In x-raymammography, some of the components that play significant role to early diagnosis are the x-ray source, the breast composition as well as the composition of the x-ray converter. Various studies have previously investigated separately the influence of breast characteristics and detector configuration on the optimization of mammographicimaging systems. However, it is important to examine the combined effect of both components in improving the signal transfer properties in mammography systems of the mammograms. In the present study, the authors compared and evaluated x-ray converters using software breast models and realistic mammographicspectra in terms of: (a) zero-frequency detective quantum efficiency (DQE) and (b) sensitivity. The impact of x-ray converter thickness on contrast threshold (CTH) for observer assessment, based on the Rose model, was demonstrated as well.Methods:
Monte Carlo techniques were applied to simulate the x-ray interactions within the software breast phantoms and thereafter within the detective medium. Simulations involved: (a) two mammographicx-ray spectra: 28 kV Mo, 0.030 mm Mo, and 32 kV W, 0.050 mm Rh of different entrance surface air kerma (ESAK: 3–7 mGy), (b) realistic breast models (dense and fatty) and (c) x-ray converter materials most frequently considered in investigations on energy integrating digital mammographydetectors: the Gd2O2S:Tb granular phosphor, the CsI:Tl structured phosphor, and the a-Se photoconductive layer. Detector material thickness was considered to vary in the range from 50 mg/cm2 up to 150 mg/cm2.Results:
The Monte Carlo study showed that: (a) the x-ray beam becomes less penetrating after passing through dense breasts leading to higher values of zero-frequency DQE of the x-ray imaging converters and improved CTH values in all cases considered, (b) W/Rh target/filter combination results in improved CTH values at higher ESAK values, and (c) a-Se shows higher zero-frequency DQE values than the phosphor-based converters, Gd2O2S:Tb and CsI:Tl. However, thicker layers of CsI:Tl could be comparable to a-Se layers achieving approximately 27.6% CTH improvement at a thickness of 150 mg/cm2.Conclusions:
The present Monte Carlo investigation indicates that in the energy range employed in mammography, an upper limit, approximately 100 mg/cm2, should be considered in the development of thicker a-Se converters. On the other hand, above this thickness value, CsI:Tl converter could improve its imaging performance.
II. MATERIALS AND METHODS
II.A. Breast models
II.B. X-ray converters
II.C. Monte Carlo code
II.C.1. Monte Carlo simulation of x rays in breast models
II.C.2. Monte Carlo simulation within the x-ray converters
II.D. Imagining performance of the x-ray converters
II.D.1. Zero-frequency detective quantum efficiency
II.D.2. Detector sensitivity
II.D.3. Contrast threshold
III. RESULTS AND DISCUSSION
III.A. Energy and angular distribution of x-rays exiting the breast phantoms
III.B. Intrinsic parameters within the x-ray converter
III.C. Imaging performance of x-ray converters
III.C.2. Zero-frequency DQE
III.C.3. DQE considerations in digital mammography
III.C.4. Contrast threshold
Data & Media loading...
Article metrics loading...
Full text loading...