Multi-modality radiotherapy in cervical cancer Impact on the 3D dose distribution

This thesis describes the impact of image and dose accumulation uncertainties on the cumulative 3D dose distribution in multimodality radiotherapy for cervical cancer.
Image distortions for brachytherapy planning on 3T magnetic resonance (MR) images were evaluated with an MRI-only method. We conclude that it is safe to use 3T MRI scanners for applicator reconstruction during brachytherapy planning, because the dosimetric implications of these image distortions were clinically insignificant.
We investigated the added value of non-rigid registration for the evaluation of cumulative external beam radiotherapy (EBRT) and brachytherapy (BT) dose volume histogram (DVH) parameters, for both the planned and delivered EBRT dose. With the current planning strategies, dose addition with non-rigid registration is not necessary and simple addition of DVH parameters is an adequate method to assess the toxicity from EBRT and BT.
We investigated uncertainties related to non-rigid registration when accumulating the dose to the rectum of multiple BT fractions. We showed that, due to incorrect registration, dose addition with non-rigid registration is problematic for doses that have steep gradients such as brachytherapy doses.
With the advent of novel techniques, such as MRI-guided radiotherapy, assessing the 3D dose distribution will become more important. For these new techniques with a high gradient dose distribution, it may be necessary to evaluate the cumulative dose using non-rigid registration. However, for the accumulation of 3D dose distributions with very steep gradients new registration algorithms/strategies need to be developed that take into account biomechanical properties.