Comparison of conventional and Monte Carlo dose calculations for prostate treatments

Monte Carlo (MC) calculations are rapidly finding their place in clinical dose assessments. We investigated conformal prostate dose distributions as calculated by MC, and compared them to several analytical dose calculations. The treatment distributions for twenty prostate cancer patients, treated with 18 MV 3D conformal radiation therapy, were retrospectively assessed. The BEAM code based on EGSnrc was used to model the beam from which phase space files were used as input into the XVMC algorithm. This was compared to conventional treatment planning system calculations (CADPLAN) with and without inhomogeneity corrections. Results indicate that the CADPLAN generalized Batho Power Law, modified Batho Power Law, and equivalent tissue-air ratio methods contain inaccuracies in calculated dose to 95 % of the prostate planning target volume of 3.5 %, 3.3 %, and 2.9 %, respectively. The greatest discrepancies in the organs at risk were seen in the bladder where the inhomogeneity correction methods all predicted that 50 % of the prescribed dose covered an average of 8.2 % more of the bladder volume than that predicted from the MC calculation. Water equivalent MC and water equivalent CADPLAN calculations revealed important discrepancies on the same order as those between heterogeneous MC and heterogeneous CADPLAN calculations. The data indicate that the effect of inhomogeneities is greater in the target volume than the organs at risk, and that accurately modeling the dose deposition process is important for each patient geometry, and may have a greater impact on the dose distribution in the prostate region than correcting an analytical algorithm for the presence of inhomogeneities.