Home
Search
Collections
Journals
About
Contact us
My IOPscience
An x-ray CT polymer gel dosimetry prototype: II. Gel characterization and clinical application
This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2012 Phys. Med. Biol. 57 3155 (http://iopscience.iop.org/0031-9155/57/10/3155) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 142.104.63.231 The article was downloaded on 02/05/2012 at 16:56
Please note that terms and conditions apply.
IOP PUBLISHING
PHYSICS IN MEDICINE AND BIOLOGY
Phys. Med. Biol. 57 (2012) 3155–3175
doi:10.1088/0031-9155/57/10/3155
An x-ray CT polymer gel dosimetry prototype: II. Gel characterization and clinical application H Johnston 1 , M Hilts 1,2 , J Carrick 1 and A Jirasek 1 1
Department of Physics and Astronomy, University of Victoria, Victoria BC V8W 3P6, Canada Medical Physics, BC Cancer Agency—Vancouver Island Centre, Victoria BC V6R 2B6, Canada 2
E-mail:
[email protected]
Received 27 September 2011, in final form 3 April 2012 Published 1 May 2012 Online at stacks.iop.org/PMB/57/3155 Abstract This article reports on the dosimetric properties of a new N-isopropylacrylamide, high %T, polymer gel formulation (19.5%T, 23%C), optimized for x-ray computed tomography (CT) polymer gel dosimetry (PGD). In addition, a new gel calibration technique is introduced together with an intensity-modulated radiation therapy (IMRT) treatment validation as an example of a clinical application of the new gel dosimeter. The dosimetric properties investigated include the temporal stability, spatial stability, batch reproducibility and dose rate dependence. The polymerization reaction is found to stabilize after 15 h post-irradiation. Spatial stability investigations reveal a small overshoot in response for gels imaged later than 36 h post-irradiation. Based on these findings, it is recommended that the new gel formulation be imaged between 15–36 h after irradiation. Intra- and inter-batch reproducibility are found to be excellent over the entire range of doses studied (0–28 Gy). A significant dose rate dependence is found for gels irradiated between 100–600 MU min−1 . Overall, the new gel is shown to have promising characteristics for CT PGD, however the implication of the observed dose rate dependence for some clinical applications remains to be determined. The new gel calibration method, based on pixel-by-pixel matching of dose and measured CT numbers, is found to be robust and to agree with the previously used region of interest technique. Pixel-by-pixel calibration is the new recommended standard for CT PGD. The dose resolution for the system was excellent, ranging from 0.2–0.5 Gy for doses between 0–20 Gy and 0.3–0.6 Gy for doses beyond 20 Gy. Comparison of the IMRT irradiation with planned doses yields excellent results: gamma pass rate (3%, 3 mm) of 99.3% at the isocentre slice and 93.4% over the entire treated volume. (Some figures may appear in colour only in the online journal)
0031-9155/12/103155+21$33.00
© 2012 Institute of Physics and Engineering in Medicine
Printed in the UK & the USA
3155
3156
H Johnston et al
1. Introduction X-ray computed tomography (CT) polymer gel dosimetry (PGD) has been under development as a technique for three-dimensional (3D) radiation therapy (RT) treatment verification for the past decade. PGD is an attractive method for verifying complex dose distributions (Baldock et al 2010) as polymer gels are tissue equivalent and inherently 3D. Briefly, polymer gels contain radiosensitive chemicals that polymerize in response to absorbed radiation dose and spatial dose information is extracted from the gels using imaging. A variety of imaging techniques can be utilized, including magnetic resonance imaging (Maryanski et al 1993, De Deene 2009, Ceberg et al 2010), optical CT (Gore et al 1996, Oldham et al 2008, Doran 2009), CT (Hilts et al 2000a, Jirasek 2009, Chain et al 2011), ultrasound (Mather et al 2002, Crescenti et al 2010), or Raman spectroscopy (Rintoul et al 2003, Rahman et al 2010). Of these, CT is a highly promising dose read-out technique as it offers a well-developed, easy to use tool that is readily available in every RT clinic. Despite the potential of CT PGD, the clinical utility of the technique has been hindered by poor dose resolution that results from low signal-to-noise and low contrast-to-noise CT images. As a result, research efforts have focused on improving dose resolution through identifying new PGD formulations with enhanced dose sensitivity and developing post processing techniques to improve CT image quality. Recipe optimization studies have identified new dose-sensitive monomers (Senden et al 2006, Koeva et al 2008) and determined methods to improve crosslinker solubility (Koeva et al 2009, Jirasek et al 2009, 2010, Chain et al 2011). A significant improvement in dose sensitivity was found using isopropanol as a cosolvent to increase the concentration of N,N′ -methylenebisacrylamide (BIS) in the gel system (Jirasek et al 2010). More recently, a new high %T formulation without cosolvent was found to exhibit comparable dose sensitivity to the isopropanol recipe (Chain et al 2011). This new formulation, consisting of 15% N-isopropylacrylamide (NIPAM), 4.5% BIS and 5 mM tetrakis hydroxymethyl phosphonium chloride (THPC) as antioxidant, shows great promise for use in CT PGD, and characterization of this dosimeter is the focus of this work. Research in post-processing techniques has shown adaptive mean (AM) filtering to be an optimal method for reducing stochastic image noise (Hilts et al 2005). A companion article to the current work discusses post-processing in more detail and introduces a new remnant artefact removal (RAR) technique that further enhances gel CT images by reducing unwanted structured image noise (Jirasek et al 2012). The present study characterizes the dosimetric properties of the new high %T PGD formulation recently developed by Chain et al (2011). The temporal and spatial stability of the new gel formulation are examined to determine the reliability of gel response. Both intraand inter-batch reproducibility are investigated for doses between 0 and 28 Gy. Dose rate dependence is also examined. Tissue equivalence was confirmed by a separate group using Monte Carlo (MC) modelling studies (Gorjiara et al 2012). Temperature dependencies are not investigated explicitly as gel temperature was carefully controlled during irradiation and imaging, and recent work suggests radiation induced increases in the temperature of high concentration NIPAM gels do not affect their dose response (Sedaghat et al 2011). Energy dependence is not investigated as all irradiations are performed using 6 MV photon beams. In order to benefit from the sensitivity and large dynamic range of the new gel formulation, we test and utilize a new pixel-by-pixel calibration method. This technique offers advantages over the traditionally used region of interest (ROI) techniques (Hilts et al 2000a) where mean values of CT number and dose are extracted from corresponding user defined, uniform dose ROIs in a gel CT image and planned dose distribution. The new pixel-by-pixel technique has the potential to be automated, fast and user-independent, and eliminates potential inaccuracies
X-ray CT PGD: gel characterization and prototype application
(a)
3157
(b)
Figure 1. Head and neck CT PGD phantom used to irradiate 1 L gels. (a) The phantom consists of an anthropomorphic head cast and acrylic base plate with a gel holder (black). (b) The head and neck phantom at the treatment unit with the cast in place and filled with water.
due to discrepancies between calibration and test gels (De Deene et al 2006). In moving away from ROI based calibration, this work follows in the footsteps of work by Oldham et al (1998) that utilized a percentage depth dose for gel calibration. With research focused on technique development, there have been very few reports on clinical applications of CT PGD in the literature (Audet et al 2002, Hilts et al 2000b, Ghavami et al 2010). The study of Audet et al (2002) is the most complete and although this work illustrated the early promise of CT PGD, it also highlighted the dose resolution limitations of the technique. To illustrate the enhanced potential for clinical application offered by the next generation CT PGD system introduced here, an example application for an intensity-modulated radiation therapy (IMRT) treatment validation is presented in this work. 2. Materials and methods 2.1. Gel preparation All polymer gels were manufactured as described in a recent companion work (Jirasek et al 2012). For the temporal stability study, the final gel solution was poured into 20 mL scintillation vials (Wheaton Scientific, Millville, NJ) and sealed in cylindrical acrylic phantoms as described in Hilts et al (2004). For all remaining studies, including the IMRT treatment validation, completed gel solutions were transferred to 1 L, cylindrical, polyethylene terephthalate jars (Modus Medical Devices, London, ON). Gels were wrapped in aluminum foil to prevent photo-initiated polymerization and refrigerated until solid for at least 1 hour for gel vials and 6 h for 1 L jars. When utilized, blank gels were manufactured as above using 5% gelatin, 0.5% NIPAM and 0.5% BIS (no THPC). 2.2. Head and neck phantom A customized head and neck phantom was used for irradiation of all 1 L gels. The phantom, shown in figure 1, was designed and constructed specifically for performing clinical CT PGD dose measurements. It is optimized for low CT image noise and artefacts, and to provide accurate and reproducible localization throughout the 3D dosimetry process. A removable clear, molded perspex head (1 mm thick) is fastened to an acrylic base plate. Plastic screws
3158
H Johnston et al
are used to minimize both dose heterogeneity effects during irradiation and CT artifacts at gel read-out. A precision machined gel holder (Delrin) secures the 1 L cylinder within the phantom. Once screwed on, the phantom head is water tight and the phantom can be filled with water for irradiation. The design also allows for removal of the head while maintaining gel position, a feature required to optimally reduce image noise during CT read-out (Hilts et al 2011). Finally, the phantom can be precision positioned on the treatment and CT couches using a locking bar or alternately using a clinical mask-type immobilization system. 2.3. Treatment planning and irradiation All polymer gels were irradiated with 6 MV photons using a Varian Clinac 21EX linear accelerator (Varian Medical Systems, Palo Alto, CA), between 5 and 7 h after fabrication (unless otherwise indicated). Gels were maintained at a consistent temperature during and after irradiation to ensure reproducible reaction kinetics across all experiments (Salomons et al 2002). Machine dose rate was 400 MU min−1 unless otherwise noted. All treatment planning was performed using the Eclipse treatment planning system (TPS) (Varian Medical Systems, Palo Alto, CA) running version 10.0 of the analytical anisotropic algorithm. 2.3.1. Gel characterization. For the temporal stability study, gel vials were irradiated individually, within their cylindrical acrylic phantoms, using a set-up described previously (Hilts et al 2004). A uniform dose of 10 Gy was delivered to each vial using two parallelopposed, 10 × 10 cm2 fields. For the spatial stability study, the head and neck phantom, with anthropomorphic cast removed, was used to position the gel at planning CT and radiation delivery. A treatment plan was designed to deliver a uniform dose of 10 Gy to half the gel container. The plan consisted of two parallel opposed, 20 × 8 cm2 wedged fields. The axis of each field was perpendicular to the cylinder axis. The beam arrangement and isocentre dose distribution for the plan are shown in figures 2(a) and (d), respectively. Gels for the batch reproducibility and dose rate studies were positioned at planning CT and radiation delivery using the head and neck phantom with the cast in place and filled with water. Each gel was irradiated with the treatment plan designed for gel calibration described in section 2.3.2 (figures 2(b) and (e)). Intra-batch reproducibility was examined locally by delivering the calibration plan to one batch of gel at the bottom of the 1 L cylinder, and globally by delivering the calibration plan to a second batch of gel at both ends of the cylinder. In this case, where two separate dose distributions were delivered to a single gel, irradiation isocentres were spaced at least 6.5 cm apart to minimize scattered radiation between neighbouring distributions. For the inter-batch study, the plan was delivered to three independently fabricated gel batches at identical locations along the length of each cylinder. Dose rate was examined by delivering the calibration plan to two regions of a 1 L gel batch (as described for global intra-batch irradiations) at 100 and 600 MU min−1 , respectively. All dose responses curves were measured from 1 L gels, thereby avoiding any possible discrepancies in dose response due to container size (Dumas et al 2007). 2.3.2. Gel calibration and IMRT validation. Treatment plans for gel calibration and IMRT treatment validation were designed using a planning CT scan of the head and neck phantom with the head cast secured and filled with water. Figure 2(b) shows the simple beam arrangement used for the gel calibration plan, consisting of three intersecting 3 × 3 cm2 fields. The plan was designed to produce a full range of doses that could be accurately predicted by the TPS. The resulting, planned dose distribution at isocentre is illustrated in figure 2(e).
X-ray CT PGD: gel characterization and prototype application
3159
(a)
(b)
(c)
(d)
(e)
(f)
Figure 2. Treatment plan beam arrangements and calculated dose distributions (isocentre) for the spatial stability study ((a) and (d)), gel calibration, batch reproducibility and dose rate studies ((b) and (e)) and IMRT treatment validation ((c) and (f)).
The prescription dose was 26 Gy, with a maximum dose of 27.6 Gy (106.3%). Based on commissioning data, the expected accuracy of the TPS dose calculation for the geometry shown in figure 2(b) is better than 2%. This was further validated by MC dose calculation using the Vancouver Island MC system, described in detail in Bush et al (2008). Doses computed using MC and the TPS were found to agree within 2.5% for all doses used for gel calibration except for a small region (10% of pixels at ∼10 Gy) which showed discrepancies of up to 4%. Agreement for high doses (>20 Gy) was within 1.5%. The beam arrangement for the IMRT treatment plan is illustrated in figure 2(c). The plan consisted of five fields spread over anterior gantry angles, designed to treat a concave PTV, the inferior half of which is wrapped around an avoidance structure. Figure 2(f) illustrates the planned dose distribution (isocentre slice). The plan prescription dose was 22 Gy, with a maximum dose of 24.8 Gy (112.5%). Gel calibration and IMRT treatment plans were delivered to the same gel dosimeter. This streamlined the dosimetry process and ensured calibration and treatment irradiations were performed on phantoms of the same size, thereby avoiding any discrepancy in dose response due container size (Dumas et al 2007). Treatment isocentres were placed with maximum longitudinal (superior–inferior) separation in order to minimize the scattered dose (
