Skip to ContentSkip to Navigation
KVI - Center for Advanced Radiation TechnologyOnderzoek en onderwijsMedical PhysicsResearch projects

Converting Dual Energy CT images into proton stopping powers

The agreement between the actual and planned dose distributions in charged particle therapy of tumours critically depends on accurate charged particle stopping powers in the tissues traversed by the charged particles, which are determined by the electron densities with corrections due to the elemental composition and molecular structure. The current uncertainty in the proton range caused by the uncertainties of the stopping powers deduced from CT-images amounts, depending on the complexity of the tumour surroundings, up to about 3 %, i.e. 3 – 6 mm for typical depths of tumours. This uncertainty imposes significant constraints on the treatment planning process (e.g. limitation of possible fields due to critical organs, sub-optimal reduction of dose in healthy tissue)

Particle therapy treatment plans are based on CT-information. Apart from all kind of artefacts due to the non-monochromatic photon spectrum used in CT, the detection method currently used and the reconstruction techniques, the information contained in the CT image is a combination of electron density and elemental composition of the tissue. Using a priori knowledge on the anatomy and correlations between density and elemental composition the two contributions can be disentangled to a significant extent, thus allowing the determination of the stopping powers. However, the stopping powers obtained in this way are a limiting factor for the accuracy of particle therapy, in particular when the charged particles pass through materials with significant differences in density or elemental composition. Typical examples of this are the treatment of tumours in the head and neck region, in the lungs and in the lower abdomen. Image reconstruction artefacts for the growing group of patients with prostheses containing ceramic and/or metallic parts complicates the situation even more. In the head and neck region metallic inlays in teeth, present in almost any patient pose serious problems.

The aim of this research project is to study the potential of dual energy and spectral CT imaging to generate stopping powers with sub-percent accuracy and to develop the conversion algorithms needed to achieve this, so that the stopping power data is no longer a limiting factor for the quality of treatment plans in charged particle therapy. Dual energy CT is currently already commercially available but is not yet used for treatment planning; spectral CT is under development for applications such as cardio-vascular imaging using contrast agents.

The project encompasses extensive simulations of and experiments on phantoms of precisely known geometry and composition that in the final stage of project will be representative for the complex geometries encountered in practice. CT-data of the phantoms are generated experimentally with single and dual energy and possibly experimental spectral CT-scanners. From these data the stopping power of the charged particles is deduced using different algorithms for the conversion of the CT-data into electron density and effective atomic number and subsequently stopping power. The stopping powers are also obtained directly from the known composition and geometry of the phantoms. Experimental stopping powers are determined from proton stopping power measurements on the phantoms. Comparison with the calculated values obtained via the different routes make it possible to identify the origin of discrepancies and thus to improve the modelling.

Gammex 467 phantom in a dual source CT system for dual energy CT imaging at Siemens Medical Solutions, Forchheim, Germany
Gammex 467 phantom in a dual source CT system for dual energy CT imaging at Siemens Medical Solutions, Forchheim, Germany
Water phantom for measuring proton stopping powers in the proton beam line at KVI-CART.
Water phantom for measuring proton stopping powers in the proton beam line at KVI-CART.
Last modified:12 September 2018 10.18 a.m.