Accuracy assessment of pencil beam and Monte Carlo dose calculation algorithms during IMRT and VMAT delivery

Abstract

Intensity Modulated Radiation Therapy (IMRT) and Volumetric Modulated Arc Therapy (VMAT) are techniques that have been widely used over the last years for treating cancer due to better target coverage and the lower dose to the critical organs. This study assesses the accuracy of the Pencil Beam (PB) and the Monte Carlo (MC) calculation algorithms commonly used to determine the dose distribution in such applications. IMRT and VMAT plans were created for ten prostate patients using a commercially available treatment planning system (Monaco, Elekta Inc.). All plans were optimized to be clinically acceptable and the final dose calculation was performed using a) the PB and b) the MC calculation algorithm. The plans were evaluated in the central coronal plane using the I'mRT Matrixx (IBA Dosimetry GmbH) 2D-array placed in a solid water phantom a) by setting all control points to zero gantry degree and b) by leaving the beam orientation as planned. Evaluation was performed with OmniPro-I'mRT software using the gamma index method with acceptance criteria of 3% dose difference and 3 mm distance to agreement taking into account all points with dose values greater than the 10% of the maximum dose. Results show that PB calculation algorithm can accurately determine the dose distribution applied to the I'mRT Matrixx 2D array when all control points were set to zero gantry degree for both IMRT and VMAT plans (average passing rates of 96.6 and 94.7% of the points passed the 3%/3 mm acceptance criteria, for IMRT and VMAT, were observed, respectively). On the other hand, considerable differences between the different calculation algorithms were observed when the actual (i.e. the planned) beam orientation was used for the plan verification. In this case, the PB algorithm was found adequate for dose determination only for the IMRT plan verification (an average passing rate of 91% was observed). For the VMAT plans, MC was found significant superior compared to the PB dose calculation algorithm (average passing rates of 73.5 and 93% for the PB and the MC algorithms, were observed, respectively).