Abstract:
The presence of oxygen deprived (hypoxic) tissue regions in human tumours is associated with a reduced effectiveness of different treatment modalities. This effect is especially important in radiotherapy (RT) using photons with energies in the MeV range. In order to achieve high control probabilities with RT also for hypoxic tumours, it has been proposed to increase the radiation dose in hypoxic regions. For this purpose a three-dimensional assessment of the distribution and degree of hypoxia is required. This is possible with positron emission tomography (PET) in combination with the tracer 18F-Fluoromisonidazole (FMISO), as its accumulation is increased at low oxygen partial pressures (PO2).
FMISO PET has a number of fundamental limitations: Its low spatial resolution (millimetres) results in large averaging effects. Moreover, FMISO accumulates in vital cells only and its diffusive distribution in tissue is slow. In this thesis, the effects of these limitations on clinical FMISO PET imaging were investigated using computer simulations. The coupled diffusion-reaction processes of O2 and FMISO on the microscopic length scale were incorporated in a mathematical model. The resulting partial differential equations were treated numerically using the finite element method. The flux of molecules across vessel membranes into and out of tissue was modelled by biologically realistic boundary conditions. For this purpose vessel distributions were created based on tumour tissue microsections. Applying this method, clinically relevant questions regarding the relationship between microscopic O2 distributions and the resulting macroscopic FMISO PET signals.