Establishing clinical cancer treatments with MR-guided adaptive radiotherapy

Abstract

The advances in image guided radiotherapy (IGRT) lead to an significant increase in treatment quality of cancer patients. However the guidance of radiotherapy based on X-Ray shows limitations and never translated into online adaptive radiotherapy. Therefore the new concept of MR-guided radiotherapy (MRgRT) was introduced, combining an linear accelerator with a magnetic resonance imaging (MRI). In this hybrid MR-linac system the higher soft tissue contrast of the MRI in combination with an optimization on the anatomy of the day promises increased treatment quality and new standard of practice. However, a MRI has before only supplied additional offline information for the annotation of tumor tissue and an online adaptive workflow was not part of clinical treatments. Therefore the challenges in regard to the magnetic field effect on dose distributions, algorithms and measurement devices as well as the problems due to the online adaptive workflow have not been investigated. In this work, the medical physics basis of MR-guided radiotherapy (MRgRT) was researched and concluded in safe clinical treatments of patients. In a first step the technical drawbacks were researched in a simulation study comparing the possible dose distribution for esophageal cancer between the MR-linac system and a modern state of the art linear accelerator. This simulation study showed the dosimetric drawbacks of the MR-linac system due to loss of conformality and size limitations. In addition we showed, that greater focus has to be put on mid- and low-dose areas and that the corresponding OARs must be included into optimization. However, we showed as well that by adaptation of the optimization, the clinical parameters can be fulfilled leading to feasibility of treatment plans. In a second step the only existing algorithm for calculation of dose distributions at the 1.5 T MR-Linac was compared to a detailed research Monte Carlo (MC) system, due to possible approximations on behalf of the magnetic field and attenuating layers of the MRI. In this investigation the detailed MC-simulation was designed to incorporate the material specific information of the accelerator and MRI and their corresponding cross sections for an exact dose calculation. With an iterative comparison to experimental data the linac specific free parameters were derived and the independent accelerator head model finalized for a validation of patient treatments. This validation showed, with gamma criterion of 3 mm and 3% a passing rate of 99.83 % and therefore validated that the approximations do not correspond to a significant change in dose distribution. In a third step new quality assurances were developed to allow treatment of patients on the MR-linac system. For an experimental measurement of dose deposition a novel workflow was developed and integrated, as the conventional systems showed high influences by the static magnetic field. Based on the characterization, a 2D ionization chamber inside a hexagonal phantom can accurately measure the applied dose and, by comparison with the respective simulation, assure correct machine calibration. For an comprehensive check of online adaptive plans within a treatment workflow an independent secondary dose calculation was established. Based on an independent build Monte Carlo accelerator head and a 3D comparison of dose distributions this secondary dose calculation can verify clinical plans in a median time of 1:23 min. In this comparison of dose distributions a gamma criterion validates the machine parameters and assures correct dose over the full patient anatomy. In the final step the acquired knowledge from the previous studies and developed quality assurances allowed first treatments of patients with online adaptive MRguided radiotherapy. In this dissertation the first worldwide MR-guided treatment of partial breast stands representative to the multitude of different tumor entities initiated at the MR-linac system. Equivalent to the investigations on ERE, ESE and geometric distortion in partial breast, further tumor specific magnetic field effects were evaluated for an assessment of potential risks. With this work the basis for a safe treatment with MR-guided radiotherapy (MRgRT) was developed. Therefore, as methods and respective handling of magnetic field effects are now established, new clinical concepts are being developed starting the next phase of MRgRT improving patient treatments in radiotherapy by deviating from current clinical practice.