MR-guided radiotherapy: magnetic field dose effects

Abstract

At the UMC Utrecht, together with Elekta Oncology and Philips Research, we are developing a combined system of a 1.5 Tesla MRI scanner and a 6 MV linear accelerator for cancer treatment. In contrast to present online imaging methods, superior soft-tissue contrast will be achieved. The system will enable patient positioning based on the tumour itself. In the design, the patient will be irradiated in the presence of a 1.5 Tesla magnetic field. This will affect the dose distribution, due to the Lorentz force acting on the secondary electrons. These so-called magnetic field dose effects have been investigated using the Monte Carlo code GEANT4. In homogeneous media, the magnetic field causes a reduced build-up distance and a shifted, asymmetric penumbra. At tissue-air boundaries dose increase of 40% is shown, due to electrons returning into the phantom by arc-shaped trajectories. This phenomenon has been called ‘Electron Return Effect’ (ERE). The ERE will take effect at the distal side of the treatment beam and at the proximal side of interior air cavities within the patient. The ERE in the latter case can be compensated by using opposing beams. We validated the simulation results of GEANT4 by comparison to dose measurements at 0, 0.6 and 1.3 Tesla. We demonstrated that both the reduced build-up distance and the ERE are highly dependent on surface orientation. If the treatment beam extends over the edges of a phantom, the so-called lateral ERE causes a dose increase of 50% relative to the central axis dose. We showed that even for oblique incidence opposing beams still have a compensating effect. We also investigated the ERE at the distal side, the lateral ERE, ERE at air cavities and ERE at tissue-lung transitions for magnetic field values of 0.2, 0.75, 1.5 and 3 Tesla. Results show that the ERE is reduced by lower magnetic field strengths, in particular for small irradiation fields, the lateral ERE, small cavities and lung tissue. However, for large irradiation fields (10 cm) and large interior air cavities (3 cm), the ERE reaches considerable levels of exit dose increase for all magnetic field values. For air cavities within the patient near the target, multiple beams, although not necessarily opposing, do have a compensating effect on the ERE dose increase. The remaining dose inhomogeneities can be dealt with by IMRT. We designed an inverse treatment planning approach, to calculate optimized IMRT dose distributions in the presence of a magnetic field. We used this method to calculate optimized IMRT treatment plans for a prostate cancer, a laryngeal cancer and an oropharyngeal cancer at B = 0 and 1.5 Tesla. Results hardly showed any differences between B = 0 and 1.5 T in terms of target coverage and sparing of organs at risk. This thesis shows that radiotherapy treatment in the presence of a magnetic field is feasible.