Absolute calibration of the Gamma Knife® Perfexion™ and delivered dose verification using EPR/alanine dosimetry Amaury Hornbeck, Tristan Garcia, Marguerite Cuttat, and Catherine Jenny Citation: Medical Physics 41, 061708 (2014); doi: 10.1118/1.4873686 View online: http://dx.doi.org/10.1118/1.4873686 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/41/6?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Fast transit portal dosimetry using density-scaled layer modeling of aSi-based electronic portal imaging device and Monte Carlo method Med. Phys. 39, 7593 (2012); 10.1118/1.4764563 Calibration of the Gamma Knife Perfexion using TG-21 and the solid water Leksell dosimetry phantom Med. Phys. 38, 1685 (2011); 10.1118/1.3557884 Calibration of helical tomotherapy machine using EPR/alanine dosimetry Med. Phys. 38, 1168 (2011); 10.1118/1.3553407 Calibration of the Gamma Knife using a new phantom following the AAPM TG51 and TG21 protocols Med. Phys. 35, 514 (2008); 10.1118/1.2828187 Consequences of the spectral response of an a-Si EPID and implications for dosimetric calibration Med. Phys. 32, 2649 (2005); 10.1118/1.1984335
R Absolute calibration of the Gamma Knife PerfexionTM and delivered dose verification using EPR/alanine dosimetry
Amaury Hornbecka) and Tristan Garciaa) CEA, LIST, Laboratoire National Henri Becquerel, 91191 Gif-sur-Yvette Cedex, France
Marguerite Cuttat and Catherine Jenny Radiotherapy Department, Medical Physics Unit, University Hospital Pitié-Salpêtrière, 75013 Paris, France
(Received 30 July 2013; revised 4 April 2014; accepted for publication 15 April 2014; published 12 May 2014) R Purpose: Elekta Leksell Gamma Knife (LGK) is a radiotherapy beam machine whose features are not compliant with the international calibration protocols for radiotherapy. In this scope, the Laboratoire National Henri Becquerel and the Pitié-Salpêtrière Hospital decided to conceive a new LKG dose calibration method and to compare it with the currently used one. Furthermore, the accuracy of the dose delivered by the LGK machine was checked using an “end-to-end” test. This study also aims to compare doses delivered by the two latest software versions of the Gammaplan treatment planning system (TPS). Methods: The dosimetric method chosen is the electron paramagnetic resonance (EPR) of alanine. Dose rate (calibration) verification was done without TPS using a spherical phantom. Absolute calibration was done with factors calculated by Monte Carlo simulation (MCNP-X). For “end-to-end” test, irradiations in an anthropomorphic head phantom, close to real treatment conditions, are done using the TPS in order to verify the delivered dose. Results: The comparison of the currently used calibration method with the new one revealed a deviation of +0.8% between the dose rates measured by ion chamber and EPR/alanine. For simple fields configuration (less than 16 mm diameter), the “end-to-end” tests showed out average deviations of −1.7% and −0.9% between the measured dose and the calculated dose by Gammaplan v9 and v10, respectively. Conclusions: This paper shows there is a good agreement between the new calibration method and the currently used one. There is also a good agreement between the calculated and delivered doses especially for Gammaplan v10. © 2014 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4873686]
Key words: Gamma Knife, EPR, alanine, dosimetry, calibration 1. INTRODUCTION In radiotherapy treatment centers, high-energy photon beams calibrations follow international protocols, in particular the AAPM TG-51 or the IAEA TRS-398.1, 2 In the past 15 years, technological evolutions have provided better solutions to treat patients more specifically and more precisely with a growing number of particular radiotherapy and radiosurgery devices. The Elekta Leksell Gamma Knife (LGK) Perfexion model (PFX) is a stereotactic radiosurgery equipment. It differs from conventional radiotherapy machines in terms of clinical indications, mechanical design, and beam characteristics. Indeed, it contains 192 sources of 60-cobalt distributed over a crown-shaped collimator at distances of 48.1–51.9 cm from the unit center point (UCP). Table I illustrates the big difference between these parameters and the required international calibration protocols. Thus, one cannot calibrate LGK PFX beams using the latest protocols. Even if other calibration methods have been suggested,3, 4 for example, Alfonso et al.,5 there is no consensus about the way to verify calibration and delivered dose. However, the AAPM Task Group 178 is working on recommendations for calibration and quality insurance of stereotactic radiosurgery devices, which are expected by the end of 2014.6 061708-1
Med. Phys. 41 (6), June 2014
The Laboratoire National Henri Becquerel (LNHB) which is the French National Metrology Laboratory for ionizing radiations decided to study the feasibility of an absolute calibration and delivered dose verification of LGK using electron paramagnetic resonance (EPR)/alanine dosimetry. This technique has already been proved to be adequate dosimeters for radiotherapy and brachytherapy.7–10 Moreover, the LNHB has already carried out such studies for innovative radiotherapy techniques like Cyberknife or Tomotherapy.11, 12 Thanks to the shape and dimensions of alanine dosimeters, they can be used for small beams. Alanine dosimeters have interesting advantages: measures are at solid state and nondestructive, their responses are almost independent on the dose rate and energy of radiotherapy photon beams and their signal is stable in time (signal decreases less than 1% per year).13 First, calibration curves are made at LNHB. Second, beam calibration accuracy is verified by dose rate measurements, without using Gamma Knife’s treatment planning system (TPS). The currently used method is compared to an absolute calibration method proposed by the LNHB and based on MC simulations. Finally, Gamma Knife’s TPS “Gamma plan” calculated doses are compared to EPR/alanine measured dose, using an anthropomorphic head phantom.
0094-2405/2014/41(6)/061708/10/$30.00
© 2014 Am. Assoc. Phys. Med.
061708-1
061708-2
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-2
TABLE I. Main differences between reference conditions and LGK PFX irradiation conditions.
2. MATERIALS 2.A. Calibration curve establishment
2.A.1. Irradiation facilities: Reference beams
Calibration curves are performed with two different reference beams of the LNHB: The Co60 γ -ray: TRS 398 reference conditions for Co60 (i.e., at 80 cm source axis distance or SAD, 5 cm depth in water with a 10 × 10 cm2 field size at reference plane), the dose rate is known with an uncertainty of 0.48% (k = 1). The linear accelerator (LINAC) Saturn 43: TRS 398 reference conditions for 6 MV x-rays (i.e., at 100 cm SAD, 10 cm depth in water with a 10 × 10 cm2 field size at reference plane), the dose rate is known with an uncertainty of 0.38%. 2.A.2. Large water phantom
A large water phantom is used with reference beams. It is a Polymethyl methacrylate (PMMA) tank of 30 × 30 × 30 cm3 filled with distilled water. The phantom is set on a mechanical system allowing submillimetric translation of phantom and dosimeters to insure a precise positioning within the three axes (x, y, z). Micrometric screws are used to measure the source-phantom and the source-detector distances so that the relative position of the dosimeter is known within 10 μm accuracy. 2.A.3. Alanine dosimeters and EPR spectrometer
Alanine dosimeters are made up with four alanine pellets R . Each of included in a cylindrical container made of Delrin them measures 4.8 mm diameter, 3 mm height, 67.5 ± 0.1 mg R . For LGK’s mass and are provided by Synergy Health Medical Physics, Vol. 41, No. 6, June 2014
calibration verification, two pellets are placed in the phantom, without container. The alanine dosimeters signals are measured on a Bruker ELEXSYS E500 EPR spectrometer with an ER 4119 HS resonator, which operates in X-band. The gain of the spectrometer depends on the amplitude of the signal. The sweep time is set to 21 s, the time constant is set to 82 ms, and the conversion time is set to 28 ms. 2.B. Absolute calibration of the LGK
2.B.1. LGK PFX
The LGK aims to treat intracranial tumors and lesions in a unique session. The PFX model uses 192 sources of 60cobalt distributed as a crown shape. Sources are located in eight sectors. Each sector can independently align their 24 sources in front of a 4, 8, or 16 mm collimator channel, or set them in blocked position. All beams converge to a unique geometrical point called UCP. 2.B.2. Spherical acrylonitrile butadiene styrene (ABS) phantom
The 15.9 cm diameter Elekta spherical ABS phantom is usually used by medical physicists to perform the reference dose rate verification of the PFX, according to the Elekta protocol. It is composed of three parts: two hemispheres and one flat plate drilled so that the measurement point of the ionization chamber matches with the UCP of the GammaKnife. Flat plate can be changed allowing different kinds of detectors to be placed in the sphere (ion chamber, radiochromic films, etc.). For this study, a plate has been manufactured by the LNHB for two alanine pellets to be placed at its center (Fig. 1). Whatever the dosimeter, the phantom is assembled
061708-3
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-3
F IG . 1. ABS phantom disassembled: two hemispheres and the flat plate.
with two holding cylinders and fixed on the treatment table via a mechanical system composed by two aluminum plinths as shown in Fig. 2.3
tion chamber volume is 0.125 cm3 with an inner diameter of 5.5 mm which is suitable for LGK PFX use. 2.C. Delivered dose verification (“end-to-end” test)
2.B.3. Dosimetric material
2.B.3.a. Alanine dosimeters. Because of the geometrical restraint of the ABS phantom, only two pellets can be put in the center of the phantom. To do that, a no-drilled flat plate of the ABS phantom is customized to handle two pellets (Fig. 1). 2.B.3.b. Ionization chamber and electrometer. Quality controls and reference dose rate determination are made by the medical physicist using an ionization chamber PTW 31010 associated with a PTW Unidos E electrometer. The ioniza-
2.C.1. Elekta’s TPS Gammaplan v9 and v10
Pitié-Salpêtrière Hospital’s TPS (Gammaplan) has recently been updated (v9 to v10). In order to study dose calculations of both versions, “end-to-end” tests are done for a couple of configurations using v9 then v10. Both versions considered tissues as water and do not take heterogeneities into account. The v10 introduces an evolution of the waterbased algorithm, which, for example, implies new values for
F IG . 2. ABS phantom, including two alanine pellets, fixed on the LGK PFX table. Medical Physics, Vol. 41, No. 6, June 2014
061708-4
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
F IG . 3. Anthropomorphic head phantom fixed, as for a head patient, on a frame, which is fixed on the LGK PFX table.
061708-4
dosimeter amplitude. The EPR readout process involves measurements at different angles α (0◦ for α = 1, 120◦ for α = 2, and 240◦ for α = 3). This is made to take into account a possible response variation with angle. The pellet EPR signal is the average of the central peak amplitude of those three angles. The dosimeter EPR signal is the mean value of the four pellets EPR signals. EPR readout parameters are the ones described by Garcia 2009 except that five sweeps per angle are performed instead of 7.17 The gain is set according to the signal. Correction factors due to pellet mass, gain, and irradiation temperature are applied. So the corrected amplitude is calculated as below ⎛ ⎞ 3 1 ⎜ ⎟ Ai,α 4 ⎟ 1 ⎜ ⎜ 3 α=1 ⎟ ◦ (1 + (20 −T ) × 0.14%) ADosi,corr = ⎜ ⎟ ⎟ 4 i=1 ⎜ mi .10G/20 ⎝ ⎠ 1 = Ai corr , 4 i=1 4
dose profiles (full width at half maximum values, penumbra, relative isodose volumes, output factors, etc.). 2.C.2. Anthropomorphic head phantom
In order to meet the objectives detailed above, an anthropomorphic head has been designed to host an alanine dosimeter and the LNHB developed stands to fix it on a stereotactic frame in order to be as close as possible from real patient treatment conditions, as shown in Fig. 3. 2.C.3. Dosimetric material
Dosimetric material used for delivered dose verifications is alanine dosimeter (four pellets inside a container). 3. METHODS 3.A. Calibration curve establishment and alanine readouts
3.A.1. Irradiation of alanine dosimeters
For each reference beam, 20 alanine dosimeters (two dosimeters per dose) are placed in a 30 × 30 × 30 cm3 water phantom according to reference conditions of the TRS 398 and are irradiated from 3 to 60 Gy. 3.A.2. EPR readout
Alanine EPR spectrum shows five peaks (Fig. 6). The central peak spectrum amplitude measurement has been chosen in accordance with different authors.14–16 EPR measurements are performed at regulated room temperature (23 ± 1 ◦ C) and humidity (50 ± 10% HR) and made two or three weeks after irradiation of the dosimeters to insure signal stabilization. EPR signal of each pellet is analyzed. Tubes in Suprasil quartz with 5 mm internal diameter are used for maintaining the pellet in the readout cavity. Raw amplitude of EPR signal of each pellet is measured. Equation (1) shows how to get to corrected Medical Physics, Vol. 41, No. 6, June 2014
(1)
where Ai,α : amplitude for the pellet i measured for angle alpha; ADosi,corr (u.a.): mean amplitude of the whole dosimeter corrected by:
r G: gain applied to the EPR signal; r mi (mg): weight of the i pellet; r (20−T) × 0.14%: irradiation temperature correction.18, 19
This method is used to read out alanine dosimeter for the calibration curve but also every dosimeter used in the presented work. The only difference will be the number of pellets which are used: four pellets for calibration curves, two for the calibration verification measurements, and three for “end-to-end” tests. 3.A.3. Calibration curve D = f (EPR amplitude)
20 alanine dosimeters are irradiated from 3 to 60 Gy at the LNHB. Indeed, the LHNB uses 60 Co irradiators and linac with precise delivered doses (u = 0.49% and u = 0.38%, respectively). Both are used for calibration curve establishment to confirm their equivalence, which has already been proven in previous internal works made by the LNHB. Each dosimeter is irradiated in a water phantom, according to the TRS 398 reference conditions, for their signal to be easily related to an absorbed dose into water, with low uncertainties. Linear regression is calculated between EPR amplitudes with their associated uncertainties17 and delivered dose with associated uncertainties (dosimeter positioning, irradiation time, etc.). LNHB alanine calibration curves had been validated as suitable for use of alanine in nonreference conditions.12 Uncertainties are calculated for k = 1, using the following expression:20 u = ucf 2 + ud 2 + ut 2 + ureg 2 , (2)
061708-5
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
where ucf is the uncertainty on the conversion factors applied to determine dose rate values under standard conditions (type B), ud is the uncertainty on dosimeter positioning (type B), ut is the uncertainty on irradiation time (type B), ureg is the uncertainty on linear regression curve slope (type A). 3.B. Absolute calibration of the LGK
3.B.1. Routine reference dose rate verification
The ionization chamber is put in the middle of the spherical ABS phantom which is mounted on the table. The LGK is set so that all the sources are in front of the 16 mm collimators channels. In those conditions, ion chamber measurements are performed for 60 s of irradiation, and the measurement is reproduced seven times for a good statistic. Thanks to the calibration factor of the ion chamber, the dose rate D˙ w,i (where “i” is the number of the measurement) can be obtained after being corrected by kT,P, kpol , and kQ,Q0 factors (ambient temperature and pressure, polarity, and beam quality). The formalism proposed by Alfonso et al.5 is currently used at the Pitié-Salpêtrière Hospital for measurements in Tomotherapy and Gamma Knife. This formalism supposes the use of f ,fmsr f ,fref and kQpcsr correction factors, taking into acthe kQmsr msr ,Q pcsr ,Qmsr count the differences between the conditions of field size, geometry, phantom material, quality, and the use of composite fields. Then, the reference dose rate D˙ w is the average of all those measurements corrected by all those factors. In that way, the medical physicist can check that the reference dose rate of measurements is compliant with the one used in the TPS and with the French legislation that requires 2% maximal deviation valuable for classical radiotherapy devices.21 Equation (3) is used to determinate the dose rate f
D˙ W,i =
,f
ref Mi .ND,W .kT ,P .kpol .kQ,Q0 .kQmsr msr ,Q
dt
,
(3)
where: M: charge measurements in [nC]. ND,W : calibration factor in terms of absorbed dose to water for a dosimeter at a reference beam quality Q0 . kQ,Q0 = 1 due to the use of 60 Co beams. f ,fref kQmsr is set equal to 1 because constructor’s calibration msr ,Q protocol is not based on Alfonso’s formalism. 3.B.2. Dose rate verification with alanine dosimeter
Alanine pellets are put in the middle of the ABS phantom in a specifically drilled plate that can receive two pellets. As they are completely surrounded by ABS material to insure homogeneity of the environment, the dose measured in such conditions is an absorbed dose in alanine inside an ABS environment. To increase statistic quality, nine measurements are done for four irradiations durations (4, 8, 11, and 15 min). An absolute calibration implies a method allowing expressing results in terms of absorbed dose to water, inside a water environment, and in the TRS 398 reference conditions. Thus, ionMedical Physics, Vol. 41, No. 6, June 2014
061708-5
ization chamber readouts can be compared with EPR/alanine readouts. To get to such an expression, the LNHB developed the following relations: Step 1: In the reference conditions, determination of the calibration coefficient in terms of dose in water to an absorbed dose in alanine. This step aims to take into account the change of alanine measurement configuration (four pellets in the container for the calibration curve vs two pellets in the ABS phantom) and the attenuation of the alanine itself
Dalanine DW . , (4) NDalanine = ADosi,corr DW,tank MC where:
r NDalanine : calibration coefficient; r DW : absorbed dose to water, delivered by reference beam;
r ADosi,corr : average EPR signal corresponding to the absorbed dose delivered by LNHB reference beams, in TRS 398 reference conditions, for calibration curve establishment;
r Dalanine = 0.976 ± 0.002 which is the absorbed dose Dw,tank MC ratio in alanine and in water, taking the delrin container into account, calculated by MCNP-X code. The geometry set for MCNP-X calculation of this ratio is: -Source: 1 beam of 60-Co, mean energy 1.25 MeV, beam orientation: horizontal. -Phantom: Water cube (30 × 30 × 30 cm3 ). The center of the cube placed on the 60 Co beam axis. -Skin-source distance: 95 cm. -Detector: alanine for Dalanine and water for Dw,tank , the center of the detector placed on the beam axis at 100 cm from the source (5 cm depth in water). Step 2: Determination of a calibration coefficient in terms of absorbed dose to alanine in the ABS spherical phantom. This step allows a transition from the reference conditions to the LGK conditions. Furthermore, because of the complexity of a whole LGK geometrical configuration, the problem has been considered symmetrical in spherical dimensions, so the simulation includes only one 1.25 MeV photon beam
Dsphere , (5) NDsphere = NDalanine Dalanine MC where:
r NDSphere : calibration coefficient in the ABS spherical r
phantom.
Dsphere Dalanine
= 1.005 ± 0.003 which is the absorbed dose MC ratio in the ABS spherical phantom and in alanine (calculated by MC simulation).
The geometry set for MCNP-X calculation of this ratio is: -Source: 1 beam of 60-Co, mean energy 1.25 MeV, beam orientation: horizontal
061708-6
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-6
3.C. Delivered dose verification (“end-to-end” test)
3.C.1. Test “end-to-end”
The “end-to-end” test aims to reproduce the different patient treatment process in order to evaluate the delivered dose by the LGK via the TPS in real conditions with the following steps:
r CT-scan acquisition; r treatment plan calculation; r LGK treatment irradiation. Images are acquired by a CT-scan, two different treatments are planned (Fig. 5):
F IG . 4. Monte Carlo simulation of secondary electrons around the ABS phantom (black sphere) irradiated by a Cobalt 60 beam.
-Phantom: ABS sphere (diameter 16 cm). The center of the cube placed on the 60 Co beam axis. -Skin-source distance: 50 cm. -Detector: alanine for Dalanine and ABS for Dsphere . The center of the detector is the same as the center of the phantom. The simulation is done for up to 108 histories, its result is shown in Fig. 4. Step 3: in ABS spherical phantom, from the EPR signal of alanine dosimeters (calibrated in terms of absorbed dose to water) to the absorbed dose into water, in LGK PFX conditions. Equations (4) and (5) development is DW ADosi,corr
Dsphere Dalanine . . .Amoy,corr , Dw,tank MC Dalanine MC
Dsphere = NDsphere .Amoy,corr =
(6)
where: Dsphere : absorbed dose in LGK conditions brought back to TRS 398 reference conditions; Amoy,corr : average EPR signal corresponding to the two pellets for LGK irradiation; Step 4: Brought into reference conditions, the reference dose rate can now be determined, using the correct irradiation duration (4, 8, 11, or 15 min) dDsphere . D˙ sphere = dt
(7)
Two weeks after irradiation in LGK, EPR spectra are acquired and the corresponding dose is obtained thanks to the calibration. Corrected measured dose rates are compared to the reference dose rate, measured by ion chamber. Medical Physics, Vol. 41, No. 6, June 2014
-Plan 1: A single shot (centered on the dosimeter) is performed with all sectors aligned with the 16 mm collimators channels. To improve measurements statistics, three dosimeters are irradiated for Dmax = 39.60 Gy (Gammaplan v9), three for Dmax = 39.50 Gy (Gammaplan v10), one for Dmax = 118.80 Gy (Gammaplan v9), and one for Dmax = 118.50 Gy (Gammaplan v10). -Plan 2: A double-shot (16 and 8 mm) is performed: Three dosimeters are used for Dmax = 36.40 Gy (Gammaplan v9), three for Dmax = 37.10 Gy (Gammaplan v10), one for Dmax = 109.20 Gy (Gammaplan v9), and one for Dmax = 111.30 Gy (Gammaplan v10). In this precise case, the maximum dose Dmax will be used, even though neurosurgeons usually work with the D50 . But the maximum dose actually calculated in the PTV is not always the same value than the maximal prescript dose Dmax . So the difference will be made between Dmax and Dmax,PTV , which are, respectively, the maximum prescript dose and the maximum calculated dose in the PTV. Finally, Dmax,PTV will be compared to alanine measurements. For each plan, Dmax,PTV is obtained from dose profiles measured by Gamma Plan inside the PTV. To verify if the Dmax,PTV covered the four alanine pellets of the dosimeter, dose-volume histograms and dose profiles are studied. It appeared that dose profiles are not adapted for the entire dosimeter, consequently, only three pellets are used (Fig. 5) for PTV, whereas the bottom pellet was discarded. Two weeks after irradiation in LGK PFX, EPR spectra are acquired and corresponding doses are obtained thanks to the calibration curve. After been projected on the calibration curve, alanine measured doses are able to be compared to the average maximal doses calculated by the TPS in the PTV. 4. RESULTS 4.A. Calibration curve establishment and alanine readouts
From EPR readouts corresponding to dose values at the reference point, calibration curve is established by linear regression (Fig. 6). The calibration curve is linear (R2 = 1) and uncertainties on EPR signal decrease while the absorbed dose increases. EPR signals of alanine can now be related to an absorbed dose in water.
061708-7
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-7
F IG . 5. Planning of the treatment: two different plans showing the dosimeter pellet 2, 3, and 4 entirely covered by the isodoses at 50% of the maximum dose.
F IG . 6. Calibration curve alanine EPR signal as a function of absorbed dose in water, in TRS 398 reference conditions for Co60 beam, with associated uncertainties. Medical Physics, Vol. 41, No. 6, June 2014
061708-8
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-8
TABLE II. Test “end-to-end” results: Absorbed dose to water measured by alanine vs prescripted dose (TPS). EPR
Gammaplan v9
Sector configuration
Alanine dosimeter
Absorbed dose (Gy)
Uncertainty (%)
Isodose 100% (Gy)
Uncertainty (%)
Disparity EPR/TPS (%)
Plan 1
v9-1 v9-2 v9-3 v9-7 v9-4 v9-5 v9-6 v9-8 v10-1 v10-2 v10-3 v10-7 v10-4 v10-5 v10-6 v10-8
38.7 38.9 38.7 117.5 35.7 35.9 35.9 108.1 39.0 39.1 39.0 117.9 36.7 36.7 36.8 111.1
1.2 1.2 1.2 0.7 1.3 1.3 1.3 0.7 1.3 1.3 1.3 1.4 1.3 1.0 1.0 1.0
39.6 39.6 39.6 118.8 36.4 36.4 36.4 109.2 39.5 39.5 39.5 118.5 37.1 37.1 37.1 111.3
2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
− 2.2 − 1.9 − 2.4 − 1.1 − 1.8 − 1.4 − 1.4 − 1.0 − 1.2 − 1.1 − 1.3 − 0.5 − 0.9 − 1.0 − 0.9 − 0.2
Plan 2
Gammaplan v10
TPS
Plan 1
Plan 2
4.B. Absolute calibration of the LGK
Dose rate verification means to compare measurements with ionization chamber (reference dose rate) and alanine EPR signal, projected to the calibration curve and brought back to reference conditions. 4.B.1. Reference dose rate verification
Seven measurements are done with PTW 31010 ion chamber. The correction of the temperature and pressure kT,P = 1.01456, the correction of the polarity kpol = 1, kQ,Q0 = 1 due f ,fref fpcsr ,fmsr to the use of 60 Co beams and kQmsr .kQpcsr ,Qmsr is set equal to msr ,Q 1. The calibration factor of the PTW 31003 is Nw = 0.3083 Gy nC−1 . The mean reference dose rate measured with ion¯˙ −1 ization chamber is D W,IC = 2.84 ± 0.01 Gy min . The slight difference (0.2%) from the reference dose rate used in the ˙ W,TPS = 2.84 ± 0.06 Gy min−1 is within tolerances. TPS, D 4.B.2. Dose rate verification with alanine dosimeter
According to the method developed in Sec. 3.B.2, the measured dose rates by EPR/alanine dosimetry lie with an average of 2.86 ± 0.06 Gy min−1 , that is to say, +0.8% relative to ion chamber measurements. Here, the LNHB Linacmade calibration curve has been used. ¯˙ The average dose rate measured by alanine is D W,alanine −1 = 2.86 ± 0.06 Gy min corresponds to a +0.8 % deviation ˙ W,TPS , the TPS dose rate, which is within measurement from D uncertainties. 4.C. Delivered dose verification (“end-to-end” test)
Alanine measurements are done for both versions of Gammaplan v9 and v10. For each TPS version, four measurements are done for plan 1 and four other measurements Medical Physics, Vol. 41, No. 6, June 2014
for plan 2 (see Sec. 3.C.1 for more details). The calibration curve made in the 60-cobalt beam has been used. The aim is to verify the consistency between calculated dose and the dose actually delivered but also to test differences between both versions for simple beam configurations. Table II illustrates the results of the “end-to-end” test. All measured doses are lower than TPS ones, the average measured doses by EPR/alanine compared to the calculated doses for both plans show an average disparity of −1.7% for Gammaplan v9 and an average disparity of 0.9% for Gammaplan v10 (Table II).
5. DISCUSSIONS 5.A. Absolute calibration
Calibration verification has been made by comparison between ion chamber dose rate and reference dose rate based on and EPR/alanine measurements. The dose rate measured with alanine is close to the dose rate measured with ion chamber by +0.8%; nevertheless, it is important to note that the holders of the ABS phantom attenuate beams of sectors 3 and 7, as shown in Fig. 2. This attenuation was quantified by McDonald et al.3 as almost 1% and is not taken into account for reference dose rate used in the TPS for dose calculations. Besides, this effect is not visible in the dose rate comparison because the same holder is used for both detectors. However, during a patient (or anthropomorphic phantom) treatment, beams of sectors 3 and 7 are not attenuated, because the frame holder geometry is different from the ABS phantom holder. As a consequence, the actual measured dose was expected to be slightly higher than the calculated one. If the effect of ABS phantom holders of 1% would be taken into account, the disparities would be −2.7% and −1.9% for Gamma plan v9 and v10, respectively.
061708-9
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
So the dose rate is expressed as a dose rate into water measured in the ABS phantom taking into account there is no correction factor to consider the difference of attenuation between ABS and water. So the TPS calculated doses are set with a dose rate value measured by the ion chamber as an absorbed dose into water, but actually attenuated by ABS. This effect is not taken into account neither for the clinical measurements nor for LGK acceptance tests. A newer spherical solid water phantom provides by the manufacturer is made in water equivalent material and has a different holder system which does not attenuate cobalt beams.3 It would be interesting to perform such a study using this phantom. But the ABS phantom was the only one available in this work. So to quantify this effect, a correction factor has been calculated by MC simulation using the MCNP-X. For the simulation, the number of histories is set between 2 × 108 and 5 × 108 . The calculated correction factor is
DW = 1.015 ± 0.003. Dsphere MC If this factor had been taken into account in the measurement, the average ionization chamber dose rate would have been
M.ND,W .kT ,P .kpol DW ˙ . . (8) DW = dt Dsphere MC Thus, dose rate should be equal to 2.88 Gy min−1 . Hence, the average difference between alanine and ionization chamber measurements should be −0.7%. 5.B. Delivered dose verification
All measured doses are lower than calculated doses for both plans. Regarding the discussion of Sec. 5.A, the measured dose was expected to be higher than the calculated one. Nevertheless, the effect of the attenuation due to the ABS phantom holder is not obvious. This effect seems to be compensated by the fact that heterogeneities of the anthropomorphic head phantom made by bone and tissue (water equivalents), and eventually sinus cavities are not taken into account by the TPS, which considers everything as water for calculation. However, this study showed very good agreements between EPR/alanine measurements and TPS calculated dose, delivered doses of Gammaplan v10 seem to be closer to measured dose. Besides, for both versions, high dose irradiations (about 120 Gy) measurements are closer to calculated doses comparatively to lower doses (about 30 Gy). 6. CONCLUSION The LGK PFX calibration and delivered dose verification carried out in this study show good agreement between beam calibrations in the LGK conditions compared to TRS 398 reference conditions. “End-to-end” tests lead us to two conclusions: first, it shows good results for calculated dose by TPS, compared to measured doses. Second, it shows the good agreements between both latest versions of LGK PFX available TPS, Gammaplan v9 and v10. In light of those Medical Physics, Vol. 41, No. 6, June 2014
061708-9
results, the basic version of Gamma plan seems suitable for precise calculation of delivered dose in homogeneous part of the brain. However, it could be interesting to proceed to similar studies with the “Convolution” add-on of Gammaplan, which take into account heterogeneous areas. Finally, this study confirmed former studied results about alanine dosimeters to be well-suited for radiotherapy dosimetric applications.
ACKNOWLEDGMENTS The authors would like to thank the LNHB colleagues and in particular, Frank Delaunay, Jean Gouriou, and Nicolas Perichon for methodology pieces of advice, MC simulations, and also Eric Leroy for technical help. Besides, the authors would like to express our gratitude to Dr. Charles-Ambroise Valery, Neurosurgeon at Pitié-Salpêtrière Hospital, Medical Director of the Gamma Knife Unit, and LGK medical staff for their support. The authors extend their thanks to Laurence Roussillon-Constanty, Armand Nachef, and Venkat Narra for English corrections. a) Authors
to whom correspondence should be addressed. Electronic addresses: [email protected] and [email protected] 1 P. R. Almond, P. J. Biggs, B. M. Coursey, W. F. Hanson, M. S. Huq, R. Nath, and D. W. O. Rogers, “AAPM’s TG-51 protocol for clinical reference dosimetry of high-energy photon and electron beams,” Med. Phys. 26, 1847–1870 (1999). 2 P. Andreo, D. T. Burns, K. Hohlfeld, M. S. Huq, T. Kanai, F. Laitano, V. G. Smythe, and S. Vynckier, “Absorbed dose determination in external beam radiotherapy: An international code of practice for dosimetry based on standards of absorbed dose to water,” IAEA Technical Report Series No. 398 (IAEA, Vienna, Austria, 2000). 3 D. McDonald, C. Yount, N. Koch, and M. Ashenafi, “Calibration of the Gamma Knife Perfexion using TG-21 and the solid water Leksell dosimetry phantom,” Med. Phys. 38, 1685–1693 (2011). 4 J. Novotny, Jr., J. P. Bhatnagar, H.-T. Chung, J. Johansson, G. BedR narz, M. Lijun, and M. Huq, “Assessment of variation in Elekta plastic spherical-calibration phantom and its impact on the Leksell Gamma R Knife calibration,” Med. Phys. 37, 5066–5071 (2010). 5 R. Alfonso et al., “Reference dosimetry of small and nonstandard fields,” Med. Phys. 35, 5179–5186 (2008). 6 See more details on Task Group’s web site at http://www.aapm.org/org/ structure/?committee_code=TG178. 7 F. Kuntz, B. Chabanais, D. Karamanoukian, J. P. Delpech, and E. Marchioni, “Dosimétrie RPE - alanine, étude de faisabilité et applications possibles,” J. Chim. Phys. 95, 796–799 (1998). 8 V. Nagy, S. V. Sholom, V. V. Chumak, and M. F. Desrosiers, “Uncertainties in alanine dosimetry in the therapeutic dose range,” Appl. Radiat. Isot. 56, 917–929 (2002). 9 M. Anton, D. Wagner, J.-J. Selbach, T. Hackel, R. M. Hermann, C. Friedrich, and H. Vorwerk, “In vivo dosimetry in the urethra using alanine/ESR during 192 Ir HDR brachytherapy of prostate cancer—A phantom study,” Phys. Med. Biol. 54, 2915–2931 (2009). 10 T. Garcia, M. Anton, and P. Sharpe, “Comparison of alanine dosimetry systems for absorbed dose to water measurements in gamma- and x-radiation at radiotherapy levels,” Metrologia 49(Tech. Suppl. 06004), 1–18 (2012). 11 T. Garcia, T. Lacornerie, R. Popoff, V. Lourenço, and J.-M. Bordy, “Dose R verification and calibration of the Cyberknife by EPR/alanine dosimetry,” Radiat. Meas. 46, 952–957 (2011). 12 N. Perichon, T. Garcia, P. François, V. Lourenço, C. Lesven, and J.M. Bordy, “Calibration of helical tomotherapy machine using EPR/alanine dosimetry,” Med. Phys. 38, 1168–1177 (2011). 13 O. F. Sleptchonok, V. Nagy, and M. F. Desrosiers, “Advancements in accuracy of alanine dosimetry system. Part 1: The effects of environmental humidity,” Radiat. Phys. Chem. 57, 115–133 (2000).
061708-10
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-10
14 D. F. Regulla and U. Deffner, “Dosimetry by ESR spectroscopy of alanine,”
18 M. F. Desrosiers, M. Lin, S. L. Cooper, Y. Cui, and K. Chen, “A study of the
Int. J. Appl. Radiat. Isot. 33, 1101–1114 (1982). 15 F. J. Ahlers and C. C. J. Schneider, “Alanine ESR dosimetry: An assessment of peak-to-peak evaluation,” Radiat. Prot. Dosim. 37, 117–122 (1991). 16 P. H. G. Sharpe, K. Rajendran, and J. P. Sephton, “Progress towards an alanine/ESR therapy level reference dosimetry service at NPL,” Appl. Radiat. Isot. 47, 1171–1175 (1996). 17 T. Garcia, M. Lin, I. Pasquié, and V. Lourenço, “A methodology for choosing parameters for ESR readout of alanine dosimeters for radiotherapy,” Radiat. Phys. Chem. 78, 782–790 (2009).
irradiation temperature coefficient for L-alanine and DL-alanine dosimeters,” Radiat. Prot. Dosim. 120, 235–237 (2006). 19 V. Nagy, J. M. Puhl, and M. F. Desrosiers, “Advancements in accuracy of the alanine dosimetry system, Part 2. The influence of the irradiation temperature,” Radiat. Phys. Chem. 57, 1–9 (2000). 20 JCGM100, Guide to the Expression of Uncertainty in Measurement (GUM) (JCGM, Sèvres, France, 1995). 21 “AFSAPSS décision du 27 juillet 2007 fixant les modalités du contrôle de qualité interne des installations de radiothérapie externe,” J. Offic. Républ. Française texte n◦ 12 (2007).
Medical Physics, Vol. 41, No. 6, June 2014
R Absolute calibration of the Gamma Knife PerfexionTM and delivered dose verification using EPR/alanine dosimetry
Amaury Hornbecka) and Tristan Garciaa) CEA, LIST, Laboratoire National Henri Becquerel, 91191 Gif-sur-Yvette Cedex, France
Marguerite Cuttat and Catherine Jenny Radiotherapy Department, Medical Physics Unit, University Hospital Pitié-Salpêtrière, 75013 Paris, France
(Received 30 July 2013; revised 4 April 2014; accepted for publication 15 April 2014; published 12 May 2014) R Purpose: Elekta Leksell Gamma Knife (LGK) is a radiotherapy beam machine whose features are not compliant with the international calibration protocols for radiotherapy. In this scope, the Laboratoire National Henri Becquerel and the Pitié-Salpêtrière Hospital decided to conceive a new LKG dose calibration method and to compare it with the currently used one. Furthermore, the accuracy of the dose delivered by the LGK machine was checked using an “end-to-end” test. This study also aims to compare doses delivered by the two latest software versions of the Gammaplan treatment planning system (TPS). Methods: The dosimetric method chosen is the electron paramagnetic resonance (EPR) of alanine. Dose rate (calibration) verification was done without TPS using a spherical phantom. Absolute calibration was done with factors calculated by Monte Carlo simulation (MCNP-X). For “end-to-end” test, irradiations in an anthropomorphic head phantom, close to real treatment conditions, are done using the TPS in order to verify the delivered dose. Results: The comparison of the currently used calibration method with the new one revealed a deviation of +0.8% between the dose rates measured by ion chamber and EPR/alanine. For simple fields configuration (less than 16 mm diameter), the “end-to-end” tests showed out average deviations of −1.7% and −0.9% between the measured dose and the calculated dose by Gammaplan v9 and v10, respectively. Conclusions: This paper shows there is a good agreement between the new calibration method and the currently used one. There is also a good agreement between the calculated and delivered doses especially for Gammaplan v10. © 2014 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4873686]
Key words: Gamma Knife, EPR, alanine, dosimetry, calibration 1. INTRODUCTION In radiotherapy treatment centers, high-energy photon beams calibrations follow international protocols, in particular the AAPM TG-51 or the IAEA TRS-398.1, 2 In the past 15 years, technological evolutions have provided better solutions to treat patients more specifically and more precisely with a growing number of particular radiotherapy and radiosurgery devices. The Elekta Leksell Gamma Knife (LGK) Perfexion model (PFX) is a stereotactic radiosurgery equipment. It differs from conventional radiotherapy machines in terms of clinical indications, mechanical design, and beam characteristics. Indeed, it contains 192 sources of 60-cobalt distributed over a crown-shaped collimator at distances of 48.1–51.9 cm from the unit center point (UCP). Table I illustrates the big difference between these parameters and the required international calibration protocols. Thus, one cannot calibrate LGK PFX beams using the latest protocols. Even if other calibration methods have been suggested,3, 4 for example, Alfonso et al.,5 there is no consensus about the way to verify calibration and delivered dose. However, the AAPM Task Group 178 is working on recommendations for calibration and quality insurance of stereotactic radiosurgery devices, which are expected by the end of 2014.6 061708-1
Med. Phys. 41 (6), June 2014
The Laboratoire National Henri Becquerel (LNHB) which is the French National Metrology Laboratory for ionizing radiations decided to study the feasibility of an absolute calibration and delivered dose verification of LGK using electron paramagnetic resonance (EPR)/alanine dosimetry. This technique has already been proved to be adequate dosimeters for radiotherapy and brachytherapy.7–10 Moreover, the LNHB has already carried out such studies for innovative radiotherapy techniques like Cyberknife or Tomotherapy.11, 12 Thanks to the shape and dimensions of alanine dosimeters, they can be used for small beams. Alanine dosimeters have interesting advantages: measures are at solid state and nondestructive, their responses are almost independent on the dose rate and energy of radiotherapy photon beams and their signal is stable in time (signal decreases less than 1% per year).13 First, calibration curves are made at LNHB. Second, beam calibration accuracy is verified by dose rate measurements, without using Gamma Knife’s treatment planning system (TPS). The currently used method is compared to an absolute calibration method proposed by the LNHB and based on MC simulations. Finally, Gamma Knife’s TPS “Gamma plan” calculated doses are compared to EPR/alanine measured dose, using an anthropomorphic head phantom.
0094-2405/2014/41(6)/061708/10/$30.00
© 2014 Am. Assoc. Phys. Med.
061708-1
061708-2
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-2
TABLE I. Main differences between reference conditions and LGK PFX irradiation conditions.
2. MATERIALS 2.A. Calibration curve establishment
2.A.1. Irradiation facilities: Reference beams
Calibration curves are performed with two different reference beams of the LNHB: The Co60 γ -ray: TRS 398 reference conditions for Co60 (i.e., at 80 cm source axis distance or SAD, 5 cm depth in water with a 10 × 10 cm2 field size at reference plane), the dose rate is known with an uncertainty of 0.48% (k = 1). The linear accelerator (LINAC) Saturn 43: TRS 398 reference conditions for 6 MV x-rays (i.e., at 100 cm SAD, 10 cm depth in water with a 10 × 10 cm2 field size at reference plane), the dose rate is known with an uncertainty of 0.38%. 2.A.2. Large water phantom
A large water phantom is used with reference beams. It is a Polymethyl methacrylate (PMMA) tank of 30 × 30 × 30 cm3 filled with distilled water. The phantom is set on a mechanical system allowing submillimetric translation of phantom and dosimeters to insure a precise positioning within the three axes (x, y, z). Micrometric screws are used to measure the source-phantom and the source-detector distances so that the relative position of the dosimeter is known within 10 μm accuracy. 2.A.3. Alanine dosimeters and EPR spectrometer
Alanine dosimeters are made up with four alanine pellets R . Each of included in a cylindrical container made of Delrin them measures 4.8 mm diameter, 3 mm height, 67.5 ± 0.1 mg R . For LGK’s mass and are provided by Synergy Health Medical Physics, Vol. 41, No. 6, June 2014
calibration verification, two pellets are placed in the phantom, without container. The alanine dosimeters signals are measured on a Bruker ELEXSYS E500 EPR spectrometer with an ER 4119 HS resonator, which operates in X-band. The gain of the spectrometer depends on the amplitude of the signal. The sweep time is set to 21 s, the time constant is set to 82 ms, and the conversion time is set to 28 ms. 2.B. Absolute calibration of the LGK
2.B.1. LGK PFX
The LGK aims to treat intracranial tumors and lesions in a unique session. The PFX model uses 192 sources of 60cobalt distributed as a crown shape. Sources are located in eight sectors. Each sector can independently align their 24 sources in front of a 4, 8, or 16 mm collimator channel, or set them in blocked position. All beams converge to a unique geometrical point called UCP. 2.B.2. Spherical acrylonitrile butadiene styrene (ABS) phantom
The 15.9 cm diameter Elekta spherical ABS phantom is usually used by medical physicists to perform the reference dose rate verification of the PFX, according to the Elekta protocol. It is composed of three parts: two hemispheres and one flat plate drilled so that the measurement point of the ionization chamber matches with the UCP of the GammaKnife. Flat plate can be changed allowing different kinds of detectors to be placed in the sphere (ion chamber, radiochromic films, etc.). For this study, a plate has been manufactured by the LNHB for two alanine pellets to be placed at its center (Fig. 1). Whatever the dosimeter, the phantom is assembled
061708-3
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-3
F IG . 1. ABS phantom disassembled: two hemispheres and the flat plate.
with two holding cylinders and fixed on the treatment table via a mechanical system composed by two aluminum plinths as shown in Fig. 2.3
tion chamber volume is 0.125 cm3 with an inner diameter of 5.5 mm which is suitable for LGK PFX use. 2.C. Delivered dose verification (“end-to-end” test)
2.B.3. Dosimetric material
2.B.3.a. Alanine dosimeters. Because of the geometrical restraint of the ABS phantom, only two pellets can be put in the center of the phantom. To do that, a no-drilled flat plate of the ABS phantom is customized to handle two pellets (Fig. 1). 2.B.3.b. Ionization chamber and electrometer. Quality controls and reference dose rate determination are made by the medical physicist using an ionization chamber PTW 31010 associated with a PTW Unidos E electrometer. The ioniza-
2.C.1. Elekta’s TPS Gammaplan v9 and v10
Pitié-Salpêtrière Hospital’s TPS (Gammaplan) has recently been updated (v9 to v10). In order to study dose calculations of both versions, “end-to-end” tests are done for a couple of configurations using v9 then v10. Both versions considered tissues as water and do not take heterogeneities into account. The v10 introduces an evolution of the waterbased algorithm, which, for example, implies new values for
F IG . 2. ABS phantom, including two alanine pellets, fixed on the LGK PFX table. Medical Physics, Vol. 41, No. 6, June 2014
061708-4
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
F IG . 3. Anthropomorphic head phantom fixed, as for a head patient, on a frame, which is fixed on the LGK PFX table.
061708-4
dosimeter amplitude. The EPR readout process involves measurements at different angles α (0◦ for α = 1, 120◦ for α = 2, and 240◦ for α = 3). This is made to take into account a possible response variation with angle. The pellet EPR signal is the average of the central peak amplitude of those three angles. The dosimeter EPR signal is the mean value of the four pellets EPR signals. EPR readout parameters are the ones described by Garcia 2009 except that five sweeps per angle are performed instead of 7.17 The gain is set according to the signal. Correction factors due to pellet mass, gain, and irradiation temperature are applied. So the corrected amplitude is calculated as below ⎛ ⎞ 3 1 ⎜ ⎟ Ai,α 4 ⎟ 1 ⎜ ⎜ 3 α=1 ⎟ ◦ (1 + (20 −T ) × 0.14%) ADosi,corr = ⎜ ⎟ ⎟ 4 i=1 ⎜ mi .10G/20 ⎝ ⎠ 1 = Ai corr , 4 i=1 4
dose profiles (full width at half maximum values, penumbra, relative isodose volumes, output factors, etc.). 2.C.2. Anthropomorphic head phantom
In order to meet the objectives detailed above, an anthropomorphic head has been designed to host an alanine dosimeter and the LNHB developed stands to fix it on a stereotactic frame in order to be as close as possible from real patient treatment conditions, as shown in Fig. 3. 2.C.3. Dosimetric material
Dosimetric material used for delivered dose verifications is alanine dosimeter (four pellets inside a container). 3. METHODS 3.A. Calibration curve establishment and alanine readouts
3.A.1. Irradiation of alanine dosimeters
For each reference beam, 20 alanine dosimeters (two dosimeters per dose) are placed in a 30 × 30 × 30 cm3 water phantom according to reference conditions of the TRS 398 and are irradiated from 3 to 60 Gy. 3.A.2. EPR readout
Alanine EPR spectrum shows five peaks (Fig. 6). The central peak spectrum amplitude measurement has been chosen in accordance with different authors.14–16 EPR measurements are performed at regulated room temperature (23 ± 1 ◦ C) and humidity (50 ± 10% HR) and made two or three weeks after irradiation of the dosimeters to insure signal stabilization. EPR signal of each pellet is analyzed. Tubes in Suprasil quartz with 5 mm internal diameter are used for maintaining the pellet in the readout cavity. Raw amplitude of EPR signal of each pellet is measured. Equation (1) shows how to get to corrected Medical Physics, Vol. 41, No. 6, June 2014
(1)
where Ai,α : amplitude for the pellet i measured for angle alpha; ADosi,corr (u.a.): mean amplitude of the whole dosimeter corrected by:
r G: gain applied to the EPR signal; r mi (mg): weight of the i pellet; r (20−T) × 0.14%: irradiation temperature correction.18, 19
This method is used to read out alanine dosimeter for the calibration curve but also every dosimeter used in the presented work. The only difference will be the number of pellets which are used: four pellets for calibration curves, two for the calibration verification measurements, and three for “end-to-end” tests. 3.A.3. Calibration curve D = f (EPR amplitude)
20 alanine dosimeters are irradiated from 3 to 60 Gy at the LNHB. Indeed, the LHNB uses 60 Co irradiators and linac with precise delivered doses (u = 0.49% and u = 0.38%, respectively). Both are used for calibration curve establishment to confirm their equivalence, which has already been proven in previous internal works made by the LNHB. Each dosimeter is irradiated in a water phantom, according to the TRS 398 reference conditions, for their signal to be easily related to an absorbed dose into water, with low uncertainties. Linear regression is calculated between EPR amplitudes with their associated uncertainties17 and delivered dose with associated uncertainties (dosimeter positioning, irradiation time, etc.). LNHB alanine calibration curves had been validated as suitable for use of alanine in nonreference conditions.12 Uncertainties are calculated for k = 1, using the following expression:20 u = ucf 2 + ud 2 + ut 2 + ureg 2 , (2)
061708-5
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
where ucf is the uncertainty on the conversion factors applied to determine dose rate values under standard conditions (type B), ud is the uncertainty on dosimeter positioning (type B), ut is the uncertainty on irradiation time (type B), ureg is the uncertainty on linear regression curve slope (type A). 3.B. Absolute calibration of the LGK
3.B.1. Routine reference dose rate verification
The ionization chamber is put in the middle of the spherical ABS phantom which is mounted on the table. The LGK is set so that all the sources are in front of the 16 mm collimators channels. In those conditions, ion chamber measurements are performed for 60 s of irradiation, and the measurement is reproduced seven times for a good statistic. Thanks to the calibration factor of the ion chamber, the dose rate D˙ w,i (where “i” is the number of the measurement) can be obtained after being corrected by kT,P, kpol , and kQ,Q0 factors (ambient temperature and pressure, polarity, and beam quality). The formalism proposed by Alfonso et al.5 is currently used at the Pitié-Salpêtrière Hospital for measurements in Tomotherapy and Gamma Knife. This formalism supposes the use of f ,fmsr f ,fref and kQpcsr correction factors, taking into acthe kQmsr msr ,Q pcsr ,Qmsr count the differences between the conditions of field size, geometry, phantom material, quality, and the use of composite fields. Then, the reference dose rate D˙ w is the average of all those measurements corrected by all those factors. In that way, the medical physicist can check that the reference dose rate of measurements is compliant with the one used in the TPS and with the French legislation that requires 2% maximal deviation valuable for classical radiotherapy devices.21 Equation (3) is used to determinate the dose rate f
D˙ W,i =
,f
ref Mi .ND,W .kT ,P .kpol .kQ,Q0 .kQmsr msr ,Q
dt
,
(3)
where: M: charge measurements in [nC]. ND,W : calibration factor in terms of absorbed dose to water for a dosimeter at a reference beam quality Q0 . kQ,Q0 = 1 due to the use of 60 Co beams. f ,fref kQmsr is set equal to 1 because constructor’s calibration msr ,Q protocol is not based on Alfonso’s formalism. 3.B.2. Dose rate verification with alanine dosimeter
Alanine pellets are put in the middle of the ABS phantom in a specifically drilled plate that can receive two pellets. As they are completely surrounded by ABS material to insure homogeneity of the environment, the dose measured in such conditions is an absorbed dose in alanine inside an ABS environment. To increase statistic quality, nine measurements are done for four irradiations durations (4, 8, 11, and 15 min). An absolute calibration implies a method allowing expressing results in terms of absorbed dose to water, inside a water environment, and in the TRS 398 reference conditions. Thus, ionMedical Physics, Vol. 41, No. 6, June 2014
061708-5
ization chamber readouts can be compared with EPR/alanine readouts. To get to such an expression, the LNHB developed the following relations: Step 1: In the reference conditions, determination of the calibration coefficient in terms of dose in water to an absorbed dose in alanine. This step aims to take into account the change of alanine measurement configuration (four pellets in the container for the calibration curve vs two pellets in the ABS phantom) and the attenuation of the alanine itself
Dalanine DW . , (4) NDalanine = ADosi,corr DW,tank MC where:
r NDalanine : calibration coefficient; r DW : absorbed dose to water, delivered by reference beam;
r ADosi,corr : average EPR signal corresponding to the absorbed dose delivered by LNHB reference beams, in TRS 398 reference conditions, for calibration curve establishment;
r Dalanine = 0.976 ± 0.002 which is the absorbed dose Dw,tank MC ratio in alanine and in water, taking the delrin container into account, calculated by MCNP-X code. The geometry set for MCNP-X calculation of this ratio is: -Source: 1 beam of 60-Co, mean energy 1.25 MeV, beam orientation: horizontal. -Phantom: Water cube (30 × 30 × 30 cm3 ). The center of the cube placed on the 60 Co beam axis. -Skin-source distance: 95 cm. -Detector: alanine for Dalanine and water for Dw,tank , the center of the detector placed on the beam axis at 100 cm from the source (5 cm depth in water). Step 2: Determination of a calibration coefficient in terms of absorbed dose to alanine in the ABS spherical phantom. This step allows a transition from the reference conditions to the LGK conditions. Furthermore, because of the complexity of a whole LGK geometrical configuration, the problem has been considered symmetrical in spherical dimensions, so the simulation includes only one 1.25 MeV photon beam
Dsphere , (5) NDsphere = NDalanine Dalanine MC where:
r NDSphere : calibration coefficient in the ABS spherical r
phantom.
Dsphere Dalanine
= 1.005 ± 0.003 which is the absorbed dose MC ratio in the ABS spherical phantom and in alanine (calculated by MC simulation).
The geometry set for MCNP-X calculation of this ratio is: -Source: 1 beam of 60-Co, mean energy 1.25 MeV, beam orientation: horizontal
061708-6
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-6
3.C. Delivered dose verification (“end-to-end” test)
3.C.1. Test “end-to-end”
The “end-to-end” test aims to reproduce the different patient treatment process in order to evaluate the delivered dose by the LGK via the TPS in real conditions with the following steps:
r CT-scan acquisition; r treatment plan calculation; r LGK treatment irradiation. Images are acquired by a CT-scan, two different treatments are planned (Fig. 5):
F IG . 4. Monte Carlo simulation of secondary electrons around the ABS phantom (black sphere) irradiated by a Cobalt 60 beam.
-Phantom: ABS sphere (diameter 16 cm). The center of the cube placed on the 60 Co beam axis. -Skin-source distance: 50 cm. -Detector: alanine for Dalanine and ABS for Dsphere . The center of the detector is the same as the center of the phantom. The simulation is done for up to 108 histories, its result is shown in Fig. 4. Step 3: in ABS spherical phantom, from the EPR signal of alanine dosimeters (calibrated in terms of absorbed dose to water) to the absorbed dose into water, in LGK PFX conditions. Equations (4) and (5) development is DW ADosi,corr
Dsphere Dalanine . . .Amoy,corr , Dw,tank MC Dalanine MC
Dsphere = NDsphere .Amoy,corr =
(6)
where: Dsphere : absorbed dose in LGK conditions brought back to TRS 398 reference conditions; Amoy,corr : average EPR signal corresponding to the two pellets for LGK irradiation; Step 4: Brought into reference conditions, the reference dose rate can now be determined, using the correct irradiation duration (4, 8, 11, or 15 min) dDsphere . D˙ sphere = dt
(7)
Two weeks after irradiation in LGK, EPR spectra are acquired and the corresponding dose is obtained thanks to the calibration. Corrected measured dose rates are compared to the reference dose rate, measured by ion chamber. Medical Physics, Vol. 41, No. 6, June 2014
-Plan 1: A single shot (centered on the dosimeter) is performed with all sectors aligned with the 16 mm collimators channels. To improve measurements statistics, three dosimeters are irradiated for Dmax = 39.60 Gy (Gammaplan v9), three for Dmax = 39.50 Gy (Gammaplan v10), one for Dmax = 118.80 Gy (Gammaplan v9), and one for Dmax = 118.50 Gy (Gammaplan v10). -Plan 2: A double-shot (16 and 8 mm) is performed: Three dosimeters are used for Dmax = 36.40 Gy (Gammaplan v9), three for Dmax = 37.10 Gy (Gammaplan v10), one for Dmax = 109.20 Gy (Gammaplan v9), and one for Dmax = 111.30 Gy (Gammaplan v10). In this precise case, the maximum dose Dmax will be used, even though neurosurgeons usually work with the D50 . But the maximum dose actually calculated in the PTV is not always the same value than the maximal prescript dose Dmax . So the difference will be made between Dmax and Dmax,PTV , which are, respectively, the maximum prescript dose and the maximum calculated dose in the PTV. Finally, Dmax,PTV will be compared to alanine measurements. For each plan, Dmax,PTV is obtained from dose profiles measured by Gamma Plan inside the PTV. To verify if the Dmax,PTV covered the four alanine pellets of the dosimeter, dose-volume histograms and dose profiles are studied. It appeared that dose profiles are not adapted for the entire dosimeter, consequently, only three pellets are used (Fig. 5) for PTV, whereas the bottom pellet was discarded. Two weeks after irradiation in LGK PFX, EPR spectra are acquired and corresponding doses are obtained thanks to the calibration curve. After been projected on the calibration curve, alanine measured doses are able to be compared to the average maximal doses calculated by the TPS in the PTV. 4. RESULTS 4.A. Calibration curve establishment and alanine readouts
From EPR readouts corresponding to dose values at the reference point, calibration curve is established by linear regression (Fig. 6). The calibration curve is linear (R2 = 1) and uncertainties on EPR signal decrease while the absorbed dose increases. EPR signals of alanine can now be related to an absorbed dose in water.
061708-7
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-7
F IG . 5. Planning of the treatment: two different plans showing the dosimeter pellet 2, 3, and 4 entirely covered by the isodoses at 50% of the maximum dose.
F IG . 6. Calibration curve alanine EPR signal as a function of absorbed dose in water, in TRS 398 reference conditions for Co60 beam, with associated uncertainties. Medical Physics, Vol. 41, No. 6, June 2014
061708-8
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-8
TABLE II. Test “end-to-end” results: Absorbed dose to water measured by alanine vs prescripted dose (TPS). EPR
Gammaplan v9
Sector configuration
Alanine dosimeter
Absorbed dose (Gy)
Uncertainty (%)
Isodose 100% (Gy)
Uncertainty (%)
Disparity EPR/TPS (%)
Plan 1
v9-1 v9-2 v9-3 v9-7 v9-4 v9-5 v9-6 v9-8 v10-1 v10-2 v10-3 v10-7 v10-4 v10-5 v10-6 v10-8
38.7 38.9 38.7 117.5 35.7 35.9 35.9 108.1 39.0 39.1 39.0 117.9 36.7 36.7 36.8 111.1
1.2 1.2 1.2 0.7 1.3 1.3 1.3 0.7 1.3 1.3 1.3 1.4 1.3 1.0 1.0 1.0
39.6 39.6 39.6 118.8 36.4 36.4 36.4 109.2 39.5 39.5 39.5 118.5 37.1 37.1 37.1 111.3
2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
− 2.2 − 1.9 − 2.4 − 1.1 − 1.8 − 1.4 − 1.4 − 1.0 − 1.2 − 1.1 − 1.3 − 0.5 − 0.9 − 1.0 − 0.9 − 0.2
Plan 2
Gammaplan v10
TPS
Plan 1
Plan 2
4.B. Absolute calibration of the LGK
Dose rate verification means to compare measurements with ionization chamber (reference dose rate) and alanine EPR signal, projected to the calibration curve and brought back to reference conditions. 4.B.1. Reference dose rate verification
Seven measurements are done with PTW 31010 ion chamber. The correction of the temperature and pressure kT,P = 1.01456, the correction of the polarity kpol = 1, kQ,Q0 = 1 due f ,fref fpcsr ,fmsr to the use of 60 Co beams and kQmsr .kQpcsr ,Qmsr is set equal to msr ,Q 1. The calibration factor of the PTW 31003 is Nw = 0.3083 Gy nC−1 . The mean reference dose rate measured with ion¯˙ −1 ization chamber is D W,IC = 2.84 ± 0.01 Gy min . The slight difference (0.2%) from the reference dose rate used in the ˙ W,TPS = 2.84 ± 0.06 Gy min−1 is within tolerances. TPS, D 4.B.2. Dose rate verification with alanine dosimeter
According to the method developed in Sec. 3.B.2, the measured dose rates by EPR/alanine dosimetry lie with an average of 2.86 ± 0.06 Gy min−1 , that is to say, +0.8% relative to ion chamber measurements. Here, the LNHB Linacmade calibration curve has been used. ¯˙ The average dose rate measured by alanine is D W,alanine −1 = 2.86 ± 0.06 Gy min corresponds to a +0.8 % deviation ˙ W,TPS , the TPS dose rate, which is within measurement from D uncertainties. 4.C. Delivered dose verification (“end-to-end” test)
Alanine measurements are done for both versions of Gammaplan v9 and v10. For each TPS version, four measurements are done for plan 1 and four other measurements Medical Physics, Vol. 41, No. 6, June 2014
for plan 2 (see Sec. 3.C.1 for more details). The calibration curve made in the 60-cobalt beam has been used. The aim is to verify the consistency between calculated dose and the dose actually delivered but also to test differences between both versions for simple beam configurations. Table II illustrates the results of the “end-to-end” test. All measured doses are lower than TPS ones, the average measured doses by EPR/alanine compared to the calculated doses for both plans show an average disparity of −1.7% for Gammaplan v9 and an average disparity of 0.9% for Gammaplan v10 (Table II).
5. DISCUSSIONS 5.A. Absolute calibration
Calibration verification has been made by comparison between ion chamber dose rate and reference dose rate based on and EPR/alanine measurements. The dose rate measured with alanine is close to the dose rate measured with ion chamber by +0.8%; nevertheless, it is important to note that the holders of the ABS phantom attenuate beams of sectors 3 and 7, as shown in Fig. 2. This attenuation was quantified by McDonald et al.3 as almost 1% and is not taken into account for reference dose rate used in the TPS for dose calculations. Besides, this effect is not visible in the dose rate comparison because the same holder is used for both detectors. However, during a patient (or anthropomorphic phantom) treatment, beams of sectors 3 and 7 are not attenuated, because the frame holder geometry is different from the ABS phantom holder. As a consequence, the actual measured dose was expected to be slightly higher than the calculated one. If the effect of ABS phantom holders of 1% would be taken into account, the disparities would be −2.7% and −1.9% for Gamma plan v9 and v10, respectively.
061708-9
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
So the dose rate is expressed as a dose rate into water measured in the ABS phantom taking into account there is no correction factor to consider the difference of attenuation between ABS and water. So the TPS calculated doses are set with a dose rate value measured by the ion chamber as an absorbed dose into water, but actually attenuated by ABS. This effect is not taken into account neither for the clinical measurements nor for LGK acceptance tests. A newer spherical solid water phantom provides by the manufacturer is made in water equivalent material and has a different holder system which does not attenuate cobalt beams.3 It would be interesting to perform such a study using this phantom. But the ABS phantom was the only one available in this work. So to quantify this effect, a correction factor has been calculated by MC simulation using the MCNP-X. For the simulation, the number of histories is set between 2 × 108 and 5 × 108 . The calculated correction factor is
DW = 1.015 ± 0.003. Dsphere MC If this factor had been taken into account in the measurement, the average ionization chamber dose rate would have been
M.ND,W .kT ,P .kpol DW ˙ . . (8) DW = dt Dsphere MC Thus, dose rate should be equal to 2.88 Gy min−1 . Hence, the average difference between alanine and ionization chamber measurements should be −0.7%. 5.B. Delivered dose verification
All measured doses are lower than calculated doses for both plans. Regarding the discussion of Sec. 5.A, the measured dose was expected to be higher than the calculated one. Nevertheless, the effect of the attenuation due to the ABS phantom holder is not obvious. This effect seems to be compensated by the fact that heterogeneities of the anthropomorphic head phantom made by bone and tissue (water equivalents), and eventually sinus cavities are not taken into account by the TPS, which considers everything as water for calculation. However, this study showed very good agreements between EPR/alanine measurements and TPS calculated dose, delivered doses of Gammaplan v10 seem to be closer to measured dose. Besides, for both versions, high dose irradiations (about 120 Gy) measurements are closer to calculated doses comparatively to lower doses (about 30 Gy). 6. CONCLUSION The LGK PFX calibration and delivered dose verification carried out in this study show good agreement between beam calibrations in the LGK conditions compared to TRS 398 reference conditions. “End-to-end” tests lead us to two conclusions: first, it shows good results for calculated dose by TPS, compared to measured doses. Second, it shows the good agreements between both latest versions of LGK PFX available TPS, Gammaplan v9 and v10. In light of those Medical Physics, Vol. 41, No. 6, June 2014
061708-9
results, the basic version of Gamma plan seems suitable for precise calculation of delivered dose in homogeneous part of the brain. However, it could be interesting to proceed to similar studies with the “Convolution” add-on of Gammaplan, which take into account heterogeneous areas. Finally, this study confirmed former studied results about alanine dosimeters to be well-suited for radiotherapy dosimetric applications.
ACKNOWLEDGMENTS The authors would like to thank the LNHB colleagues and in particular, Frank Delaunay, Jean Gouriou, and Nicolas Perichon for methodology pieces of advice, MC simulations, and also Eric Leroy for technical help. Besides, the authors would like to express our gratitude to Dr. Charles-Ambroise Valery, Neurosurgeon at Pitié-Salpêtrière Hospital, Medical Director of the Gamma Knife Unit, and LGK medical staff for their support. The authors extend their thanks to Laurence Roussillon-Constanty, Armand Nachef, and Venkat Narra for English corrections. a) Authors
to whom correspondence should be addressed. Electronic addresses: [email protected] and [email protected] 1 P. R. Almond, P. J. Biggs, B. M. Coursey, W. F. Hanson, M. S. Huq, R. Nath, and D. W. O. Rogers, “AAPM’s TG-51 protocol for clinical reference dosimetry of high-energy photon and electron beams,” Med. Phys. 26, 1847–1870 (1999). 2 P. Andreo, D. T. Burns, K. Hohlfeld, M. S. Huq, T. Kanai, F. Laitano, V. G. Smythe, and S. Vynckier, “Absorbed dose determination in external beam radiotherapy: An international code of practice for dosimetry based on standards of absorbed dose to water,” IAEA Technical Report Series No. 398 (IAEA, Vienna, Austria, 2000). 3 D. McDonald, C. Yount, N. Koch, and M. Ashenafi, “Calibration of the Gamma Knife Perfexion using TG-21 and the solid water Leksell dosimetry phantom,” Med. Phys. 38, 1685–1693 (2011). 4 J. Novotny, Jr., J. P. Bhatnagar, H.-T. Chung, J. Johansson, G. BedR narz, M. Lijun, and M. Huq, “Assessment of variation in Elekta plastic spherical-calibration phantom and its impact on the Leksell Gamma R Knife calibration,” Med. Phys. 37, 5066–5071 (2010). 5 R. Alfonso et al., “Reference dosimetry of small and nonstandard fields,” Med. Phys. 35, 5179–5186 (2008). 6 See more details on Task Group’s web site at http://www.aapm.org/org/ structure/?committee_code=TG178. 7 F. Kuntz, B. Chabanais, D. Karamanoukian, J. P. Delpech, and E. Marchioni, “Dosimétrie RPE - alanine, étude de faisabilité et applications possibles,” J. Chim. Phys. 95, 796–799 (1998). 8 V. Nagy, S. V. Sholom, V. V. Chumak, and M. F. Desrosiers, “Uncertainties in alanine dosimetry in the therapeutic dose range,” Appl. Radiat. Isot. 56, 917–929 (2002). 9 M. Anton, D. Wagner, J.-J. Selbach, T. Hackel, R. M. Hermann, C. Friedrich, and H. Vorwerk, “In vivo dosimetry in the urethra using alanine/ESR during 192 Ir HDR brachytherapy of prostate cancer—A phantom study,” Phys. Med. Biol. 54, 2915–2931 (2009). 10 T. Garcia, M. Anton, and P. Sharpe, “Comparison of alanine dosimetry systems for absorbed dose to water measurements in gamma- and x-radiation at radiotherapy levels,” Metrologia 49(Tech. Suppl. 06004), 1–18 (2012). 11 T. Garcia, T. Lacornerie, R. Popoff, V. Lourenço, and J.-M. Bordy, “Dose R verification and calibration of the Cyberknife by EPR/alanine dosimetry,” Radiat. Meas. 46, 952–957 (2011). 12 N. Perichon, T. Garcia, P. François, V. Lourenço, C. Lesven, and J.M. Bordy, “Calibration of helical tomotherapy machine using EPR/alanine dosimetry,” Med. Phys. 38, 1168–1177 (2011). 13 O. F. Sleptchonok, V. Nagy, and M. F. Desrosiers, “Advancements in accuracy of alanine dosimetry system. Part 1: The effects of environmental humidity,” Radiat. Phys. Chem. 57, 115–133 (2000).
061708-10
Hornbeck et al.: Calibration, delivered dose verification of the Gamma Knife
061708-10
14 D. F. Regulla and U. Deffner, “Dosimetry by ESR spectroscopy of alanine,”
18 M. F. Desrosiers, M. Lin, S. L. Cooper, Y. Cui, and K. Chen, “A study of the
Int. J. Appl. Radiat. Isot. 33, 1101–1114 (1982). 15 F. J. Ahlers and C. C. J. Schneider, “Alanine ESR dosimetry: An assessment of peak-to-peak evaluation,” Radiat. Prot. Dosim. 37, 117–122 (1991). 16 P. H. G. Sharpe, K. Rajendran, and J. P. Sephton, “Progress towards an alanine/ESR therapy level reference dosimetry service at NPL,” Appl. Radiat. Isot. 47, 1171–1175 (1996). 17 T. Garcia, M. Lin, I. Pasquié, and V. Lourenço, “A methodology for choosing parameters for ESR readout of alanine dosimeters for radiotherapy,” Radiat. Phys. Chem. 78, 782–790 (2009).
irradiation temperature coefficient for L-alanine and DL-alanine dosimeters,” Radiat. Prot. Dosim. 120, 235–237 (2006). 19 V. Nagy, J. M. Puhl, and M. F. Desrosiers, “Advancements in accuracy of the alanine dosimetry system, Part 2. The influence of the irradiation temperature,” Radiat. Phys. Chem. 57, 1–9 (2000). 20 JCGM100, Guide to the Expression of Uncertainty in Measurement (GUM) (JCGM, Sèvres, France, 1995). 21 “AFSAPSS décision du 27 juillet 2007 fixant les modalités du contrôle de qualité interne des installations de radiothérapie externe,” J. Offic. Républ. Française texte n◦ 12 (2007).
Medical Physics, Vol. 41, No. 6, June 2014
