Quantitative diffusion-weighted imaging in breast and liver tissue

    Research output: ThesisThesis fully internal (DIV)

    1004 Downloads (Pure)


    Diffusion-weighted imaging (DWI) is a biomedical application of MRI, based on diffusion (random walk) of hydrogen protons in the human body. By visual inspection of DWI, and other MRI images, the radiologist determines whether there is any pathology present in the organ examined.

    Visual assessment is subjective. Quantitative analysis provides an objective alternative for assessing the diffusion of pathology by expressing it in a number in comparison with healthy tissue. The benefit of this approach is that it does not depend on the reader.

    Two models for quantitative analysis of diffusion in the liver and breast were investigated. The intravoxel incoherent motion (IVIM) model separates diffusion into slow and fast diffusion. Fast diffusion is present in highly perfused tissues, such as tumors. The second model describes only one parameter, the apparent diffusion coefficient (ADC). Benefits and drawbacks of both models were investigated in DWI studies of the liver. Diffusion is dependent on the hepatic fat fraction, and the ADC furthermore on the measurement location in the liver.

    A novel method was introduced for semi-automatic selection and analysis of breast lesions. This provides better discrimination between benign and malignant breast lesions compared to the ADC. Moreover, the method appeared independent of the reader.

    In daily practice, a group of patients suspected of breast cancer is subject to an invasive procedure to provide a diagnosis. The number of invasive procedures may be reduced in the future when IVIM is implemented in the diagnostic algorithm.
    Original languageEnglish
    QualificationDoctor of Philosophy
    Awarding Institution
    • University of Groningen
    • Oudkerk, Matthijs, Supervisor
    • Sijens, Paul, Co-supervisor
    Award date2-Nov-2016
    Place of Publication[Groningen]
    Print ISBNs978-90-367-9097-0
    Electronic ISBNs978-90-367-9096-3
    Publication statusPublished - 2016

    Cite this