Page 58 - Fisica In Medicina n° 1/2017

^Äëíê~Åí=cáëáÅá= =Commissioning of the Monte Carlo dose calculation algorithm for a CyberKnife systemA. MICALI*, C. SIRAGUSA*, A. BROGNA*, I. BONAPARTE*, M. C. ANGIOCCHI*, F. MIDILI*, E. MONGELLI*, I.IELO**U.O.C. di Fisica Sanitaria – A.O.U. Policlinico “G. Martino” - MessinaIntroductionAccurate dose calculation is essential in SBRT to understand dose levels to target volumes and organs at risk, especially for treatment planning involving heterogeneous patient anatomy. The Monte Carlo (MC) is indicated as the "gold standard" of dose calculation algorithms. It predicts the absorbed dose by simulating the radiation transport and it takes into account the electronic disequilibrium. In this work, MC dose calculation method for SBRT treatment planning was commissioned on a G4 CyberKnife (CK) for the fixed and the variable IRIS collimation systems.Materials and MethodsTPRs, OCRs and OFs were measured in water at different depth using the PTW 60017 diode. The commissioning procedure was performed using the Gaussian method, which simulates the source distribution through the FWHM of a Gaussian function. The source energy spectrum was determined from the measured central axis PDDs in water for a 60 mm cone defined at SSD=80 cm. Then, an iterative procedure, which involved the Energy Correction Factors (ECF) generation, was used to compare measured and simulated TPRs (1% uncertainty level). The fluence distribution is obtained from the measured profile at 25 mm depth without any cones. Then, an iterative procedure, which involved the Collimator Correction Factors (CCF) generation, was used to compare measured and simulated OCRs (1% uncertainty level). Finally measured and calculated OFs (0,2% uncertainty level) were compared.ResultsIn order to obtain the source distribution, a comparison between measured and simulated penumbra values for 5 and 60 mm cones at 100 mm depth was performed. FWHM=2,6 mm corresponds to the best agreement with the data for both the collimation systems. Satisfactory energy distributions were found when measured and simulated average TPRs were in agreement within 1% in all the49