I Jornadas RSEF / IFIMED de Física Médica

Optimized Monte-Carlo based dose calculation for low energy X-rays intraoperative radiation therapy

No programado

Parque Científico. Salon de Actos Edificio de Cabecera (IFIC-Valencia)

Descripción

Introduction: Intraoperative Radiation Therapy with low energy X-rays (XIORT) is increasingly used in oncology (ex: INTRABEAM®, Carl Zeiss, and Axxent®, Xoft), predominantly for breast cancer treatments with spherical applicators [1], but also for other clinical applications such as kyphoplasty [2] with needle applicators and superficial intraoperative radiotherapy [3] with flat and surface applicators. This study proposes a fast and precise method to calculate Monte Carlo (MC) dose distribution from a previously stored full database of Monte Carlo monochromatic phase space files (PHSP) and depth dose profiles (PDD) for spherical, needle, flat and surface applicators. Methods: A detailed geometry of the device which describes the observed general features of the dose produced by standard INTRABEAM® applicators has been simulated with penEasy [4]. A set of monoenergetic PHSP, covering the range of the particle spectrum coming from the device, and the corresponding PDD, have been generated and stored, one for each energy up to 50 keV. This requires a computing time of several CPU-days. Each PHSP is binned and parametrized [5] in terms of the relevant variables to make them easy to manipulate. A generic spectrum composed by a Bremsstrahlung tail and characteristic X-rays, whose general features were derived from a realistic MC simulation of the X-ray source, is fine-tuned by means of a genetic algorithm [6] until it describes the experimental PDD of any given applicator. The previously simulated monoenergetic PHSP are combined with weights given by the derived energy spectrum, to build the resulting PHSP optimized to describe the dose distribution of the considered applicator. These two phases are performed for each individual applicator. From the final optimized PHSP, the dose is computed either by penEasy or by an in-house hybrid (MC and analytical) algorithm [7] which takes into account condensed history simulations of both photoelectric and Compton interactions for X-rays up to 50 keV. The whole process was validated against MC simulations as well as with radiochromic film dose measurements both in water and heterogeneous phantoms (bone, lung, air) for the spherical, needle, surface and flat applicators. Results: Building the PHSP file optimized to a particular depth-dose curve in water only takes a few minutes in a single core (i7@2.5 GHz), for all the applicators considered in this work. From that PHSP file, the hybrid Monte Carlo code is able to compute dose distributions within 5 minutes. For all the applicators, dose distributions computed with the proposed strategy are in good agreement with the Monte Carlo simulations performed with penEasy. Gamma index calculation in water shows than 95% of the voxels fulfill the dose distance criteria of 2%/1mm. Concerning the heterogeneous phantoms, more than 90% of the voxels fulfill the gamma index for 2%/1mm. Conclusion: The dose calculation process presented in this work is fast, flexible and accurate enough for XIORT planning. This method is implemented in Radiance® (GMV SA, Spain), a IORT Treatment Planning System for all INTRABEAM® (Carl Zeiss) applicators. [1] J.S. Vaidya et al. 2010. TARGIT-A trial. Lancet, 376, 91-102. [2] F. Schneider et al. 2011. Int J Radiat Oncol Biol Phys.81(4):1114-9 [3] F. Schneider et al. 2014. J Appl. Clin. Med. Phys. 15, 4502. [4] J. Sempau et al. 2011. Med. Phys. 38(11), 5887. [5] E. Herranz, et al. 2015. Phys. Med. Biol. 60(1):375-401. [6] C. Fernández-Ramírez et al. 2008. Phys. Rev. C 77(6), p. 065212. [7] M. Vidal et al. 2014. Radiother. Oncol. 111(S.1):117-118.