14-15 Jun 2018 Lyon (France)
A silicon pixel detection system for helium-beam radiography
Tim Gehrke  1, 2, 3@  , Carlo Amato  1, 2, 4  , Maria Martisikova  1, 2  
1 : Department of Medical Physics in Radiation Oncology, German Cancer Research Center (DKFZ)
2 : Heidelberg Institute for Radiation Oncology (HIRO), National Center for Radiation Research in Oncology (NCRO)
3 : Department of Radiation Oncology, Heidelberg University Hospital
4 : Department of Physics, University of Pisa

A full exploitation of the potential of ion-beam therapy sets increased demands on dedicated quantitative imaging methods. Ion-beam radiography (iRad) is a promising modality due to its potential for measuring projections of the actual stopping power distribution of the patient in treatment position. However, iRad is not yet implemented in clinical applications. Current research topics include the development of dedicated detection systems and the investigation of different ion types as imaging radiation.

We built a small prototype of a fully pixelated silicon detection system, based on the Timepix technology developed at CERN. The system consists of two tracker units and a thin energy deposition detector that also enables the identification of ions. The performance of the system was experimentally assessed for imaging with helium ions and protons (αRad & pRad) at the HIT facility in Heidelberg, Germany. A special focus of the work was a comprehensive experimental comparison between αRad and pRad.

The developed methods of αRad and pRad were shown to provide a thickness resolution of 0.6 % (resolving a 1 mm thickness difference at an overall thickness of 160 mm) in head-sized PMMA phantoms at diagnostic imaging dose levels. For αRad the identification of the primary ions was found to be crucial to reach this thickness resolution. The ion identification enabled the rejection of image noise that originates from secondary hydrogen ions. This measure increased the contrast-to-noise ratio (CNR) by at least 150 % with respect to an image formation without ion identification. Evaluating the spatial resolution (SR) showed that αRad provides radiographs with an average SR of 0.46 lp/mm (MTF10%) for head-sized objects. This correponds to an improvement of the SR by 55 % compared to pRad performed with the same detection system under identical conditions. Deploying ion identification, the CNR of the measured αRads was shown to be equal to pRad at similar imaging doses.

In conclusion, the built prototype system for iRad reached clinically desired image qualities in simple PMMA phantoms. Furthermore, this study showed in experiments that αRad provides a better SR than pRad without disadvantages in terms of imaging dose or CNR.


Online user: 1