Dose rate response of Digital Megavolt Imager detector for flattening filter‐free beams

Abstract In this study we investigated the dose rate response characteristics of the Digital Megavolt Imager (DMI) detector, including panel saturation, linearity, and imager ghosting effects for flattening filter‐free (FFF) beams. The DMI detector dose rate response characteristics were measured as a function of dose rate on a Varian TrueBeam machine. Images were acquired at dose rates ranging from 400 to 1400 MU/min for 6XFFF and 400 to 2400 MU/min for 10XFFF. Line profiles and central portal doses derived from the images were analyzed and compared. The linearity was verified by acquiring images with incremental Monitor Unit (MU) ranging from 5 to 500 MU. Ghosting effects were studied at different dose rates. Finally, for validation, test plans with optimal fluence were created and measured with different dose rates. All test plans were analyzed with a Gamma criteria of 3%‐3 mm and 10% dose threshold. Our study showed that there was no panel saturation observed from the profile comparison even at the maximum dose rate of 2400 MU/min. The central portal doses showed a slight decrease (1.013–1.008 cGy/MU for 6XFFF, and 1.020–1.009 cGy/MU for 10XFFF) when dose rate increased (400–1400 MU/min for 6XFFF, and 400–2400 MU/min for 10XFFF). The linearity of the DMI detector response was better than 0.5% in the range of 20–500 MU for all energies. The residual image was extremely small and statistically undetectable. The Gamma index measured with the test plans decreased from 100% to 97.8% for 6XFFF when dose rate increased from 400 to 1400 MU/min. For 10XFFF, the Gamma index decreased from 99.9% to 91.5% when dose rate increased from 400 to 2400 MU/min. We concluded that the Portal Dosimetry system for the TrueBeam using DMI detector can be reliably used for IMRT and VMAT QA for FFF energies.

The Electronic Portal Imaging Devices (EPID) is a very useful device for routine clinical use because of its prompt setup, easy data acquisition, and high resolution. The Portal Dosimetry is an application integrated on the ARIA electronic medical record system (Varian Medical Systems VMS, Palo Alto, CA, USA) that allows IMRT or VMAT field verification. The dosimetry process consists of three steps: (a) Fluence prediction using Eclipse Portal Dose Calculation (PDC) algorithm, (b) Fluence acquisition using dosimetry mode on the linac, and (c) Fluence comparison using the Portal Dosimetry module in Aria. The Portal Dosimetry adds a fast and efficient workflow for the verification of IMRT and VMAT plans. 1 Since the option of removing the flattening filter (FF) in the linacs for IMRT and VMAT treatments was introduced in 2010, there has been a lot of interest generated in using flattening filter-free (FFF) beams which give the benefit of reduced headscatter and hence, reduced dose outside the field. 2 These beams also deliver dose faster than flattened beams, which would be beneficial for hypofractionated treatments by reducing treatment time and potential intrafractional organ motion. 3 The Digital Megavolt Imager (DMI) detector is now a standard MV imaging detector installed on all Varian TrueBeam machines. The DMI detector offers not only the possibility to image large field size (43 9 43 cm 2 ) but also the images with a higher pixel resolution (1280 9 1280) than the older detectors (i.e., IDU20, aS1000). The detector area used for dosimetry measurements (integrated images) is a little smaller than the complete imaging size (40 9 40 cm 2 with 1190 9 1190 pixel). 4 With faster readout electronics and a higher pixel capacitance, the DMI detector allows for much higher dose rate than the older detectors (i.e., IDU20, aS1000). It has been adapted by Varian for use in FFF beams at any source-to-detector distance. 5 Since the dose rates for FFF beams are up to six times higher than for conventional flattened beams, portal images taken at maximum FFF dose rate may saturate the images. The EPID saturation occurs when no additional photocurrent outputs from the photodiode as the incident optical power increases. There are a number of parameters that can affect the saturation limit. However, the detailed information of the hardware and software components is proprietary and unknown to the public. Several studies investigated the feasibility of using Portal Dosimetry to FFF beams for IMRT and VMAT plan verification. 6 Varian TrueBeam STX 1.6 and an aS1000 model. 7 Miri et al. studied the Portal Dosimetry of a TrueBeam with an aS1200 panel for flattened filter (FF) and flattening filter-free (FFF) beams. They studied the linearity of dose-response with MU, the imager lag, and the effectiveness of backscatter shielding. They concluded that significant improvements were observed in the dosimetric response of the aS1200 imager compared to previous imaging detectors (IDU20, aS1000). 5 To the best of our knowledge, the study of the dose rate response characteristics for the DMI detector has not been reported in the literature. The purpose of this work was to study the panel saturation as a function of the dose rate, including dose linearity and imager ghosting effects of the DMI detector for FFF beams.

2.A | Portal dosimetry commissioning
The Portal Dosimetry was used to acquire delivered dose images, while Eclipse was used to compute a corresponding dose distribution. Both images were viewed and quantitatively compared in the Portal Dosimetry. The PDC was used to calculate portal dose images for fields containing fluences as part of a pre-treatment verification for IMRT and VMAT planning. 11 The portal imager was calibrated in dosimetric acquisition mode at isocenter according to the official Portal Dosimetry calibration procedure. 12 The procedure includes applying a dark and flood-field correction, an absolute calibration, and a beam profile correction.
The absolute calibration is defined using 100 MU to correspond to 1 calibrated unit (CU) with a 10 9 10 cm 2 field. The beam profile correction was made using diagonal profiles measured in a water phan-

2.B | Profile and central point dose measurement as function of dose rate
The DMI detector dose rate response characteristics were studied as a function of dose rate. Images were acquired at dose rates from 400 to the maximum of 1400 and 2400 MU/min for 6XFFF and 10XFFF, respectively, in six steps, delivering a 100 MU 20 9 20 cm 2 field. Line profiles and central portal doses derived from the images were compared and analyzed. The central portal doses were measured as the average of pixel value in the center of the field, using the output factor tool of the Portal Dosimetry.

2.C | The linearity of the DMI detector response
The linearity of the DMI detector response as a function of MU was investigated by delivering a 10 9 10 cm 2 field with MU ranging from 5 to 500 for both of the unflattened modes (6XFFF and 10XFFF) and the conventional 6X mode, for clinical dose rates at 1400, 2400, and 400 MU/min, respectively.

2.D | The dose rate response of ghosting effects
The dose rate response of ghosting effects was studied by measuring the imager lag after delivering 500 MU at different dose rates for all energies. A 10 9 10 cm 2 treatment field was acquired immediately after exposing the DMI detector using a periphery region outside the central 10 9 10 cm 2 , but within the 20 9 20 cm 2 treatment field, was carefully checked by comparing the portal doses in the same region that was acquired after a long time interval without image acquisitions (see Fig. 3 for illustration). The residual image (RI) is then where PD Before and PD After are the portal doses before and after 20 9 20 cm 2 treatment field is delivered, respectively. PD Center is the portal dose at filed center for 10 9 10 cm 2 field size. An average of 10 cmx 10 cm 10 cmx 10 cm 20 cmx 20 cm 100 cm SID, 500 MU F I G . 3. The residual image (ghost) was studied by measuring imager lag after 500 MU delivery in the periphery region (four points in red). RI calculated based on four points at 8 cm away from the center was used to determine the residual signals.
Finally, plans with pyramid-shaped fluence were created. Portal images of the field were acquired at 6XFFF and 10XFFF energies delivered at different dose rates. All test plans were analyzed with a gamma criteria of 3%-3 mm and 10% dose threshold, which is a typical gamma criteria for IMRT QA in our institution.   Table 1. The Gamma agreement index was decreased from 100% to 97.8% when dose rate increased from 400 to 1400 MU/ min for 6XFFF, and from 99.9% to 91.5% when dose rate increased from 400 to 2400 MU/min for 10XFFF. As the data shown in Fig. 6 for 10XFFF, only a slight decrease (about 1%) of central portal dose

| CONCLUSION
The Portal Dosimetry utilizing EPID is an existing technique that has been shown to work well with standard, flattened radiotherapy beams. With the emerging use of high-dose-rate FFF radiotherapy there is an associated need to verify these treatments efficiently. This study has shown that the DMI panel saturation was clinically insignificant even at the maximum dose rate of 2400 MU/min. The linearity of the DMI detector response was better than 0.5% in the range of 20-500 MU for all energies. The residual image (ghost) was extremely small and statistically undetectable. With faster readout electronics and a higher pixel capacitance, the DMI detector allows for a much higher dose rate than the older detectors (i.e., IDU20, aS1000) without saturation. Therefore, we conclude that the Portal Dosimetry system for the TrueBeam with the DMI detector can be reliably used for IMRT and VMAT pretreatment QA verification for FFF energies.

CONF LICTS OF INTEREST
The authors do not have any conflicts of interest to declare.