Monte-Carlo simulation to optimize SPECT-hardware dedicated to in-beam control of particle therapy

Therapeutic irradiation with ions and protons is superior to a treatment with gammas with respect to tumour conformity of dose and damage of normal tissue. On the other hand, mispositioning of the patient or density changes in the treated volume may easily compromise the success of treatment.
For this reason, non-invasive, in-situ dose verification is necessary. The only clinically proven method up to now is Positron Emission Tomography [Eng04]. This method does not allow direct dose quantification due to limited angle artefacts and washout in the patient. Another promising approach is in-beam SPECT: the detection of prompt gamma-rays following nuclear interaction between the ions and the atoms of the penetrated tissue. Detection systems based on Compton-scattering and pair-conversion are now under investigation. So far, Compton-cameras have been used for diagnostic imaging in nuclear medicine and astronomy. In telescopes also pair-creation cameras are established [Zog04]. In contrast to this, in-beam dose verification systems have to manage the specific energy range and low count rates due to the patient dose.

A prototype of a Compton-camera is under construction at OncoRay [Kor10, Fie11, Kor11]. It consists of a scatter layer made from CZT and an absorption layer (LSO).
From the energy deposited in the detector planes a cone surface of all possible directions of the incident photon can be reconstructed [Sch11]. The expected energy distribution of the prompt gammas is calculated by means of simulations based on treatment plan data [Fie11].
To optimize the setup simulations are required. Therefore, the Monte-Carlo framework GEANT4 is used for:

(1) Analysis of the efficiency in dependence of the geometry;
(2) Study of background caused by backscattered photons and other secondary particles;
(3) Development of filters for event selection;
(4) Study parameters influencing the spatial resolution of the reconstructed image (multiple Compton-scattering inside a detector layer, escape of energy, intrinsic radioactivity of the detector material, Doppler-broadening);
(5) Creation of test data sets as input for the reconstruction.

For photon energies above 7 MeV the pair production is the dominant electromagnetic process in CZT. To use these events the Compton-camera might be combined with a pair-conversion camera by adding thin silicon layers in between to track the path of the electron and the positron produced during pair conversion. Exploiting this information the direction of the incident photon can be deduced.
Unfortunately, the angular resolution is heavily degraded especially for photons with rather low incident energies by small-angle scattering of the secondary particles and by the recoil of the nuclei when a pair production takes place [Gol11]. Both effects were studied whereas the latter is not included in the GEANT4 code version 9.3. This simulation serves as basis for the development of an iterative reconstruction algorithm dedicated to in-beam SPECT with a pair production camera. Apart from the analysis of efficiency, angular resolution and noise, the combination of a pair production and a Compton-camera is a challenging task.

Results on simulations for optimizing the setup and the refinement of the reconstruction algorithm for a clinical dose verification system will be presented.

[Eng04] W. Enghardt et al., Nucl. Instr. and Meth. A 525 (2004)
[Fie11] F. Fiedler et al., Nucl. Sci. Symp. IEEE (2011), accepted
[Gol11] C. Golnik et al., Nucl. Sci. Symp. IEEE (2011), accepted
[Kor10] T. Kormoll et al., Nucl. Instr. and Meth. A (2010)
[Kor11] T. Kormoll et al., Nucl. Sci. Symp. IEEE (2011), accepted
[Ric10] M.-H. Richard et al., IEEE Transactions on Nuclear Science (2010)
[Sch11] S. Schöne et al., Proc. Fully 3D Meeting (2011)
[Zog04] A. Zoglauer et al., Nucl. Sci. Symp. Conf. Rec. IEEE (2004)