Improved thermal modeling of the prostate by 3D perfusion imaging

Full text: In hyperthermia the Pennes' bio-heat or heat-sink model is often the preferred thermal model due to its simplicity and usefulness. Knowledge about the local volumetric perfusion is important in this model, however, it is mostly not known. Consequently, assumptions about organ specific perfusion rates have to be made. Furthermore, the Pennes' bioheat model fails to consider the significant impact of the discrete vasculature. For this reason, discrete vessel models have been developed that allow separate description of the vasculature up to the capillaries. Acquisition of the vasculature at such detailed level has proved to be difficult. However, lately with the advent of sophisticated dynamic contrast enhanced (DCE) imaging techniques, it has become possible to image complete 3D perfusion maps and detailed angiograms. This information can be employed for further refinement in thermal modelling. In this work this fact will be demonstrated for regional hyperthermia treatment of the prostate. A comparison is made between thermal simulations using Pennes' bioheat heatsink model for a homogeneous prostate perfusion versus measured perfusion maps of patients. Furthermore, the effect of preheating of blood in the vessels supplying the prostate will be investigated using a detailed discrete vessel model of the pelvis. Perfusion imaging showed that prostate tumors often are characterized by an elevation in perfusion. Consequently, a thermal model which assumes a mean homogeneous prostate perfusion, results in too optimistic tumor temperature. Pre-heating of blood in the supplying arteries was found to be marginal and has therefore only a small effect on the prostate temperature. Refs. 3 (author)