Tomographic imaging with carbon ion beams

The sensitivity of ion therapy to inaccuracies in treatment planning and delivery procedures demands the development of imaging techniques capable of accurate assessment of the Relative (to water) ion Stopping Power (RSP). The investigated tomographic carbon ion CT (iCT) imaging setup consists of a prototype integration-mode detector working as range telescope using active scanning ion beam delivery. Extensive Monte Carlo simulations using the FLUKA code were performed and compared to experimental data. The influence of acquisition parameters was investigated to optimize the imaging dose. A dedicated post-processing method relying on the linear decomposition of the integrated signal was applied to solve ambiguities in the presence of material inhomogeneities due to the finite beam size. As comparison the performance of a single-particle list-mode detector was examined. For raw iCT images of a tissue-equivalent phantom, an average median error of 2.1% was observed, consistent with experimental data. An optimization of the acquisition parameters reduced the imaging dose exposure down to below 30 mGy without severe RSP accuracy degradation. The signal decomposition showed a considerable benefit reducing the average RSP error to around 1%, being comparable to the list-mode performance for
¹²
C ions. Similar observations were made for simulations using as patient model an X-ray CT of a head and neck patient case. This study quantitatively supports the benefit of carbon iCT for range verification, showing the performance of different detector modes for direct RSP assessment at acceptable dose levels.