Measurement of skin surface dose distributions in radiation therapy using poly(vinyl alcohol) cryogel dosimeters

Abstract In external beam radiation therapy (EBRT), skin dose measurement is important to evaluate dose coverage of superficial target volumes. Treatment planning systems (TPSs) are often inaccurate in this region of the patient, so in vivo measurements are necessary for skin surface dose estimation. In this work, superficial dose distributions were measured using radiochromic translucent poly(vinyl alcohol) cryogels. The cryogels simultaneously served as bolus material, providing the necessary buildup to achieve the desired superficial dose. The relationship between dose to the skin surface and dose measured with the bolus was established using a series of oblique irradiations with gantry angles ranging from 0° to 90°. EBT‐2 Gafchromic film was placed under the bolus, and the ratio of bolus‐film dose was determined ranging from 0.749 ± 0.005 to 0.930 ± 0.002 for 0° and 90° gantry angles, respectively. The average ratio over 0–67.5° (0.800 ± 0.064) was used as the single correction factor to convert dose in bolus to dose to the skin surface. The correction factor was applied to bolus measurements of skin dose from head and neck intensity‐modulated radiation therapy (IMRT) treatments delivered to a RANDO phantom. The resulting dose distributions were compared to film measurements using gamma analysis with a 3%/3 mm tolerance and a 10% threshold. The minimum gamma pass rate was 95.2% suggesting that the radiochromic bolus may provide an accurate estimation of skin surface dose using a simple correction factor. This study demonstrates the suitability of radiochromic cryogels for superficial dose measurements in megavoltage photon beams.

increase electron fluence, increasing the dose deposited in the superficial tissues. [1][2][3][4][5][6] Other factors influencing the surface dose distribution include electron contamination from the linear accelerator, obliquity, field size, beam modifiers, air gap, and delivery technique. [7][8][9][10][11][12][13][14] Accurate knowledge of dose to superficial tissues is necessary to ensure that shallow targets receive the prescribed dose while the dose to normal tissue is within tolerance. 15,16 However, this is confounded due to the inaccuracy of most TPSs in the buildup region. 17 Modern radiotherapy TPSs are able to calculate skin dose within AE25%. [18][19][20][21][22][23] Most TPSs estimate skin surface dose by extrapolating measured data with fitting functions. 24,25 Monte Carlo simulation is capable of calculating the dose in the buildup region accurately [26][27][28] but the use of these systems is limited in the clinic due to the computational requirements. 29 Therefore, in vivo measurements are desirable to verify the skin surface dose.
Several dosimeters are currently used in radiotherapy for surface skin dose estimation. Thermoluminescent detectors (TLDs), 14,23,[30][31][32] diodes, 33,34 and metal oxide semiconductor field effect transistors (MOSFETs) 29,[35][36][37] may be used to produce low resolution surface dose distributions. Radiographic or radiochromic film may be used to quantitate the distribution of surface dose in two dimensions. [38][39][40][41] Radiochromic film has several advantages, such as tissue equivalency, self development, and high spatial resolution. 42 However, film is difficult to form to surfaces that contain both convex and concave regions, which complicates dosimetry. 40 In these situations, a more flexible material is desirable.
Gel dosimeters may be used to measure skin dose at individual points with small fields and steep dose gradients treatments such as dosimetry of IMRT and stereotactic radiosurgery. Such techniques involve the delivery of a high radiation dose using small size radiation beams. [43][44][45] Gel dosimeters such as cryogels are flexible and can easily conform to the skin over large, complex, curved regions. Furthermore, as suggested by Chu et al. 46 without further investigation, poly(vinyl alcohol) cryogels (PVA-C) based dosimeters may be simultaneously employed as a dosimetric bolus to provide an accurate estimation of skin surface dose. PVA-C is flexible and stable, and, when loaded with a radiosensitive material such as ferrous benzoic xylenol orange (FBX), is capable of recording dose in two and three dimensions. 46,47 These cryogels may be used as both a buildup material and act as an in vivo dosimeter to monitor the treatment delivery. These cryogels were used to monitor chest wall radiation therapy treatment; it can quantitate uncertainties in setup, and breathing irregularities during left breast or chest wall deep inspiration breathing hold (DIBH) technique. 48,49 And therefore, the PVA-C-based dosimeter may used as a dosimetric bolus for simultaneous skin dose boosting and measurement during radiotherapy to provide an accurate estimation of superficial dose distribution.
The dose estimated using TPS at depths of 0.5-1.0 cm lacks the accuracy typically desired for radiotherapy targets. Thus, a dosimetric bolus material would be useful in simultaneously increasing dose to superficial targets and in ensuring these areas receive the prescribed dose. For megavoltage photon beams, the dose increases up to 60% within the first 0.5 cm depth, making the measured surface dose sensitive to the buildup thickness. 25,29,36 Due to the rapid dose build up, the dose will not be uniform throughout the 0.5 cm thick bolus material, and not equal to the actual surface dose. In this work, a correction factor is determined from the dose measured in the gel dosimeter to the surface dose. Eyadeh et al. (2014) 50 described a FBX-PVA-C material that may be used as radiochromic bolus readable in two dimensions using a simple camera system. The electron density of the FBX-PVA-C material is 1.05 gm/cm 3 , which is almost water equivalent. The material is translucent, allowing visualization of underlying skin marks to assist in patient and beam positioning. The cryogel dosimeter is deformable with good stability and sensitivity, no significant dose rate or energy dependence for this dosimeter; the cryogel signals are constant after 2 hr post exposure with 2.8 9 10 À4 mm À1 cGy À1 rate of base line drift post exposure. 50 The purpose of this work was to evaluate the ability of FBX-PVA-C radiosensitive bolus material to measure skin dose during radiotherapy. The concept is generated using clinical IMRT fields delivered to the head and neck region of a RANDO phantom (Phantom Laboratory, Salem, NY, USA).

2.A. | Translucent FBX PVA-C dosimeter preparation
In this study, translucent FBX-PVA-C was used as radiochromic bolus. A detailed description of its production was described elsewhere. 50 Briefly, PVA concentration of 15% by weight was selected to optimize sensitivity, fabrication time, sturdiness, and ease of han- The hydrogels were subjected to three cycles of 18 hour freezing at À80°C and 6 hr thawing at room temperature. The finished cryogels were removed from their molds and cut to size as necessary. The cryogel is flexible and can conform to most parts of a patient's body.

2.B | Radiochromic bolus read out apparatus
The radiochromic bolus samples were imaged pre-and post irradiation using the equipment shown in Fig. 1

2.C | Calibration of radiochromic bolus and film
The relationship between bolus absorption coefficient and delivered dose was established using a Varian iX linear accelerator (Varian Inc., Palo Alto, CA, USA) with 6 MV photons and 20 9 20 cm 2 field size. The 0.5 cm thick samples of 7 9 7 cm 2 radiochromic bolus were sandwiched at isocenter between two 5.6 cm slabs of polystyrene. Doses ranging from 100 to 4000 cGy were applied with a dose rate of 633 cGy/min. The expected doses in cGy for this arrangement were computed using the Pinnacle v9.2 TPS (Philips, Amsterdam, Netherlands).
The same procedure was employed to relate optical density and dose in 7 9 7 cm 2 pieces of EBT-2 Gafchromic film (lot #A052810-01) (International Specialty Products, Wayne, NJ, USA). Doses ranging from 100 to 1500 cGy were applied with a dose rate of 633 cGy/min. The expected doses in the film were also computed with Pinnacle 9.2. As described later, film was used as the gold standard measurement of skin dose.
The process of film marking, read out, and analysis was consistent with the manufacturer's recommendations. 51 EBT-2 film was also read prior to irradiation to obtain the background optical density; the net optical density of each irradiated film was obtained by subtracting the background optical density. The EBT-2 film was scanned using an Epson 11000XL Scanner (Proscan, Avision, Australia) and analyzed using Film QA TM Pro (Ashland, Wayne, NJ, USA).
The described film calibration curve above was used for the film response. The red channel data were used during film analysis at a resolution of 150 DPI; color correction was disabled. All films were read out approximately 24 hr after their irradiation. Calibration irradiations and subsequent read out were performed at room temperature. Subsequent comparisons of radiochromic bolus and film were also performed in the Film QA software suite.

2.D | Calibration of radiochromic bolus for skin surface dosimetry
Open field irradiations with normal and oblique incidence were used to examine the relationship between the dose distribution recorded by the radiochromic bolus and the true surface dose, which was estimated using EBT-2 film. The configuration of these measurements is shown in Fig. 2. A 0.5 cm thick, 7 9 7 cm 2 radiochromic bolus sample and F I G . 1. In-house 2D optical imaging apparatus. The lens is 61 cm away from the LED array, and the light box is 15 9 15 cm 2 . Excess area on the light surface was masked using black construction paper to improve the dynamic range of the system. F I G . 2. Schematic of the radiochromic bolus and EBT-2 film irradiation. A 3 9 3 cm² field was formed using the jaw collimator and 1000 MU were delivered with a rate of 600 MU/min. The procedure was repeated with gantry rotations ranging from 0°to 90°. Optical density and absorption coefficient measured in film and bolus, respectively, were converted to doses using the respective calibration curves. Finally, the ratio between the film and radiochromic bolus measured dose distributions was obtained to serve as a correction factor for scaling radiochromic bolus dose to dose on the skin surface. The two tangential POP static beams (0°/180°and 90°/270°) were delivered to the 0.5 cm radiochromic bolus and film stack on the surface of the RANDO phantom. The beams were 3 9 3 cm 2 open fields static beam arrangements delivered to the neck with 6 MV photons and 250 MU with a rate of 600 MU/min. Figure 3 shows the placement of the radiochromic bolus.
The two clinical step-and-shoot IMRT plans were delivered as closely to the original planned conditions as possible. The larynx treatment included three step-and-shoot fields, with a total of 276 MU delivered at 400 MU/min. The head and neck treatment employed nine fields to deliver multiple dose levels to superficial disease and various neck nodes with total of 593 MU at 400 MU/min.

3.A | Calibration of radiochromic bolus for skin surface dosimetry
The relationship between absorption coefficient and dose delivered to the 0.5 cm slabs of radiochromic bolus in its linear range was (3.00 AE 0.04) 9 10 À4 mm À1 cGy À1 , which is consistent with our previous measurements. 49,50 Figure 4 shows the irradiated radiochromic bolus and EBT-2 film following exposures of 1000 MU at gantry angles of 0°, 45°, and 90°.
Measurements correlating the dose distribution in bolus to the expected distribution at the underlying surface (estimated using Gafchromic EBT-2 film) indicated that surface dose increased with gantry angle. Dose increases because beam enters obliquely and experiences a larger path length. This is consistent with previous investigations that reported increasing surface doses with incident beam angles with a rapid increase beyond about 45°incidence. 14,52,53 Film QA Pro software (Ashland, Wayne, NJ, USA) was used to visually align the radiochromic bolus and film dose maps. The pointby-point ratio was computed from these aligned images and is shown in Fig. 5. The mean ratios for the irradiated areas are summarized in Table 1. The goal of measuring these ratios was to develop a simple approach to convert dose in bolus to dose on skin. A simpler approach would be to use a single correction factor independent of the angle of incidence. The range of ratios shown in Table 1 F I G .
3. An example of radiochromic bolus definition. The bolus is 0.5 cm thick. seems to suggest this approach may not be feasible. However, if we restrict ourselves to a range of angles, 0°-67.5°for example, we find that the average ratio between radiochromic bolus and film is 0.800 AE 0.064. Using this correction factor, the agreement between the film and the corrected bolus was improved at all gantry angles.
This factor was used to correct all subsequent bolus images.
The mean ratio between radiochromic bolus and film for normal incidence can be estimated by integrating over the percent depth    Kry et al. (2011) reported an overall 22% difference in surface dose between TPS and TLDs. 23 These types of discrepancies arise in model-based TPSs due to challenges in modeling electron contamination and regions of electronic disequilibrium. 25 With the application of the correction factor, good agreement of inline and crossline profiles between the corrected bolus and Gafchromic film was observed at all gantry angles, with average differences ranging from 1.4 to 1.9%. As an example, the central inline (yaxis) and crossline (x-axis) axes of the absolute dose distributions measured at 0°and 90°gantry angles with Gafchromic film, radiochromic bolus, and corrected radiochromic bolus using the derived scaling factor of 0.800 are shown in Figs. 6 and 7, respectively. The corresponding dose profiles were calculated using the TPS, and these are also shown in Figs. 6 and 7.
This suggests that a 0.5 cm radiochromic cryogel should be able to predict dose deposited at the bolus-skin interface. Although the irradiation geometry was quite simple, it was dosimetrically more challenging than realistic treatment geometries where exit dose from some beams may help to mitigate the large differences in buildup seen due to beam obliquity.

3.B | Validation of skin surface dosimetry using radiochromic bolus
The correction factor of 0.800 was applied to the radiochromic bolus measurements for the four treatment plans. This correction factor The mean ratio between surface dose and the dose measured in the radiochromic bolus at different gantry angles ranging from 0°to 90°.

ACKNOWLEDG MENTS
This work has been supported by the NSERC Discovery Grant program and the Yarmouk University Physics Department.

CONFLI CT OF INTEREST
The authors report there are no conflicts of interest.
T A B L E 2 Gamma pass rates for comparisons of corrected radiochromic bolus and Gafchromic film measurements (3%/3 mm, 10% threshold) for different field arrangements. POP, parallel-opposed-pair; IMRT, intensity-modulated radiation therapy.