• hadron therapy
• helium beam radiography
• range monitoring
• spatial resolution

The presented thesis deals with ion-beam radiography for the beam range monitoring in hadron therapy. The research work was done at the Deutsches Krebsforschungszentrum (DKFZ) in Heidelberg (Germany), during an Erasmus mobility program.
Introduction - In the treatment of malignant tumors the usage of ionizing radiation (radiotherapy) plays an important role, beside other techniques like surgery or chemotherapy. Hadron therapy is the branch of radiotherapy that uses accelerated protons and heavier ions in order to deliver the sufficient dose required to kill the tumor cells in the target volume. The depth-dose profile for ion beams is called Bragg curve and is characterized by low dose deposition in the entrance region and a high dose peak close to the end of beam range. Compared to the depth-dose profile of X-ray radiation used in conventional photon therapy, the Bragg curve shows high potential in providing a dose distribution more conformal to the target volume. Exploiting it for sparing healthy tissues requires an accurate treatment planning together with a precise positioning of the patient with respect to the beam source. Nowadays, the hadron therapy treatments are planned using stopping power (SP) maps of the patient obtained by the conversion of the Hounsfield units from X-ray computer tomography (CT) images. Because of the different physical interactions that ions and photons undergo while travelling through matter, the conversion between Hounsfield units and SP is not always unambiguous and can introduce errors (up to 3-3.4 %) in the prediction of the beam range. It has been shown that the accuracy of range prediction can be improved by using ion-beam radiography (iRad). This promising tool allows the direct measurement of the integrated SP of the target and can also be used for monitoring the correct positioning of the patient. One of the main challenges related to iRad is the poor spatial resolution (SR) due to the multiple Coulomb scattering (MCS) that ions undergo while travelling through matter.
Recently, in the German Cancer Research Center (DKFZ) in Heidelberg (Germany), an apparatus for iRad has been developed. In this thesis the candidate has implemented a novel technique to improve SR and the new performance of the apparatus have been investigated.
Methods - The iRad apparatus is a single-ion tracking system composed of thin silicon pixelated detectors, with a sensitive area of 2 cm^2 and pixel-size of 55 micrometer. The apparatus consists of two tracking systems (one in front and one in the rear of the target) and a downstream energy detection system. A helium-ion beam is used as imaging radiation because it undergoes less MCS than a proton beam. The energy of the incoming ion beam (imaging radiation) is chosen such that the steep rising part of the Bragg curve of the beam is placed on the energy detector. In this way, small changes in the SP along particle path are detected as big changes in the energy deposited in the energy detector.
The raw data of the detectors are processed and signals not due to helium ions are rejected. In order to track single ions along all the detection system, signals from the different detectors are matched by using a self-written matching routine. Track informations are used to extrapolate both the position and the direction of ions when impinging and exiting the phantom. The path of each ion inside the target is estimated with most likely path methods and images are reconstructed through a back-propagation algorithm. Radiographies of an edge phantom are used to evaluate the SR and the contrast-to-noise ratio (CNR) of the images. The imaged phantom consists of a PMMA parallelepiped 160mm thick (typical head size) with a 2mm air edge in the middle.
In order to increase the SR, a novel technique is proposed: to decrease the MCS that ions undergo in matter, the energy of the helium beam is increased from 677MeV to 882MeV in four steps. For beam energies > 677MeV, an energy degrader (plate of high density material called build-up material (BUM)) is placed between the rear tracking system and the energy detection system in order to. compensate for the deeper position of the Bragg peak. For each energy step investigated, the SR, the CNR and the absorbed dose are evaluated through simulations and/or measurements. An increase of SR and a decrease of CNR and absorbed dose are expected in relation to the increase of the beam energy.
Results - The performance of the apparatus in reconstructing particles path have been improved by using a most likely path approach. The detection of undesired signals (fragments or detector artifacts) is characterized for all the detectors and all the beam energies used. A decrease of detector artifacts by 28% is found when the beam energy is increased from 677MeV to 882MeV. The performance of the matching algorithm is studied as function of the intensity of the helium beam and a mean matching efficiency of 57% is found.
For a helium-ion beam energy of 677MeV, the measured SR is 0.52(1) lp/mm. By increasing the beam energy up to 882MeV, an SR increase equal to the 33(8)% of the initial value is found. The CNR dependence as function of the increasing beam energy is investigated for radiographies reconstructed using the same number of helium ions and for radiographies with the same absorbed dose in the phantom. In case of a fixed number of helium ions (10e5), the CNR is found to decrease by 49(9)% and the absorbed dose to decrease by 35(3)%. In case of a fixed amount of absorbed dose (350 microGy), the CNR decreases by 33(18)%.
For the first time, a calibration curve for the apparatus is established, allowing direct measurement of the integrated SP. Performance of the apparatus is tested by imaging an air cavity with a diameter of 6mm in a PMMA phantom 160mm thick. The measured and expected values for the integrated SP are compared and a maximum deviation of the 0.9% is found.
Discussion - The novel technique proposed in this thesis for SR improvement exploits higher beam energy and suitable energy degrader. The techniques has been proven to increase SR by 33(8)% when the helium beam energy is increased from 677MeV to 882MeV. For radiographies reconstructed with the same number of ions, the absorbed dose is found to decrease by 35(3)%.
The CNR decrease associated to the technique leads still to high values of CNR (>5), even when a small inhomogeneity (1.2% of the phantom thickness) is imaged.
After making the described improvements on the apparatus, a direct measurement of integrated SP has shown an accuracy below 0.9%. Such a level of accuracy confirms iRad as an attractive tool for supporting the treatment planning in hadron therapy, where the error associated to beam range prediction can currently be up to 3.4%.