ESTRO 37 Abstract book

S71

ESTRO 37

PV-0144 Are gantry angular speed variations relevant to dose delivery accuracy? A. Onses 1 , D. Jurado-Bruggeman 1 1 Institut Catala d'Oncologia, Medical Physics, Girona, Spain Purpose or Objective In many modern radiotherapy treatments, high dose per fraction is often delivered using VMAT. In order to reduce the treatment time, the overall number of arcs should be kept as low as possible. Therefore, high dose rate achievable by the clinac often implies variations in the gantry rotational speed. We studied how the gantry rotational acceleration can affect the accuracy of the delivery. Material and Methods To quantify the grade of affectation suffered by the dose delivery caused by the variations in gantry speed, we prepared a set of single arc plans at maximum dose rate which gantry speed was modified by changing the metersetweight (cumulating fraction of total dose) of each control point. In each plan the gantry angular speed was forced to be a step function, varying between two values. (Figure 1a, 1b) We tested these changes in two sets of plans: in the first set the MLC was static (so we just test the accuracy of variations in gantry speed) and in the other set a 1 cm MLC gap swept the field. The collimator was set to 0º so the movement of MLC was parallel to the gantry rotation movement. Using an in house script written in Matlab we modified the amplitude and the frequency of speed changes along a single arc, we also modified the plans where the MLC was dynamic for achieving constant leaf speed despite of variations in gantry speed. All the plans were imported to Varian Eclipse TPS (v.13). Verification plans of each modified plan were created, calculated using Acuros XB algorithm, and delivered in a Varian Clinac iX with Millenium 120 MLC. The speed and acceleration limits for this clinac model are 4.8 deg/s and 0.75 deg/s 2 , so the plans were tuned to perform acceleration between zero and the machine limit. The accuracy of delivered dose distribution was measured using ArcCheck, and the gamma 2%/2mm and 3%/3mm passing rates were calculated, as we do in clinical practice. Moreover, the dynalog files of delivered plans were analyzed using MATLAB to obtain the actual gantry speed and the differences between the expected and actual position of the leaves.

Conclusion Dose reconstruction including the effects of dynamic target rotations was demonstrated for VMAT prostate delivery. The CTV dose was much more susceptible to dynamic prostate rotations than to static rotations. The calculations can be integrated with real-time 6DoF motion monitoring for on-the-fly 6DoF dose reconstruction during treatment delivery. Although inclusion of both rotations and translations is straightforward, only rotations were presented here to illustrate the residual dose errors after the translational real-time corrections currently applied in ongoing protocols with prostate gating and MLC tracking.

Results Both sets of analyzed plans showed no differences in dose distribution measurements using the detector array. When compared against the reference plans with no gantry acceleration, the gamma passing rates where the same in all cases. These results are compatible with a