Characterisation and application of three-dimensional silicone-based radiochromic dosimetry in 1.5 T magnetic resonance imaging-guided radiotherapy

The magnetic field in magnetic resonance imaging-guided radiotherapy (MRgRT) systems influences the three-dimensional (3D) dose deposition and hence the delivered dose distributions. The aim of this study was to investigate the dose-rate dependency and dose response of silicone-based radiochromic dosimeters for photon irradiation in the precense of a magnetic field using a 1.5 T MRgRT system. Additionally, the study aimed to provide a proof of the concept of radiotherapy treatment based on a treatment plan calculated on magnetic resonance imaging rather than a computed tomography (CT) scan. The delivered dose was read out in 3D with an optical CT scanner and the dose distribution was verified using gamma analysis. We found an insignificant dose-rate dependency for dose rates ranging from 3.2 to 5.1 Gy/min and a linear dose response up to 20 Gy. A 3D 3%/3mm gamma analysis showed a pass rate of 95.1%. The dosimeter showed clinical potential for 3D dose verification of MRgRT delivery.


Introduction
Trajectories of secondary electrons are affected by the magnetic field inherently present in magnetic resonance imaging-guided radiotherapy (MRgRT) systems, resulting in changes of the three dimensional (3D) dose distribution [1,2].Therefore, current dosimetry systems and quality assurance procedures used for conventional radiotherapy (RT) must be revisited and appropriately adjusted to be applicable for these hybrid imaging/treatment machines.
During the recent years, there has been increasing interest in 3D dosimetry due to the high conformity of new treatment modalities such as intensity-modulated radiation therapy and volumetric modulated arc therapy [3,4].A radiochromic silicone-based 3D dosimeter was first introduced in 2015 [5][6][7][8] and has been characterized for conventional megavoltage X-ray RT [8] and megavoltage X-ray RT in the presence of a magnetic field using a 0.35 T MRgRT system [9].Furthermore, it has been characterized for proton therapy [10] and has shown potential for high resolution dose verification in three dimensions [8].The dosimeter is deformable and can be used to investigate the influence of deformation during irradiation [11].Furthermore, it can be molded into complex phantom geometries with embedded air cavities to investigate dose deposition at air-tissue interfaces which is of special interest for MRgRT systems due to the electron return effect [12].
However, the magnetic field strength will affect the severity of any magnetic field dose effect [13] and hence, characterization of the dosimeter in the presence of the magnetic field originating from a 1.5 T MRgRT system is a prerequisite for using the dosimeter in such systems.Therefore, the aim of this study was to characterise the dose-rate dependency and dose response of silicone-based radiochromic dosimeters in the presence of a magnetic field.And additionally, the study aimed to provide a proof of the concept of RT treatment based on a treatment plan calculated on magnetic resonance imaging (MRI) rather than a computed tomography (CT) scan using a 1.5 T MRgRT system.The delivered dose was read out in 3D using an optical CT scanner and the dose distribution was verified using gamma analysis.

Fabrication
Dosimeters were fabricated from 93.2% silicone elastomer (SE), 5% curing agent (CA), 0.26% leucomalachite green (LMG) and 1.5% chloroform (all in weight percentages).SE and CA came from the commercially available SYLGARD 184 silicone elastomer (DOW) kit.LMG was dissolved in chloroform before SE and CA were added and thoroughly mixed using an electric mixer.Air bubbles were removed from the mixture using a vacuum desiccator before dosimeters were cast and left to cure protected from light at room temperature.Two batches were fabricated.Batch 1: 60 cuvettes (1 cm × 1 cm × 4.5 cm) curing for 62 hours (from casting to pre-irradiation read-out); Batch 2: a cylindrical dosimeter (diameter 5 cm and height 5 cm) curing for 69 hours.

Irradiation conditions
All dosimeters were irradiated at Odense University Hospital using an Elekta Unity (Stockholm, Sweden) 1.5 T MRgRT system with a 7 MV flattening-filter-free beam quality, source-to-axis distance of 143.5 cm and a pulse repetition frequency (PRF) of 275 Hz.Cuvettes were irradiated with a 10 cm x 10 cm field with a gantry angle of 90 degrees and centred in the radiation field by placing them standing on a water-equivalent RW3 (PTW, Freiburg, Germany) platform.The dosimeters were placed between RW3 build-up and backscatter slabs with a constant source-to-surface distance of 133.5 cm.While backscatter slabs of 5 cm were used for all irradiations, the build-up slabs varied among 4.5, 9.5 and 14.5 cm to obtain different dose-rate conditions.
The dosimeters were irradiated to doses of 2, 5, 10, 15 and 20 Gy and the number of monitor units needed to deliver a given dose was calculated on a virtual water phantom in the treatment planning system (TPS -Monaco, v. 5.51.10).The number of photon pulses (#pulses) was measured for each irradiation with a TF930 frequency counter (Aim-TTi, GB), and local dose rates were given as (dose × PRF) / #pulses which corresponded to 3.2, 4.1 and 5.1 Gy/min for build-up slabs of 14.5, 9.5 and 4.5 cm, respectively (electron gun duty cycle was 100%).Two dosimeters were irradiated concurrently with a total of four dosimeters irradiated per dose level for each dose rate (60 in total).The alignment of the dosimeters was verified with electronic portal imaging device (EPID) images acquired while the treatment beam was on (i.e., no additional dose given).
The cylindrical dosimeter was placed standing on a polystyrene platform to locate the irradiation isocentre within the cylinder.A T1 weighted MRI reference scan was acquired and imported to the TPS and the external boundary of the dosimeter was delineated and the volume was assigned a relative electron density of 1.0.An ellipsoid of 0.66 cm 3 was defined as target, and inverse treatment planning was performed using a half-arc 5-field equally spaced step-and-shoot beam configuration with a 10 Gy target mean prescribed dose.The treatment plan was exported to the Elekta Unity system and the reference image was matched to the online MRI scan based on a rigid translation registration.The body outline of the dosimeter was verified on the online MRI and the online plan was created using the same field configuration as the reference plan; see Bernchou et al. [14] for a more detailed description.The online plan was exported for data analysis with an 1 mm dose grid resolution and statistical uncertainty of 1% in the Monte Carlo dose results.

Pre-and post-irradiation optical read-out
The cuvette-sized dosimeters were read out optically at 627 nm using a spectrophotometer (Spectroquant Pharo 100) 9 hours pre-and 5 hours post-irradiation.The dose responses were given as the slope of the difference in linear attenuation coefficient (Δα = αpost -αpre) plotted against dose.
The cylindrical dosimeter was read out 7 hours pre-and 5 hours post-irradiation at 635 nm with the Vista TM 16 Optical CT scanner (Modus Medical Devices, Canada) using 1000 2D projections over 360-degrees rotation.Data reconstruction was performed with software provided by the manufacturer as an integral part of the optical CT scanner using an OSC-TV algorithm with a voxel volume of 1.0 mm 3 ; see Valdetaro et al. for a more detailed description [10].

Gamma Analysis
The optical CT read-out and the exported TPS calculation were normalized and compared using a global 3%/3 mm as well as 2%/2 mm 3D gamma analysis.The measured and simulated data were aligned by applying an image registration in MATLAB.Creep-up effects (i.e.non-irradiation induced gradual increase in absorption over time) [15] in the 3D dosimeter result in a background offset.A gamma analysis was performed with an offset subtracted, the same for all voxels, from the optical CT read-out prior to normalization.The offset was chosen so that voxel values within the low-dose bath of the normalized optical CT read-out were comparable to voxel values of the low-dose bath of the normalized TPS dataset.The data was normalized by selecting an identical high dose region in both datasets.Only voxels at 10% or more of the maximum value in the TPS dataset were included in the gamma analysis.Furthermore, slices of the measured dataset containing many artefacts were disregarded.

Figure 1:
Left: For cuvette exposures, red, yellow and green correspond to dose rates of 3.2, 4.1 and 5.1 Gy/min, respectively.Data points represent the mean value and the error bars represent the 95% confidence intervals of the four dosimeters irradiated per dose level for each dose rate.R 2 > 0.99 for all three regressions.Right: Normalized dosimeter dose distributions in the frontal plane (perpendicular to the longitudinal axis of the cylinder) for optical CT read-out and TPS dose calculation.The magnetic field is perpendicular to the cylindrical axis.The white line represents the 10% cut off.

Discussion
In the presence of a magnetic field originating from a 1.5 T MRgRT delivery system, cuvette-sized silicone-based radiochromic dosimeters showed a linear dose response up to 20 Gy with no significant variation of response as a function of dose rate in the range from 3.2 to 5.1 Gy/min.Acceptable gamma pass rates were obtained when subtracting an offset from the 3D read-out.
The dose response for the applied radiochromic dosimeter vary across studies [16,17], attributed to factors such as a slightly different curing temperature (room temperature varies with season of the year), read-out wavelength, batch-to-batch dependence, ambient light exposure, time between pre-and post-irradiation due to fading (i.e.temporal absorption changes for irradiated dosimeters) and creep-up effects [15] as well as different curing times [16].To assess if the magnetic field affects the dose response, dosimeters prepared from the same batch must be irradiated concurrently under the same conditions with and without the presence of a magnetic field.
The insignificant dose-rate dependency is in line with previous findings in the absence of a magnetic field based on a previous established protocol [7,16] as well as for a 0.35 T MRgRT system (9% CA, 89.2% SE, 0.26% LMG and 1.5% chloroform) [9].
The limitations of the study include the method of image registration based on optical CT readout and the TPS dose matrix instead of embedded markers in the dosimeter that could be aligned with the MRI DICOM images.Furthermore, it is an approximation to apply the same offset for all voxels for the optical CT read-out prior to normalization as temporal changes of optical density are dose dependent [15].This emphasizes the importance of minimizing the time between irradiation and read-out as this reduces creep-up and fading effects [18].Additionally there is a lack of explicit connection between the cuvette-sized and cylindrical dosimeters.Further experiments must be conducted to assess volume and surface effects (i.e. if dosimetric properties depend on the volume and surface area) and hence, if dosimetric properties of the cuvette-sized dosimeters can be directly translated into larger samples.
In conclusion, this study of the silicone-based radiochromic dosimeter in the presence of a magnetic field using a 1.5 T MRgRT system showed an insignificant dose-rate dependency in the range from 3.2 to 5.1 Gy/min and a linear dose response up to 20 Gy.A treatment plan calculated on an MRI reference image was verified in 3D with a 3%/3 mm gamma pass rate of 95.1%.The dosimeter showed clinical potential and applicability for 3D dose verification for photon irradiation of MRgRT delivery.