Characterization of an inorganic scintillator for small‐field dosimetry in MR‐guided radiotherapy

Abstract Introduction Aim of this study is to dosimetrically characterize a new inorganic scintillator designed for magnetic resonance‐guided radiotherapy (MRgRT) in the presence of 0.35 tesla magnetic field (B). Methods The detector was characterized in terms of signal to noise ratio (SNR), reproducibility, dose linearity, angular response, and dependence by energy, field size, and B orientation using a 6 MV magnetic resonance (MR)‐Linac and a water tank. Field size dependence was investigated by measuring the output factor (OF) at 1.5 cm. The results were compared with those measured using other detectors (ion chamber and synthetic diamond) and those calculated using a Monte Carlo (MC) algorithm. Energy dependence was investigated by acquiring a percentage depth dose (PDD) curve at two field sizes (3.32 × 3.32 and 9.96 × 9.96 cm2) and repeating the OF measurements at 5 and 10 cm depths. Results The mean SNR was 116.3 ± 0.6. Detector repeatability was within 1%, angular dependence was <2% and its response variation based on the orientation with respect to the B lines was <1%. The detector has a temporal resolution of 10 Hz and it showed a linear response (R2 = 1) in the dose range investigated. All the OF values measured at 1.5 cm depth using the scintillator are in accordance within 1% with those measured with other detectors and are calculated using the MC algorithm. PDD values are in accordance with MC algorithm only for 3.32 × 3.32 cm2 field. Numerical models can be applied to compensate for energy dependence in case of larger fields. Conclusion The inorganic scintillator in the present form can represent a valuable detector for small‐field dosimetry and periodic quality controls at MR‐Linacs such as dose stability, OFs, and dose linearity. In particular, the detector can be effectively used for small‐field dosimetry at 1.5 cm depth and for PDD measurements if the field dimension of 3.32 × 3.32 cm2 is not exceeded.


| INTRODUCTION
Magnetic resonance-guided radiotherapy (MRgRT) represents to date one of the most promising techniques in the framework of personalized cancer care, offering high soft tissue contrast imaging and allowing precise radiation delivery. [1][2][3] The hybrid machines designed for MRgRT combine linear accelerators with an onboard magnetic resonance (MR) scanner and differ in terms of main architecture of the system and magnetic field (B) strength.
The two systems currently available for clinical practice use a Linac. [4][5][6] The introduction of these hybrid machines has led to the necessity of new dosimetry systems for radiation beam quality controls, whose response would be constant in the presence of B and at the same time accurate for small-field dosimetry. 7 Furthermore, there is a growing interest in detectors able to provide a response characterized by high temporal resolution, to start exploring the possibility of in-vivo applications in MRgRT.
The first detectors used for absolute dose measurements in the presence of B were ion chambers, for which a new formalism was introduced, including correction factors to take into consideration the dependence of the detector response from its orientation with respect to the B field lines. 8,9 Alternative systems for relative dosimetry were also tested, such as diamonds, diodes, and radiochromic films, with remarkable results in different experiences using MR-Linac systems. [10][11][12] Thanks to their physical properties, optical fiber-based detectors appear to be particularly promising for applications of relative dosimetry in MRgRT, since they offer a real-time response, are potentially accurate for small-field dosimetry and are characterized by a light yield constant in the presence of a B with known strength. 13 Even if the dosimetric properties of such detectors have been widely investigated for external beam radiotherapy (EBRT) and brachytherapy, limited experiences in MRgRT setting are today reported. [14][15][16] The aim of this study was to dosimetrically characterize a new inorganic scintillator designed for MRgRT in the presence of 0.35 T B and to evaluate its clinical feasibility for small-field dosimetry of MR-Linac systems.

2.A | The detector
DoseWire Series 200 (DoseVue, N. V, Diepenbeek, Belgium) is an inorganic scintillator detector consisting of a hemisphere of 0.5 mm radius coupled to an optical fiber. The scintillating material is based on europium-doped yttrium oxide and emits in the 600-650 nm window. This emission band helps increasing signal to noise ratio (SNR) thanks to the high scintillator light yield and the reduced presence of stem effect at these wavelengths.
The detector has a sensitive volume of 0.00026 cc and an effective point of measurement (EPOM) located at 3r/8, 'where r is the radius of the semi-sphere. The scintillator has density of 3.4 g/cm3 and effective atomic number equal to 30.79.
The full system is designed as a 4-channel device, allowing realtime dose measurements at four locations simultaneously, with a maximum sampling frequency of 10 Hz.
The digital signal is then sent to a controlling computer, where a web-based software interface displays the cumulative number of counts per channel. Manufacturer data reported that the system has a response independent of dose rate (variation < 1% in a range of 600-1400 MU/ min) and a signal stable for cumulative dose up to 500 Gy (maximum variation equal to 0.6%). These data were confirmed in a recent experience performed on a standard linac. 17 The system also shows good stability in terms of response with respect to the temperature variation, with variations within 1% in the range 17°-28°C. 18 Such range can be extended up to 15-40°C, thanks to literature studies performed on detectors with similar scintillator composition that showed 1.3% variation between 15°C and 40°C. 19

2.B | Experimental measurements
The physical characterization of the detector was realized in the presence of 0.35 T B"/>, using a hybrid 6 MV low tesla MR-Linac  The detector has been characterized in terms of reproducibility, dose linearity, angular response, time-dependent luminescence, and dependence by energy, field size, and B orientation.
A preliminary dose calibration was performed by exposing the detector to a 9.96 × 9.96 cm 2 field and delivering 100 MU, adopting the experimental setup reported in Fig. 2.
The detector was placed in a water tank, at 1.5 cm depth. The source to axis distance (SAD) is 90 cm, the source to surface distance (SSD) was 88.5 cm and the source to detector distance (SDD) was the same as the SAD.
In this experimental setup, 1 MU corresponds to a dose value of 1 cGy according to the machine calibration that was performed following the TG-51 protocol. 20 A detector holder was realized in-house, to ensure stable positioning of the detector. The MR-Linac dose rate was 600 MU/min.
The number of counts generated by the scintillator was associated to the dose value measured by a 0.125 cc ion chamber (PTW 31010 Semiflex, Freiburg, Germany) placed in the same experimental conditions to calibrate the scintillator in dose values.
The detector was then characterized in terms of SNR to by varying the sampling rate, time-dependent luminescence, response dependence from B orientation, angular dependence, reproducibility, and dose linearity with the same calibration experimental setup. SNR was evaluated by delivering 100 MU thrice with a 9.96 × 9.96 cm 2 field and repeating the measurements using a bare fiber (i.e., a fiber without scintillator detector) to estimate the noise level. SNR measurements were repeated at three different sampling rates: 2, 5, and 10 Hz.

ficient.
Time-dependent luminescence was investigated using the experimental scheme proposed by Kertscher et al: the raw detector signal was acquired during two 5000 MU irradiations (corresponding to 50 Gy), with a time interval of 500 s between the two acquisitions. 21 The raw signal variation was measured during both the irradiations:  PDD curves were acquired at SSD equal to 78 cm, using the scintillator and the ion chamber and comparing the results with those calculated using the MC TPS. The following depths were considered: 10, 15, 20, 30, 50, 70, 100, and 150 mm. Figure 3 shows the detector raw signal as a function of time obtained during calibration and the scintillator response to variations of the MU, once the detector is calibrated.

| RESULTS
The SNR was equal to 116.2 ± 0.6 at 1 Hz, 116.3 ± 0.6 at 5 Hz, and 116.1 ± 0.7 at 10 Hz. The whole raw data are reported in Table S1.
The response variation with varying sampling rates is within 0.5%. A temporal resolution of 10 Hz was chosen for all the measurements considering the delivery of more than 10 MU.
In the graph of the dose linearity, the error bars did not exceed the box sizes. The R 2 correlation coefficient was equal to 0.999.
As regards time-dependent luminescence, Fig. S2 reports the raw signal acquired in function of time for the two 5000 MU acquisitions. With respect to the signal intensity registered after 1 Gy, an increase in scintillation was observed in both the irradiations equal to 0.3%/0.6% after 24 Gy and 0.5%/0.8% after 50 Gy.
The response dependence with respect to the orientation of the detector in the B lines is shown in Table 1.
The detector signal repeatability was below 0.4%, as expressed by the coefficient of variation. The response variation when changing the detector orientation with respect to the B lines, was below 1% (−0.7% at 90°and −0.5% at 270°). The measurements related to the angular dependence of the detector are reported in Fig. 4, together with a visual representation of the detector orientation with respect to the beam axis.
The angular response of the detector is within 2% for all the gantry angles considered. Figure 5 reports the OF values measured at 1.5 cm depths using the scintillator, the ion chamber, the microdiamond, and calculated with MC simulation. The OF measured using the ion chamber is not available for field sizes smaller than 2.49 cm.
All the OFs measured at 1.5 cm depth using the scintillator are in accordance within 1% with those measured using the other detectors and calculated by MC simulation.

| DISCUSSION
In this work, the physical characterization of a new inorganic scintillator designed for small-field dosimetry in MRgRT has been successfully performed.
T A B L E 1 Variability of detector response when changing its orientation with respect to the magnetic field (B) force lines.  The dose measurements performed using these systems were found to be in agreement with ion chamber measurements within 1% for 6-25 MV photon beams and for 8-21 MeV electron beams. [23][24][25][26] Some recent experiences have investigated the properties of organic scintillators in the presence of low-and high-tesla B, concluding that these detectors can be effectively used for OF measurements in MRgRT. 27,28 However, plastic scintillators emit in the blue spectral region (400-450 nm) and their signal is usually heavily contaminated by the stem effect, due to the Cerenkov radiation and the fluorescence light induced in the optical fiber during the irradiation. 29,30 Several techniques were developed to reduce the stem effect, mainly focused on the use of an additional background fiber to estimate and suppress the Cerenkov radiation. 31,32 The recent interest toward the use of inorganic scintillators is justified by the fact that these materials show higher light yield, emitting in a spectral region where the aforementioned contaminating effects are significantly less prominent. 33 The results of this study confirm that the inorganic scintillators show high SNR independent of the sampling rate chosen and they are characterized by a highly reproducible and linear response which is linear with the dose.

| CONCLUSION
The basic physical characterization of a new inorganic scintillator designed for MRgRT application has been successfully performed.
In the present form, this inorganic scintillator can represent a valuable tool for quality controls and relative dosimetry on MR-Linacs, such as dose stability during the different treatment days, OFs, and dose linearity check.
In particular, the detector can be effectively used for small-field dosimetry at 1.5 cm depth and for PDD measurements if the field dimension of 3.32 × 3.32 cm 2 is not exceeded.
F I G . 6. Percentage depth dose measured using the scintillator at two different field sizes and calculated using the Monte Carlo treatment planning system (TPS).
Thanks to its high temporal resolution and its independence of B

SUPPORTING IN FORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.  Table S1 . SNR in function of sampling rate.  but also those corrected using the correction factors for ion chamber (IC) and microdiamond (MD). The percentage difference between the measured values and the calculated ones are also reported. No correction factors are applied for scintillator. Table S4. OF values calculated using the TPS and measured using the scintillator at 5 depth. In table are present not only the raw values but also those corrected using the numerical model. The percentage difference between the measured values and the calculated ones are also reported. Table S5. OF values calculated using the TPS and measured using the scintillator at 10 depth. In table are present not only the raw values but also those corrected using the numerical model. The percentage difference between the measured values and the calculated ones are also reported. | 251