z-logo
Premium
SU‐FF‐I‐35: Scatter Correction For Digital Tomosynthesis
Author(s) -
Sechopoulos I,
Suryanarayanan S,
Vedantham S,
Karellas A
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 - 10.1118/1.1997515
Subject(s) - imaging phantom , tomosynthesis , projection (relational algebra) , contrast to noise ratio , mammography , nuclear medicine , optics , image quality , aperture (computer memory) , physics , iterative reconstruction , computer science , artificial intelligence , medicine , algorithm , acoustics , cancer , breast cancer , image (mathematics)
Purpose: To investigate post acquisition scatter correction for digital tomosynthesis breast imaging. Method and Materials: Images of a composite phantom that was fabricated for evaluating digital breast tomosynthesis and used in a previous contrast‐detail (CD) study [Suryanarayanan et al., Acad Radiol 7: 1085–1097, 2000] were used to test the scatter correction method. These images were acquired using a prototype full‐field digital mammography (FFDM) system (GE Medical Systems, Milwaukee, WI) without an anti‐scatter grid. The phantom comprised of a centrally placed CD insert (MedOptics, Tucson, AZ), blocks of cluttered paraffin and polymethyl methacrylate (PMMA), and beeswax surrounding it to provide a total phantom thickness of 54 mm. A set of 7 projection images of the phantom were acquired over an angular range of ± 18° at 6 0 intervals at 26 kVp, MoMo, and 32 mAs/view. The projection data sets were corrected for scatter using the scatter correction technique described by Trotter et al. [Proc. SPIE, vol. 4682: 469–478, 2002] and processed with an adaptive noise filter. The projection images were then reconstructed using back‐projection and iterative restoration methods using Tuned Aperture Computed Tomography (TACT) [Webber et al., J. Digit. Imaging, 13: 90–97, 2000] software (developed by R.L. Webber, Wake Forest University, NC). The contrast‐to‐noise (CNR) ratio, signal‐to‐noise ratio (SNR), and % contrast were computed for one of the targets (2.32 mm diameter and 0.24 mm depth). Results: The uncorrected projection data set reconstructed with back‐projection resulted in CNR = 3.0, SNR = 30.4, and % contrast = 11.1, while the scatter corrected and processed projection images yielded CNR = 10.2, SNR =60.7, and %contrast = 20.2. Conclusion: The results of this study indicate improved SNR, CNR, and % contrast after scatter correction in tomosynthesis. We are currently implementing and evaluating other scatter correction and reconstruction methods for digital tomosynthesis.

This content is not available in your region!

Continue researching here.

Having issues? You can contact us here