Dosimetric evaluation of photon dose calculation under jaw and MLC shielding

The accuracy of photon dose calculation algorithms in out-of-field regions is often neglected, despite its importance for organs at risk and peripheral dose evaluation. The present work has assessed this for the anisotropic analytical algorithm (AAA) and the Acuros-XB algorithms implemented in the Eclipse treatment planning system. Specifically, the regions shielded by the jaw, or the MLC, or both MLC and jaw for flattened and unflattened beams have been studied.

The accuracy in out-of-field dose under different conditions was studied for two different algorithms. Measured depth doses out of the field, for different field sizes and various distances from the beam edge were compared with the corresponding AAA and Acuros-XB calculations in water. Four volumetric modulated arc therapy plans (in the RapidArc form) were optimized in a water equivalent phantom, PTW Octavius, to obtain a region always shielded by the MLC (or MLC and jaw) during the delivery. Doses to different points located in the shielded region and in a target-like structure were measured with an ion chamber, and results were compared with the AAA and Acuros-XB calculations. Photon beams of 6 and 10 MV, flattened and unflattened were used for the tests.

Good agreement between calculated and measured depth doses was found using both algorithms for all points measured at depth greater than 3 cm. The mean dose differences (± 1SD) were -8% ± 16%, -3% ± 15%, -16% ± 18%, and -9% ± 16% for measurements vs AAA calculations and -10% ± 14%, -5% ± 12%, -19% ± 17%, and -13% ± 14% for Acuros-XB, for 6X, 6 flattening-filter free (FFF), 10X, and 10FFF beams, respectively. The same figures for dose differences relative to the open beam central axis dose were: -0.1% ± 0.3%, 0.0% ± 0.4%, -0.3% ± 0.3%, and -0.1% ± 0.3% for AAA and -0.2% ± 0.4%, -0.1% ± 0.4%, -0.5% ± 0.5%, and -0.3% ± 0.4% for Acuros-XB. Buildup dose was overestimated with AAA, while Acuros-XB gave results more consistent with measurements. From RapidArc plan analysis the average difference between calculation and measurement in the shielded region was -0.3% ± 0.4% and -2.5% ± 1.2% for AAA and Acuros-XB, respectively, relative to the mean target dose value (1.6% ± 2.3%, -12.7% ± 4.0% if relative to each local value). These values were compared with the corresponding differences in the target structure: -0.7% ± 2.3% for AAA, and -0.5% ± 2.3% for Acuros-XB.

The two algorithms analyzed showed encouraging results in predicting out-of-field region dose for clinical use.