Proton therapy is a relatively new technique in the field of radiation oncology. The advantage of using protons can be illustrated by the depth-dose relation, which results in a more concentrated dose at a specific depth and thus potentially less dose in the surrounding healthy tissue compared to conventional photon therapy. However, this depth-dose relation also makes the dose delivery very sensitive to small geometric uncertainties. Because of this sensitivity, high accuracy of the quality assurance (QA) is essential.
QA can be used to test robustness of the system when dealing with small geometric uncertainties such as air gaps. Most of the QA protocols use phantoms to test the clinical treatment protocol or the complete treatment chain. To accurately simulate a patient, the phantom should resemble the human anatomy as well as tissue composition its interaction properties with ionizing radiation. In order to be optimally test the accuracy of the treatment system, the phantom should include small air gaps, density gradients in soft and bone tissue-substitutes and millimeter-scale structures. Currently phantoms are produced using casting techniques, which limits the possibilities to include small features or density gradients. This causes the phantoms available to lack the level of detail required for proton therapy QA.
One important source of errors in current treatment planning is the usage of a mono energy CT-scan of the patient for treatment planning, which measures the photon attenuation and converts this to Hounsfield Units (HU). For proton therapy, the HU is converted to the proton stopping power ratio (SPR) of the tissue compared to water, using a HU-SPR conversion model. Since there is no one-to-one relation between HU and SPR, this inevitably leads to errors, for example when tissues with the same HU have a slightly different SPR. Therefore dual-energy CT (DECT) has been proposed as a replacement for conventional mono energetic CT in proton treatment planning. The information acquired by the DECT is used in the Bethe formula to calculate the SPR directly, making the HU-SPR calibration curve obsolete, thus improving the accuracy of treatment planning.
The goal of this study is to explore the possibilities of designing an anthropomorphic phantom with small geometric features using a 3D-printer. By using a 3D-printer we can print the structure of the phantom at the millimeter-scale. Multiple materials can be mixed while printing, making it easier to adjust material properties such as density. Since DECT could improve proton treatment planning, we will aim to use materials that will be compatible with DECT-based HU-SPR conversion methods, so that these materials will be treated similar to that of human tissue by the treatment system. This way, the phantom should have the same SPR as human tissue, as well as covering small geometric details.