TG‐51 reference dosimetry for the Halcyon™: A clinical experience

Abstract Halcyon™ is a single‐energy (6 MV‐FFF), bore‐enclosed linear accelerator. Patient setup is performed by first aligning to external lasers mounted to the front of the bore, and then loading to isocenter through pre‐defined couch shifts. There is no light field, optical distance indicator or front pointer mechanism, so positioning is verified through MV imaging with kV imaging scheduled to become available in the future. TG‐51 reference dosimetry was successfully performed for Halcyon™ in this imaging‐based setup paradigm. The beam quality conversion factor, k Q, was determined by measuring %dd(10)x three ways: (a) using a Farmer chamber with lead filtering, (b) using a Farmer chamber without lead filtering, and (c) using a PinPoint chamber without lead filtering. Values of k Q were determined to be 0.995, 0.996, and 0.996 by each measurement technique, respectively. Halcyon™'s 6 MV‐FFF beam was found to be broader than other FFF beams produced by Varian accelerators, and profile measurements at d max showed the beam to vary less than 0.5% over the dimensions of our Farmer chamber's active volume. Reference dosimetry can be performed for the Halcyon™ accelerator simply, without specialized equipment or lead filtering with minimal dosimetric impact. This simplicity will prove advantageous in clinics with limited resources or physics support.

dosimetry of medical linear accelerators. 1 The TG-51 protocol for reference dosimetry is flexible when it comes to calibration conditions: source-to-axis distance (SAD) or source-to-surface distance (SSD) setup, maximum dose depth (d max ) or a user-defined depth. In contrast, the protocol is stringent in terms of measurement conditions for specification of the beam quality conversion factor (k Q ).
These conditions are 100-cm SSD and 10 9 10 cm 2 field size, therefore this protocol cannot be directly applied to machines incapable of fulfilling these conditions, such as Tomotherapy or CyberKnife.
Technically, the Halcyon TM accelerator is TG-51 compliant, as it is able to create 10 9 10 cm 2 at 100-cm SSD, however, typical setup with the aid of light-projected crosshairs and front-pointer positioning tools is not possible, posing setup challenges unique to Halcyon TM . Therefore, positioning verification of both the water-tank and stage for the Imaging and Radiation Oncology Core (IROC) optically stimulated luminescent detector (OSLD) irradiation must be performed entirely with megavoltage (MV) imaging as kV imaging is not currently available.
Use of FFF beams has expanded rapidly in recent years. Halcyon TM 's 6 MV-FFF beam profile and monitor unit (MU) rate are different compared to FFF beams produced by the TrueBeam â (Varian Medical Systems). Ion recombination correction factors, P ion , are larger for FFF beams compared to flattened beams 2 in part due to the increase in dose rate, however, the Halcyon TM 's 800 MU/min dose rate is closer to the 600 MU/min dose rate offered by Varian's flattened fields, than to the 1400 MU/min or greater dose rate offered in TrueBeam â 's FFF modes. Regardless, the recombination correction needs to be benchmarked for Halcyon TM . Furthermore, use of Farmer chambers to perform measurements of FFF beams, in particular for percent depth dose (PDD) measurements, may lead to inaccuracies due to averaging over the ion chamber active volume 3 leading to an overestimate of the PDD at 10 cm. The beam profile from Halcyon TM does not exhibit a peak as pronounced as that observed for the TrueBeam â accelerator, and the use of a Farmer chamber needs to be validated.
The original TG-51 report 1 stipulates that for photon beams with energies 10 MV and above, PDD measurements must be performed with a 1-mm lead foil placed approximately 30 or 50 cm from the water surface. This procedure accounts for contamination electrons that may affect the dose at d max . The measured %dd(10) Pb is corrected to determine the PDD due to photons alone, %dd(10) X , by means of eqs. (13) and (14) provided in the report. The addendum to the TG-51 report 4 specifically addresses the applicability of the protocol to FFF beams and states that the energy threshold for lead filtering applies only to beams with flattening filters, and as FFF beams

2.B | Setup
As there are no collimating jaws, the Halcyon TM beam is shaped entirely with two independently functioning multi-leaf collimators  During loading, the couch shifts from an external laser virtual isocenter to the radiation isocenter. This shift is verified through the Machine Performance Check (MPC), which must be performed daily before the Halcyon TM can be used. For the measurements performed in this work, the ion chamber was initially aligned to the virtual isocenter as indicated by the external positioning laser system, and then loaded into the bore (Fig. 1). Similarly, the IROC-supplied stage for OSLD irradiation was initially set to the external lasers. Following loading, SSD and ion chamber positioning were fine-tuned based on MV images acquired at gantry angles 0°and 90°. Imaging was performed in Service Mode using the "Intermediate" user profile, which is the only profile in Service Mode with access to absolute dose calibration. High-quality MV imaging was selected from the XI tab and the number of frames acquired as well as the imaging dose were increased to improve image quality.

2.C | Calibration procedure
The setup for Halcyon TM in the treatment planning system (TPS) and system settings allows for three possible calibration point definitions: Calibration was performed using the Farmer chamber positioned at a depth of 10 cm in the water tank. The PDD curve measured with the Farmer chamber was used to correct back to d max and the beam was calibrated to 1 cGy/MU at d max under reference conditions (10 9 10 cm 2 field, at 100-cm SSD).
The IROC-provided OSLD and stage were setup such that the stage was positioned at 100-cm SSD and the acrylic block was placed on top of the stage with the OSLD centered along the beam axis. A 10 9 10 cm 2 field was used to deliver 100 MU to the OSLD.

| RESULTS
Relative profiles in the crossplane and inplane directions for a 10 9 10 cm 2 field delivered at isocenter with 0.9 cm buildup are displayed in Fig. 2

3.B | Reference dosimetry
Examples of the orthogonal MV images used for chamber positioning are shown in Fig. 4. Based on imaging, the chamber position was adjusted with lateral and vertical couch shifts. All shifts were less than 2 mm in magnitude.
Values of P ion and P pol were determined as described by the TG-51 protocol, by recording the charge produced in the Farmer chamber at 10 cm depth for a 10 9 10 cm 2 field delivered at SSD = 100 cm using electrometer biases of À300, +150, and +300 V. Three readings at each polarity were obtained and averaged to determine P ion , P pol , and M raw . The values of P ion and P pol for the Halcyon TM are compared against those determined for a Varian TrueBeam â linear accelerator (6 MV-FFF and 6 MV beams) and for a Varian Clinac â linear accelerator (6 MV only) in Table 1. P ion differs between the filtered and unfiltered beams by 0.0035 while P pol matches for both beams. and (14) for use in the TG-51 protocol, it is recommended that for %dd(10) Pb < 71% or 73%, depending on the position of the lead, the measurement can be performed without the lead given that electron contamination has a negligible effect on %dd(10) at these energies.
For FFF beams, we know this is not the case. 7 Given the successful implementation of TG-51 reference dosimetry for unflattened beams used in clinics throughout the world for over a decade, the impact T A B L E 1 P ion and P pol for PTW 30013 Farmer chamber in Halcyon TM , TrueBeam â and Clinac â beams. During couch loading, it was observed that the fixed speed of couch resulted in substantial agitation of the water within the water tank. Beyond requiring tens of minutes to still the water, we found that an over-full water tank may spill onto the couch and gantry cover below, therefore, we recommend the water tank not be overfilled and that towels be kept close at hand during couch loading to quickly attend to spills.

| CONCLUSION
TG-51 reference dosimetry was successfully implemented on the Halcyon TM linear accelerator using orthogonal MV imaging to accurately position the ion chamber at the radiation isocenter. Despite profile differences in the 6 MV-FFF beams produced by the Halcyon TM and TrueBeam â accelerators, TG-51 correction factors were the same for both machines, while Halcyon's broad profile means that Farmer chambers can be used for PDD acquisition without concerns of volume averaging. Additionally, values of k Q calculated from PDDs acquired with and without lead filtering differed by 0.001, corresponding to differences in output of only 0.1%.

CONFLI CT OF INTEREST
The authors have no conflicts of interest to declare.