Fachbereich Veterinärmedizin



    Limitations of the permeability-limited compartment model in estimating vascular permeability and interstitial volume fraction in DCE-MRI (2011)

    Zeitschriftenartikel / wissenschaftlicher Beitrag
    Carreira, Guido Correia
    Gemeinhardt, Ole
    Gorenflo, Rudolf
    Beyersdorff, Dirk
    Franiel, Tobias
    Plendl, Johanna (WE 1)
    Lüdemann, Lutz
    Magnetic resonance imaging : an international journal of basic research and clinical applications; 29(5) — S. 639–49
    ISSN: 0730-725x
    DOI: 10.1016/j.mri.2011.02.018
    Pubmed: 21546193
    Institut für Veterinär-Anatomie

    Koserstr. 20
    14195 Berlin
    Tel.+49 30 838 53555 Fax.+49 30 838-53480

    Abstract / Zusammenfassung

    Dynamic contrast-enhanced magnetic resonance imaging commonly uses compartment models to estimate tissue parameters in general and perfusion parameters in particular. Compartment models assume a homogeneous distribution of the injected tracer throughout the compartment volume. Since tracer distribution within a compartment cannot be assessed, the parameters obtained by means of a compartment model might differ from the actual physical values. This work systematically examines the widely used permeability-surface-limited one-compartment model to determine the reliability of the parameters obtained by comparing them with their actual values. A computer simulation was used to model spatial tracer distribution within the interstitial volume using diffusion of contrast agent in tissue. Vascular parameters were varied as well as tissue parameters. The vascular parameters used were capillary radius (4 and 12 μm), capillary permeability (from 0.03 to 3.3 μm/s) and intercapillary distances from 30 to 300 μm. The tissue parameters used were tortuosity (λ), porosity (α) and interstitial volume fraction (v(e)). Our results suggest that the permeability-surface-limited compartment model generally underestimates capillary permeability for capillaries with a radius of 4 μm by factors from ≈0.03 for α=0.04, to ≈ 0.1 for α=0.2, to ≈ 0.5 for α=1.0. An overestimation of actual capillary permeability for capillaries with a radius of 12 μm by a factor of ≥1.3 was found for α=1.0, while α=0.2 yielded an underestimation by a factor of ≈0.3 and α=0.04 by a factor of ≈ 0.03. The interstitial volume fraction, v(e), obtained by the compartment model differed with increasing intercapillary distances and for low vessel permeability, whereas v(e) was found to be estimated approximately accurately for P=0.3 μm/s and P=3.3 μm/s for vessel distances <100 μm.