Room: AAPM ePoster Library
Mixed beam radiotherapy (MBRT) utilizes intensity modulated photon and electron apertures collimated by the photon MLC and was shown to be dosimetrically superior than VMAT for superficial targets. This work aims to improve and evaluate computational efficiency of a Monte Carlo (MC) treatment planning process (TPP) for MBRT towards clinical usability and to validate the deliverability of generated plans.
Following features are implemented in the MC-TPP: use of pre-calculated pre-patient phase-spaces for beamlet MC dose calculation, adapted statistical uncertainties of beamlet dose distributions, sparse representation of beamlet dose distributions, voxel merging, normal tissue objectives, dynamic stopping criteria for optimization, particle recycling for final MC dose calculation and automatic xml-file generation. The impact of these features on computational efficiency and plan quality is evaluated by generating MBRT plans for a bladder and a head and neck case. For validation purposes, the enhanced MC-TPP is applied to an academic case of a water slab phantom and the generated xml-file is delivered on a TrueBeam to an EBT3-film. Furthermore, a log-file based dose re-calculation is performed. Both, log-file re-calculation and measurement are compared to the final dose calculation.
The computation time for optimization, beamlet and final dose calculation are reduced from 1h to 10min, 8h to 25min and 30min to 5min, respectively. Moreover, RAM-memory space for optimization is reduced from 17.5GB to 7GB. All these improvements have no substantial impact on treatment plan quality. The film measurement agreed with dose calculation with a passing rate >99% for a 3%/2mm gamma analysis with 10% threshold. The discrepancies between final calculated and log-file re-calculated dose are within statistical uncertainty of <1%.
The computational efficiency of a MC-TPP for MBRT and the deliverability of generated plans are successfully improved and validated, respectively. The MC-TPP is deemed to be adequate for clinical use.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by grant 200021_185366 of the Swiss National Science Foundation and Varian Medical Systems.