorphic breast phantom images, producing various lesion configurations which were blindly scored. It was found that the AEC table 1 improves the low contrast FOM by 11% to 20% for the breast thicknesses of 4.5-6 cm. However, for the 7 cm thick breast, the AEC table 1 decreases the low contrast FOM by 17%. For microcalcifications, the AEC table 1 improves the FOM slightly for the breast thicknesses of 4.5--6 cm and decreases it by 18% at a thickness of 7 cm. The lesion simulation showed enhanced contrast due to the AEC table 1 for the breast thicknesses of 5 cm, 6 cm, and 7 cm, but the enhancement gradually reduces as the thickness increases. The lesion reading showed that the microcalcification detection was scored significantly higher from the AEC table 1 for the thicknesses 5 cm, 6 cm, and 7 cm. The corresponding improvement of mass detection scores was also observed but not consistently significant over the thickness range.