A Computerised treatment planning system for synchrotron radiotherapy

Synchrotron Radiotherapy (SyncRT) is a pre-clinical technique that delivers high flux synchrotron X-rays to in vivo and in vitro cancer models. The synchrotron beam is characterised by a mean energy of approximately 100 keV and a small source size on the order of 10s-100s of microns. These physical properties of the synchrotron X-ray beam permit the delivery of radiation fields which are minimally divergent over distances of 1-2 metres.

The very high flux synchrotron beam can deliver dose rates on the order of 100s to 1000s of Grays per second in water. Microbeam Radiotherapy (MRT) exploits the unique nature of the synchrotron X-rays by collimating the beam into an array (or lattice) of spatially fractionated planar beamlets that are typically 50 𝜇m wide and separated 400 𝜇m centre-to-centre.

In animal studies, healthy tissues have shown remarkable tolerance to very high doses of microbeam radiation (called peak doses) when the doses between the microbeams (called valley doses) are kept relatively low. Conversely, cancerous tissue does not tolerate the MRT dose distribution and, therefore, tumour control may be obtained while sparing the surrounding healthy tissue. SyncRT has been used to treat xenograft cancers in mice and rats. Normal tissue tolerance to SyncRT has also been studied in piglets, rabbits and other small-animal models. Veterinary trials on dogs bearing spontaneous tumours are underway at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France.

SyncRT on humans has been limited to low-dose rate Synchrotron Broadbeam Radiotherapy (SBBR) using a monochromatic spectrum. In order for human clinical trials in high dose rate SBBR and MRT to commence, synchrotron radiotherapy beamlines must transition from a research environment into a clinical environment that best reflects the current standards and protocols for conventional radiotherapy. This shift from research to clinical mode requires considerations for patient safety and positioning, image guidance, record-and-verify systems and treatment planning. In particular, a robust Treatment Planning System (TPS) that can accurately predict the synchrotron dose, and includes the features and functionality familiar to the radiation oncology community is essential.

The Eclipse TPS from Varian Medical Systems, Inc. is widely implemented in cancer treatment centres around the world. As commercial clinical software, Eclipse has had decades of research and development dedicated towards patient treatment planning in radiotherapy. Furthermore, Eclipse provides research-licensed workstation for physicists and clinical scientists to develop and implement custom dose calculation algorithms. In this thesis, we used the Eclipse research license to develop a TPS for SyncRT. The TPS includes all of the useful features for treatment planning including Computed Tomography (CT) import/export capabilities, organ and tumour contouring, beam configuration, dose prescribing, and critically, an accurate dose calculation algorithm for MRT.

The workflow using Eclipse for SyncRT is virtually identical to the clinical setting. The treatment plans produced by the TPS are delivered on the synchrotron beamline by simply inputting the calculated plan’s Monitor Unit (MU)s into the beamline’s controls system. For SyncRT the MU is the exposure time required to deliver the prescribed synchrotron dose to the target. Dose calculation is achieved via a Hybrid Monte Carlo (MC) photon tracking simulation paired with a convolution algorithm for analytically calculated electron dose kernels. The Hybrid algorithm is able to calculate SyncRT dose distributions in patient CT datasets, with calculation times comparable to MC-based algorithms in the clinic.

Extensive dosimetry experiments were performed at the Australian Synchrotron’s Imaging and Medical Beamline (IMBL) in order to validate the TPS for SyncRT veterinary and clinical trials. Dosimetric measurements were acquired using PTW PinPoint ionisation chambers, microDiamond detectors, and HD-V2 radiochromic film in Solid Water and heterogeneous tissue substitute phantoms. The SyncRT treatment planning calculations compared well with the dosimetric measurements to within 5% for SBBR doses.

In addition, a near-complete MC model of the IMBL was developed and validated to within 5% for SBBR doses and to within 7.5% for MRT field sizes of 20 × 20 mm2. The MC model was developed as a tool for verifying treatment planning dose calculation algorithms in SyncRT, especially for MRT, where dosimetry is onerous owing to the highly non-uniform dose distribution. For MRT, the TPS was compared against the validated MC model of the IMBL and showed excellent agreement in water to within 3%. The TPS also compared well with the MC model for heterogeneous phantoms to within 5-8% for materials that are radiologically close to water (muscle, adipose tissue, soft tissue, and skin). Regions of dense cortical bone showed significant disagreement due to limitations in the CT calibration method used to convert Hounsfield Unit (HU)s to material compositions, and these materials should be carefully considered in MRT treatment planning.

The Eclipse SyncRT TPS achieves clinical grade treatment planning on synchrotron bio-medical beamlines and is suitable for veterinary trials on the IMBL at the Australian Synchrotron. Furthermore, the TPS is easily adapted for other beamlines and is now available on the ESRF’s ID17 bio-medical beamline, where it is expected to facilitate ongoing veterinary trials in MRT.