Abstract
The radiotherapy treatment of prostate cancer has improved substantially over the last years. It is now possible to accurately deliver complex dose distributions to the prostate, while limiting the dose to the surrounding healthy tissue. Nevertheless, while the treatment of prostate cancer has improved, there are still patients dying from
... read more
the disease. Evidence is emerging that prostate tumor recurrences mainly arise at the location of the primary tumor. This suggests that patients would benefit from an extra boost dose at the primary tumor lesion. To minimize the irradiation of healthy tissue, a focal boost to the gross tumor volume (GTV) inside the prostate has been proposed. Such an approach requires a proper delineation of the GTV. Diffusion weighted-imaging and dynamic contrast-enhanced MRI have a high sensitivity and specificity for tumor detection. How to use DWI and DCE-MRI for the purpose of tumor delineation prior to radiotherapy is the main question of this thesis. To this end, first was investigated if DWI and DCE-MRI provide similar information about the presence of tumor on a voxel level. In some patients the overlap between the two techniques is high. However, in most patients this overlap is low. Therefore, we came up with pragmatic approaches on how to deal with the perceived inconsistencies. In order to be able to validate MR based tumor delineations, we came up with a registration approach between MR images and histopathology, which is accurate enough to perform these validations. This registration method was used to validate a logistic regression model that we created in order to predict the tumor presence on a voxel level. This model had a good performance in predicting the presence of tumor on a voxel level. Nevertheless, some tumor voxels were not detected by the model. A further investigation into these voxels showed that mainly the voxels with low cell density and low vessel density were not detected. One could argue that the tumor regions with low cell and vessel density are also the least aggressive tumor regions. Many patients that are scheduled for radiotherapy treatment are also treated with hormonal therapy. This therapy may affect the conspicuity of the tumor tissue. Therefore, we investigated the effect of hormonal treatment on the conspicuity of prostate cancer. In patients who used hormonal treatment longer than 3 months prior to MR imaging, the tumor conspicuity is lower compared to patients who used hormonal treatment for a shorter time. In these patients different thresholds need to be set to distinguish tumor from healthy tissue. In conclusion, in this thesis some of the important questions regarding the use of functional MR imaging for focal boost strategies in the prostate have been answered. Furthermore, approaches were presented to face challenges which arise when using DWI and DCE-MRI in a radiotherapy setting. In this way, this work may help to shift from the more classical homogeneous prostate dose distributions towards personalized approaches in which each prostate is treated based on its individual tumor characteristics.
show less