Feasibility of portal dosimetry for flattening filter-free radiotherapy

Abstract

The feasibility of using portal dosimetry (PD) to verify 6 MV flattening filter-free (FFF) IMRT treatments was investigated. An Elekta Synergy linear accelerator with an Agility collimator capable of delivering FFF beams and a standard iViewGT amorphous silicon (aSi) EPID panel (RID 1640 AL5P) at a fixed SSD of 160 cm were used. Dose rates for FFF beams are up to four times higher than for conventional flattened beams, meaning images taken at maximum FFF dose rate can saturate the EPID. A dose rate of 800 MU/min was found not to saturate the EPID for open fields. This dose rate was subsequently used to characterize the EPID for FFF portal dosimetry. A range of open and phantom fields were measured with both an ion chamber and the EPID, to allow comparison between the two. The measured data were then used to create a model within The Nederlands Kanker Instituut's (NKI's) portal dosimetry software. The model was verified using simple square fields with a range of field sizes and phantom thicknesses. These were compared to calculations performed with the Monaco treatment planning system (TPS) and isocentric ion chamber measurements. It was found that the results for the FFF verification were similar to those for flattened beams with testing on square fields, indicating a difference in dose between the TPS and portal dosimetry of approximately 1%. Two FFF IMRT plans (prostate and lung SABR) were delivered to a homogeneous phantom and showed an overall dose difference at isocenter of ~0.5% and good agreement between the TPS and PD dose distributions. The feasibility of using the NKI software without any modifications for high-dose-rate FFF beams and using a standard EPID detector has been investigated and some initial limitations highlighted.

Figure 1

The planned dose distribution from the TPS (top row) for a prostate and SABR lung plan. A gamma analysis using 3%/3 mm is shown in the bottom row comparing the measured dose to the planned dose for each site.

Figure 2

Profiles across the center of the EPID image showing only five different dose rates for clarity. The pixel size is 0.25 mm at isocenter. The high dose rates exhibit saturated profiles which become less saturated until, at 800 MU/min, the images are no longer saturated. Note that the 800 MU/min (thick blue line) and 700 MU/min (dashed red) lines are superimposed.

Figure 3

Examples of the characterization measurements showing the open‐field size series (left) and phantom thickness series (right). The plots show the sum of the central $16\times 16$ pixels of the EPID images (blue circles) and the ion chamber measurements (red squares). The pixel size is 0.25 mm at isocenter.