You have requested a machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Neither SPIE nor the owners and publishers of the content make, and they explicitly disclaim, any express or implied representations or warranties of any kind, including, without limitation, representations and warranties as to the functionality of the translation feature or the accuracy or completeness of the translations.
Translations are not retained in our system. Your use of this feature and the translations is subject to all use restrictions contained in the Terms and Conditions of Use of the SPIE website.
6 March 2013Model observer detectability as a substitute for contrast detail analysis in routine digital mammography quality control
This work investigated the substitution of CDMAM contrast detail (c-d) analysis with detectability (d’) from a nonprewhitening
eye filter model observer (NPWE) for routine quality control (QC). Routinely acquired QC data for 53
systems were analyzed: 13 computed radiography (CR) and 40 integrated detector (DR) systems. For a given system,
threshold gold thickness from the c-d analysis (T) was calculated from 16 images and compared against d’ calculated for
0.1 and 0.25 mm diameter discs. The d’ data were calculated from the routine 50 mm PMMA AEC image and the
measured pre-sampling detector modulation transfer function (MTF). Threshold gold thickness T and d’ were plotted as
function of MGD and compared. The Fuji CR and the Agfa CR systems had the highest T values compared to the other
systems. The Hologic systems were found to have a low value of T, compared to the other systems, for both disc
diameters. The NPWE results reflected the performance seen for T data for the majority of the systems with the
exception of the Fuji CR and Konica CR systems. The Hologic systems gave unexpectedly low d’ results or
unexpectedly low T values. The correspondence between the two quality indices was examined with the Pearson
correlation statistical test. This test was not applicable to the GE Essential systems because all systems are grouped
together at the same working point so the result of r is about 0. For all other groups of systems the test gave good results
(larger than -0.65).
The alert did not successfully save. Please try again later.
Elena Salvagnini, Kim Lemmens, Hilde Bosmans, Lara Struelens, Nicholas W. Marshall, "Model observer detectability as a substitute for contrast detail analysis in routine digital mammography quality control," Proc. SPIE 8668, Medical Imaging 2013: Physics of Medical Imaging, 866821 (6 March 2013);