Nuclear methods for real-time range verification in proton therapy based on prompt gamma-ray imaging

Accelerated protons are excellent candidates for treating several types of tumours. Such charged particles stop at a defined depth, where their ionisation density is maximum. As the dose deposit beyond this distal edge is very low, proton therapy minimises the damage to normal tissue compared to photon therapy. Nonetheless, inherent range uncertainties cast doubts on the irradiation of tumours close to organs at risk and lead to the application of conservative safety margins. This constrains significantly the potential benefits of proton over photon therapy and limits its ultimate aspirations. Prompt gamma rays, a by-product of the irradiation that is correlated to the dose deposition, are reliable signatures for the detection of range deviations and even for three-dimensional in vivo dosimetry. In this work, two methods for Prompt Gamma-ray Imaging (PGI) are investigated: the Compton camera (Cc) and the Prompt Gamma-ray Timing (PGT). Their applicability in a clinical scenario is discussed and compared. The first method aspires to reconstruct the prompt gamma ray emission density map based on an iterative imaging algorithm and multiple position sensitive gamma ray detectors. These are arranged in scatterer and absorber plane. The second method has been recently proposed as an alternative to collimated PGI systems and relies on timing spectroscopy with a single monolithic detector. The detection times of prompt gamma rays encode essential information about the depth-dose profile as a consequence of the measurable transit time of ions through matter. At Helmholtz-Zentrum Dresden-Rossendorf (HZDR) and OncoRay, detector components are characterised in realistic radiation environments as a step towards a clinical Cc. Conventional block detectors deployed in commercial Positron Emission Tomography (PET) scanners, made of Cerium-doped lutetium oxyorthosilicate - Lu2SiO5:Ce (LSO) or Bismuth Germanium Oxide - Bi4Ge3O12 (BGO) scintillators, are suitable candidates for the absorber of a Cc due to their high density and absorption efficiency with respect to the prompt gamma ray energy range (several MeV). LSO and BGO block detectors are compared experimentally in clinically relevant radiation fields in terms of energy, spatial and time resolution. On a different note, two BGO block detectors (from PET scanners), arranged as the BGO block Compton camera (BbCc), are deployed for simple imaging tests with high energy prompt gamma rays produced in homogeneous Plexiglas targets by a proton pencil beam. The rationale is to maximise the detection efficiency in the scatterer plane despite a moderate energy resolution. Target shifts, increase of the target thickness and beam energy variation experiments are conducted. Concerning the PGT concept, in a collaboration among OncoRay, HZDR and IBA, the first test at a clinical proton accelerator (Westdeutsches Protonentherapiezentrum Essen) with several detectors and heterogeneous phantoms is performed. The sensitivity of the method to range shifts is investigated, the robustness against background and stability of the beam bunch time profile is explored, and the bunch time spread is characterised for different proton energies. With respect to the material choice for the absorber of the Cc, the BGO scintillator closes the gap with respect to the brighter LSO. The reason behind is the high energies of prompt gamma rays compared to the PET scenario, which increase significantly the energy, spatial and time resolution of BGO. Regarding the BbCc, shifts of a point-like radioactive source are correctly detected, line sources are reconstructed, and one centimetre proton range deviations are identified based on the evident changes of the back projection images. Concerning the PGT experiments, for clinically relevant doses, range differences of five millimetres in defined heterogeneous targets are identified by numerical comparison of the spectrum shape. For higher statistics, range shifts down to two millimetres are detectable. Experimental data are well reproduced by analytical modelling. The Cc and the PGT are ambitious approaches for range verification in proton therapy based on PGI. Intensive detector characterisation and tests in clinical facilities are mandatory for developing robust prototypes, since the energy range of prompt gamma rays spans over the MeV region, not used traditionally in medical applications. Regarding the material choice for the Cc: notwithstanding the overall superiority of LSO, BGO catches up in the field of PGI. It can be considered as a competitive alternative to LSO for the absorber plane due to its lower price, higher photoabsorption efficiency, and the lack of intrinsic radioactivity. The results concerning the BbCc, obtained with relatively simple means, highlight the potential application of Compton cameras for high energy prompt gamma ray imaging. Nevertheless, technical constraints like the low statistics collected per pencil beam spot (if clinical currents are used) question their applicability as a real-time and in vivo range verification method in proton therapy. The PGT is an alternative approach, which may have faster translation into clinical practice due to its lower price and higher efficiency. A proton bunch monitor, higher detector throughput and quantitative range retrieval are the upcoming steps towards a clinically applicable prototype, that may detect significant range deviations for the strongest beam spots. The experimental results emphasise the prospects of this straightforward verification method at a clinical pencil beam and settle this novel approach as a promising alternative in the field of in vivo dosimetry.