Use of a 2-Dimensional Ionization Chamber Array to Measure Head Leakage of a Varian Truebeam ® Linear Accelerator

The purpose of this work is to evaluate the use of a two-dimensional (2D) planar ion chamber array to characterize leakage radiation from the head of the linear accelerator. Ion chamber arrays provide a benefit over a singular ion chamber measurement as they allow for the measurement of a larger area in order to isolate the point of maximum leakage dose and the small size of each individual ion chamber minimizes volume-averaging effects. A Varian Truebeam ® undergoing acceptance testing was used for all measurements. The gantry was wrapped in Portal Pack for Localization (PPL) radiographic film in order to isolate the location of maximum leakage. A calibration curve was developed and used to determine dose-to-film. An Ion Chamber Profiler (IC Profiler™) manufactured by Sun Nuclear Corporation was used to confirm measurements by the PPL film. All measurements were normalized to leakage at 100 cm from the target relative to the central axis. Three points were investigated with the IC Profiler, including the top of the gantry, the Varian logo, and the side of the gantry. For the three locations, respectively, the PPL film and the IC profiler were measured 0.142% and 0.131%, 0.036% and 0.030%, and 0.014% and 0.019%. The good agreement between the PPL film and the IC Profiler provides confidence in the use of a more efficient and accurate ion chamber array for head leakage measurements.

of the treatment head is evaluated through the measurement of leakage radiation. According to the National Council on Radiation Protection (NCRP) report 49, the recommended magnitude of leakage radiation, defined at 1 meter from the source, should not exceed 0.1% of the exposure along the primary axis at the 1 meter from the source for a megavoltage linear accelerator [1]. A more recent report from the NCRP defines leakage radiation as radiation measured outside of the maximum field size produced. The leakage radiation limit is 0.2% of the absorbed dose rate on the central axis and at the same treatment depth [2].
Leakage radiation from the accelerator head is of clinical importance as peripheral dose for structures with low dose tolerances may affect patient treatments. Peripheral dose consists of three sources of radiation, including patient scatter, collimator scatter, and head leakage [3]. At greater distances from the treatment field, around 20 cm (the distance depends on treatment field parameters, beam modulation, and beam energy), head leakage dominates peripheral dose [4]. Leakage measurements performed during accelerator acceptance can be used to characterize collimator and head scatter doses, and confirm that the head leakage falls within NCRP recommendations [5].
In addition, measurements of head leakage are important to assess the shielding effectiveness of a linear accelerator vault. Radiation outside of the primary beam is dominated by head leakage, and the thickness of the secondary barrier is directly affected by the magnitude of this radiation [6].
According to a recent survey, only about 54% of the sixty-five physicists that participated perform a leakage survey during acceptance testing (https://www.surveymonkey.com/results/SM-YRG2SBGJ/). The low number of leakage measurements is mainly due to no outlined method and the cost and availability of detectors. As a result, the manufacturer's specifications of leakage are often used for shielding calculations and the improvement of the head construction and the shipment of the head as one unit (instead of building the head in the vault) for modern linear accelerators improves the consistency and accuracy of the manufacturer's specifications.
Previous studies have looked to characterize head leakage from various linear accelerators. Most studies determined leakage dose at cardinal angles using Monte Carlo [7] and measurements using point detectors such as radiographic film, thermoluminescent detectors, diodes, and ionization chambers [3] [4] [8] [9] [10]. The conventional method for determining head leakage is to wrap the gantry in film to isolate the point of maximum leakage, and use and ion chamber to verify the dose [11].
The challenges of using an ion chamber to measure leakage dose include 1) a standard farmer chamber could be larger than the leakage radiation identified by the film, resulting in partial volume effects, and 2) positioning the farmer chamber in the maximum leakage spot requires an iterative approach to ensure that the chamber is capturing maximum leakage dose. An ion chamber array has the advantage over a point detector because it can measure dose in a plane and each

Oncology portal pack for localization (PPL) film manufactured by Carestream
Health (Rochester, NY) was used to determine the location on the gantry head where the maximum leakage was present. According to the manufacturer, PPL film has a responsive range of (0.25 to 5) cGy and an approximate saturation exposure > 10 cGy. A calibration curve was produced for the PPL film by delivering (1, 1.5, 2, 2.5, 3, 4, 6, 8, 10) cGy to film ( Figure 1). Post irradiation, the films were exposed at the NYU department of radiation safety, and read out using a clamshell densitometer manufactured by Fluke Biomedical (Solon, OH).
Calculations of the leakage of the linear accelerator head were performed using Equation (1), where D Leakage and D Norm are the measured leakage dose and central axis dose used for output normalization, respectively. The following sections described the methods for measuring leakage and central axis output.
PPL films were wrapped around the linear accelerator gantry head ( Figure 2) and 300 MU were delivered, which was calculated to be the optimal dose based on the manufacturer's leakage measurements, the distance from the target to the PPL film, and the responsive range of the film. Dose was delivered using an energy of 15 MV, a gantry angle of 90˚, jaws set to a minimum field size of (0.5 × 0.5) cm 2 , and with the multi leaf collimators (MLCs) fully extended. The location and orientation on the gantry head for each PPL film was recorded during the irradiation. The maximum optical density was measured on each film using the densitometer and the maximum dose was determined using the calibration curve, which was used to calculate the dose at 1 meter away from the gantry head. The dose was normalized to 300 cGy, which was the maximum dose deli-

Results
A total of 35 films were placed along the gantry head, and their locations were determined using coordinate system with angular and azimuthal angles (Figure 4).  A subset of the PPL films that showed a visible exposure and their associated leakage dose location on the gantry head are shown in Table 1. The dose at 100 cm from the target was calculated using the r-squared falloff and the leakage at

Discussion
Characterization of leakage radiation from a linear accelerator is important because of its impacts on peripheral dose to the patient, shielding calculations, and evaluation of the construction of the head of a linear accelerator. Current methods of leakage determination using film can be time consuming, labor intensive, and costly. As a result, this work looked to verify the accuracy of a 2-D ion chamber array to measure leakage radiation by comparing leakage radiation measurements from the array to measurements completed using film.
Previous work has used point detectors to measure leakage radiation, typically at the cardinal angles. These measurements are limited based on the construc-tion of the detector and the placement of the detector relative to the gantry head.
The IC Profiler provides a 2-D array of detectors that cover an area of (32 × 32) cm 2 . Ion chambers are placed along the principal and diagonal axes of the IC Profiler, which inherently provides limitations in measurements outside of those axes, but the IC profiler allows for improved measurement coverage compared with point measurements from ion chambers. There has been limited work published on the energy-dependent response of the PPL film in megavoltage beams. Comparable radiographic films were investigated in work performed by Muench et al. [13], and it was found that variable response occurs at energies < 100 KeV [14]. The relative contribution of low energy photons to leakage measurements is anticipated to be low, as leakage radiation is dominated by thick-target bremsstrahlung at large angles, and the leakage energy spectrum is similar to a useful megavoltage beam [15]. 15 MV photons were used in this work and it is expected that the majority of the leakage measurement originated from photon penetration from the linear accelerator head where less shielding was present. In addition, this work calculated leakage at a distance of 100 cm from the target of a linear accelerator using PPL film by using an r-squared falloff factor. Based on the nature of leakage radiation, this was assumed the most accurate representation of dose falloff. This work validated the use of a commercial detector array for the measurement of head leakage by comparing the percent leakage to PPL film. Results showed good agreement between the IC Profiler and the PPL film, with a maximum difference of 0.008%.
There are two main considerations for medical physicists when characterizing leakage radiation from the gantry head, 1) isolating the point of maximum leakage, and 2) accurately measuring the leakage radiation to determine if the accelerator meets NCRP recommendations. Ideally, a physicist will wrap the gantry head in film to visually determine the location of maximum leakage and confirm the dose using ion chamber measurements. The use of an ion chamber array presented in this work will not be able to isolate the point of maximum leakage, but will provide a more efficient method for determining the dose at 100 cm when compared with an ionization chamber.
Based on the results of the presented survey, only 54% of physicists perform leakage measurements during acceptance testing. Instead of not performing leakage measurements, it would be recommended to use an ion chamber array to determine dose at 100 cm from the target at a set of locations where there is less shielding on the gantry head. These measurements points can be determined using schematics from the manufacturer, and in some cases, previous measurements performed by the manufacturer. This method acknowledges the robustness of modern linear accelerator construction and installation, and that there are few locations on the gantry head construction that could produce leakage radiation when a linear accelerator is installed.