Proton therapy has received increasing recognition as a preferred cancer treatment modality due to its superior dose properties, which allows for placement of the dose-intense Bragg peak of protons anywhere inside the body and may also include intensity modulation. However, the ability to predict the Bragg peak location is currently limited due to inaccuracies in X-ray CT based calculation of the 3D proton stopping power map. Proton CT, registering the entry and exit path of individual protons and reconstructing a proton stopping power map relative to water using the energy loss of each proton, is a promising imaging modality that will overcome this limitation. It may also provide a novel low-dose imaging modality that can give information not only on small density differences in tissues but also its atomic composition. In this presentation, an overview of the proton CT imaging method currently under development will be given with emphasis on fast and efficient reconstruction algorithms, which are crucial for the practical implementation of proton CT imaging. Our recent research in this area has focused on investigating and computationally fine-tuning the implementation of block-iterative projections (BIP) and string-averaging projections (SAP) in combination with total variation (TV) superiorization techniques that stir the iteration of the reconstructed proton relative stopping power maps towards a superior solution. Lastly, an outlook of how these developments could also advance the state-of-the-art in intensity modulated proton therapy (IMPT) and other inversely planned radiation therapy modalities will be given.
Joint work with Scott N. Penfold, Yair Censor
Back to Mathematical Problems, Models and Methods in Biomedical Imaging