Theoretical and experimental validation of treatment planning for narrow MLC defined photon fields.

In intensity modulated radiotherapy (IMRT), the use of small fields where electronic equilibrium does not exist is becoming more common and presents difficulties for both the measurement and calculation of dose to such fields. Pinnacle(3) (Version 6.2b) allows the user to specify a total minimum open area for each IMRT segment, which can result in sub-segments with widths of only a few millimetres. The dose for 6 MV narrow MLC defined fields between 0.1 and 3 cm in width was investigated using Kodak extended dose range film (EDR2), ionization chamber and MOSFET dosimeters and BEAMnrc Monte Carlo calculations, and these results were used to determine the accuracy of Pinnacle(3) dose calculation for narrow MLC segments. The incident fluences calculated by Pinnacle(3) and BEAMnrc were also compared. Results show that if a fluence and dose grid resolution of 0.1 cm is used, Pinnacle(3) dose agrees with the EDR2 and BEAMnrc to within 5% for field widths between 0.5 and 3.0 cm. However, Pinnacle(3) will underestimate the dose by up to 45% for the 0.1 and 0.3 cm wide fields. It is shown that the source size in the Pinnacle(3) beam model and both the fluence and dose grid resolutions have a significant effect on the accuracy of dose calculation for field widths of 1.0 cm and less. For single segment fields, Pinnacle(3) agrees with EDR2 and BEAMnrc to within 0.1 cm at the field edges and underestimates the penumbra width by up to 0.08 cm. Results for multiple segment fields showed that an MLC transmission of 1.7% and a 0.06 cm inward shift of MLCs prior to beam delivery gave the closest agreement between Pinnacle(3) and measurement. The multiple segment fields also revealed a pattern of low dose troughs of up to 7% in the Pinnacle(3) dose profiles.

Photon dose calculations in homogeneous media for a treatment planning system using a collapsed cone superposition convolution algorithm.

This work reports on some of the initial tests that were conducted during the commissioning of a commercially available 3D treatment planning system. The system (Pinnacle 3.0d-u1) uses a collapsed cone implementation of the superposition convolution algorithm. Calculated and measured dose in homogeneous media were compared for wedged and unwedged fields for both symmetric and asymmetric collimator settings. Results show agreement of 2% or 2 mm in most cases. Where larger differences were found, further investigation was undertaken to explain these differences. These tests demonstrate the correct behaviour of the collapsed cone implementation of the algorithm in homogeneous media and its ability to characterize the beam.

Beam profiles in the nonwedged direction for dynamic wedges.

One feature of the dynamic wedge is the improved flatness of the beam profile in the nonwedged direction when compared to fixed wedges. Profiles in the nonwedged direction for fixed wedges show a fall-off in dose away from the central axis when compared to the open field profile. This study will show that there is no significant difference between open field profiles and nonwedged direction profiles for dynamically wedged beams. The implications are that the dynamic wedge offers an improved dose distribution in the nonwedged direction that can be modelled by approximating the dynamically wedged field to an open field. This is possible as both the profiles and depth doses of the dynamically wedged fields match those of the open fields, if normalized to dmax of the same field size. For treatment planning purposes the effective wedge factor (EWF) provides a normalization factor for the open field depth dose data set. Data will be presented to demonstrate that the EWF shows relatively little variation with depth and can be treated as being independent of field size in the nonwedged direction.