|Year : 2018 | Volume
| Issue : 2 | Page : 292-299
A novel quantification method for low-density gel dosimeter
Hasan Ali Nedaie1, Farideh Pak2, Vahid Vaezzadeh3, Ehsan Eqlimi4, Abas Takavar4, Hamid Reza Saligheh Rad5, Mohammad Amin Mosleh Shirazi6, Mona Mirheydari7
1 Department of Medical Physics and Biomedical Engineering, Faculty of Medicine, Tehran University of Medical Sciences; Department of Radiotherapy Oncology, Cancer Research Centre, Cancer Institute; Tehran, Iran
2 Department of Radiation Science, School of Allied Medical Sciences, Tehran University of Medical Sciences, Tehran, Iran
3 Department of Radiotherapy Oncology, Cancer Research Centre, Cancer Institute, Tehran University of Medical Sciences, Tehran, Iran
4 Department of Medical Physics and Biomedical Engineering, Faculty of Medicine, Tehran University of Medical Sciences, Tehran, Iran
5 Department of Medical Physics and Biomedical Engineering, Faculty of Medicine, Tehran University of Medical Sciences; Research Center for Molecular and Cellular Imaging, Tehran, Iran
6 Department of Radiotherapy and Oncology, Medical Imaging Research Center, Shiraz University of Medical Sciences, Shiraz, Iran
7 Department of Physics, Kent State University, Kent, OH, USA
|Date of Web Publication||8-Mar-2018|
Dr. Farideh Pak
Department of Radiation Science, School of Allied Medical Sciences, Tehran University of Medical Sciences, West Taleghani Avenue, Poursina Street, Tehran 14174
Source of Support: None, Conflict of Interest: None
Aim: Low signal-to-noise ratio (SNR) images of lung-like (low-density [LD]) gel dosimeters, compared to unit-density (UD) gels, necessitate the use of different quantification methods.
Setting and Design: In this study, a new method is introduced based on noise correction and exponential (NCEXP) fitting. The feasibility of NCEXP method for quantifying dose absorption in LD gels is evaluated.
Materials and Methods: Sensitivity, dose resolution, detectable dynamic range, and correlation of the calibration curve for both UD and LD gel dosimeters are the parameters, which we analyze to investigate the consequences of new method. Results of NCEXP method are compared to maximum likelihood estimation of rician distribution (MLE-R) and variable echo number (VAREC) quantification methods.
Results: Dose response of LD gel dosimeter shows wider detectable dynamic range as compared to UD gel. Using NCEXP method for both LD and UD dosimeter gels, a more sensitive calibration curve with a superior dose resolution is obtained. The advantage of new quantification method is more significant for LD dosimeter gel analysis, where SNR decreases as a result of higher absorbed doses (≥10 Gy). Despite the inverse effect of the VAREC method on detectable dose range of UD gel, no specific changes are observed in dynamic dose range of LD gel dosimeter with different quantification methods. The correlations obtained with different methods were approximately of the same order for UD and LD gels.
Conclusion: NCEXP method seems to be more effective than the MLE-R and VAREC methods for quantification of LD dosimeter gel, especially where high-dose absorption and steep-dose gradients exist such as those in intensity-modulated radiation therapy and stereotactic radiosurgery.
Keywords: Fitting algorithm, low-density polymer gel dosimeter, lung equivalent gel, quantification method
|How to cite this article:|
Nedaie HA, Pak F, Vaezzadeh V, Eqlimi E, Takavar A, Saligheh Rad HR, Mosleh Shirazi MA, Mirheydari M. A novel quantification method for low-density gel dosimeter. J Can Res Ther 2018;14:292-9
|How to cite this URL:|
Nedaie HA, Pak F, Vaezzadeh V, Eqlimi E, Takavar A, Saligheh Rad HR, Mosleh Shirazi MA, Mirheydari M. A novel quantification method for low-density gel dosimeter. J Can Res Ther [serial online] 2018 [cited 2020 May 28];14:292-9. Available from: http://www.cancerjournal.net/text.asp?2018/14/2/292/209956
| > Introduction|| |
Recent advancement in computer-based techniques has made radiotherapy treatment plans complex in three dimensions (3D). Hence, interests in the development of 3D radiation dose-measuring devices have increased. Gel dosimeters have proven to be reliable and precise radiation detectors to measure dose distributions in 3Ds, especially in dosimetry situations where steep-dose gradients exist, for example, in intensity-modulated radiation therapy (IMRT) and stereotactic radiosurgery (SRS)., Development in gel dosimeters allows evaluating the accuracy of patient treatments with complex field shapes and large-dose gradients delivered by advanced radiation systems.
In most applications of gel dosimetry, density of used gel is close to water.,,, These gels are excellent radiation detector devices for the evaluation of dose absorption in homogeneous media. Verification of adequate dose for different targets in the body has become of researchers interest, which led to introducing a polymer gel with the possibility of simulating different electron densities and inhomogeneous tissues like lung.,, Haraldsson et al. made a low-density (LD) gel dosimeter using styrofoam beads distributed within a polymer dosimeter gel. However, compared to unit-density (UD) gel, the dynamic range of the measured dose for LD gel was reduced, and linearity was reported to be only between 2 Gy and approximately 8 Gy. In 2013, De Deene et al. made a heterogeneous phantom consisting of UD and LD polymer dosimeter gel. A significant deviation was observed in depth-dose profile of LD gel with an overestimation of measured dose from treatment planning system calculated. Although usage of LD gel dosimeters in new techniques can improve dosimetry systems in different treatment situations extensively, a narrow detectable dynamic dose range and inconsistency of the measured dose with dose profile are two drawbacks that limit their clinical use.
Dose absorption in a gel phantom is correlated to the spin–spin relaxation rates (R2 =1/T2), which can be extracted by applying an exponential fitting algorithm to a train of echo signals of multiple spin echo images. Accuracy and precision of R2 calculations are mainly dependent on applied fitting algorithm and echo numbers. When echo signal intensity decreases to the background signal offset level for a long echo time (TE), applying such echoes results in anomalous R2 values. To solve the problem, some algorithms were proposed that used echoes whose signal intensities were greater than a preset threshold., The frequently used method maximum likelihood estimation-R (MLE-R) suggests using signal-to-noise ratio (SNR) as a threshold, and just echoes with SNR over three are approved for R2 calculation. In variable echo number (VAREC) method directed by Watanabe et al., threshold was defined as standard deviation of the Gaussian noise times a multiplier α. Both of these methods were applied for analysis of UD dosimeter gels, which had homogenous structure and higher SNR, in comparison to LD dosimeter gels. Lung-like gels experience lower SNR as a result of their lower density and inhomogeneous structure. Using the same imaging protocol with long TE, the measured signal intensity (SM) in LD gel images decreases very quickly as compared to UD polymer gel. However, it never becomes zero and varies randomly around σ√π/2 (background signal). A nonzero mean distribution of noise in the absence of signal causes SM not to decay exponentially with increasing TE; thus, the exponential quantification methods can produce anomalous T2 and consequently R2 values.
Miller et al. stated that, in low-SNR images, signal separation from noise can be achieved by computing power images from traditional magnitude data, over region of interest (ROI). It was shown that, if all the images in a T2 series convert to the corrected power images, the noise will average to zero and an exponential fit to the data from the power images yields a time constant of half of the true T2 value over an ROI. He et al. applied a noise correction method for T*2 measurement. They suggested that the T*2 decay curve in an ex vivo heart can be fitted by a monoexponential model and accurate T*2 measurements can be obtained with proper noise correction. Based on the work done by Miller et al., a method for voxelwise T2 calculation in low-SNR images of articular cartilage was established by Raya et al.
Miller and Joseph applied a noise correction algorithm for analysis of heterogeneous media, which is independent of signal strength, although the feasibility was not investigated.
To calculate R2 in low-SNR images of LD gel dosimeters, analysis should be performed using a new quantification method which takes into account the noise distribution. In this work, we have established a new quantification method based on a noise-corrected algorithm to calculate R2 values from heterogeneous and low-SNR images of lung-equivalent gel dosimeters. The performance of the method is evaluated in terms of dose resolution, sensitivity, and detectable dynamic dose range of LD polymer dosimeter gels. Results are compared with MLE-R and VAREC quantification methods.
| > Materials and Methods|| |
In this case, MAGAT gel dosimeter was chosen due to its high sensitivity, to compensate the expected loss in SNR as a result of the lower density. The gel was prepared in a laboratory condition, using protocol presented by Hurley et al.
MAGAT gel consisted of 86% deionized water, 8% gelatin (300 Bloom, Sigma-Aldrich), 6% methacrylic acid (purity grade approximately 99%, Sigma-Aldrich), and 50 mM of tetrakis (hydroxymethyl) phosphonium chloride (THPC) (technical grade 80% in water, Sigma-Aldrich) in weight percentage.
Gelatin was soaked in 80% of deionized water for 10 min at room temperature to expand and become uniform. Then, the mixture was heated under magnetic stirring until the temperature reached 50°C and a clear solution resulted. Methacrylic acid was added to the mixture after cooling it down to 35°C. Ten minutes later, a solution of antioxidant was prepared with THPC and 20% remaining of deionized water and added to the solution.
Two sets of gel dosimeters were prepared (LD and UD). The prepared UD polymer dosimeter gel solution was split into 12 testing vials. To prepare LD gels, the solution was transfused into the vials containing polystyrene spheres (Styrofoam™ spheres, Isopan, Regensburg, Germany). The polystyrene spheres' diameter varied between approximately 0.8 and 1.3 mm. To prevent possible photopolymerization, the gel samples were stored in a card box. Samples were left in a conventional refrigerator at 4°C for about 1 h to solidify and then transferred into a cupboard.
Irradiations were run approximately 24 h postpreparation of dosimeter gel on a Varian2100 C/D linear accelerator (Varian Medical Systems, Palo Alto, CA, USA). Test vials were placed in a rectangular in-house built water bath, which was designed to simultaneously expose multiple gel vials to different doses. For dose–response evaluations, vials were irradiated with 6 MV photons to deliver absorbed doses between 1 and 20 Gy. In this work, it is assumed that the dose absorption of calibration vial is close to the dose absorption in water. As the photon beam width is relatively large (15 cm × 30 cm) and sample vial has small dimension, the introduced error is restricted to a few percent.
The radiation was conducted perpendicular to the length of the test tubes. To prevent any dose gradient in gels along the diameter of the test tube, they were turned 180° halfway through the irradiation.
To calculate dose from R, the equation, R, = Rbg+ b.D was used to fit the calibration data, where Rbg, b, and D are background R, slope, and absorbed dose. respectively. One sample was left unirradiated for background measurement.
The irradiated gels were kept in the card box in magnetic resonance imaging (MRI) room temperature until the imaging was done.
Magnetic resonance imaging evaluation
Since R2 values are affected by gel temperature at MRI time, samples were transferred to MRI room 24 h before the procedure to start. This prevents samples from any temperature fluctuation.
Gels' MRI were taken using Siemens Magnetom Avanto 1.5 T scanner (Siemens Medical Solutions, Erlangen, Germany). Vials were allocated in the head-coil and T2 was determined using a multiple spin echo sequence. For all the measurements, a time to repeat of 4000 ms was used, with 32 TE ranging from 20 to 640 ms with increment rate of 20, and for each scan, two acquisitions were averaged (NEX = 2).
T2 relaxation data were transferred to a personal computer and processed to calculate R2 image “maps” using software developed in-house using MATLAB. R2 values were calculated by fitting an exponential T2 decay curve on signal intensities of corresponding pixels in the base images versus TE.
Maximum likelihood estimation-R method
In the conventional fitting algorithm, the R2 value is calculated by fitting the signal intensities of corresponding pixels in images versus TE with an exponential T2 decay curve:
where S0 is the noise-free signal intensity at zero TE and S stands for the noise-free signal intensity at TE.
As mentioned, when signal intensity of echoes drop to the background signal offset level for a long TE, applying such echoes in R2 calculation results in anomalous R2 values. In the MLE-R method, echoes with SNR <3 were excluded from R2 calculation.
Where no true signal is present, the average value measured over an ROI (SBG) is related to standard deviation (σ) of noise:
SBG was calculated as a mean of the pixel values in a 4 cm × 4 cm square outside the polymer gel phantoms but inside the field of view in echo images.
Many point method was used in calculations for all quantification methods.
Since the MLE-R method is a common method of R2 calculation in a gel dosimetry system, the results of the other methods were compared with its results as a reference.
Variable echo number method
In the VAREC method for R2 calculation, a threshold was defined as standard deviation (σ) times a multiplier (α). When the average signal intensity of three signals beyond the Ith echo was equal to or smaller than the threshold, then number of used echoes used was set to I for R2 estimation. To determine the maximum number of echoes, the VAREC technique was recommended with 2–3 times the standard deviation of the Gaussian noise as the threshold (α = 2–3). In this study, α = 2 was implemented following Watanabe and Kubo.
Noise corrected and exponential method
In low-SNR images, the nonzero mean distribution of noise causes the measured signal intensity (SM) not to decay exponentially with increasing TE.
For an ROI, the average power value (SM2) is given by:
where S represents the true signal and the variance 2σ2 can be estimated as the power of the background signal.
Offset () level for a long TE:
As can be seen from equation 3, the noise is apparently additive to the power signal and can be corrected by simple subtraction:
In this case, S is an exponential and converting all echo images in a T2 series to corrected images and exponential fit to these images can yield a time constant of the true R2 value in low-SNR images.
In the new quantization method, noise correction and exponential (NCEXP) method was performed on each echo images of gel vials with different absorbed doses.
R2 values were calculated using NCEXP, MLE-R, and VAREC methods implied in MATLAB code. Detectable dose range, correlation, R2–dose sensitivity, and also dose resolution of polymer gel response were calculated to evaluate the performance of the fitting algorithms. Correlation analysis was performed using the Pearson test.
Dynamic detectable dose range of polymer gel dosimeters was described by fitting quasilinear increase of R2 versus absorbed dose between 1 Gy and the saturation dose linearly.R2–dose sensitivity was defined as the slope of this linear fit and was calculated using the following equation:
Performance of a polymer gel dosimeter cannot be assessed based on the R2–dose sensitivity only and should be coupled with the ability of measurement technique to provide small uncertainties. The uncertainty in dose related to the uncertainty of R2 in polymer gel dosimetry can be calculated bydose resolution (). Dose resolution is defined as 'the minimal separation between two absorbed doses such that they may be distinguished by a given level of confidence, p' For 95% confidence level, dose resolution is related to the standard uncertainty of R2:
where u (R2) is the standard uncertainty of a specific R2, approximated as equal to the standard deviation of R2.
| > Results|| |
[Figure 1] shows the quantitative R2 image obtained from two sets of LD (two upper rows) and UD gel (two lower rows) dosimeters. While the LD gels show nonhomogeneous structure, a homogeneous image of UD gels can be seen. The entire phantom was scanned at the same time. Signal intensities for all of the gel dosimeter were averaged over a small circular area inside the vials.
|Figure 1: The spatial arrangement of low- and unit-density polymer gel vials used for magnetic resonance imaging scans. The delivered doses to each vial are also indicated|
Click here to view
Dose response of the unit- and low-density gel dosimeters
Dose–response curves of LD and UD dosimeter gels are shown in [Figure 2]. [Figure 2]a shows the results of nonthresholding while considering all echo signals for R2 calculation. In UD dosimeter gel, it is seen that R2 values increase slowly by increasing delivered dose up to 10 Gy, while there is only an increase up to 2 Gy in the dose response of LD gel. [Figure 2]b shows R2 values calculated by the MLE-R method. A threshold of SNR = 3 was applied in these calculations. With this quantification method, detectable dose range of LD gel increased to 20 Gy, while no difference in detectable dose range of UD gel response was observed.
|Figure 2: Dose–response curve of low- and unit-density gel dosimeters for (a) nonthresholding and (b) maximum likelihood estimation-R methods|
Click here to view
Variation in dose response of low-density gel dosimeter with different quantification methods
[Figure 3]a shows R2 values as a function of dose for nonthresholding, NCEXP, MLE-R, and VAREC methods in LD gel. The results of the nonthresholding method show divergent R2 values for the whole measured dose range. R2 values obtained by the NCEXP, MLE-R, and VAREC methods show a quasilinear trend over the measured dose range with different slopes. As it is shown in [Table 1], R2–dose sensitivity for NCEXP increased about 14% (doses ≤20 Gy) and 9% (doses ≤10 Gy) while decreased about 8% (doses ≤20 Gy) and 18% (doses ≤10 Gy) for the VAREC method, comparing to values measured by MLE-R method. The correlations obtained with the different methods were similar (with 0.6% deviation).
|Figure 3: (a) R2values as a function of absorbed doses for the nonthresholding, noise correction and exponential, maximum likelihood estimation-R, and variable echo number methods in low-density gel. (b) Replot of (a) for doses ≤10 Gy and (c) for doses in 10 Gy–20 Gy interval|
Click here to view
|Table 1: R2–dose sensitivity and correlation in low- and unit-density gel dosimeters for the nonthresholding, noise correction and exponential, maximum likelihood estimator-R, and variable echo number fitting methods|
Click here to view
[Figure 3]b and c are replots of [Figure 3]a for doses below 10 Gy and between 10 and 20 Gy, respectively. In [Figure 3]b, R2 values of the NCEXP and MLE-R methods agree within the measured standard deviations of <5% (except at 4 and 10 Gy). For doses larger than 10 Gy, the uncertainty of R2 values increase to >5% (range 7%–13%).
Variation in dose response of unit-density gel with different quantification methods
[Figure 4]a shows R2 values as a function of absorbed dose in UD gel dosimeter for nonthresholding, NCEXP, MLE-R, and VAREC methods. Results of the nonthresholding method show small increase in R2 values as a function of dose, for doses below 10 Gy. [Figure 4]b shows data, like [Figure 4]a, for the NCEXP, MLE-R, and VAREC methods, which only presents data for doses below 10 Gy. R2 values calculated with the VAREC method inaccurately decreased for doses >6 Gy. A quasilinear trend is achieved for data fitted using NCEXP and MLE-R methods which is true for doses smaller than 10 Gy. Results from these two methods are in agreement within 5% of standard deviation. R2 values vary more randomly for doses >10 Gy. From [Table 1], it can be seen that the R2–dose sensitivity of the UD gel measured by NCEXP is greater than the MLE-R method (≤10 Gy).
|Figure 4: (a) R2values as a function of dose absorption in unit-density gel for the nonthresholding, noise correction and exponential, maximum likelihood estimation-R, and variable echo number methods.(b) A replot of (a) for dose ≤10 Gy with the VAREC, maximum likelihood estimation-R, and noise correction and exponential methods|
Click here to view
Dose resolution of unit- and low-density gel dosimeters
In [Figure 5]a, is plotted as a function of dose absorption for LD gel with different quantification methods. Within the measured dose range for LD gels, NCEXP method gives a lower than other quantification methods. Interestingly, dose resolution by both methods VAREC and MLE-R has almost the same value in a dose range smaller than 14 Gy. However, except 8 Gy and doses over 14 Gy, dose resolution values obtained by MLE-R method is slightly smaller than those calculated by VAREC method. [Figure 5]b shows the corresponding results for the UD gel. The dose resolution of the three methods are consistent in a dose range smaller than 4 Gy; however, in a dose range from 4 to 8 Gy, there is small decrease for dose resolution calculated by NCEXP method, where there is an increased in analyzed values obtained by VAREC method in a dose range over 6 Gy.
|Figure 5: Dose resolution of (a) low-density and (b) unit-density gel as a function of dose absorption, with different quantification methods (variable echo number, maximum likelihood estimation-R, and noise correction and exponential)|
Click here to view
| > Discussion|| |
The present work involves studying the effect of quantification methods on dose response for UD and LD gel dosimeters. Furthermore, a new method (NCEXP) based on NCEXP fitting is introduced for LD gel analysis. The effects of the MLE-R, VAREC, and NCEXP methods on detectable dynamic dose range, R2–dose sensitivity, and dose resolution of these two types of gels were determined.
Although both UD and LD dosimeter gels have same compounds and were prepared simultaneously, the LD dosimeter gel shows wider dynamic detectable dose range as compared to UD gels. In doses higher than 10 Gy, R2 values of LD dosimeter gel increases with an increase in delivered dose, whereas dose response of UD dosimeter gel is saturated. The only difference between LD and UD gels is inclusion of styrofoam beads in LD polymer dosimeter gels. The oxygen carried by these beads to the gel structure can be responsible for discrepancy. Sedaghat et al. stated that oxygen and antioxidants both act as radical scavengers that affect the amount of polymer formed in gel dosimeter. They provided evidence that antioxidants not only scavenge oxygen but also react with other radicals in dosimeter gel. The conventional protocol for MAGAT dosimeter gel contains 5 mM of antioxidant. However, to remove any probability of oxygen contamination in LD gel, 50 mM of THP is used for gel preparation based on Haraldsson et al.'s study. According to the results, prohibiting role of excess amount of antioxidant was evident in response of UD gel. However, in LD gel, the oxygen which is released by the styrofoam beads may react with extra amount of antioxidant and somehow neutralized its inverse effect on polymerization process. Since amount of unreacted antioxidant left in gel dosimeter will have an impact on polymerization reaction, this would be an extreme coincidence if there is same ratio of unreacted THPC in all gel vials. Further investigation is needed to explore possibility of other sources of discrepancy.
The dose response of UD gel is found to be linear approximately in a dose range smaller than 10 Gy, which agrees well with the results reported by De Deene et al. and Sedaghat et al.,
Dose response of LD gel increases over the measured dose range (20 Gy) and also a sharp increase in R2 values was seen in 0–2 Gy interval, which is in contrast with the findings of Haraldsson et al. They reported a linear dose response just between 2 and 8 Gy. Although same steps of gel preparation, irradiation, and imaging were followed in both experiments, differences may have arisen from using different fitting methods and threshold application in data processing. As it was mentioned before, weak echo signals can degrade R2 values significantly. In Haraldsson et al.'s study, a two-point fit method was applied for R2 extraction with no report of thresholding, while we use a many-point fit method in R2 calculation and those echoes with SNR <3 were excluded from data analysis (which would end up with more accurate analysis).
[Figure 3] shows that using different fitting algorithms had no effect on detectable dynamic dose range of LD gel while their effect on R2–dose sensitivity was quite visible [Table 1]. For doses <10 Gy, LD dosimeter sensitivity increases about 9% with NCEXP method and reduces about 18% by VAREC method. Dose sensitivity of doses <20 Gy increases 14% with NCEXP fitting method while decreases about 8% by VAREC method. To define the performance of dosimeter gel in detecting a range of dose absorption, besides reporting R2–dose sensitivity, uncertainty of the measurement technique should also be included. An increase in sensitivity is insignificant if standard deviation is also increasing. For dose resolution calculation, both gel sensitivity and standard deviation in fit values are taken into account. It evaluates intrinsic dosimetric precision in terms of dose sensitivity values and SNR scanning. Comparing R2–dose sensitivities in a wide dose range shows the validity of NCEXP method in contrast to other methods investigated in this work. Results of dose resolution [Figure 5]a support this statement. NCEXP method is based on eliminating statistical noise from the received signal, which was interpreted by Miller and Joseph as a necessary step to be performed over low-SNR images before the fit. As dose absorption in gel increases, image's SNR decreases. In other words, gels' MRI with high-dose absorption lie within low-SNR images, comparing to gel images with a lower level of dose absorption. The performance of the NCEXP method in different dose ranges can be explained by considering the SNR with various dose absorptions. At higher doses, the contribution of statistical noise in the received signal is more significant. It seems that, by removing the effect of noise from the received signals, the T2 curve will descend more quickly with lower uncertainty, which leads to a better differentiation between dose levels. This can result in a higher R2–dose sensitivity and also a better dose resolution.
It is expected that VAREC method shows weaker performance in R2 dose sensitivity as it has a lower threshold compared to MLE-R method. However, [Figure 5]a shows that both techniques result in very close values for dose resolution. A lower threshold in the VAREC method leads to contribution of further number of echoes in R2 calculation. Having more data points for fitting leads to more precise parameter estimation and uncertainty reduction. Thus, it is apparent that the more uncertainty with the MLE-R method cancels out its higher sensitivity compared to the VAREC method. Results demonstrate the discrepancy between the analysis using and R2–dose sensitivity. The R2–dose sensitivity of the UD gel with the NCEXP increased by 11% compared to MLE-R method for dose range <10 Gy. It shows that, even in the UD gels with higher SNR, noise can reduce the sensitivity of dosimeter gel and result in degradation of 3D map of R2.
[Figure 4]a shows that the detectable dose range of UD gel was less using the VAREC method (up to 6 Gy). This is in good agreement with the data reported by Watanabe and Kubo. They also reported saturation of detectable dose in lower doses with VAREC method.
| > Conclusions|| |
Sensitivity, dose resolution, detectable dynamic range, and correlation of the calibration curve for both UD and LD gel dosimeters were the parameters, which were analyzed to evaluate the performance of the NCEXP, MLE-R, and VAREC quantification methods in this study. Dose response of LD gel dosimeter showed wider detectable dynamic dose range comparing to UD gel. Although a more sensitive calibration curve with a superior dose resolution was obtained by the NCEXP method in both LD and UD dosimeter gels, its effect was more significant for LD dosimeter gels analysis where SNR decreases as a result of higher dose absorptions (≥10 Gy). Despite the inverse effect of VAREC method on detectable dose range of UD gels, no specific variations with different quantification methods on the dynamic dose range of LD gels were observed. The correlations obtained with different methods were approximately of the same order for both UD and LD gels.
NCEXP method seems to be more effective than the MLE-R and recently introduced methods for LD gel dosimetry systems, especially when high-dose absorption and steep dose gradients exist, like those in IMRT and SRS.
The assistance of Mr. Shojaee Moghadam from Payambaran Hospital for MRI of polymer gel dosimeters is acknowledged.
Financial support and sponsorship
This study was supported by Grant No. 25920 from Tehran University of Medical Sciences.
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Baldock C, De Deene Y, Doran S, Ibbott G, Jirasek A, Lepage M, et al.
Polymer gel dosimetry. Phys Med Biol 2010;55:R1-63.
Hassani H, Nedaie HA, Zahmatkesh MH, Shirani K. A dosimetric study of small photon fields using polymer gel and Gafchromic EBT films. Med Dosim 2014;39:102-7.
McJury M, Oldham M, Cosgrove VP, Murphy PS, Doran S, Leach MO, et al.
Radiation dosimetry using polymer gels: Methods and applications. Br J Radiol 2000;73:919-29.
Farajollahi AR, Pak F, Horsfield M, Myabi Z. The basic radiation properties of the N-isopropylacrylamide based polymer gel dosimeter. Int J Radiat Res 2014;12:347-54.
Venning AJ, Nitschke KN, Keall PJ, Baldock C. Radiological properties of normoxic polymer gel dosimeters. Med Phys 2005;32:1047-53.
Gorjiara T, Hill R, Bosi S, Kuncic Z, Baldock C. Water equivalence of NIPAM based polymer gel dosimeters with enhanced sensitivity for x-ray CT. Radiat Phys Chem 2013;91:60-9.
Sellakumar P, Samuel EJ, Supe SS. Water equivalence of polymer gel dosimeters. Radiat Phys Chem 2007;76:1108-15.
Lee CC, Wu JF, Chang KP, Chu CH, Wey SP, Liu HL, et al
. The use of normoxic polymer gel for measuring dose distributions of 1, 4 and 30 mm cones. Radiat Phys Chem 2014;104:221-4.
De Deene Y, Vandecasteele J, Vercauteren T. Low-density Polymer Gel Dosimeters for 3D Radiation Dosimetry in the Thoracic Region: A Preliminary Study. Paper Presented at: Journal of Physics: Conference Series; 2013.
De Deene Y, Vergote K, Claeys C, De Wagter C. Three dimensional radiation dosimetry in lung-equivalent regions by use of a radiation sensitive gel foam: Proof of principle. Med Phys 2006;33:2586-97.
Haraldsson P, Karlsson A, Wieslander E, Gustavsson H, Bäck SA. Dose response evaluation of a low-density normoxic polymer gel dosimeter using MRI. Phys Med Biol 2006;51:919-28.
Deene YD, Walle R, Achten E, Wagter CD. Mathematical analysis and experimental investigation of noise in quantitative magnetic resonance imaging applied in polymer gel dosimetry. Signal Processing 1998;70:85-101.
De Deene Y, Baldock C. Optimization of multiple spin-echo sequences for 3D polymer gel dosimetry. Phys Med Biol 2002;47:3117-41.
Watanabe Y, Kubo H. A variable echo-number method for estimating R2
in MRI-based polymer gel dosimetry. Med Phys 2011;38:975-82.
Watanabe Y, Perera GM, Mooij RB. Image distortion in MRI-based polymer gel dosimetry of gamma knife stereotactic radiosurgery systems. Med Phys 2002;29:797-802.
Raya JG, Dietrich O, Horng A, Weber J, Reiser MF, Glaser C. T2
measurement in articular cartilage: Impact of the fitting method on accuracy and precision at low SNR. Magn Reson Med 2010;63:181-93.
Miller AJ, Joseph PM. The use of power images to perform quantitative analysis on low SNR MR images. Magn Reson Imaging 1993;11:1051-6.
He T, Gatehouse PD, Kirk P, Mohiaddin RH, Pennell DJ, Firmin DN. Myocardial T*2 measurement in iron-overloaded thalassemia: An ex vivo
study to investigate optimal methods of quantification. Magn Reson Med 2008;60:350-6.
Hurley C, Venning A, Baldock C. A study of a normoxic polymer gel dosimeter comprising methacrylic acid, gelatin and tetrakis (hydroxymethyl) phosphonium chloride (MAGAT). Appl Radiat Isot 2005;63:443-56.
De Deene Y, Vergote K, Claeys C, De Wagter C. The fundamental radiation properties of normoxic polymer gel dosimeters: A comparison between a methacrylic acid based gel and acrylamide based gels. Phys Med Biol 2006;51:653-73.
Senden RJ, De Jean P, McAuley KB, Schreiner LJ. Polymer gel dosimeters with reduced toxicity: A preliminary investigation of the NMR and optical dose-response using different monomers. Phys Med Biol 2006;51:3301-14.
Stigler SM. Francis Galton's Account of the Invention of Correlation. Statist Sci 1989;4:73-9.
Baldoc C, Lepage M, Bäck SA, Murry PJ, Jayasekera PM, Porter D, et al.
Dose resolution in radiotherapy polymer gel dosimetry: Effect of echo spacing in MRI pulse sequence. Phys Med Biol 2001;46:449-60.
Sedaghat M, Bujold R, Lepage M. Severe dose inaccuracies caused by an oxygen-antioxidant imbalance in normoxic polymer gel dosimeters. Phys Med Biol 2011;56:601-25.
Gustafsson H. Radiotherapy gel dosimetry: Development and application of normoxic polymer gels. Medical Radiation Physics, Malmö: Lund University; 2004.
De Deene Y, Hurley C, Venning A, Vergote K, Mather M, Healy BJ, et al.
Abasic study of some normoxic polymer gel dosimeters. Phys Med Biol 2002;47:3441-63.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]