• DocumentCode
    2632179
  • Title

    Modeling of tumor conspicuity in hepatic CT images: combined compartment and vascular models

  • Author

    Bézy-Wendling, Johanne ; Kretowski, Marek ; Coupe, Pierrick

  • Author_Institution
    Laboratoire Traitement du Signal et de l´´Image, Rennes I Univ., France
  • fYear
    2004
  • fDate
    15-18 April 2004
  • Firstpage
    444
  • Abstract
    The aim of this work is to simulate dynamic CT scan of the liver, in normal tissue and in a hypervascular tumor. Two models are developed: a physiological vascular model for the main vessels, until arterioles and venules, and a compartment model for parenchyma enhancement. Combining these two models allows us to compute locally the contrast product concentration, all along the propagation, after injection in the hepatic artery and the portal vein. In the second step, a density representation of the organ is created and CT scans are simulated by using the standard reconstruction algorithm - filtered backprojection. As a final step, enhancement curves are extracted from the obtained images, showing very good agreement with real hepatic enhancement in CT.
  • Keywords
    blood vessels; computerised tomography; image enhancement; image reconstruction; image representation; liver; medical image processing; physiological models; tumours; arterioles; compartment model; contrast product concentration; density organ representation; filtered backprojection; hepatic CT images; hepatic artery; hypervascular tumor; liver; main vessels; normal tissue; parenchyma enhancement; physiological vascular model; portal vein; reconstruction algorithm; tumor conspicuity; venules; Arteries; Blood; Computational modeling; Computed tomography; Computer science; Liver neoplasms; Pathology; Portals; Solid modeling; Veins;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on
  • Print_ISBN
    0-7803-8388-5
  • Type

    conf

  • DOI
    10.1109/ISBI.2004.1398570
  • Filename
    1398570