Room: AAPM ePoster Library
Auto-modeling in RayStation does not optimise some MLC parameters which, however, can majorly impact the output for highly modulated plans. The purpose is to determine the MLC transmission, y-position gain, tongue and groove and leaf tip width required to model the photon fluence of an Elekta Agility accelerator.
Dosimetric measurements were made using three different beam geometries. An initial evaluation of the results was performed with portal imaging to assess variability. The measurements were then repeated with radiochromic films inside solid water once for each energy on two matched accelerators for a more representative setup of the patient geometry. For each geometry, metrics measurable on the dose distribution, such as transmission and attenuation, were defined. Averaged metrics obtained from films were used to evaluate model parameters. The optimal model parameter value is such that the difference between the measured and calculated metrics is minimized. Dosimetric analysis were performed with an in-house program.
All presented values are for 6, 10 and 18 MV, respectively. Values of 0.00490, 0.00474, and 0.00630 were obtained for MLC transmission, and of 0.089 cm, 0.072 cm, and 0.108 cm for tongue and groove. Values of 0.378 cm, 0.400 cm, and 0.380 cm were obtained for leaf tip width, and of -0.0043, -0.0041, and -0.0029 for MLC y-position gain. The difference between two metrics measured for a given accelerator and energy were less than the difference between metrics measured for the same energy from two matched accelerators.
Four model parameters were evaluated using dosimetric measurements, and the values obtained are consistent both with the expected ranges from the RayStation documentation and with the physical properties of the accelerators. Because beam modeling can be an underdetermined problem, this method facilitates model optimisation as it reduces the number of parameters to estimate.