Technical note: TROG 15.01 SPARK trial multi‐institutional imaging dose measurement

Abstract Purpose The Trans‐Tasman Radiation Oncology Group (TROG) 15.01 Stereotactic Prostate Adaptive Radiotherapy utilizing Kilovoltage intrafraction monitoring (SPARK) trial is a multicenter trial using Kilovoltage Intrafraction Monitoring (KIM) to monitor prostate position during the delivery of prostate radiation therapy. KIM increases the accuracy of prostate radiation therapy treatments and allows for hypofractionation. However, an additional imaging dose is delivered to the patient. A standardized procedure to determine the imaging dose per frame delivered using KIM was developed and applied at four radiation therapy centers on three different types of linear accelerator. Methods Dose per frame for kilovoltage imaging in fluoroscopy mode was measured in air at isocenter using an ion chamber. Beam quality and dose were determined for a Varian Clinac iX linear accelerator, a Varian Trilogy, four Varian Truebeams and one Elekta Synergy at four different radiation therapy centers. The imaging parameters used on the Varian machines were 125 kV, 80 mA, and 13 ms. The Elekta machine was measured at 120 kV, 80 mA, and 12 ms. Absorbed doses to the skin and the prostate for a typical SBRT prostate treatment length were estimated according to the IPEMB protocol. Results The average dose per kV frame to the skin was 0.24 ± 0.03 mGy. The average estimated absorbed dose to the prostate for all five treatment fractions across all machines measured was 39.9 ± 2.6 mGy for 1 Hz imaging, 199.7 ± 13.2 mGy for 5 Hz imaging and 439.3 ± 29.0 mGy for 11 Hz imaging. Conclusions All machines measured agreed to within 20%. Additional dose to the prostate from using KIM is at most 1.3% of the prescribed dose of 36.25 Gy in five fractions delivered during the trial.


| INTRODUCTION
Intrafraction motion during prostate radiation therapy treatment can reduce the dose to the prostate and increase the dose to organs at risk. Prostate position during treatment can be monitored using a variety of methods, including megavoltage (MV) imaging, 1 ultrasound, 2 Calypso electromagnetic guidance, 3 the BrainLAB ExacTrac x-ray system, 4 the Cyberknife platform, [5][6][7] and Navotek radioactive fiducials. 8 Adjustments to either the patient or the beam position can then be made during treatment to correct for the motion observed.
Many of the methods available for prostate position monitoring require expensive equipment and expertise. Kilovoltage Intrafraction Monitoring (KIM) uses the gantry mounted kV imager available on modern linear accelerators and software installed on a framegrabber computer. KIM determines the position of the prostate in three dimensions from 2D kV projections using a probability density function. 9,10 KIM has successfully been used to measure prostate displacement during treatment in retrospective 11 and interventional studies. 12 The TROG 15 Carlo simulations for a linear accelerator at a single center. 13,14 This work presents a simple method to assess the KIM kV imaging dose, which could be easily performed at multiple centers. The method uses standard, widely available equipment and an adaptation of the IPEMB protocol 15 for kV dose measurement in air. Imaging dose delivered during the use of KIM by the different linear accelerator models used in the SPARK trial was then measured and assessed at all centers participating in SPARK.

| ME TH OD
The method used for measurement of the kV imaging dose is based on the IPEMB protocol for kV dose measurements in air. 15 This method was chosen as dose measurements needed to be acquired at four geographically separated centers with differing equipment available. The only required equipment is a 0.6 cc ionization chamber and holder with an electrometer, along with aluminum sheets for half-value layer (HVL) measurements. For Varian linacs, dose measurements were taken using fluoroscopy mode at 125 kV, 80 mA, and 13 ms per frame. Measurements on the Elekta machine were taken using continuous acquisition mode at 120 kV, 80 mA, and 12 ms per frame.

2.A | Ion chamber measurements
The beam quality was measured by finding the half value layer (HVL) of aluminum using a narrow 2 9 2 cm 2 square field to reduce scatter. A farmer type ion chamber (without buildup cap) was placed at the isocenter 100 cm from the source and aligned to the center of the field using kV projections. The ion chamber remained in the same position for all HVL measurements. The attenuating material was placed 50 cm from the source. Care was taken to avoid scatter into the chamber by retracting the couch as far as possible and by extending the kV detector panel away from the ionization chamber. The equipment setup used for this measurement is shown in Fig. 1.
Dose was then measured at isocenter in air using the same setup (minus the holder for the attenuating material) and exposure settings for a field size of 6 9 6 cm 2 (8 9 8 cm 2 for the Synergy). The field F I G . 1. Experimental set up for ionization chamber measurements showing position of ionization chamber, kV source, and kV detector.
| 359 size was chosen as it represents the field size used for KIM. Images were taken for 30 s in fluoroscopy mode at 11 fps. The number of frames measured was estimated from the time over which readings were taken multiplied by the number of frames acquired per second.
The average ion chamber reading per frame was calculated from three readings.

2.B | Calculation of dose in patient
The dose to water at the surface was calculated using eq. 1: where D w;z¼0 is the dose to water at surface, M is the ion chamber reading per frame corrected for temperature and pressure, N k is the chamber air kerma calibration factor, l en q is the mass energy absorption coefficient ratio (water to air) interpolated from the IPEMB protocol using the measured HVL and B w is the back scatter factor, also from the IPEMB protocol. 15 Crocker et al. 13 found that for a cohort of 22 patients set up for treatment with the center of their prostate at isocenter, the median source to surface distance (SSD) was 84.7 cm and the median PTV depth was at 15.3 cm below the surface. Percentage depth dose (PDD) at 100 cm SSD within a CIRS phantom setup for a 10 9 10 cm 2 field at 15 cm depth was measured in that study to be 10%. 13 Therefore, an inverse square law (ISL) factor was first applied to determine the dose to the water at surface at an SSD of 84.7 cm from the measured dose to water surface at isocenter (100 cm). Then the percentage depth dose was used to determine the dose at depth 15.3 cm at isocenter within the phantom (eq. 2): Where D w;z¼0;iso is the dose to surface at isocenter found using equation 1 above, ISL is the inverse square law factor, PDD is the percentage depth dose factor and FSF is the field size factor. This method makes the assumption that the dose delivered to the isocenter is uniform throughout the whole prostate and that the prostate is of uniform depth in the patient at all gantry (kV source) angles during VMAT treatment.
The dose delivered per treatment, D iso; treatment ; was calculated using eq. 3: where fps is the number of kV frames acquired per second (images can be acquired at 1, 5, or 11 Hz) and t treatment is the estimated treatment length.

2.B | Estimation of skin dose
The dose delivered to each section of skin was estimated by assuming a cylindrical patient geometry with a radius of 15.3 cm. The dose per frame to skin was determined using eq. 4: where D w;z¼0;iso is the dose delivered to the surface at isocenter calculated using eq. 1 and ISL is the inverse square law factor used to scale the dose at isocenter to the dose delivered at a point 15.3 cm above isocenter (1.39).
Taking into account beam divergence and assuming fractions are delivered using two partial arcs, each of 280°, the skin dose delivered to a section of skin by a 6 9 6 cm 2 kV beam can be estimated using Equation 5: D skin ¼ t exposed :fps:D w;z¼0 (5) where t exposed is the time in seconds for which each point of skin is exposed to the kV beam, fps is the number of kV frames acquired per second and D w;z¼0 is the dose per frame delivered to the skin surface calculated using eq. 4.
The value for t exposed can be found using the beam on time and the width of the beam at the skin surface. The width is determined using trigonometry as shown in eq. 6: where w is the width of the beam at the skin surface, SSD is the source to surface distance and x is the field size.
The time that each patch of skin is exposed to the beam was then determined using eq. 7: where t exposed is the time that each patch of skin is exposed to the beam, w is the width of the beam at the skin surface, c is the patient circumference, t total is the time taken to deliver all arcs making up the entire treatment and angle is the angle subtended by each individual arc delivered. In this case, for a field size of 6 9 6 cm 2 , SSD of 83.7 cm, treatment time of 314 s, arc angle of 280°and assuming a patient radius of 15.3 cm, the resulting t exposed was found to be 21.4 s.

| RESULTS
The half value layer (HVL), dose per frame at isocenter at 15.3 cm depth and absorbed dose to the prostate at different imaging frequencies delivered by each of the seven linear accelerators measured using an ion chamber are listed in Table 1.
The estimated dose delivered to a point on the skin by each machine at each frame rate appears in Table 2. The results are for the whole course of treatment (five treatment fractions).
The uncertainty of the measurements was estimated based on the limit of reading of the electrometers used, uncertainty in thickness of materials used for HVL measurements, and uncertainties resulting from estimation of the number of frames measured. The uncertainty in the measured doses is approximately 7%.

| DISCUSSION
Imaging dose delivered during the use of kilovoltage intrafraction monitoring for prostate SBRT treatments was measured at four different radiation therapy centers and on seven linear accelerators from two vendors. Limited comparisons between imaging dose delivered by different machines or at different centers exist in the literature, and such reports focus on the dose given for CBCT scans. 16,17 The method presented here allows for simple, quick measurement of kV imaging dose at geographically separated centers. The standardized procedure used minimal equipment, namely a 0.6 cc ionization chamber, a chamber holder, an electrometer and aluminum sheets, all of which is readily available at most radiation therapy centers. The imaging dose delivered during the use of this localization method can be reduced significantly by reducing the imaging frequency. Imaging at 1 Hz has been shown to provide sufficient localization accuracy for prostate treatments when using KIM. 10 Lowering the imaging frequency requires methods to reduce the level of MV scatter on the kV detector. This can be achieved by a dark frame readout immediately prior to kV image acquisition. 19 Reducing the kV field size based on individual patient seed placement following the method outlined by Crocker et al. can also provide reductions in imaging dose. 13 Further reductions in imaging dose could be obtained using both patient-specific and gantry anglespecific kV and mAs settings. For example, the task of finding fiducial markers in an anterior-posterior kV image where the anatomical pathlength is short and the absence of bony anatomy within the beam requires much lower dose than achieving the same task for lateral beams. The maximum entrance dose per projection measured for KIM in this study is slightly lower than the lowest Cyberknife value, however, at least three times as many images are likely to be taken when using KIM during a VMAT treatment, and so the overall surface dose is likely to be on the same order of magnitude. It should be noted, however, that the skin dose during a KIM VMAT treatment will be spread around the patient as the gantry moves, while the Cyberknife is likely to deliver a more concentrated dose to just some regions of the skin.