Dosimetric validation of a magnetic resonance image gated radiotherapy system using a motion phantom and radiochromic film

Abstract Purpose Magnetic resonance image (MRI) guided radiotherapy enables gating directly on the target position. We present an evaluation of an MRI‐guided radiotherapy system's gating performance using an MRI‐compatible respiratory motion phantom and radiochromic film. Our evaluation is geared toward validation of our institution's clinical gating protocol which involves planning to a target volume formed by expanding 5 mm about the gross tumor volume (GTV) and gating based on a 3 mm window about the GTV. Methods The motion phantom consisted of a target rod containing high‐contrast target inserts which moved in the superior‐inferior direction inside a body structure containing background contrast material. The target rod was equipped with a radiochromic film insert. Treatment plans were generated for a 3 cm diameter spherical planning target volume, and delivered to the phantom at rest and in motion with and without gating. Both sinusoidal trajectories and tumor trajectories measured during MRI‐guided treatments were used. Similarity of the gated dose distribution to the planned, motion‐frozen, distribution was quantified using the gamma technique. Results Without gating, gamma pass rates using 4%/3 mm criteria were 22–59% depending on motion trajectory. Using our clinical standard of repeated breath holds and a gating window of 3 mm with 10% target allowed outside the gating boundary, the gamma pass rate was 97.8% with 3%/3 mm gamma criteria. Using a 3 mm window and 10% allowed excursion, all of the patient tumor motion trajectories at actual speed resulting in at least 95% gamma pass rate at 4%/3 mm. Conclusions Our results suggest that the device can be used to compensate respiratory motion using a 3 mm gating margin and 10% allowed excursion results in conjunction with repeated breath holds. Full clinical validation requires a comprehensive evaluation of tracking performance in actual patient images, outside the scope of this study.

each case and the static treatment plan using the gamma 16 metric. This methodology has been frequently used to evaluate gating systems [17][18][19][20][21][22][23] . The purpose of this study was to establish a technical performance baseline of the respiratory gating which can inform clinical gating protocols as well as be used as a machine performance benchmark by other institutions.
The MRIdian can gate treatments based on planar images acquired during treatment. The gating is based on tumor tracking by deformable image registration. Real-time images can be acquired in a single sagittal plane at 4 Hz or in three sagittal planes at 2 Hz.
Because the additional 0.25 s of beam-off latency required when using three-plane imaging, our institutional policy is to not use three-plane imaging, and it was not evaluated in this study. Our institution's clinical gating protocol uses the MRIdian's gating capabilities as follows. The gating target is contoured on the volumetric scan used for treatment planning. The gating target is usually the gross tumor volume (GTV), but could be a critical organ at risk or a surrogate structure. For best performance, the gating target contour should correspond to a strong contrast boundary in the image. For every treatment fraction, the target contour is transferred either rigidly or deformably to a volumetric pretreatment setup image. The target contour is optionally updated manually. Both the setup image and the simulation image are acquired under voluntary breath hold of 17 or 25 s duration depending on the imaging field of view. Subsequently, a two-dimensional cine preview scan is acquired with a frame rate of four frames per second in a single sagittal plan centered on the tumor. The cine preview scan is acquired under free breathing conditions. Each frame of the preview scan is deformably registered to the volumetric setup image by the MRIdian software.
Contours are propagated from the setup image to the best-matching cine frame, called the key frame. During treatment, the now twodimensional gating target contour is propagated from the key frame to each cine image in real time. If the propagated target contour exceeds a preset boundary by a configurable percentage of its area, the treatment beam is held. The boundary corresponds either to a numerical expansion relative to the gating target's initial location, or to another contour from the planning volumetric image (e.g., the planning target volume).
Our clinical protocol uses repeated breath holds throughout the treatment duration. The MRIdian system tracks the target and enables beam only when the target is within a predefined boundary; repeated breath holds are used to maximize the duty cycle and minimize dosimetric error due to beam hold latency. A margin of 3 mm around the gross tumor volume (GTV) position in the volumetric setup scan is used as the gating boundary, and a 10% excursion of the tracked GTV contour outside the boundary is allowed. The 10% allowed excursion is used to make the gating procedure robust to random errors in contour tracking arising from image noise.
Repeated breath holds are used by preference if the patient is able to perform them at consistent states of inhalation, with coaching if needed. If the patient is not able to perform consistent repeated breath holds, free breathing is used at the discretion of the attending physician. If the gating efficiency is sufficiently low that the patient may not be able to tolerate the treatment time, the gating margin may be relaxed from 3 mm to 5 mm at the discretion of the attending physician.

2.A | Motion phantom
Respiratory motion was simulated using a MRI-compatible phantom that was capable of one-dimensional motion in the superior-inferior direction (Fig. 1). The phantom was developed by CIRS, Inc (Norfolk, VA, USA), with input from ViewRay and is commercially available from CIRS. The phantom consists of a stationary body structure filled with background contrast material and four small spheres of high-contrast material. A 6.5 cm diameter target rod was made to move within the body structure by a stepper motor. The target rod included three regions of high-contrast material (a small sphere, a small cuboid, and a large cylinder) within and adjacent to a background contrast region. In standard configuration, the target rod includes an ion chamber insert. A custom target rod with a film insert was developed by CIRS to our specification. A cross-section of the target rod is shown in Fig. 2. The entire body structure, excluding a 1 cm acrylic supporting shell, is approximately water equivalent.

2.B | Treatment planning
A treatment plan was produced using the ViewRay treatment planning system and delivered on the MRIdian system. The planning target volume (PTV) was a 3 cm diameter sphere. The plan contained 15 coplanar beams and the dose distribution was typical for a stereotactic body radiotherapy plan, scaled to a maximum dose 3.75 Gy per fraction to lie within the dynamic range of EBT3 film. In addition, a 2.1 cm 9 1.05 cm reference beam delivering 2.4 Gy per fraction at the film surface was added at a 6.5 cm longitudinal isocenter shift relative to the PTV isocenter. The reference beam was used for film spatial registration and dose calibration as described below. Figure 2a shows the dose distribution at the surface of the film.

2.C | Radiochromic film dosimetry
Gafchromic EBT3 radiochromic film (Ashland, Inc, Bridgewater, NJ, USA) was loaded in the phantom and carefully aligned to a raised fiducial reference line in the insert. To minimize the uncertainty in alignment due to cutting the film, the side of the film that was not cut was aligned to this reference line. FilmQA Pro 2015 (Ashland, Inc, Bridgewater, NJ, USA) with the triple-channel algorithm 24 was used for all film analysis. Calibration films from the same lot of film used for experiments were exposed at nine dose levels between 0 and 500 cGy using the MRIdian machine. Film was read using an Epson 10000 XL scanner (Epson America, Inc., Long Beach, CA, USA) at least 24 hr after irradiation. Delivered dose distributions were compared to the planned dose distributions using the gamma technique 16 , with 3 mm distance to agreement and 3%, 4%, and 5% global dose difference tolerance, and regions with dose less than 30% of the maximum dose were suppressed. A 2.1 cm 9 1.05 cm reference beam at a distance of 6.5 cm from the PTV isocenter was delivered with the phantom at rest after | 165 every motion trajectory. The reference beam was used to spatially register the film to the treatment plan in the software used for film dosimetry FilmQA. The reference beam was also used to extract a relative dose scaling factor to account for inter-scan differences in film calibration. The spatial registration and the dose scaling factor were derived by iterative maximization of the gamma passing rate in an ROI including only the reference beam.

2.D | Breathing waveforms
An artificial breathing waveform at two frequencies and three actual tumor trajectories were used for phantom motion. The artificial breathing wave was a cos 6 function, which was used to represent an idealized human breathing pattern which was asymmetric with more time spent at exhalation than inhalation. Trajectory amplitude was 2 cm peak-to-peak and both 10 and 15 breaths per minute frequencies were used. The actual tumor trajectories were derived from real-time MRI imaging acquired at four frames per second during abdominal radiotherapy treatments on the MRIdian machine, and are shown in Fig. 3. One trajectory was acquired during repeated voluntary breath hold at exhalation, and the other two were acquired during free breathing. To test the robustness of gating to rapid breathing, sped-up versions of each trajectory were also tested. Trajectory properties are listed in Table 1. Trajectories were looped throughout the duration of treatment delivery.

| RESULTS
For the static delivery, 96.5% of points passed the gamma criteria at 3%/3 mm tolerance. Dosimetric accuracy relative to the static treatment plan under different motion trajectories and with different gating parameters is shown in Table 2. Using 3 mm gating window and 10% allowed excursion, the breath hold trajectory at actual speed had 97.8% gamma pass rate at 3%/3 mm criteria, and all of the patient tumor motion trajectories at actual speed had at least 95% gamma pass rate at 4%/3 mm. Dosimetric accuracy for the artificially sped-up trajectories is shown in Table 3. Accuracy was noticeably less for the artificially sped-up trajectories and for the sinusoidal trajectories. Considering all trajectories, increased accuracy is observed for 3 mm versus 5 mm gating margin (P < 0.004) and for 5% allowed excursion versus 10% allowed excursion (P = 0.005). Statistical testing was performed with three-way ANOVA using margin size, allowed excursion, and breathing trajectory as grouping variables.

| DISCUSSION
The gamma pass rate for our preferred clinical gating protocol (repeated breath holds, 3 mm gating margin, 10% excursion tolerance) was as good as the gamma pass rate of a static delivery using 3%/ 3 mm criteria, indicating that under such a protocol, phantom motion is completely compensated. For both free-breathing measured tumor trajectories, 3 mm gating margin and 10% excursion tolerance resulted in pass rates above 95% for 4%/3 mm criteria. Our standard MRI-gated protocol uses a 5 mm PTV margin, so a residual 2 mm of margin is available to account for motion-unrelated uncertainty.
Increasing the gating margin to 5 mm resulted in a statistically significantly reduction in accuracy, however, accuracy was still adequate in most cases, compared to an overall target dose uncertainty goal of 5%.
In the case of free breathing tumor trajectories that were artificially sped up to 1.0 and 1.6 s per breath, as well as the cos 6