Validation of PTV margin for Gamma Knife Icon frameless treatment using a PseudoPatient® Prime anthropomorphic phantom

Abstract The Gamma Knife Icon allows the treatment of brain tumors mask‐based single‐fraction or fractionated treatment schemes. In clinic, uniform axial expansion of 1 mm around the gross tumor volume (GTV) and a 1.5 mm expansion in the superior and inferior directions are used to generate the planning target volume (PTV). The purpose of the study was to validate this margin scheme with two clinical scenarios: (a) the patient’s head remaining right below the high‐definition motion management (HDMM) threshold, and (b) frequent treatment interruptions followed by plan adaptation induced by large pitch head motion. A remote‐controlled head assembly was used to control the motion of a PseudoPatient® Prime head phantom; for dosimetric evaluations, an ionization chamber, EBT3 films, and polymer gels were used. These measurements were compared with those from the Gamma Knife plan. For the absolute dose measurements using an ionization chamber, the percentage differences for both targets were less than 3.0% for all scenarios, which was within the expected tolerance. For the film measurements, the two‐dimensional (2D) gamma index with a 2%/2 mm criterion showed the passing rates of ≥87% in all scenarios except the scenario 1. The results of Gel measurements showed that GTV (D100) was covered by the prescription dose and PTV (D95) was well above the planned dose by up to 5.6% and the largest geometric PTV offset was 0.8 mm for all scenarios. In conclusion, the current margin scheme with HDMM setting is adequate for a typical patient’s intrafractional motion.

treatment system with integrated cone beam computed tomography (CBCT) and a high-definition motion management (HDMM) system has extended Gamma Knife treatment options to single-or multi-fractional treatment regimens 5 using frameless mask-based immobilization, and it may be used for large tumor volumes and cases involving a postoperative surgical cavity. [6][7][8] The inherent ability of the integrated CBCT system to localize the acquired image in stereotactic space enables rigid co-registration with planning images [magnetic resonance imaging (MRI) or CT], thereby eliminating the need for a stereotactic head frame. The HDMM system monitors a patient's intrafractional motion, in which an infrared marker placed on the patient's nose is used as an external surrogate to track the intracranial target motion. 8 In clinic, a planning target volume (PTV) margin is applied to account for intrafractional motion, and the HDMM is set to an alert threshold of 1.5 mm. 8,9 To generate a PTV, uniform axial expansion of 1 mm around the gross target volume (GTV), and a 1.5 mm expansion in the superior and inferior directions were used. This margin recipe is based on the assumption that the mechanical accuracy of the mask-based Gamma Knife delivery systems have uncertainties comparable with linac-based treatment. The increased margin in the superior-inferior direction accounts for greater rotational motion uncertainty about the lateral axis (X coordinates in LGK) and uncertainty from the slice thickness in the same direction.
The planning goal is to achieve a coverage of 100% of the GTV volume (D 100 ) and more than 95% of the PTV volume (D 95 ) by the prescription dose, which is similar to linac-based brain radiosurgery.
Although intracranial target motion is expected to be lower than for the external surrogate, studies have shown that with targets located superiorly and farther from the pivot point of head movement (close to the ears), the target motion caused by a patient's intrafractional motion could exceed that of the external surrogate. 9,10 Hence, careful evaluation of the target by a patient's intrafractional motion is required to validate the PTV margins.
Volumetric evaluation allows an assessment of the target coverage through a measured cumulative dose volume histogram (DVH). 11 However, there have been limited studies focused on the volumetric evaluation of the dose distribution with Gamma Knife Icon frameless treatments and the impact of patient motion.
The purpose of the current study was to validate the PTV margin scheme for two clinical scenarios: the patient's head remains below but close to the HDMM tolerance threshold (1.5 mm), and frequent multiple treatment interruptions induced by a large pitch head motion (X coordinates in LGK) followed by plan adaptation. To mimic these clinical scenarios, a remote-controlled pitch-adjustable head assembly 12  was used, which a 3D-printed anatomical replica is created using the CT image of a human head. The hollow phantom with internal anatomical bony structures can be filled with water and has different inserts to hold (a) an ionization chamber for point dose measurements, (b) films for sagittal or coronal 2D plane dosimetric evaluation, and (c) a cylindrical gel insert for 3D dosimetric evaluation. The ionization chamber insert is a polymethyl methacrylate (PMMA) plug of 2 mm thick and 120 mm long and is manufactured to fit the ionization chamber. The film cassette is made of solid water with rectangular dimensions of 70 mm width and 145 mm length, and it has four metal pins for registration purposes. The gel insert consists of PMMA cylinder with dimensions of 140 mm length and 74 mm diameter.
The head phantom was aligned in a neutral position using two lateral rods secured to the Gamma Knife Icon head holder; these rods can be pivoted at both ears of the phantom. Pitch rotation was accomplished by pushing the head phantom by the linear slider that was part of a translation assembly with a stepper motor (XSlide, Velmex, Bloomfield, NY) that was remotely controlled from the console area ( Fig. 1). 12

2.B | CT imaging
Prior to treatment planning, the head phantom was positioned in the Gamma Knife Icon head holder and immobilized using a Moldcare head cushion (Alcare, Tokyo, Japan). CT images of the immobilized head phantom along with the Gamma Knife Icon head holder and Moldcare cushion were acquired using a Phillips Brilliance Big-Bore (Phillips, Amsterdam, Netherlands) scanner with a slice thickness of 1 mm. Three CT images of the water-filled head phantom with the ionization chamber, film, and without inserts were acquired as reference images for point dose, 2D, and 3D dosimetric planning.

2.C | Gamma Knife Icon treatment planning
Leksell Gamma Plan (Ver 11.1.1; Elekta AB, Stockholm, Sweden) was used for treatment planning. Two targets (PTV1 and PTV2) were defined within the intracranial space of the head phantom.  The convolution algorithm with a dose grid size of 1 mm 3 was chosen for the absolute dose measurement comparison to better include the tissue heterogeneities within the skull and head phantom assembly. The CT image of the water-filled phantom without inserts was used to map CT versus density for treatment planning. A total beam-on time of 14.9 min was used.

2.D | Simulated clinical scenarios
To evaluate the PTV margin for typical frameless treatment scenarios related to the patient's intrafractional motion, one reference scenario and two clinical scenarios were simulated. The details are as follows.

2.D.1 | Reference scenario
Ideal scenario without intrafractional motion or interruption of treatment.

2.D.2 | Treatment scenario 1
Head displacement right below the HDMM threshold level but without plan adaptation  CBCT image was re-acquired and co-registered to CBCT ref . The adapted plan was then reviewed and the treatment continued until the next interruption. The simulated pitch angles were chosen following retrospective analysis of patients (n = 50) who underwent frameless Gamma Knife therapy at the institution, demonstrating an average pitch angle of 0.3°± 1.1°and a maximum of 4.5°.

2.E | Dosimetric measurements
The phantom was irradiated after CBCT for the phantom initial setup and plan adaptation. Because each CBCT scan contributed no more than 2.5 mGy to the total dose, CBCT doses were not subtracted from the measurements. In order to evaluate the dose calculation accuracy within Gamma Plan, the plans were generated using both TMR 10 and convolution algorithm by placing a single 16 mm shot with 2 Gy at 50% IDL at the center of the sensitive volume of the ionization chamber. The two planned doses were compared to the ionization chamber measurements.

2.E.2 | Film
The purpose of the film measurement was to evaluate the current PTV margin scheme by comparing the planar dose distribution to the plans for the clinical scenarios. EBT3 GAFchromic film (Ashland Inc., Wayne, NJ) sandwiched between film slabs was positioned in the sagittal plane of the head phantom prior to each irradiation. Analysis of the dose profiles and gamma index between the calculated and measured dose distributions were performed by RTsafe. The film was calibrated using a single-channel protocol with the red color channel. 15 The irradiated films were digitized with an EPSON flatbed color scanner (Perfection V850 Pro, Nagano, Japan) using the scanning parameters described by Makris et al. 16 After the film scans, the net optical density values were converted into the absolute dose values. The spatial resolution of the film was 0.169 mm.
The gamma index criteria (dose difference/distance to agreement [DTA]) of 2%/1 mm, 2%/2 mm, 2%/3 mm, and 10% low-dose threshold were used to evaluate the correlation between the treatment planning system (TPS) dose and film measurement.

2.E.3 | Polymer gel
The purpose of the gel measurements was to evaluate the current PTV margin scheme by comparing the 3D dose distributions and dose-volume histograms (DVH) to the plans depending on the clinical scenarios. In this study, vinylpyrrolidone-based (VIP) polymer gels were used and a detailed description of the VIP formulation and manufacturing process can be found in the literature. 13,14,17 Immediately after irradiation, the polymer gels were equilibrated to the MRI scanner room temperature and MRI scans were acquired The gel measurements were normalized to the mean R2 value in a central homogeneous area (4 mm radius circle) within each PTV at the sagittal plane, and the same normalization was applied to the TPS-calculated dose distributions. Quantitative comparison of the DVH, geometric offset and 3D gamma index were performed with the measured and calculated datasets. 19,20 The gamma index criteria were 2%/1 mm and 2%/2 mm with a 10% low-dose threshold.

3.A.1 | Dose calculation between TMR 10 and convolution algorithm
The planned dose for a single 16 mm shot was compared with the ionization chamber measurements. The TMR 10 plan showed an overestimation of 5.7% compared with the ionization chamber measurements. The convolution dose calculations, which take into account the tissue inhomogeneity from the bony structures of the skull, head holder, and phantom assembly, showed closer agreement, within 1.3% of the ionization chamber measurements ( Table 1). The absolute dose was calculated using the method described by AAPM Task Group 21, 21 and the temperature-pressure (P TP ), polarity (P pol ), and ionization (P ion ) correction factors were obtained from measurements using the same head phantom. The energy-dependent correction factor (k q ) was estimated to be 1.001, assuming waterequivalent homogeneity within the phantom.  On the basis of these observations, to better estimate the target doses and to account for tissue inhomogeneity within the skull, the convolution algorithm was used to generate plans for the treatment scenarios.

3.A.2 | Measurements of three scenarios with PTV1 and PTV2
The phantom was aligned using the CBCT image guidance and irradiated according to each treatment scenario, including PTV1 and PTV2. The dose to the ionization chamber was measured and compared with the mean dose to the chamber-sensitive volume in the treatment plan (
The results showed a dramatically reduced passing rate at 2%/1 mm in all scenarios. This is attributed to the uncertainty of the film dosimetry protocol, as Makris et al. reported the film-to-CT image registration uncertainty to be in the order of 1.5 mm. 16,22 The gamma passing rates with the 2%/2 mm criteria for the ref-

3.C | Polymer gel measurements
Image registration between the post-irradiation MRI and planned TPS data using the structures within the gel phantom was performed to align each target to its planned location. The geometric accuracy of 3D gel dosimetry is in the order of 1 mm since the final dose grid derived from the gel measurements after post-imaging analysis has a resolution of 1 mm 3 . Figure 2(b) shows image registration of post-irradiation MRI images and planned TPS data. This is to demonstrate the coincidence of each treated target to its planned location.

3.C.1 | DVH comparison
Comparison between the planned and measured relative dose distributions was presented in terms of cumulative DVHs for both PTV1, GTV2 and PTV2. The dose distributions were normalized to the corresponding D 50 metric (the dose received by at least 50% of the volume) for each target. The measured D 95 values of PTVs were higher than the planned values by up to 5.6% for all three scenarios. The measured D 100 values of GTV2 were also higher by up to 2.6% than the planned value in all scenarios (Table 4).

3.C.2 | Geometric offset
Center-to-center offsets were measured independently for each target by comparing the difference in the 3D center-of-mass of each target between the gels (polymerized area) and plans (high-dose area). The center-of-mass was calculated by averaging the distributions of the centers-of-mass derived by various ranges of dose thresholds, taking into account the dose gradient of each target. 17,23 Table 5 shows the largest geometric offset (0.8 mm) caused by 1.5 mm HDMM head rotation without plan adaptation in scenario 1. The 3D gamma passing rate with 2%/2 mm criteria for all three scenarios was >98%, as shown in Table 6. T A B L E 3 Film two-dimensional (2D) gamma index, comparing with the treatment planning system (TPS)-calculated dose distributions using 2%/1 mm, 2%/2 mm, and 2%/3 mm passing criteria with 10% low-dose threshold. The percentage difference between the ionization chamber-measured and planned doses using the convolution algorithm were within 3%, which is the anticipated overall measurement uncertainty due to source calibration, measurement setup, and the use of an inhomogeneous phantom instead of a water phantom for the absolute dosimetry.

Passing criteria
For the film analysis with 2%/2 mm criteria and the gel analysis

| CONCLUSION S
Three different measurement methods (ionization chamber, film, and gel) were used with an anthropomorphic head phantom to validate the PTV margin scheme for Gamma Knife Icon frameless treatments. Compared to two other dosimeters, the 3D volumetric analysis with the gel dosimeter was able to show the adequate GTV/PTV coverage in various clinical scenarios, therefore, the current margin scheme with HDMM setting is sufficient for a typical patient's intrafractional motion.

ACKNOWLEDG MENTS
The authors thank Erica Goodoff in Scientific Publications, Research Medical Library at the University of Texas MD Anderson Cancer Center for editing this manuscript.

CONFLI CT OF INTEREST
The authors declare no conflict of interest. PTV, planning target volume; TPS, treatment-planning system.
T A B L E 5 Geometric offset between the centers-of-mass of planned and measured (Gel) dose distributions for PTV1 and PTV2.