Premium
SU‐GG‐I‐131: Non‐Primary to Primary Ratio Measurements in a Digital Breast Tomosynthesis System
Author(s) -
Shen S,
Mainprize J,
Mawdsley G,
Yaffe M
Publication year - 2008
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 - 10.1118/1.2961529
Subject(s) - tomosynthesis , optics , image quality , flat panel detector , glare , detector , projection (relational algebra) , isotropy , materials science , mammography , nuclear medicine , physics , medicine , mathematics , computer science , breast cancer , image (mathematics) , artificial intelligence , layer (electronics) , cancer , composite material , algorithm
Purpose: For a breast tomosynthesis system using a stationary detector, a standard mammographic anti‐scatter grid cannot be used. As a result, scatter and off‐focal radiation are expected to be high and will negatively impact reconstructed image quality. In addition, phosphor‐based systems can suffer from glare which will further degrade the image. The goal of this project is to quantify the “non‐primary” sources of signal as a function of projection angle to provide information for potential image correction in the tomosynthesis reconstruction. Method and Materials: The scatter, off‐focal and glare components will be evaluated in terms of a non‐primary to primary ratio (NPR). Using a series of tantalum apertures of increasing size, we measured the primary and off‐focal components, and scatter/glare in an open beam. The NPR was then evaluated over typical tomosynthesis projection angles (0° to 25°) for a range of thicknesses of plastic phantoms. Results: The off‐focal to primary ratio is measured to be 0.091. The NPR at 0° was calculated as 0.44, 0.82, and 1.2 for 2, 4 and 6 cm phantoms. A small angular dependence is seen in the NPR, increasing at higher angles. The NPR can be reasonably predicted by an empirical model of NPR versus effective thickness with a maximum error in NPR of 0.022 (4.9% error) over a range of 0° to 20°. At 25° the NPR appeared to increase sharply (5%–9.5% greater than the empirical model). Conclusion: The NPR appears to be weakly dependent on projection angle up to about 20°. Further investigation will involve extending the measurements to a range of kV and filter combinations, and to elucidate the mechanism for increased NPR at highly oblique angles (>20°). Conflict of Interest (only if applicable): Our lab has a research agreement with GE Healthcare regarding several topics in breast imaging.