Encrypted login | home

Program Information

Implications of Monte Carlo Dose Calculation for Structures of Very Small Volume Compared with Measurement

J McKay

J McKay1*, J Grimm2 , K Osterman3 , I Das3 , J Xue3 , (1) Erlanger Health System, Chattanooga, TN, (2) Johns Hopkins University, Baltimore, MD, (3) NYU Langone Medical Center, New York, NY


MO-RAM-GePD-TT-5 (Monday, July 31, 2017) 9:30 AM - 10:00 AM Room: Therapy ePoster Theater

Purpose: Dose calculation with Monte Carlo method may have large statistical errors for a small volume due to limited number of histories included. This study investigated the uncertainty of Monte Carlo (MC) dose calculation in relation to field size and target volume size.

Methods: A stereotactic heterogeneous phantom containing a small simulated lung tumor (1 cm diameter) at the center has been constructed to validate the Monte Carlo dose calculation in the Mulitplan TPS. The measurement was performed with an Exradin W1 plastic scintillator with a single AP beam from CyberKnife VSI. Dose was calculated with MC algorithm using a 0.5% uncertainty for each cone size. The Exradin W1 has a very small sensitive volume of 1 mm × 3 mm cylindrical detector. MC dose was calculated for the contoured W1 sensitive volume measuring 9 voxels as well as an expanded 1 cm spherical target. The mean dose for each calculation is also compared with the measurement.

Results: The central voxel uncertainty of MC dose calculation was plotted against cone sizes for two calculation volumes. The central voxel uncertainty with the additional 1 cm target structure is much smaller than for the 9-voxel volume due to the adequate trace histories included. Additionally, dose homogeneity within the 9-voxel target is reduced with the expanded target structure. The uncertainty increases linearly with field size. Both calculations resulted in a mean dose within 3% of measurement to the 9-voxel structure.

Conclusion: For the validation test of dose calculation using a very small sensitive volume, uncertainty can be reduced by expanding the target volume around detector and forcing an increase in number of history traces. We suggested a practical solution for the reliable comparison of calculation with measurement.

Contact Email: