Validation of a Monte Carlo simulation for Microbeam Radiation Therapy on the Imaging and Medical Beamline at the Australian Synchrotron

Andrew Dipuglia,Anatoly B Rosenfeld,Iwan M Cornelius,Marco Petasecca,Stéphanie Corde,Susanna Guatelli,Matthew Cameron,Andrew W Stevenson,Michael L F Lerch,Jeremy A Davis

doi:10.1038/s41598-019-53991-9

Andrew Dipuglia, Anatoly B Rosenfeld + Show 8 more

Open Access

https://doi.org/10.1038/s41598-019-53991-9

Copy DOI

Abstract

Microbeam Radiation Therapy (MRT) is an emerging cancer treatment modality characterised by the use of high-intensity synchrotron-generated x-rays, spatially fractionated by a multi-slit collimator (MSC), to ablate target tumours. The implementation of an accurate treatment planning system, coupled with simulation tools that allow for independent verification of calculated dose distributions are required to ensure optimal treatment outcomes via reliable dose delivery. In this article we present data from the first Geant4 Monte Carlo radiation transport model of the Imaging and Medical Beamline at the Australian Synchrotron. We have developed the model for use as an independent verification tool for experiments in one of three MRT delivery rooms and therefore compare simulation results with equivalent experimental data. The normalised x-ray spectra produced by the Geant4 model and a previously validated analytical model, SPEC, showed very good agreement using wiggler magnetic field strengths of 2 and 3 T. However, the validity of absolute photon flux at the plane of the Phase Space File (PSF) for a fixed number of simulated electrons was unable to be established. This work shows a possible limitation of the G4SynchrotronRadiation process to model synchrotron radiation when using a variable magnetic field. To account for this limitation, experimentally derived normalisation factors for each wiggler field strength determined under reference conditions were implemented. Experimentally measured broadbeam and microbeam dose distributions within a Gammex RMI457 Solid Water® phantom were compared to simulated distributions generated by the Geant4 model. Simulated and measured broadbeam dose distributions agreed within 3% for all investigated configurations and measured depths. Agreement between the simulated and measured microbeam dose distributions agreed within 5% for all investigated configurations and measured depths.

Highlights

® measured broadbeam and microbeam dose distributions within a Gammex RMI457 Solid Water phantom were compared to simulated distributions generated by the Geant[4] model
At the Australian Synchrotron (AS), Microbeam Radiation Therapy (MRT) is delivered via a brilliant, synchrotron-generated x-ray beam, with a photon spectral energy range extending from 40–300 keV19,20, which is highly polarised in the electron orbital plane[21,22]
Geant[4] provides the essential features needed to satisfy the requirements of MRT simulations: the ability to model pseudo time-dependent geometries required for dose delivery in MRT; a selection of photon and electron physics models both with and without polarisation that have been validated for common materials in the energy ranges relevant for radiotherapy applications[37,38]; the ability to model complex geometries via a computer-aided design (CAD) interface[39]; and the ability to import data using the DICOM interface to read-in patient-specific geometries acquired by CT scans[40,41]

Summary

Introduction

® measured broadbeam and microbeam dose distributions within a Gammex RMI457 Solid Water phantom were compared to simulated distributions generated by the Geant[4] model. Geant[4] provides the essential features needed to satisfy the requirements of MRT simulations: the ability to model pseudo time-dependent geometries required for dose delivery in MRT (due to the stationary beam requiring vertical patient translation); a selection of photon and electron physics models both with and without polarisation that have been validated for common materials in the energy ranges relevant for radiotherapy applications[37,38]; the ability to model complex geometries via a computer-aided design (CAD) interface[39]; and the ability to import data using the DICOM interface to read-in patient-specific geometries acquired by CT scans[40,41]

Methods

Results

Conclusion