• DocumentCode
    2196257
  • Title

    Monte Carlo based down-scatter correction of SPECT attenuation maps

  • Author

    Vastenhouw, Brendan ; De Jong, Hugo W A M ; Van Rijk, Peter P. ; Beekman, Freek J.

  • Author_Institution
    Image Sci. Inst., Univ. Med. Center, Utrecht, Netherlands
  • Volume
    2
  • fYear
    2002
  • fDate
    10-16 Nov. 2002
  • Firstpage
    1308
  • Abstract
    Combined acquisition of transmission and emission data in SPECT can be used for accurate correction of non-uniform photon attenuation. However, down-scatter from a higher energy isotope (e.g. Tc-99m, 140 keV) contaminates lower energy transmission data (e.g. Gd-153, 100 keV), resulting in under-estimation of reconstructed attenuation coefficients. Correction for down-scatter by subtraction of down-scatter projections (for example collected in a separate scatter window) is often not very accurate and can increase noise in the attenuation-maps. Therefore, a new correction method has been developed, that is robust to noise, uses accurate scatter modeling and does not require additional energy windows. Statistical reconstruction of the attenuation-map is used, that allows adequate incorporation of model based scatter estimates. The emission images are reconstructed using a fully 3D maximum likelihood algorithm employing Monte Carlo based scatter modeling, attenuation modeling and detector blurring modeling. The correction scheme is as follows: Initially, an approximate attenuation-map is reconstructed using down-scatter contaminated transmission data (Step 1). An emission map is reconstructed based on the approximate attenuation map (Step 2). Based on this approximate Tc-99m reconstruction and the approximate attenuation-map, down-scatter in the Gd-153 window is simulated using an accelerated Monte Carlo simulator (Step 3). This down-scatter estimate is used during reconstruction of a corrected attenuation-map (Step 4). Based on the corrected attenuation-map an improved Tc-99m image is reconstructed (Step 5). Steps 3-5 are repeated to improve the down-scatter correction. The method was tested for simulated projection data of the MCAT thorax phantom with clinically realistic noise levels, assuming a dual-head camera equipped with moving transmission line sources. Typically, four cycles through the correction scheme were required to reduce errors in the attenuation coefficients from about 50% to only a few percent In addition, artifacts due to the corrupted attenuation maps on the Tc-99m emission reconstruction completely disappeared.
  • Keywords
    Monte Carlo methods; medical image processing; single photon emission computed tomography; MCAT thorax phantom; Monte Carlo based down-scatter correction; Monte Carlo based scatter modeling; SPECT attenuation maps; approximate Tc-99m reconstruction; down-scatter projections; emission data; fully 3D maximum likelihood algorithm; reconstructed attenuation coefficients; transmission data; Attenuation; Electromagnetic scattering; Image reconstruction; Isotopes; Maximum likelihood detection; Maximum likelihood estimation; Monte Carlo methods; Noise robustness; Particle scattering; Single photon emission computed tomography;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Nuclear Science Symposium Conference Record, 2002 IEEE
  • Print_ISBN
    0-7803-7636-6
  • Type

    conf

  • DOI
    10.1109/NSSMIC.2002.1239559
  • Filename
    1239559