Evaluation of treatment planning system algorithms accuracy in prediction of hip prosthesis dose distribution using Monte Carlo simulation

Beam-hardening artifacts in CT image set of patient with a hip prosthesis cause difference between dose distributions resulted by treatment planning system (TPS) algorithms and actual dose distribution in patient body. In this study, dose distributions of TPS algorithms were compared with the results of Monte Carlo simulations of Titanium and Steal as a hip prosthesis for different photon energy.
Siemens Primus linear accelerator was modeled using the MCNPX code for 6MV and 15MV energy mode. Dose reduction due to Titanium and Steal were calculated by different algorithms of Isogray TPS and compared with the results of Monte Carlo simulations of Titanium and Steal. The accuracy of the algorithms for different depth and energy was evaluated.
In 6MV energy, for Titanium, there was 12.9%, 12.6% and 10.7% and for Steal, there was 14.3%, 14.1% and 12.1% differences between simulation results and TPS algorithms included Superposition, Collapsed cone and FFT Convolution respectively. In 15MV energy, for Titanium, there was 15%, 15.6% and 13.3% and for Steal, there was 21.7%, 22.3% and 20.1% differences between simulation results and TPS algorithms included Superposition, Collapsed cone and FFT Convolution respectively.
There was significant dose reduction due to hip prosthesis whereas the TPS algorithms do not have ability to predict the dose reduction accurately. Thus, use of Monte Carlo based algorithms recommend for dose calculation in patient with high density prosthesis.