UNIVERSIDADE ESTADUAL DE CAMPINAS
SISTEMA DE BIBLIOTECAS DA UNICAMP
REPOSITÓRIO DA PRODUÇÃO CIENTIFICA E INTELECTUAL DA UNICAMP
Versão do arquivo anexado / Version of attached file:
Versão do Editor / Published Version
Mais informações no site da editora / Further information on publisher's website:
https://aapm.onlinelibrary.wiley.com/doi/full/10.1118/1.3483780
DOI: 10.1118/1.3483780
Direitos autorais / Publisher's copyright statement:
©2010
by Wiley. All rights reserved.
DIRETORIA DE TRATAMENTO DA INFORMAÇÃO Cidade Universitária Zeferino Vaz Barão Geraldo
CEP 13083-970 – Campinas SP Fone: (19) 3521-6493 http://www.repositorio.unicamp.br
brachytherapy unit using an optical fiber based detector
a… R. A. Minamisawab兲Department of Physics and Mathematics, FFCLRP, University of São Paulo, 14040-901, Ribeirão Preto-SP, Brazil
R. A. Rubo and R. M. Seraide
INRAD, Medical School — University of São Paulo, Clinical Hospital (HCUSP), 05403-001, São Paulo-SP, Brazil
J. R. O. Rochac兲
CAISM Radiotherapy Service, CEB, State University of Campinas, 13083-881, Campinas-SP, Brazil A. Almeida
Department of Physics and Mathematics, FFCLRP, University of São Paulo, 14040-901, Ribeirão Preto-SP, Brazil
共Received 19 April 2010; revised 9 August 2010; accepted for publication 9 August 2010; published 28 September 2010兲
Purpose: Several attempts to determine the transit time of a high dose rate共HDR兲 brachytherapy
unit have been reported in the literature with controversial results. The determination of the source speed is necessary to accurately calculate the transient dose in brachytherapy treatments. In these studies, only the average speed of the source was measured as a parameter for transit dose calcu-lation, which does not account for the realistic movement of the source, and is therefore inaccurate for numerical simulations. The purpose of this work is to report the implementation and technical design of an optical fiber based detector to directly measure the instantaneous speed profile of a
192Ir source in a Nucletron HDR brachytherapy unit.
Methods: To accomplish this task, we have developed a setup that uses the Cerenkov light induced
in optical fibers as a detection signal for the radiation source moving inside the HDR catheter. As the192Ir source travels between two optical fibers with known distance, the threshold of the induced signals are used to extract the transit time and thus the velocity. The high resolution of the detector enables the measurement of the transit time at short separation distance of the fibers, providing the instantaneous speed.
Results: Accurate and high resolution speed profiles of the192Ir radiation source traveling from the safe to the end of the catheter and between dwell positions are presented. The maximum and minimum velocities of the source were found to be 52.0⫾1.0 and 17.3⫾1.2 cm/s. The authors demonstrate that the radiation source follows a uniformly accelerated linear motion with accelera-tion of 兩a兩=113 cm/s2. In addition, the authors compare the average speed measured using the optical fiber detector to those obtained in the literature, showing deviation up to 265%.
Conclusions: To the best of the authors’ knowledge, the authors directly measured for the first time
the instantaneous speed profile of a radiation source in a HDR brachytherapy unit traveling from the unit safe to the end of the catheter and between interdwell distances. The method is feasible and accurate to implement on quality assurance tests and provides a unique database for efficient computational simulations of the transient dose. © 2010 American Association of Physicists in
Medicine. 关DOI:10.1118/1.3483780兴
Key words: brachytherapy, HDR, transit time, optical fiber
I. INTRODUCTION
Accurate delivery of dose using a high dose rate 共HDR兲 brachytherapy remote after-loading unit depends on knowing the source activity at the time of treatment, the precision of the timer, and the ability of the unit to position the source at the proper dwell location along the catheter. The source ac-tivity can be estimated using a calibrated ionization chamber with a well defined geometry in periodic measurements. Nor-mally, the quality assurance tests for the dwell time
determi-nation use a stopwatch, and for dwell position, radiographs or mechanical rulers and video monitors. Although these pa-rameters are computed in the dose determination, the accu-racy of the source transit time between two dwell positions and in the travel back to the source housing are neglected by most of the HDR brachytherapy treatment planning systems for the computation of the delivered dose.1,2
To overcome this problem, several techniques aimed to measure the transit time of the radiation source have been proposed using different types of instrumentation and
obtain-ing distinct results.1–10 However, the studies available in the literature have mostly measured the average speed of the radiation source traveling between different interdwell dis-tances. Since the transit dose calculation is inversely propor-tional to the speed of the source traveling a certain distance
L1,2, accurate dose calculation in a selected position requires L approaching zero. Therefore, instead of the average
veloc-ity, the instantaneous speed profile of the source is desired for numerical simulations of transit dose. In addition, most of the proposed methods have determined the transit time of the source using indirect measurements, obtaining standard de-viations as high as 100%.8
In this article, we report the use of an optical fiber detec-tion system to directly determine the instantaneous speed profile of a 192Ir source inside a treatment catheter in a brachytherapy Nucletron HDR unit. It is well known that radiation induces light in irradiated optical fibers due to the Cerenkov effect and to luminescence.11–13 We placed two optical fiber stems in specific positions close to the brachy-therapy catheter and measured the source transit time with an oscilloscope to detect the light signal induced by the radia-tion source. This method has the advantage of providing real-time measurements with the high resolution required for this application. We have measured the instantaneous velocity profile of the radiation seed moving in two programed con-figurations of clinical relevancy: The source traveling from the HDR unit and dwelling at the end of the catheter and the source traveling between different dwell positions. Based on the instantaneous speed profiles, we extracted the accelera-tion 共or deceleration兲 of the moving radiation seed. To the best of our knowledge this measurement has never been reported.1–10 Finally, in order to compare our method to the literature, we determined the average velocity between dwell positions directly measured using the optical fiber based de-tector and those calculated from the instantaneous speed pro-files. These results may be used to assist in the transit dose calculation in radiotherapy treatment planning systems, as well as an innovative quality assurance test for Nucletron HDR units.
II. MATERIALS AND METHODS
A schematic of the experimental setup is depicted in Fig.
1. Two identical 15 m long acrylic optical fiber stems 共Lux-tec Optical Systems兲 with 1.0 mm diameter are connected to a photomultiplier共model 9829A, Electron Tubes Limited兲 in one extremity and used as the effective radiation detector. The opposite extremity of each fiber is fixed perpendicular to the catheter of the HDR unit at distinct selected positions x1
and x2. The PMT is subsequently connected to a digital
os-cilloscope. The electronic equipment was located outside the radiation room to avoid scattering radiation in the PMT and/or damage to the system. As the192Ir source moves past the optical fiber, Cerenkov light is induced by the radiation field and the current signal converted by the PMT is mea-sured in the oscilloscope. Details of the radiation induced Cerenkov effect in optical fibers are described elsewhere.11,12 The time resolution of the oscilloscope is 1 ns; therefore, the
signal acquisition is rapid enough compared to the speed of the moving source, providing real-time measurements. A typical signal measured in the oscilloscope as the source moves past a single optical fiber consists of a triangular shaped peak, where the threshold indicates that the source is directly below the stem. The transit time dt measured be-tween the peak’s thresholds induced in two parallel optical fibers is used to calculate the source velocity for a fixed distance d = x1– x2. For any value of d, the extracted speed
corresponds to the average velocity of the source traveling from x1 to x2. As d approaches the dimensions of the 192Ir
source of about 0.4 cm length, the extracted velocity corre-sponds to the instantaneous speed. To a good approximation, we consider the average velocity extracted for d = 1 cm equivalent to the instantaneous speed. Distances d equal to or shorter than 0.5 cm were avoided because of the coupling of the detection peak widths, which attributes undesired uncer-tainties for the threshold extraction. On the other hand, in-significant coupling was observed for dⱖ1 cm within the range of speed measured.
Since the HDR unit offers different dwell times and step sizes, we chose for this experiment 2 s dwell time and 5.0 mm step size. The indexer length of the catheter is given as 99.5 cm, where this distance corresponds to the final dwell position. In order to study different aspects of the radioactive source motion, we measured the transit time using different programs typically applied in routine treatments by changing the dwell positions as well as the positions of the optical fibers in the following configurations:
共i兲 The deceleration profile was obtained after programing the source to travel from the HDR source housing to the extreme dwell position at x0= 99.5 cm. That
trajec-tory was repeated while moving the optical fibers with fixed d = 1 cm from 99.5 cm to the position of
x = 84.5 cm, i.e., 15 cm from the end of the catheter.
共ii兲 In order to compare the measurement of the instanta-neous speed to the average speed in similar situation, the source was again programed to travel from the HDR source housing to the extreme dwell position at
FIG. 1. Schematic of the experimental setup used to measure the192Ir source velocity. The end position of the catheter is represented by x0, which corre-sponds to the indexed position 99.5 cm.
5408 Minamisawa et al.: Instantaneous speed HDR 5408
x0= 99.5 cm. However, the fiber separation d was
var-ied from 1 to 15 cm, with one of the stems fixed at the dwell position x0.
共iii兲 The instantaneous speed between interdwell positions was measured when the source was programed to ac-celerate from a dwell position near the end of the cath-eter and then decelerate to stop at 99.5 cm.
共iv兲 Finally, in order to compare our results to the source average velocities reported in the literature using dif-ferent measurement techniques, we varied the fiber separation d from 1 to 10 cm and programed the source to start from rest at one fiber and stop at the other. The second fiber was fixed at 99.5 cm.
In order to compare 共iii兲 to 共iv兲, the values of instanta-neous speed vi共x兲 from 共ii兲 were used to calculate the
aver-age speed具Vx1典 through the following relation:
具Vx1典 =
冕
X1 99.5 vi共x兲dx冕
X1 99.5 dx , 共1兲where x the positions halfway between the two fibers for the experiment共ii兲. x1is 5 or 10 cm from the end of the catheter
and dx is 1 cm.
III. RESULTS AND DISCUSSION
Figure2 displays the instantaneous speed profile and the average speed of the source traveling from the HDR unit to
x0. The average velocity is lower than the instantaneous
speed for d⬎1 cm and varies at different rate than the in-stantaneous speed as increasing d. The deviation between the average velocity and the instantaneous speed at equivalent
distances results from the integration of the entire range of instantaneous speeds for calculating the average speed 关Eq.
共1兲兴, including those in deceleration. The deviation of the average speed to those instantaneous is as high as 12.5% and therefore inappropriate for transit dose calculation. Mean-while, as a direct measurement, the instantaneous speed pro-vides the realistic motion profile of the radiation source. The source has a maximum instantaneous speed of 52.0⫾1.0 cm/s after leaving the unit housing, which is slightly different than the 50 cm/s claimed by Nucletron. The maximum average velocity is 46.7⫾0.9 cm/s, considerably lower than 52.9 cm/s reported by Wong et al.;9however, it is comparable to the result of Bastin et al.1of 45.2 cm/s mea-sured using similar programming of the treatment console.
Interestingly, the instantaneous speed of the radiation seed starts to smoothly decrease at 12 cm away from the end of the catheter until it abruptly stops at 99.5 cm. Using New-ton’s equations of motion, we obtained an accurate fitting of the instantaneous speed in Fig.2, showing that the behavior of the system is satisfied by uniformly accelerated linear mo-tion at constant deceleramo-tion of a = −113 cm/s2. Within the
resolution of our setup, the minimum speed was determined as 17.3⫾1.2 cm/s, which disagrees to those specified by Nucletron for shorter distances of 28.4 cm/s. However, the constant acceleration condition implies that the velocity of the source linearly decreases with time, ramping down to zero. Consequently, it is inappropriate to specify minimum speed, especially when limited by systematic resolution.
The instantaneous speed profiles of the source traveling between two distinct interdwell positions of 5 and 10 cm are displayed in Fig.3. For 10 cm interdwell distance, the maxi-mum speed is obviously higher than for 5 cm because of the longer distance available for initial acceleration. The profiles suggest a realistic speed behavior for the source, i.e., lower
FIG. 2. Instantaneous speed profile共open squares兲 and the average velocity 共open circles兲 of the192Ir source traveling from the HDR unit to the end of the catheter programed in configurations共i兲 and 共ii兲 as described in Sec. II, respectively. The dotted black lines indicate the maximum and minimum speeds of the source specified by Nucletron. The dashed line represents the uniformly accelerated linear motion fitting with constant acceleration of −113 cm/s2.
FIG. 3. Instantaneous speed profiles of the radiation source traveling be-tween 5共open squares兲 and 10 共open circles兲 cm interdwell distances. The measurement configuration is described in共iii兲 in Sec. II. The dotted lines indicate the average speed具Vx1典 calculated from the instantaneous velocity
profiles using Eq.共1兲. The dashed lines represent the uniformly accelerated linear motion fitting with constant acceleration of 兩a兩=113 cm/s2. The black arrow indicates the direction of the seed motion.
velocities near to the dwell positions and the highest velocity at the midpoint. The symmetry of the profiles demonstrates that the stepping motor of the source housing applies an initial torque to the radiation seed with comparable magni-tude as for the final deceleration. Interestingly, the instanta-neous speeds measured at 1 cm distance of the x0 position,
before the abrupt deceleration, are equivalent in both curves of Fig.3and matches that observed in Fig.2. The best fitting of the instantaneous speed profiles displayed in Fig.3 dem-onstrate that the motion is again uniformly accelerated with both acceleration and deceleration equal to 兩a兩=113 cm/s2 and this behavior was observed for both interdwell distances. Therefore, the acceleration of the radiation source seems to be constant for any predefined traveling distance.
The values of transit time obtained from three measure-ments were used to calculate the instantaneous speed and the standard deviation of the average value, which are included in Figs. 2 and 3. A maximum deviation of about 6% was observed only for the instantaneous speeds measured directly before or after the source dwelling. The time of the signal produced in the optical fiber fixed in the dwell position ex-tends due to the contribution of the dwell time. Therefore, the signal of a dwelled radioactive source has a rectangular shape, different from the triangular peak measured for the moving source, which attributes uncertainties for defining the threshold position. The standard deviation of the instan-taneous speed measured in any other position is always lower than 2%, supporting the accuracy of the technique.
The average speed of the source between two dwell posi-tions is shown in Fig. 4. The average speed is linearly pro-portional to the interdwell distance and decreases from 23.4⫾0.5 to 17.0⫾1.1 cm/s within the interdwell distance range of 10 cm. By applying Eq.共1兲to the results of Fig.3, we extract an average speed of具V5 cm典=19.6⫾0.5 cm/s and
具V10 cm典=24.3⫾0.4 cm/s. The data are plotted in Fig.4for
comparison. The difference of between the average speeds directly measured and extracted from the instantaneous ve-locity profiles 共Fig.3兲 are 4.5% and 3.7% for 5 and 10 cm
interdwell distances, respectively. The deviations are smaller than the one obtained from Fig. 2 because the speeds in acceleration are compensated by those in deceleration re-gime. These deviations correspond to partial errors for cal-culation of transit dose. Likewise, the application of the in-stantaneous speed profile in numerical calculations may provide more accurate results of transient dose than the av-erage velocity: Speeds faster or slower than the avav-erage ve-locity will underestimate or overestimate, respectively, the transient dose in a specific position. A comparison among the results obtained in this investigation with the literature is also depicted in Fig. 4. The results using the optical fiber based detector are in better agreement with those reported by Houdek et al.2and Wong et al.9using direct measurements obtained by the oscilloscope and video methods, respec-tively. In particular, Houdek et al.2 have shown a similar dependence of the source speed with the interdwell distance. However, the results obtained by direct measurements are in contrast with the indirect ones reported by Williamson et al.6 and Sahoo8using ionization chambers. They show the source speed independent of the interdwell distance. These discrep-ancies may be attributed to the fact that the indirect measure-ments introduce more partial errors, and therefore, direct measurements may be more accurate. The average speed de-viation in absolute values is as high as 265% approximately, and implies an equivalent error for transit dose calculations. Particularly in treatments using short dwell times, inaccurate transit dose calculation leads to significant deviation and can be estimated using a 2DMATLABsimulation associated with the formalism proposed by Bastin et al.1For example, in a typical clinical treatment using multiple catheters with the source programed to dwell for 0.1 s only at the end of the catheter, the deviation in the total delivered dose at 1 cm distance of the dwell position perpendicular to the catheter, with and without the instantaneous speed profile correction 共considering the entrance and exit of the source兲, is about 140%. Using the average speed obtained by Houdek et al.2 of 45.2 cm/s in similar simulation, the deviation decreases down to 47%. This example elucidates the impact of the instantaneous speed profile in the calculation of transit dose with respect to previous estimations. Detailed simulations and analyses of transit dose using the instantaneous speed profile are out of the scope of this work and will be presented elsewhere.
The step up presented in this work is still at the research stage and therefore time consuming for routine clinical ap-plications. However, by using multiple optical fiber detectors in parallel connected to a single PMT and a computer con-trolled oscilloscope, rapid profiling of instantaneous source speed could be measured at a cost of ⱕUS$1000. In this embodiment, quality assurance tests for calibration of transit time in HDR units may be periodically performed by the service radiotherapy staff, requiring minimum training. The
FIG. 4. Average source velocity between interdwell positions directly mea-sured共squares兲 and calculated from the instantaneous speed profiles 具V5 cm典 and具V10 cm典 共up and down triangles, respectively兲. The average speed was measured using configuration共iv兲 described in Sec. II. Results extracted from the literature are displayed for comparison. The dashed line separates indirect measurements共top兲 from direct measurement 共bottom兲.
5410 Minamisawa et al.: Instantaneous speed HDR 5410
application of the technique may provide a faster, more ac-curate, and cost-effective calibration of transit time com-pared to the service performed by the maintenance engineer. For applications in clinical treatments, we recommend that the planning system of HDR units should incorporate a da-tabase of instantaneous speed specific for the unit since me-chanical variability may be expected. Internal corrections of transit time in the planning system could then be imple-mented when relevant to the clinical situation. As the instan-taneous speed profile is independent of the dwell time, the database should contain information regarding the motion of the radioactive source during the entrance and exit of the unit safe as well as between different dwell positions. Periodical calibrations should then be performed for assuring the reli-ability of the database for transit dose corrections.
IV. CONCLUSIONS
Quality assurance tests of transit time for remote after-loading high dose rate brachytherapy units are important for the calculation of the transit dose in treatment planning. In this work, we realized the instantaneous speed profiling of a HDR radiation seed moving inside the treatment catheter, measured using a high spatial resolution optical fiber based detector. The results reported show accuracy in the determi-nation of the instantaneous speed profile of the source trav-eling from the HDR unit and between interdwell positions, accounting for realistic motion of the source. Based on these results, we demonstrate that the speed of the source changes at a constant acceleration of兩a兩=113 cm/s2. In addition, we
show for comparison that the average velocity of the source between different interdwell distances is linearly propor-tional to the traveled distance, contrary to results obtained in previous studies. Since the system enables direct measure-ment of the speed, the accuracy is enhanced compared to those observed by indirect measurements. Using the mea-surement of the instantaneous speed proposed in this article, one can provide a database for specific planning treatments and quality assurance test of the HDR unit. Further works correlating the instantaneous speed profiles with dosimetric information and simulations are needed for evaluating the method for conventional HDR brachytherapy treatments.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the support of Elec-tron Tubes Limited for supplying photomultiplier tubes 共Bury Street, Ruislip Middx HA4 7TA, United Kingdom兲 and the facilities manager of the radiotherapy service at the HU-USP共Sao Paulo, Brazil兲.
a兲
U.S. patent pending.
b兲Author to whom correspondence should be addressed. Electronic mail: r.a.minamisawa@fz-juelich.de; Present address: Forschungszentrum Jülich, Jülich D-52074, Germany.
c兲Deceased, 12/09/2009.
1K. T. Bastin, M. B. Podgorsak, and B. R. Thomadsen, “The transit dose component of high dose rate brachytherapy: Direct measurements and clinical implications,” Int. J. Radiat. Oncol., Biol., Phys. 26, 695–702 共1993兲.
2P. V. Houdek, J. G. Schwade, X. Wu, V. Pisciotta, J. A. Fiedler, C. F. Serago, A. M. Markoe, A. A. Abitbol, A. A. Lewin, P. G. Braunsch-weiger, and M. D. Skalar, “Dose determination in high dose-rate brachy-therapy,” Int. J. Radiat. Oncol., Biol., Phys. 24, 795–801共1992兲. 3R. Walstram, “Studies on therapeutic short-distance and intracavitary
gamma beam techniques,” Acta Radiol. 236, 53–56共1965兲.
4W. E. Liversage, P. Martin-Smith, and W. Ramsey, “The treatment of uterine carcinoma using the Cathetron. Part II. Physical measurements,”
Br. J. Radiol.40, 887–894共1967兲.
5A. S. Meigooni, J. F. Williamson, and E. Slessinger, “Practical quality assurance tests for positional and temporal accuracy of HDR remote af-terloaders,” Endocurie Hypertherm. Oncol. 9, 46–48共1993兲.
6J. F. Williamson, G. A. Ezzel, A. Olch, and B. R. Thomadsen, “Quality assurance for high dose rate brachytherapy,” in High Dose Rate Bracht-herapy: A Text Book, edited by S. Nag共Futura, Armonk, 1994兲, pp. 147– 212.
7B. R. Thomadsen, P. V. Houdek, R. Van der Laarse, G. Edmunson, I. K. Kolkman-Deurloo, and A. G. Visser, “Treatment planning and optimiza-tion,” in High Dose Rate Brachtherapy: A Text Book, edited by S. Nag 共Futura, Armonk, 1994兲, pp. 79–145.
8N. Sahoo, “Measurement of transit time of a remote after-loading high dose rate brachytherapy source,”Med. Phys.28, 1786–1790共2001兲.
9T. P. Y. Wong, W. Fernando, P. N. Johnston, and I. F. Bubb, “Transit dose of an Ir-192 high dose rate brachytherapy stepping source,”Phys. Med. Biol.46, 323–331共2001兲.
10C. S. G. Calcina, A. Ameida, J. R. O. Rocha, F. Chen, and O. Baffa, “Ir-192 HDR transit dose and radial dose function determination using alanine/EPR dosimetry,”Phys. Med. Biol.50, 1109–1117共2005兲.
11A. S. Beddar, T. R. Mackie, and F. H. Attix, “Water-equivalent plastic scintillation detectors for high-energy beam dosimetry: I. Physical char-acteristics and theoretical considerations,” Phys. Med. Biol.37, 1883–
1900共1992兲.
12A. S. Beddar, T. R. Mackie, and F. H. Attix, “Water-equivalent plastic scintillation detectors for high-energy beam dosimetry: II. Properties and measurements,”Phys. Med. Biol.37, 1901–1913共1992兲.
13S. F. de Boer, A. S. Beddar, and J. F. Rawlinson, “Optical filtering and spectral measurement of radiation induced light in plastic scintillator do-simetry,”Phys. Med. Biol.38, 945–958共1993兲.
