Program Information
Improving Time-Efficiency of MRT Dose Calculations by Implementing a Hybrid Dose Calculation Approach
M Donzelli1, 3*, S Bartzsch2, 3 , E Braeuer-Krisch1 , U Oelfke3 , (1) European Synchrotron Radiation Facility, Grenoble, France, (2) Technical University Munich, Munich, Germany (3) The Institute of Cancer Research, Sutton, UK
Presentations
TU-D-205-6 (Tuesday, August 1, 2017) 11:00 AM - 12:15 PM Room: 205
Purpose: The steep dose gradients present in orthovoltage microbeam radiation therapy (MRT) require dose calculations on the micrometer scale for treatment planning. A pure Monte Carlo (MC) approach to this problem leads to calculation times in the order of days even on a computing facility with hundreds of cores. A more time-efficient dose kernel convolution based approach has been proposed (Bartzsch et al., Med. Phys. 2013) and implemented. However, the assumption of local homogeneity for the transport of scattered keV photons causes systematic errors in proximity to material interfaces.
Methods: This work presents the new dose calculation algorithm hybrid dose (hyD) which makes use of MC photon transport, while calculating the microscopic dose distributions arising from secondary electrons with the kernel convolution approach. The new concept combines the advantages of both methods – precise MC photon transport and fast and precise kernel-based electron transport – without inheriting their downsides.
Results: A prototype software has been implemented and compared to full MC calculations (Geant4), showing good agreement in homogeneous and inhomogeneous phantoms. The accuracy of the calculation is the same as for full MC. A successful application of the software prototype in treatment planning for veterinary MRT trials shows the potential of hyD for future human trials. The calculation time of ~45 minutes for a full treatment plan on one single CPU (28 threads) is reduced by two orders of magnitude when compared with the equivalent MC simulation.
Conclusion: The presented algorithm is an innovative advancement for future treatment planning in MRT which reduces computing time without compromising the quality of the calculated dose distribution.
Funding Support, Disclosures, and Conflict of Interest: The authors received support from the ANR EPIRAD grant (ANR-13-BSV1-0012), Cancer Research UK (C57410/A21787), and the COST action SYRA3 (TD1205).
Contact Email: