EURADOS Report 2017-01: "Dosimetry for second cancer risk ...

EURADOS Report 2017-01 European Radiation Dosimetry Group e. V.

Neuherberg, April 2017

Dosimetry for second cancer risk estimation in radiotherapy: measurements in water phantoms R.M. Harrison, A. Di Fulvio, J-M. Bordy, S. Miljanić, L. Stolarczyk and Ž. Knežević.

ISSN 2226-8057 ISBN 978-3-943701-14-2

European Radiation Dosimetry Group e.V. EURADOS Report 2017-01 Neuherberg, April 2017

Dosimetry for second cancer risk estimation in radiotherapy: measurements in water phantoms Authors: R.M. Harrison1, A. Di Fulvio2, J-M. Bordy3, S. Miljanić4, L. Stolarczyk5 and Ž. Knežević.4 Institute of Cellular Medicine, Faculty of Medical Sciences, University of Newcastle, UK

1

Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, USA

2

Commissariat à l’Énergie Atomique et aux Énergies Alternatives (CEA), LIST, LNE/LNHB, 91191 Gif sur Yvette, France

3

4

Ruđer Bošković Institute, Bijenička 54, 10000 Zagreb, Croatia

5

Institute of Nuclear Physics, Radzikowskiego 152, Krakow, Poland ISSN 2226-8057 ISBN 978-3-943701-14-2

Imprint © EURADOS 2017 Issued by: European Radiation Dosimetry e.V. Postfach 1129 D-85758 Neuherberg Germany [email protected] www.eurados.org The European Radiation Dosimetry e.V. is a non-profit organization promoting research and development and European cooperation in the field of the dosimetry of ionizing radiation. It is registered in the Register of Associations (Amtsgericht Braunschweig, registry number VR 200387) and certified to be of non-profit character (Finanzamt Braunschweig-Altewiekring, notification from 2008-03-03). Liability Disclaimer No liability will be undertaken for completeness, editorial or technical mistakes, omissions as well as for correctness of the contents.

Contributors

Contributing members of Working Group 9 during the preparation of this report: Pawel Olko, Institute of Nuclear Physics, Krakow, Poland. Emiliano D'Agostino, Belgian Nuclear Research Institute (SCK-CEN), Mol, Belgium. Carles Domingo, Universitat Autónoma de Barcelona (UAB), Barcelona, Spain. Francesco d’Errico, Department of Mechanical, Nuclear and Production Engineering, University of Pisa, Italy. Igor Bessieres, CEA, LIST, DCSI, Gif-sur-Yvette, France. Associate contributors A. Ostrowsky, CEA, LIST, LNE/LNHB, Gif-sur-Yvette, France M. de San Pedro, Departament de Física, Universitat Autònoma de Barcelona, Spain L. Tana, Santa Chiara University Hospital, Pisa, Italy B. Poumarede, CEA, LIST, DCSI, Gif-sur-Yvette, France S. Sorel, CEA, LIST, LNE/LNHB, Gif-sur-Yvette, France D. Vermesse, CEA, LIST, LNE/LNHB, Gif-sur-Yvette, France M. Caresana, Politecnico di Milano, CESNEF, Dipartimento di Energia, Milan, Italy D. Kabat, Centre of Oncology, Kraków, Poland The authors gratefully acknowledge the invaluable contributions of staff in Institutes and hospitals where experimental work was carried out.

Contents Contents................................................................................................................................ i Abstract................................................................................................................................iii 1. Introduction: Dosimetry for second cancer risk estimation in radiotherapy ...............1 2. Project outline ..................................................................................................................4 3. Dosimeters & dosimetry systems ...................................................................................7 3.1 Photon dosimetry systems. ...........................................................................................7 3.1.1 Luminescence detectors for out-of-field dosimetry ...................................................................... 7 3.1.2 Principles of readout, annealing and measurement procedures for TL, OSL and RPL dosimeters........................................................................................................................................................... 8 3.1.3 Calibration procedures for TL, OSL and RPL dosimeters............................................................... 8 3.1.4 Dose calculations and uncertainties for TL, OSL and RPL dosimeters.................................... 10 3.1.5 Energy dependence.............................................................................................................................. 12 3.1.6. Dose dependence (linearity) ............................................................................................................. 13 3.2 Neutron dosimetry ......................................................................................................14 3.2.1 Introduction ............................................................................................................................................ 14 3.2.2. Superheated emulsions (SE) .............................................................................................................. 15 3.2.3 Solid state track detectors (radiator degrader neutron spectrometer) .................................. 18 3.3 Detector calibration procedures ..................................................................................21 4. Measurements: Water tank ............................................................................................30 4.1 Dosimeters .................................................................................................................30 4.2 Water tank ..................................................................................................................30 4.3 Reference absorbed dose values ...............................................................................31 4.4 Correction of raw results .............................................................................................32 5. Results: passive dosimeters in a water tank ................................................................33 6. Discussion: water tank results ......................................................................................42 7. Measurements: BOMAB phantom .................................................................................44 7.1 BOMAB phantom design ............................................................................................44 7.2 Radiotherapy treatment techniques ............................................................................46 7.3 Photon measurements: dosimetry methods ................................................................47 7.4 Neutron measurements: dosimetry methods...............................................................50 8. Results: BOMAB phantom .............................................................................................53 8.1 Dosimeter comparisons: photons................................................................................53 8.2 Dosimeter comparison: neutrons ................................................................................56 8.3 Out-of-field doses, photons: comparison with TPS calculations and comparison of modalities .........................................................................................................................60 i

8.3.1 Comparison of doses in different positions (pipes) in the phantom ....................................... 60 8.3.2. Comparison of different treatment modalities ............................................................................ 62 8.3.3. Comparison of treatment planning system (TPS) calculations and dosimeter measurements ................................................................................................................................................. 66 8.4 Out-of-field doses: neutrons........................................................................................73 8.4.1 Comparison of doses in different positions (pipes) in the phantom ....................................... 73 8.4.2. Comparison of different treatment modalities and energies................................................... 78 9. Discussion: BOMAB phantom results ..........................................................................85 9.1 Discussion of measurements and results (photons) ....................................................85 9.1.1. Comparison of dosimetry systems (photons) ............................................................................... 85 9.1.2. Components of out-of-field doses and their characteristics for different modalities........ 85 9.1.3. Characteristics of TPS calculations for different modalities and comparison with dosimeter measurements ............................................................................................................................. 87 9.2 Discussion of measurements and results (neutrons)...................................................89 9.2.1. Comparison of dosimetry systems (neutrons ............................................................................... 89 9.2.2.

Components of out-of-field doses and their characteristics for different modalities 90

10. Risk factors ...................................................................................................................95 11. Conclusions ................................................................................................................104

ii

Abstract This report describes dosimetry techniques used by EURADOS Working Group 9 (Radiation Dosimetry in Radiotherapy) to measure out-of-field doses for use in radiotherapy. Several passive dosimeter types were used, based on the physical principles of thermoluminescence (TLD), optically simulated luminescence (OSL) and radiophotoluminescence (RPL), in addition to an ionization chamber used as a reference standard. Superheated drop, bubble damage and track etch detectors were used specifically for neutron measurements. In this report, doses were measured in a water tank to obtain a matrix of out-of-field measurement points and in a BOMABlike phantom to simulate a prostate treatment. The experimental aspects of the project involved a collaboration between eight European institutions, the Ruđer Bošković Institute , Zagreb, Croatia, the Institute of Nuclear Physics, Krakow, Poland, the Commissariat à l’Energie Atomique, Saclay, France, the Santa Chiara University Hospital, Pisa, Italy, the Belgian Nuclear Research Institute, Mol, Belgium, the Universitat Autònoma de Barcelona, Bellaterra, Spain, the Università di Pisa, Italy and the Politecnico di Milano, Italy. After an introduction and project outline, chapter 3 describes the physical aspects and calibration of the dosimeters used and is followed in chapters 4, 5 and 6 by a description of water tank measurements at lateral distances from the isocentre up to 56 cm and at depths of 10-25 cm. TLD, OSL and RPL dosimeters were compared both within and outside the beam. Within the beam, agreement between TLDs and ionization chamber results was + 1.5% and greater discrepancies were observed for OSL and RPL dosimeters, the latter due to a high atomic number cap. Outside the beam, good agreement with ionization chamber measurements was observed for both TLDs and RPLs. However, OSL results showed an increasing overestimation of dose with distance from the isocentre and a correction was required. Chapter 7 describes the BOMAB-like phantom used for the simulation of a prostate treatment. This was undertaken using Varian Clinac 2300 linear accelerators to simulate treatments for the following conditions: (i) 6MV 4field CRT (ii) 15MV 5-field CRT (iii) 18MV 4-field CRT (iv) 6MV IMRT (v) 18MV IMRT (vi) 6MV VMAT (vii) 6MV Tomotherapy 1. Measurements were compared with treatment planning system (TPS) calculations of the doses at the measurement points and out-of-field dose underestimates were identified. Chapters 8 and 9 discuss the results of the BOMAB phantom experiments for photon absorbed doses and neutron dose equivalents. For the latter, neutron dosimeters showed consistent and repeatable responses. A non-negligible photoneutron dose was observed for a 6 MV primary photon beam. A short summary of some risk models for second cancer induction is given in Chapter 10. .

CRT: Conformal radiotherapy; IMRT: Intensity Modulated Radiotherapy; VMAT: Volumetric Modulated Arc Therapy 1

iii

iv

Dosimetry for second cancer risk estimation in radiotherapy: measurements in water phantoms

1. Introduction: Dosimetry for second cancer risk estimation in radiotherapy The induction of cancers following radiotherapy has been known for many years (NCRP, 2011) although the estimation of the probability of radiation carcinogenesis is not trivial. The overall cancer risk is influenced by the (usually non-uniform) dose to several radiosensitive organs distant from the radiotherapy target volume. The uncertainty in the radiosensitivity of a given organ is often considerable, not least because the risk factors often used are intended for low-dose radiation protection purposes (NCRP, 1993; BEIRVII, 2006; ICRP, 2007). In any case, they are influenced by the sex and age at exposure of the individual. Furthermore, conventional radiotherapy treatment planning systems do not calculate doses and risks to all radiosensitive organs, but usually only to those close to the target volume itself. The subject has attracted renewed interest recently for two main reasons: First, the prognosis for many cancers, including some in which radiotherapy is a significant component (e.g. prostate, breast) has steadily improved. For example, the 10 year survival for prostate cancer in the UK has increased from 20% to 60% over the last 30 years and the corresponding increase for breast cancer is from 40% to 70% in the same period (CRUK, 2012). This means that an increasing number of patients will survive for periods comparable to or greater than the latent period for expression of a second cancer, thus suffering a finite risk of radiocarcinogenesis. Latent periods for radiocarcinogenesis can vary from 5 years to 10 years or more, depending on the site and the individual. For example, a minimum latent period of 5 years has been adopted by BEIR (BEIRVII, 2006) for calculations of lifetime risks. Second, new modalities of photon therapy treatment delivery, such as Intensity Modulated Radiotherapy (IMRT) and Tomotherapy, utilize the principle of building up the required target dose distribution by means of a series of small field exposures, thus leading to higher beam-on times and higher leakage doses compared with previous conventional radiotherapy techniques, for the same target dose. The advantages of a more conformal target dose distribution therefore come at a cost of a higher dose “radiation bath”, where radiosensitive organs and tissues remote from the target volume may receive doses sufficiently large to lead to a significant probability of cancer induction. In fact, Hall and Wuu have suggested that the introduction of IMRT might double the incidence of second cancers (Hall and Wuu, 2003). Furthermore, there is increasing use of imaging techniques for both verification and fraction-by-fraction guidance of dose delivery (image guided radiotherapy – IGRT), involving repeated computed tomography (CT) imaging throughout the course of treatment, using kilovoltage or megavoltage x-rays. The situation may also be further complicated by the administration of adjuvant chemotherapy. Finally, proton radiotherapy is now available at an increasing number of facilities worldwide and the same rationale for the determination of out-of-field doses for this modality applies. Although this report does not address proton radiotherapy, many of the dosimetry techniques used are equally applicable to this modality. Although the benefits of radiotherapy remain and are arguably enhanced by more highly conformal techniques, a central tenet of ICRP (International Commission on Radiological Protection) guidance (ICRP, 2007) is that the benefit of human irradiation should outweigh its risks. Therefore, in order to justify an exposure, both the benefits and the risks must be evaluated and compared. Second cancer risks may, in many cases, be acceptably small. Kry et al. showed that the

EURADOS Report 2017-01

1

R.M. Harrison, A. Di Fulvio, J-M. Bordy, S. Miljanić, L. Stolarczyk and Z. Knežević

absolute risk of inducing a fatal malignancy following prostate radiotherapy was approximately 25%, for several different treatments, accelerator types and endpoint energies, although IMRT gave higher risks than conventional radiotherapy (Kry et al., 2005). Risks should nevertheless be quantified, so that robust risk-benefit judgements may be made. This is particularly relevant for radiotherapy treatments of younger patients who may have long survival prospects and for whom risk factors are higher (BEIRVII, 2006). This is illustrated in Figure 10.2, (taken from (CRUK, 2012) which shows the increasing survival over time for children aged 0-14 years who have received treatment for all forms of lymphoma. Although the determination of absolute second cancer risks is subject to considerable uncertainties, the use of risk models for relative estimates between different treatment techniques (IMRT, proton radiotherapy, Tomotherapy) may assist in the choice of the most appropriate technique to use. Whatever the difficulties and uncertainties in risk estimation, the starting point is the absorbed dose to the irradiated organs. Thus the measurement of out-of-field (sometimes referred to as peripheral) doses, from which specific organ doses may be inferred, is a crucial pre-requisite for risk estimation. The combination of out-of-field dosimetry following all types of radiotherapy may be combined in principle with doses arising from all other sources of irradiation for the radiotherapy patient, such as CT scanning and on-board imaging (OBI) for radiotherapy planning and verification, giving a complete dose specification necessary for future epidemiological studies. This long term objective is a component of the Strategic Research Agenda developed by EURADOS which outlines aspirations for dosimetry developments over the next 20 years (Rühm et al., 2014). This report is based on, and amplifies, the work presented at a workshop held at the Annual Meeting of the European Radiation Dosimetry Group (EURADOS) in Vienna in 2012 at which progress in out-of-field dosimetry for second cancer risk estimation was reviewed, based largely on the work of EURADOS Working Group 9 (Radiation Dosimetry in Radiotherapy, formerly Radiation Protection Dosimetry in Medicine).

References NCRP. 1993. Limitation of exposure to ionizing radiation. NCRP Report 116. National Council on Radiation Protection and Measurements, Bethesda, MD 20814. ISBN 0-929600-30-4 NCRP 2011. Second Primary Cancers and Cardiovascular Disease after Radiation Therapy. NCRP Report 170. National Council on Radiation Protection and Measurements, Bethesda, MD 20814. ISBN 978-0-9823843-9-8 BEIRVII 2006. Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII – Phase 2. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, National Research Council The National Academies Press, Washington DC. ICRP 2007. ICRP Publication 103: The 2007 Recommendations of the International Commission on Radiological Protection Elsevier. CRUK 2012. Cancer Research UK http://info.cancerresearchuk.org/cancerstats.

2



Hall, J.D., Wuu C.-S., 2003. Radiation-induced second cancers: The impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 56, 83-88. Kry, S.F., Salehpour M., Followill D.S., Stovall M., Kuban D.A., White R.A., Rosen I.I., 2005. The calculated risk of fatal secondary malignancies from intensity-modulated radiation therapy. Int.J.Radiation Oncology Biol. Phys. 62, 1195-1203. Rühm, W., Fantuzzi, E., Harrison, RM., Schuhmacher, H., Vanhavere, F., Alves, J., Bottollier-Depois, JF., Fattibene, P., Knežević, Ž., Lopez, MA., Mayer, S., Miljanić, S., Neumaier, S., Olko, P., Stadtmann, H., Tanner, R., Woda, C. 2014 Visions for Radiation Dosimetry over the Next Two Decades - Strategic Research Agenda of the European Radiation Dosimetry Group. EURADOS Report 2014-01 Braunschweig. ISSN 2226-8057. ISBN 978-3-943701-06-7


3


2. Project outline This work was carried out by EURADOS Working Group 9 (WG9, Radiation Dosimetry in Radiotherapy) with the invaluable help of colleagues in several European institutions. The first phase of this project involved the measurement of out-of-field doses in a water tank at the Laboratoire National Henri Bequerel, the French National Laboratory of Metrology for Ionising Radiations (CEA LIST/LNE LNHB, Saclay), using a Saturne 43 linear accelerator to compare various dosimeters under highly reproducible conditions at distances up to 43 cm from the central axis of a single square field. An ionization chamber was used as a reference standard. This was followed by similar measurements in a specially designed phantom which closely resembled a BOttle MAnnikin Absorber (BOMAB) phantom (Bush 1949, Di Fulvio et al., 2013). This phantom consisted of cylindrical and ellipsoidal water-filled poly(methyl methacrylate) (PMMA) sections arranged to approximate a realistic patient shape, but with reproducible geometry amenable to mathematical modelling. Internal out-of-field doses were measured up to 46 cm from the isocentre (that point in space about which the gantry of the linear accelerator, the treatment head, and the couch rotate). Simulations of several prostate treatments were performed using various linear accelerators at three European centres (Santa Chiara University Hospital (SCUH), Pisa, Italy, Presidio Ospedaliero Campo di Marte (POCM), Lucca, Italy and the Centre of Oncology M. Skłodowska-Curie Memorial Institute (COOK), Krakow Branch, Krakow, Poland). Several radiotherapy techniques were used, including 3-D conformal therapy (3DCRT) with multi-leaf collimators (MLC), Intensity-Modulated Radiotherapy (IMRT), Volumetric Modulated Arc Therapy (VMAT) and Tomotherapy. The experimental schedule is given in Table 2.1 and the names and affiliations of the participants in Table 2.2.

Table 2.1: Irradiation schedule for out-of-field measurements. Location

Date

Measurements

Irradiation or treatment technique

Saclay

July 2010

Water tank

Single field

Saclay

April 2011

Water tank

Single field

Pisa

July 2011

BOMAB phantom

Prostate treatment simulation

Krakow

August 2011

BOMAB phantom


Pisa

June 2012

BOMAB phantom


Krakow

Sept 2012

BOMAB phantom


Pisa

Sept 2012

BOMAB phantom


4



Participating Institutions: Ruđer Bošković Institute , Zagreb, Croatia (RBI) Institute of Nuclear Physics, Krakow, Poland (IFJ) Centre of Oncology M. Skłodowska-Curie Memorial Institute (COOK), Krakow, Poland). Commissariat à l’Énergie Atomique et aux Énergies Alternatives, Saclay, France (CEA) Santa Chiara University Hospital, Pisa, Italy (SCUH) Presidio Ospedaliero Campo di Marte (POCM), Lucca, Italy University of Newcastle upon Tyne, Newcastle (UK) Belgian Nuclear Research Institute, Mol, Belgium (SCK CEN) Universitat Autònoma de Barcelona, Bellaterra, Spain (UAB) Università di Pisa, Pisa, Italy (UNIPI) Politecnico di Milano, Milano, Italy (PoliMi) Table 2.2 Participants and Institutions Participants

Institution and address

R.M. Harrison

University of Newcastle upon Tyne, UK

J.M. Bordy, S Sorel, A. Ostrowsky, D. Vermesse

CEA, LIST, LNE/LNHB, 91191 Gif sur Yvette, France

I. Bessieres, B. Poumarede

CEA, LIST, DCSI, 91191 Gif sur Yvette, France

E. d'Agostino

Radioprotection, Dosimetry and Calibration, Belgian Nuclear Research Institute, Mol

C. Domingo, M. De San Pedro

Departament de Física, Universitat Autònoma de Barcelona, Spain

F. d'Errico, A. Di Fulvio

Department of Mechanical, Nuclear and Production Engineering, University of Pisa, Italy & Yale University School of Medicine, USA

L. Tana

Santa Chiara University Hospital, via Roma 67, Pisa, Italy

S. Miljanić , Ž. Knežević

Ruđer Bošković Institute, Bijenička 54, Zagreb, Croatia

P Olko, L. Stolarczyk

Institute of Nuclear Physics, Krakow, Poland

D. Kabat

Centre of Oncology M. Skłodowska-Curie Memorial Institute Krakow Branch Poland

M. Caresana

Centre Politecnico di Milano, CESNEF, Dipartimento di Energia, via Ponzio Milano, Italy


5


Photon dosimeters used: Ionization chamber dosimetry (CEA) Thermoluminescence dosimetry (TLD): TLD-700 & TLD-100 (RBI); TLD MTS-7 (IFJ) Optically stimulated luminescence dosimetry (OSL): (CEA) Radiophotoluminescence dosimetry (RPL): (RBI) Neutron dosimeters used: Superheated emulsion detectors: superheated drop detectors (SDD): (UNIPI and Yale) and Bubble-Damage detectors (BD-PND and BDT): (SCK CEN) Solid state nuclear track etch detectors (SSTD): polyallyldiglycol carbonate track etched detectors (PADC), (UAB) and Radiator Degrader Neutron Spectrometer (RDNS): (PoliMi)

References Bush, F. 1949 The integral dose received from a uniformly distributed radioactive isotope. Br. J. Radiol. 22, 96-102 Di Fulvio, A., Tana, L., Caresana, M., D'Agostino, E., De San Pedro, M., Domingo, C., D'Errico, F., 2013. Clinical simulations of prostate radiotherapy using BOMAB-like phantoms: Results for neutrons. Radiat Meas 57, 48-61.

6



3. Dosimeters & dosimetry systems 3.1 Photon dosimetry systems.

3.1.1 Luminescence detectors for out-of-field dosimetry Photon dosimetry methods used in this work were TL, RPL and OSL. The basic principles as well as a literature review of the use of these dosimeters for out-of-field dosimetry are described in detail in the paper by Knežević, Stolarczyk et. al. (2013). Different types of luminescence detectors were chosen for the EURADOS WG9 measurement campaign due to their high sensitivity and well-characterized dose and energy response suitable for this application. As hundreds of dosimeters for each experiment were used, their small physical size, long term stability, reproducibility and batch homogeneity were also important factors. The folowing types of luminescence detectors were used: (i) Three types of TL detectors (LiF:Mg,Ti): MTS-7, TLD-100 and TLD-700 (ii) OSL nanoDotTM detectors (Al2O3:C) (iii) RPL (GD-352M) detectors Relevant features of these detectors are given in Table 3.1. In this report, only the most important characteristics for out-of-field dose measurements will be discussed in detail. Table 3.1 Basic characteristics of detectors used by WG9 Detector and manufacturer

MTS-7 (IFJ PAN, Poland) TLD-700 (Thermo Fisher Scientific) TLD-100 (Thermo Fisher Scientific)

Material

Form

Dimensions (mm x mm)

Effective atomic number Zeff

LiF: Mg, Ti,

pellet

Φ 4.5×0.9

8.14

pellet

Φ 4.5×0.9

8.14

Modified TOLEDO 654 reader (Vinten)

pellet

Φ 4.5 × 0.9

8.14

Modified TOLEDO 654 reader (Vinten)

7

7

LiF: Mg, Ti

nat

LiF: Mg, Ti

nanoDotTM dosimetry system (Landauer Inc.)

Al2O3:C

pellet adapter

10×10×2 3 mm

11.28

RPL (GD-352M) (ATGC)

Ag activated Phosphate glass

rod holder

Φ 1.5 × 12 Φ 4.3 × 14.5

12.04

Reader

RA’94 TL ReaderAnalyser (Mikrolab, Poland)

Semiautomatic reader MicroStarTM (Landauer Inc.) Automatic reader Dose Ace (FGD-1000)

In the table, Φ 4.5×0.9 indicates a circular diameter of 4.5 mm and a thickness of 0.9 mm.


7


3.1.2 Principles of readout, annealing and measurement procedures for TL, OSL and RPL dosimeters Dosimetry with TLDs requires complex thermal annealing steps (Table 3.2), which result in reestablishing the defect equilibrium. The readout of MTS-7 was performed with an RA94 TL ReaderAnalyser with platinum heating planchet and an EMI bi-alkali 9789QB photomultiplier tube with BG-12 infrared filter. The readout of TLD-100 and TLD-700 dosimeters (Table 3.2) was carried out using a modified manual TOLEDO 654 (Vinten) reader, which enabled detailed analysis and integration of the glow curves with variable integration limits (Knežević et al., 2005).

Table 3.2 Annealing and readout parameters for TL (MTS-7, TLD-100, TLD-700) and RPL GD352M dosimeters used in WG 9 measurements. Detector Pre-irradiation annealing Temperature (°C) Time (min) Preheat Temperature (°C) Time (min) Readout Temperature (°C) Time (s) Heating rate (°C s-1)

MTS-7

TLD-100 TLD-700

RPL (GD-352M)

400 + 100 60 + 120

400 + 100 60 + 120

400 20 or 60 (>1Gy)

100 10

100 20

70 30

360 60 5

270 35 10

UV excitation

The read-out of nanoDotTM (Landauer Inc.) dosimeters with a plastic adapter, which provides protection from light, was performed using the semiautomatic reader MicroStarTM NanoDotTM system (Perks et al., 2007). This reader uses continuous wave OSL (CW-OSL), which consists of continuously illuminating the dosimeters, whilst monitoring the OSL intensity with a 1 s illumination-read period. The calibration of the MicroStarTM reader was performed in air for low (< 150 mGy) and high doses (> 150 mGy). The dose readout of the RPL GD-352M dosimeters was performed automatically by the FDG-1000 reader. After irradiating and exciting the glass with pulsed UV light (20 pulses/second), the photoluminescence signal can be observed. Given the high frequency of the pulsed interrogation beam, the signal can be read multiple times in a short time, without signal degradation. Multiple readouts also improve the statistical measurement precision (ATGC, 2007). The reader is automatically calibrated using the internal calibration element and standard irradiation element (free-in-air) and thus provides a direct measurement of the dose. After irradiation and before readout, the TL and RPL dosimeters were preheated, in order to accelerate the build-up activity, in the case of RPL, and to erase low temperature peaks, in the case of TLDs (Table 3.2).

3.1.3 Calibration procedures for TL, OSL and RPL dosimeters Absorbed dose to water Dw is the quantity of main interest in radiation therapy, since it is well defined and water is nearly tissue equivalent for photons (TRS No. 398, IAEA, Vienna, 2000). 8



However, calibration of passive solid state detectors directly in terms of Dw requires special waterproof holders, appropriate water phantoms and is more time consuming than calibration in air. During WG9 experiments, two approaches to the calibration of TL, OSL and RPL detectors in terms of Dw were applied which are easy to realize and do not require specially designed equipment. In the first approach, the calibration was performed in terms of kerma free-in-air, Kair followed by the application of fcalib, a conversion factor from Kair to Dw. In the second approach, the entire calibration was performed in a PMMA phantom, which can considered to be waterequivalent. Calibration conditions for TL, OSL and RPL dosimeters are presented in Table 3.3. 3.1.3.1 Calibration in terms of kerma in air, Kair For the measurements of the out-of-field doses in radiotherapy, TL dosimeters types TLD-100 and TLD-700 and RPL dosimeters type GD-352M were calibrated in terms of Kair. TL and RPL reference dosimeters belonging to the same batch as sample dosimeters were calibrated against a 60Co source (Table 3.3) at the Secondary Standard Dosimetry Laboratory at the RBI (Vekić et al., 2006).

Table 3.3 Calibration conditions for passive solid states detectors: TLD, RPL and OSL. Detector

MTS-7 60

Calibration source Quantity

Co Dw

Phantom fcalib conversion factor Kair to Dw Reference chamber

TLD-100 TLD-700 60

OSL nanoDotTM

60

Co Kair

60

Co Kair

Co Kair

PMMA

air

air

air

-

1.099 (TLD-100) 1.102 (TLD-700)

1.120

1.11

PTW 30010 Farmer

Field [cm x cm]

RPL (GD-352M)

PTW 30013 Farmer

Primary cavity chambers

10 x 10

10 x 10

10 x 10

10 x 10

80

100

100

100

Source to chamber distance SCD [cm]

In order to express the results of measurements in terms of Dw, conversion factors fcalib from Kair to Dw were determined experimentally for TLD and RPL detectors according to following formula:

f calib =

M air / K air M w / Dw

(3.1)

where:

Mair is the mean value of the signal for a dosimeter sample irradiated in air Mw is the mean value of the signal for a dosimeter sample irradiated in water Kair is the known kerma-in-air Dw is the known absorbed dose to water Irradiations in water were performed according to TRS 398 Code of Practice (TRS No. 398, IAEA, Vienna, 2000) at a depth of 5 cm with a 60Co source at 100 cm SCD and field size 10 x 10 cm2. Once


9


defined, the fcalib conversion factor may be applied multiple times, when calibration is performed according to the above conditions. For OSL dosimeters, an internal calibration of the OSL MicroStarTM reader, performed in air against a 60 Co source at Commissariat à l'Énergie Atomique, LIST CEA-LIST, France, was used. In this case, the correction factor from Kair to Dw was calculated using the ratio of mass energy absorption coefficients for water and air (absorption coefficients were taken from NIST data (NIST, 2008)). In this approach, measurements in a water phantom are not necessary. 3.1.3.2 Calibration in a PMMA phantom TL MTS-7 dosimeters were calibrated in a PMMA phantom (30 x 30 x 15 cm3) with a 60Co source (Table 3.3). Dw was determined with a Farmer type ionization chamber at zref = 5 cm. However, MTS7 dosimeters were calibrated at the depth of maximum dose, zmax = 0.5 cm. The central axis percentage depth-dose (PDD) data from TRS No. 398 (TRS No. 398, IAEA, Vienna, 2000) were used in order to determine Dw at zmax. Discrepancies in the determination of absorbed dose due to density variations of PMMA and to the approximate nature of the procedures for scaling depths and absorbed dose from plastic to water were neglected.

3.1.4 Dose calculations and uncertainties for TL, OSL and RPL dosimeters Luminescence detectors allow only relative dosimetry to be performed, i.e. no absolute dose measurements using the single detector are possible. The absorbed dose from measurements with TL, OSL and RPL detectors is derived by comparison with detectors irradiated with a known dose of radiation. For the dose calculations, fading, nonlinearity of dose response and energy dependence were taken into account (Izewska et al., 2008). The absorbed doses derived from TL and OSL measurements were determined using the following formula:

Dw = N cal ⋅ M ⋅ f fad ⋅ f lin ⋅ f en ⋅ f calib

(3.2)

where:

M is the signal of a sample dosimeter after background subtraction Ncal is the calibration coefficient relating the signal to a known Kair or Dw ffad is the fading correction factor flin is the dose response non-linearity correction factor fen is the energy correction factor fcalib is the conversion factor from Kair to Dw (used only if calibration was performed in terms of Kair) The absorbed dose measured with RPL detectors was calculated according to Equations (3.3), (3.4) and (3.5)

( K a ) accum = M sample ⋅ nc ⋅ H st / mst

(3.3)

( K a ) measured = ( K a ) accum − ( K a )initial

(3.4)

Dw = (( K a ) measured − ( K a ) control ) ⋅ f calib

(3.5) 10



In Equation (3.3):

(Ka)accum is the accumulated dose value nc is the reader correction factor Hst is the dose value of the standard glass mst is the readout value of the standard dosimeter Msample is the readout value for a sample dosimeter. After annealing, all RPL dosimeters were read out, in order to measure the initial dose values (Ka)initial of glass elements (subtracted “pre-dose”).

Dw was calculated according to Equations (3.4) and (3.5) where fcalib is the correction factor from Kair to Dw and Dcontrol is the background dose. In the case of TL and OSL detectors, calibration and sample detectors were irradiated on the same day, so that signal fading could be disregarded. However, for RPL dosimeters, fading is in general negligible (Rah et. al., 2009a). For all detectors used, the nonlinearity correction factor flin was considered to be negligible. For TLDs, the energy correction factor fen was assumed to be insignificant. Also for RPL dosimeters, fen was not applied (in RPL dosimeters type GD-352M an energy compensation filter is built-in). The energy correction factor fen was found to be significant for OSL detectors. The combined relative standard uncertainty uc(Dw) for Dw determined from the TL, OSL or RPL measurements is the square root of the sum of the squared individual relative uncertainties (Kirby et al., 1992; Izewska et al., 2008). The combined uncertainty for TL detector type MTS-7 is uTLD(Dw) = 2.9% for doses varying from 2 mGy to 5 Gy and uTLD(Dw) = 4.2% for doses below 2 mGy (Table 3.4). Five measurements were acquired with the GD-352M dosimeter and the standard deviation (SD) of the mean value of the data set was calculated (Table 3.4), in order to assign an uncertainty figure to RPL dosimeter measured doses. In the calculation of the combined standard uncertainty for GD352M dosimeters, the angular correction factor according to the data from the literature (Son et al., 2011) was taken into account. OSLDs are capable of providing dose estimates with an uncertainty of the order of 0.7–3.2%, depending on the readout equipment and methodology (Yukihara and McKeever, 2008). The total uncertainty estimated by Reft (2009) for nanoDotTM dosimeters is at a level of 4.7% for kilovoltage energies. The uncertainty budget for relative standard uncertainties of OSL is shown in Table 3.5. For the uncertainties due to the detector characteristics, the maximum deviation for a given parameter was found and then without excluding any data, assuming a rectangular distribution, divided by square root of 3 to get the standard uncertainties. Evaluation of the total (combined) standard uncertainty in Table 3.5 is in agreement with Reft (2009).


11


Table 3.4 Relative standard uncertainties (1 SD in %) in the individual components and the combined uncertainty uc(Dw) in the determined dose for TLD (MTS-7) and RPL (GD-352M) dosimeters Relative standard uncertainties Factor M

MTS7 (this work)

GD-352M (this work)

~ 1.7 % (2 mGy – 5 Gy)

~ 0.8 % (1 mGy – 2 Gy)

~ 3.5 % (< 2 mGy)

~ 1.9 % (< 1 mGy)

2.3 %

1.5 %

Ncal ffad

fen

flin

neglected

fang Combined standard uncertainties

-

1.2%

~ 2.9 % (2 mGy – 5 Gy)

~ 2.1 % (1 mGy – 2 Gy)

~ 4.2 % (< 2 mGy)

~ 2.7 % (< 1 mGy)

Table 3.5 Relative standard uncertainties (1 SD in %) in the individual components and the combined standard uncertainty Parameter

OSL

linearity

5/√3

Dose rate

5/√3

Batch

2/√3

Angle

0.4/√3

Energy

See correction factor

Reference – air kerma

0.4

Detector characteristics

Calibration

(conventional true value at the calibration point) 0.45 (radiotherapy doses)

Reading

0.94 (radiation protection doses)

Conversion coefficient from Ka to Dw

0.58

Correction factor (for over response to low energy)

1.16

Total standard uncertainty

4.5%

4.6%

3.1.5 Energy dependence The energy dependence is particularly important when the spectra of the photon radiation fields are difficult to assess, as in the case of the determination of dose to critical organs outside the radiation field in external beam radiotherapy. According to simulations performed by Edwards and Mountford (2004) the energy spectra outside the field edge show two main distinct regions: a broad peak below about 0.5 MeV and a lower amplitude region at higher energies from 0.5 MeV to 12



up to the maximum accelerating voltage. The energy dependence of dosimeters in photon beams is linked to their effective atomic number Zeff (Table 3.1) and the probability of photon interactions. During measurements of out-of-field doses made by EURADOS WG9, the influence of the energy response of LiF TLDs type TLD-100, TLD-700 and MTS-7 was considered negligible and the energy correction factor fen in Equation 3.2 was neglected. For the requirements of the dosimetry of scattered and secondary photons outside the target volume with nanoDotTM dosimeters, the energy dependence was instead considered nonnegligible. fen was calculated using the 2006 PENELOPE Monte Carlo code (Salvat et al., 2006). The configuration of the irradiation was simulated and the model of the entire head of the accelerator was validated, enabling precise calculations to be made of the energy fluence distribution at each point of measurement. Knowing this energy distribution, a correction factor for the energy response was applied at each point of measurement and an uncertainty was calculated for this correction (see Figure 4 in Bordy et al., 2013). Two types of glass rod RPL dosimeters are typically used for compensating the response as a function of interacting photon energy: GD-351 and GD-352M. These detectors contain a holder with a tin filter for energy compensation and are suitable for medium and low energy x-rays that are usually used in diagnostic radiology. The GD-301 and GD-302M types are mainly used for radiotherapy dose measurements and do not contain an energy compensation filter. For the measurements of out-of-field doses in the high energy photon beams in radiotherapy, RPL dosimeters type GD-352M (with a tin filter) were used and they showed good agreement in the out-of-field region with ionization chamber measurements. For dose measurements at the target volume in the high energy photon field, the reading of RPL dosimeter type GD-352M overestimates the dose and should be corrected (Bordy et al., 2013; Miljanić et al., 2013). In the high energy photon field, RPL glasses types GD-301 and GD-302M (without a tin filter), should be used (Mizuno et al., 2008; Rah et al., 2009a).

3.1.6. Dose dependence (linearity) For the measurement of out-of-field doses in radiotherapy, dose response linearity should range from low doses (~ 0.1 mGy) up to radiotherapy treatment doses of a few Gy (~ 2 Gy). The dose response of LiF TLDs depends on impurity composition, ionization density, supralinearity of individual glow peaks, emission spectrum, grain size, physical state, heating rate and annealing parameters (Horowitz, 1981). LiF:Mg,Ti detectors exhibit good linearity up to 2 Gy and supralinearity at higher doses (Gamboa-deBuen et al., 1998; Waligórski et al., 1999). The OSLD dose response depends on experimental parameters: crystal growth (Yukihara et al., 2004), the optical filters used in front of the PMT (Yukihara and McKeever, 2006), the stimulation intensities (initial OSL intensity or total OSL signal) (Yukihara and McKeever, 2008) and the dose history of the dosimeter (Edmund et al., 2006). The dose response of nanoDotTM OSL dosimeters is linear up to 2 Gy, with supralinearity occurring at higher absorbed doses (Reft, 2009). The response of RPL dosimeters has a good linear relationship when compared with the ionization chamber response for doses ranging from 0.5 to 30 Gy (the differences were within ± 2%) (Araki et al., 2004). Also in the low dose range from 0.1 mGy to 500 mGy, RPL dosimeters showed a linear dose response with coefficient of variation 0.6-4.8% (Knežević et. al., 2011). In comparison to TL and OSL dosimeters, RPL dosimeters did not show an over-response for the higher doses.


13


3.2 Neutron dosimetry

3.2.1 Introduction In external beam radiation therapy, neutrons can be produced by the interaction between highenergy photons or electrons and the accelerator structure, the treatment room and the patient. The fast neutron contamination component of radiation therapy clinical fields is well-known and contributes to the unwanted exposure of both operator and patient. Radiation protection of the operator is fulfilled by an appropriate design of the accelerator room, compliant with radiation protection guidelines (NCRP, 1984). On the other hand, the radiotherapy patient still receives an undue, non-negligible, whole body neutron dose at accelerating potentials > 8MV. Neutrons produced in the x-ray target, beam absorber, beam flattening filter, collimators and jaws show an energy distribution similar to fission neutrons, i.e., their spectrum falls in the 1–2 MeV range. Once transmitted through the accelerator head, the intensity of the neutron flux is strongly attenuated. The average energy of the spectrum also decreases, typically below 1 MeV (NCRP, 1984). The lowenergy component of the neutron field is mainly due to scattering reactions with hydrogen either in the accelerator gantry, patient couch, shielding maze or within the patient. A broad polyenergetic neutron field results, with a strong spatial dependence. Dosimetric characterization of the neutron field both within target organs and at a distance from them is one of the main goals of WG9 work, aiming to make available a robust dataset of out-offield dose measurements to be used by both radiotherapy clinics and the research community. The target volume of the radiotherapy treatment receives the highest photon dose. As a consequence, the neutron dose equivalent is also high within the target (> 1 mSv per unit of delivered photon dose, measured in Gy). However, patients are also subjected to a whole-body exposure due to neutron leakage from the accelerator. The development of different accelerator designs and treatment modalities, such as IMRT and VMAT, although greatly improving the conformal dose distribution (Wolff et al., 2009), does not address this problem. Intensity modulated and tomographic techniques precisely conform the photon beam to the target volume, by continuous adjustments of the collimators’ positions. Combinations of multiple fields, coming from different beam directions, produce a patient specific radiation dose that maximizes tumour irradiation, while also minimizing the dose to adjacent normal tissues. These treatments are relatively time inefficient and a large fraction of the primary radiation is blocked, especially when thin fan or pencil beams are used. As a consequence, patients are subjected to a larger whole-body photoneutron exposure due to radiation leakage from the accelerator. By using these new techniques, the problem of neutron contamination of radiation therapy fields, while not being reduced, may have even become more severe. A careful quantification of the unwanted dose delivered by photoneutrons, within the treatment volume and at peripheral organs, is needed for a comparison of different techniques and treatment plans and to find new strategies to minimize it. Photoneutron dosimetry in clinical fields is not an easy task since linear accelerators for radiation therapy produce non-aligned and non-expanded mixed fields. Moreover the intense pulsed photon beam easily saturates the counting electronics of many active neutron detectors, i.e. 3He or 3BF based active detectors. For all these reasons - a broad energy spectrum, a high angular spectral dependence and a strong primary photon component - only selected dosimeters can be used. Their use and readout techniques also need to be suitably assessed for this specific application. Desired characteristics are: photon insensitivity, small size to allow in-phantom dosimetry and high neutron sensitivity to 14



achieve a good accuracy in off-axis measurements, whilst not saturating within the beam. Detector response should also be independent of energy and angular distribution of the impinging neutron spectrum. Superheated emulsions (SE), or “bubble” detectors, and solid state nuclear track etch detectors meet most of these criteria. These devices have thus been selected for the WG9 measurement campaigns. First of all, both kinds of devices are insensitive to photons. They are able to work in a passive mode and provide an integral dose readout, which can be performed at the end of each irradiation. Passive operation allows a response independent of the neutron flux, so that detectors can be placed either within the target volume, where high neutron fluxes can be achieved or at a distance from the beam primary axis, where high detection sensitivity is required (minimum doses in the order of a microsievert). Importantly, no additional external electrical supply is needed during the irradiation. Both superheated emulsions and solid state nuclear track detectors have relatively small size and are thus suitable for in-phantom measurements. The features that make these two kinds of detectors suitable for photoneutron measurements will be discussed in the following sections, together with readout methods used during WG9 campaigns.

3.2.2. Superheated emulsions (SE) Superheated emulsions (Apfel, 1979) are suspensions of metastable halocarbon droplets in an inert gel matrix. Droplets vaporize into visible bubbles upon interactions with neutrons. Through emulsification within a virtually defect-free host medium, droplets are kept in a superheated liquid state, i.e. above their boiling point, but below the critical point. Neutron-induced charged particles generate vapor cavities inside the droplets. This process is irreversible when these cavities reach a critical size, and evaporated droplets form stable vapor bubbles. As the macroscopic effect of neutron interaction is the evaporation of halocarbon emulsion drops, these devices are also known as bubble detectors. These detectors are very suitable for dosimetry within medical linac fields, since their response compares well with the fluence-to-kerma equivalent conversion factor for fast neutrons. 3.2.2.1 Superheated emulsions: performance Two types of SE detectors have been used during WG9 measurement campaigns: superheated drop detectors (SDD) used by the UNIPI, Italy and bubble damage detectors (BDT and BD-PND), produced by Bubble Technology Industries (BTI ®, Ontario, Canada) and provided by SCK*CEN, Mol, Belgium. The main features of SDDs and BTI® detectors are summarized in Table 3.6.


15


Table 3.6 Comparison between SDD-UNIPI and BDP, BD-PND by BTI® SDD Energy range

800 keV – 20 MeV

BD-PND & BDT 200 keV – 15 MeV

Thermal (1/v for epithermal)

Composition Sensitivity (bubbles/μSv)

6

Halocarbon

Chlorofluorocarbon

C-318

R-12

0.2 (±0.03) and 0.5 (±0.08)

3.8 (average*)

0.6 (average*)

R-12 (plus Li compound)

For two detector sets Dose range (μSv)

1-5000 and 1-2000

1 – 5350

1.1 – 112

Readout Uncertainty (1 SD)

5%

2%

2%

Size

L 40 – Φ 18 mm

L 116 Φ 19 mm

L 123 Φ 19 mm

Recompression method

External apparatus

Built-in assembly

Built-in assembly

*average sensitivity of the used batch, single detector measurement uncertainty is ~ 20%

These two technologies are in principle very similar. The detectors have similar shapes and dimensions (Table 3.5). However, the host matrix of the drops is based on an aqueous gel in the SDDs and on a polymeric compound in BDT and BD-PND detectors. BD-PND detectors are sensitive to fast neutrons, while BDT dosimeters contain 6Li, which makes the response function to thermal neutrons resemble the 6Li (n,p) reaction cross section, decreasing as the inverse function of E1/2 with neutron energy, up to the epithermal region. SE technology was first applied to neutron dosimetry in the late 1970s (Apfel, 1979) and nowadays its use is well-established for neutron dosimetry of both patient and personnel. It was included in the ISO standard 21909, among passive personal neutron dosimetry systems. Their working principle is based on the transfer of kinetic energy by the charged secondaries, produced by neutron interactions, to the metastable droplets, causing their phase transition from liquid to vapor. Several studies confirmed the suitability of SDDs for in-phantom dose mapping, in the range between a few 100 keV to about 10 MeV. Monte-Carlo calculated depth dose equivalent distributions for monoenergetic neutrons impinging on a cylindrical tissue-equivalent phantom compare well with those measured using superheated emulsions based on dichlorofluoromethane (D'Errico et al., 2001). The detection principle of superheated emulsions is crucial for their photon insensitivity. Recoil ions generated by neutron interactions inside the emulsion lose their energy by interacting with the surrounding material in a relatively small distance of the order of tenths of microns, at room temperature (d'Errico, 2001). A high linear energy density can thus induce a phase transition, from liquid to gas, of the metastable droplets. The superheated liquid used in SDDs is octafluorocyclobutane (C-318) and its vaporization requires a linear energy transfer (LET) of approximately 300 keV/mm, over the critical radius of the bubble. This energy deposition pattern was shown to be one order of magnitude higher (d'Errico, 2001) than the LET of high energy photoelectrons (>1 keV). Photon insensitivity has also been experimentally proved for both SDD,

16



by using an intense (Waller et al., 2003).

137

Cs source, and BD-PND, irradiated with a 60Co radiotherapy treatment unit

In-phantom measurements also require point-like measurements and detectors of relative small size, compared to the phantom volume. BTI® detectors are slightly longer than SDDs (Table 3.5), nevertheless the diameter of both detectors is less than 2 cm and they are also water tight. They are thus suitable for use in water phantoms. Moreover, their cylindrical geometry guarantees a two-axis measurement isotropy. Two sets of SDDs have been used, with different amounts of C-318, resulting in two kerma equivalent response coefficients of 0.2 (±0.03) and 0.5 (±0.08) bubbles µSv-1. BTI® detector sensitivity is provided for each vial, and on average is 3.8 bubbles µSv-1 and 0.6 bubbles µSv-1 for the BD-PND and BDT used during WG9 campaign, respectively. The accuracy of sensitivity values was estimated during calibration taking into account counting statistics, source calibration uncertainties, detector temperature fluctuations, discrepancies between detector response and kerma factor. The calibration procedure is detailed in the following section. 3.2.2.2 Superheated emulsions: readout methods Integral measurements of the number of bubbles in SDDs can be carried out by at least two different methodologies: by manual or automated visual inspection and by measuring the gel volume increase. An automated readout strategy, relying on the light scattered by evaporated bubbles, has been recently developed (d'Errico et al., 2008). In this configuration, the detector vial is illuminated from the bottom by an infrared (IR) LED and the readout is carried out by three 980 nm focused planar photodiodes (Figure 3.1), recording the amount of light scattered by evaporated bubbles. It has been proved that the scattered component of the IR beam monotonically increases with the number of nucleated bubbles (d'Errico and Di Fulvio, 2011) since vapour bubbles act as scattering centers for IR light. Mono-dispersion of liquid droplets causes the evaporation of bubbles with uniform diameter, about 6 times larger than the droplets. This guarantees a uniform increase of the photodiode readout signal and allows the amount of scattered light to be related to the number of evaporated bubbles. Because of the high sensitivity of the system, which is able to discriminate a single bubble, a readout uncertainty of ~5% (1SD), with a maximum of 100 evaporated bubbles can be achieved.

Figure 3.1 Photodiode signal as a function of irradiation dose (left) and readout prototype holding a detector (right). Red light scattered by bubbles may be noticed.


17


BDT and BD-PND detectors are slightly longer than SDD (Table 3.5) and bubble size is often nonhomogeneous. A different bubble counting method was therefore developed. It is based on image analysis of 2-D images of the detectors, recorded with a Macro objective (on a Canon EOS 1100D camera). Image recognition software was developed in Matlab® (The MathWorks Inc., Natick, MA, 2009), in order to discriminate the bubbles and count them. The software applies a Hough transform to high-pass filtered images. A Hough transform is often used to recognize objects whose boundaries can be described by analytic equations, which is the circumference in this case. The reference equation is thus r2 = (x-a)2 + (y-b)2, where r is the radius and a and b are the circle center coordinates in the x-y plane. For each set of ‘a’, ‘b’ and ‘r’ parameters, a set of circumferences is allowed to exist in the Hough parameter space. Each circumference, i.e. set of parameters, is given a score, i.e. the sum of grey levels of the actual image in the pixels corresponding to the selected circumference. Only those sets whose score is higher than a set threshold are recognized as bubbles, whose radius and center coordinates are the corresponding parameters and are thus known. By setting a suitable threshold, the algorithm can be made insensitive to gaps in the boundaries, as well as to image noise, which is typically due to picture blurring. The counting method is very accurate and the maximum counting error (1 SD) made by the recognition software was 2%, for less than 200 bubbles (Figure 3.2).

Figure 3.2 Software segmentation and counting applied to an actual picture of a BD-PND bubble detector. Discrimination of evaporated bubbles from non-spherical vapour pockets is shown.

3.2.3 Solid state track detectors (radiator degrader neutron spectrometer) Solid state nuclear track detectors are thin slices of plastic materials whose chemical reactivity increases at local damage trails produced by interaction of charged particles, i.e. neutron recoils, with the polymer (Fleischer et al., 1975). These tracks can be recorded and counted by means of standard optical microscopes, after a suitable chemical etching procedure needed to enlarge the tracks. CR-39 or allyl-diglycol-carbonate is the most sensitive nuclear track detector for protons and it proves to be an effective neutron dosimeter because of its low energy threshold for chain disruption, i.e. track initiation.

18



3.2.3.1 Solid state track detectors: performance Solid state track detectors for neutrons developed at UAB, Spain and at the PoliMI, Italy, were used during WG9 in-phantom measurements for the assessment of neutron dose equivalent delivered during radiation therapy treatments. The main features of track detectors by UAB and PoliMi are summarized in Table 3.7. These two detectors, although being based on the same working principles, feature different configurations and etching strategies. As clarified in this section, the etching procedure is not only crucial to make the readout possible but also unavoidably affects the detectors’ performance. Table 3.7 Comparison between track etch detectors from the UAB, Spain and PoliMi, Italy. PADC - UAB

RDNS- PoliMi

Energy range

10 eV – 10 MeV

200 keV – 20 MeV

Convertors

Polyethylene,

Polyethylene, aluminum

(layers from top to bottom)

Makrofol polycarbonate, Polyamide nylon

Detector

PADC

PADC

10 – 2 10

1.2 10-4 – 9.9 10-4*

Minimum detectable dose

50 μSv

2.3 mSv

Readout Accuracy (1 SD)

10%

Sensitivity

-5

-4

(track density / neutron fluence) 10% 3

Size of complete dosimeter

20x20x9 mm

Etching method

Electrochemical

25x25x3.4 mm3 Chemical

*sensitivity depends on the combination of PE/aluminum layers used.

The energy transfer from ionizing charged particles to dielectric materials results in the formation of a trail of damaged molecules along the particle track. Among plastic solid state material, PADC requires a relatively low irradiation dose per chain disruption. Defects generated in PADC remain permanently on the plastic foil. A minimum value of energy loss (or linear energy transfer LET, dE/dx) is required for the damage to be detectable. Importantly, this threshold is well above the LET for electron tracks. For this reason, track detectors, like superheated emulsions, are intrinsically insensitive to high kinetic energy, low LET radiation, such as fast electrons and photons. As previously stated, in-phantom measurements require point-like measurements and detectors of relative small size. Both PADC based detectors developed at UAB and PoliMi comply with this constraint. The radiator degrader neutron spectrometer by PoliMi consists of a 25x25 mm2 PADC sheet, 1.4 mm thick, covered by an aluminum degrader foil with a radiator PE clad layer, reaching an overall thickness of 3.4 mm. UAB dosimeters consist of several layers: polyethylene, Makrofold polycarbonate, Polyamide nylon and PADC, from top to bottom. These layers are piled up on a methacrylate holder, 5 mm thick. Overall dimensions of the detector are 20x20x8.9 mm3. A well-known drawback of flat track dosimeters is their angular response. The response to thermal and epithermal neutrons typically compares well with dose equivalent. At higher energies, measured data underestimate actual dose equivalent (Luszik-Bhadra et al., 1990). This


19


underestimation is still an open issue. Its compensation may be carried out by deriving the angleenergy distribution of the impinging radiation, e.g. from the shape of the etched tracks, and subsequently applying proper weighting factors accounting for particles with different interaction angles. A flat energy response, in the widest possible neutron energy range, is typically achieved in track etch detectors by a coupling PADC with hydrogenous converters (fast neutron response) and nitrogen-containing materials (slow neutron response, through the 14N(n,p)14C reaction). This approach has been exploited by PADC developed by UAB. The response function of these detectors matches well the fluence to dose equivalent conversion coefficient in the fast neutron energy range. However, it overestimates the dose equivalent conversion coefficient in the thermal range, where it follows the 14N(n,p)14C reaction cross section (Domingo et al., 2013). Therefore PADC detectors are suitable for in-phantom photoneutron measurements assuming that their readout is weighted by an appropriate position dependent factor, which accounts for the actual impinging neutron spectrum on the detectors. Their response in terms of fluence of the UAB dosimeter varies from about 10-5 tracks per neutron at ~70 keV up to a maximum of 2 10-4 tracks per neutron at 2 MeV and practically vanishes at energies >20 MeV. The PoliMi PADC slide is covered by an aluminum degrader foil and a PE radiator clad layer, whose thicknesses vary in the range of 10-500 µm and 0.2 – 2 mm, respectively (Caresana et al., 2012). Each combination of thicknesses provides a different response function and a PADC detector can be thus used as a spectrometer in this particular configuration. A radiator degrader neutron spectrometer features a sensitivity in the range of 1.2 10-4 – 9.9 10-4 (track density per unit neutron fluence, Table 3.6), depending on the combination of PE/aluminum layers used, for an irradiation of the RDNS with a Pu–Be neutron source (Caresana et al., 2012b). 3.2.3.2 Solid state track detectors: etching and readout methods Tracks created after interactions of PADC with ionizing charged particles are more sensitive to the chemical etching than the non-damage PADC polymer chains. These tracks are actually local shallow damage, called latent-damage tracks (Fleischer et al., 1975). Etching procedures enlarge latent-damage tracks and make them visible to the standard optical microscope and easy to count. The etching procedure is thus a crucial step of the detector’s readout. Two types of etching are typically used, the electrochemical and the chemical. In both cases the PADC is submerged in a strong acid solution, whose temperature is controlled up to 100°C. The main difference between the two is whether an alternating voltage is applied across the detector between two platinum electrodes (electrochemical etching) or not (chemical etching). In electrochemical etching, used at UAB, the alternating electric field causes a tree shape track propagation into large tracks, mainly independent of the original depth of the track and thus on the energy of charged secondaries. This procedure does not allow energy discrimination, but it guarantees that low LET interacting charged particles may also be counted, because shallower tracks created by low-energy protons ( 135 mm. The positions of the dosimeters in the pipes were fixed by PMMA spacers. The term “frame” is defined as a set of pipe positions at a specified depth and at various z positions, as illustrated in Figure 4.2.

4.3 Reference absorbed dose values The reference absorbed dose values were measured with an ionization chamber (Nuclear Enterprise Type NE 2571). All these measurements were traceable to French national references in terms of absorbed dose to water for radiotherapy and were established by calorimetry in water and graphite for beam sizes ranging between 10 cm x 10 cm and 2 cm x 2 cm under standard conditions defined in IAEA TRS 398 protocol (IAEA, 2000). Three radiation qualities were used, 6, 12 and 20 MV, corresponding to a tissue-phantom ratio for 10 cm and 20 cm depth in a water phantom (TPR20,10) of 0.678, 0.751 and 0.784 respectively. The duration of irradiation, approximately 1 min, delivered a dose of 2 Gy at the isocentre reference point, for all qualities and sizes of the radiation field.


31


4.4 Correction of raw results A description of the calibration conditions for the TLD, OSL and RPL dosimeters has been given in section 3.1.3. The in-phantom photon spectra are progressively softened with increasing distance from the isocentre because of multiple scattering events. This means that the energy response of dosimeters must be considered and correction factors need to be employed if necessary. The energy response curve for OSL was measured at LNHB. The principle of the correction, as described below, could also be applied to other dosimeters. The energy spectra of the photons were calculated using the Monte Carlo code PENELOPE (Salvat et al., 2001) at each point of measurement. Then, the energy response curve of the detector in terms of absorbed dose to water, measured at LNHB, was convoluted with the distribution of the energy fluence in order to determine the correction factor (1/k) to be applied to the calibration coefficient for OSL dosimeters at each measurement point (Bordy et al 2013).

1 1 � 𝑅(𝐸 ). Ψ(𝐸 ). 𝑑𝐸 = 𝑘 Ψ𝑡𝑜𝑡 𝐸

(4.1)

Where Ψtot is the total energy fluence of the spectrum, R(E) is the response for energy E normalised to the 60Co response and Ψ(E) is the energy.

References Bordy, J.M., Bessiere, I., d'Agostino, E., Domingo, C., d'Errico, F., di Fulvio, A., Knežević, Ž., Miljanić, S., Olko, P., Ostrosky, A., Poumarede, B., Sorel, S., Stolarczyk, L., Vermersse D., Harrison, R., 2013. Radiotherapy out-of-field dosimetry: Experimental and computational results for photons in a water tank. Radiat. Meas. 57, 29-34. Salvat, F., Fernanez-Varea, J.M., Acosta, E., Sempau, J., 2001 PENELOPE – A Code System for Monte Carlo simulation of Electron and Photon Transport. Issy-les-Moulineaux: OECD Nuclear Energy Agency (available in pdf format at http://www.nea.fr TRS 398, 2000. Absorbed Dose Determination in External Beam Radiotherapy: An International Code of Practice for Dosimetry Based on Standards of Absorbed Dose to Water. IAEA, Vienna. ISSN 1011–4289.

32



5. Results: passive dosimeters in a water tank A selection of tabulated results for out-of-field dose profiles at various depths and beam qualities is given in Table 5.1 (OSL), Table 5.2 (TLD (MTS-7)), Table 5.3 (RPL(GD-352M)) and Table 5.4 (TLD-100 and TLD-700). From these data, graphical plots illustrating features of the results are given in Figures 5.1 – 5.5. Data in all plots and tables refer to an absorbed dose of 2 Gy delivered to the isocentre. Figure 5.1 shows, for OSL at 12MV, the close correspondence of out-of-field doses at various depths as a function of lateral distance. At lateral distances > 400 mm, lower doses are evident as a function of increasing depth. Figure 5.2 shows results at 6, 12 and 20 MV for OSL at a depth of 100 mm with the expected dependence on beam quality. Figures 5.3, 5.4 and 5.5 compare the dosimeters used at 6, 12 and 20 MV at a depth of 100 mm. The main observations are a small dose over-estimation of OSL at larger lateral distances and overestimation of the in-field dose when using RPL. These will be discussed in chapter 6. Table 5.5 summarises results from the PENELOPE Monte Carlo simulation and shows the percentage of the fluence and energy fluence in the beam at up to 40 cm from the beam axis for a 12 MV beam.

Figure 5.1 Out-of-field dose profiles for 100, 150, 200 and 250 mm depth, using OSL at 12 MV.


33


Figure 5.2 Out-of-field dose profiles for 6, 12 and 20 MV at 100 mm using OSL.

Figure 5.3 Out-of-field dose profiles for all dosimeters (OSL, TLD, RPL and ionization chamber reference) at 100 mm depth and 6MV.

34



Figure 5.4 Out-of-field dose profiles for all dosimeters (OSL, TLD, RPL and ionization chamber reference) at 100 mm depth and 12 MV.

Figure 5.5 Out-of-field dose profiles for all dosimeters (OSL, TLD, RPL and ionization chamber reference) at 100 mm depth and 20 MV.


35


Table 5.1 (a): Water tank results for OSL (nano Dot, Landauer), for four depths and three beam energies. The isocentre dose (d = 100 mm; D = 135 mm; z = 0 mm) is 2 Gy. The lateral distance (D) is measured from the inner face of the water tank at the level of the central axis. In this table, D = 25 – 165 mm.

OSL Depth (d) (mm)→

6MV

12 MV

20 MV

100

150

200

250

100

150

200

250

100

150

200

250

25

4.750E-02

5.464E-02

5.153E-02

4.822E-02

4.092E-02

4.377E-02

4.384E-02

4.062E-02

3.138E-02

3.244E-02

3.423E-02

3.407E-02

45

7.402E-02

8.054E-02

8.340E-02

7.083E-02

6.415E-02

6.393E-02

6.441E-02

5.979E-02

5.055E-02

5.457E-02

5.470E-02

5.287E-02

65

1.388E-01

1.398E-01

1.335E-01

1.098E-01

1.139E-01

1.144E-01

1.115E-01

9.797E-02

1.004E-01

1.036E-01

1.008E-01

1.061E-01

75

1.932E-01

2.084E-01

2.089E-01

2.175E-01

1.769E-01

1.849E-01

1.877E-01

2.087E-01

1.743E-01

1.917E-01

2.077E-01

2.400E-01

80

2.827E-01

3.685E-01

6.284E-01

6.397E-01

2.706E-01

2.843E-01

4.179E-01

6.084E-01

2.842E-01

3.426E-01

4.737E-01

6.747E-01

85

9.086E-01

1.113E+00

9.859E-01

7.579E-01

7.175E-01

9.910E-01

1.011E+00

7.689E-01

7.350E-01

1.063E+00

1.010E+00

8.765E-01

Lateral dist. (D) (mm)↓

90

1.828E+00

1.411E+00

1.018E+00

7.540E-01

1.767E+00

1.365E+00

1.119E+00

8.348E-01

1.648E+00

1.305E+00

1.127E+00

8.709E-01

105

1.833E+00

1.409E+00

1.061E+00

8.436E-01

1.954E+00

1.527E+00

1.212E+00

9.112E-01

1.875E+00

1.498E+00

1.187E+00

9.519E-01

120

1.966E+00

1.474E+00

1.099E+00

8.441E-01

1.953E+00

1.550E+00

1.193E+00

9.254E-01

1.878E+00

1.501E+00

1.199E+00

9.559E-01

135

1.892E+00

1.484E+00

1.088E+00

8.586E-01

1.880E+00

1.533E+00

1.183E+00

8.968E-01

1.971E+00

1.519E+00

1.200E+00

9.773E-01

150

1.932E+00

1.478E+00

1.094E+00

8.388E-01

1.963E+00

1.510E+00

1.217E+00

8.957E-01

1.901E+00

1.513E+00

1.222E+00

9.937E-01

165

1.836E+00

1.416E+00

1.046E+00

8.097E-01

1.948E+00

1.516E+00

1.209E+00

8.807E-01

1.891E+00

1.490E+00

1.188E+00

9.283E-01

36



Table 5.1 (b): Water tank results for OSL (nano Dot, Landauer), for four depths and three beam energies. The isocentre dose (d = 100 mm; D = 135 mm; z = 0 mm) is 2 Gy. The lateral distance (D) is measured from the inner face of the water tank at the level of the central axis. In this table, D = 180 – 561 mm.

OSL Depth (d) (mm)→

6MV

12 MV

20 MV

100

150

200

250

100

150

200

250

100

150

200

250

180

1.602E+00

1.288E+00

1.004E+00

7.750E-01

1.725E+00

1.462E+00

1.104E+00

8.481E-01

1.632E+00

1.304E+00

1.080E+00

9.105E-01

185

5.912E-01

7.900E-01

8.663E-01

7.202E-01

8.813E-01

1.227E+00

1.065E+00

8.242E-01

8.127E-01

1.115E+00

9.803E-01

8.481E-01

190

2.440E-01

2.493E-01

3.030E-01

4.687E-01

2.799E-01

4.078E-01

5.517E-01

6.620E-01

2.919E-01

3.638E-01

4.358E-01

6.339E-01

195

1.874E-01

1.720E-01

1.820E-01

1.650E-01

1.841E-01

2.029E-01

1.965E-01

2.391E-01

1.727E-01

2.013E-01

2.126E-01

2.306E-01

205

1.314E-01

1.374E-01

1.285E-01

1.117E-01

1.139E-01

1.194E-01

1.154E-01

1.030E-01

1.013E-01

1.064E-01

1.001E-01

1.024E-01

225

7.858E-02

7.887E-02

8.150E-02

7.008E-02

6.530E-02

6.955E-02

6.874E-02

6.560E-02

5.127E-02

5.881E-02

5.852E-02

5.325E-02

245

4.812E-02

5.207E-02

5.357E-02

4.567E-02

4.225E-02

4.274E-02

4.370E-02

4.129E-02

3.378E-02

3.371E-02

3.342E-02

3.330E-02

275

2.888E-02

3.029E-02

3.081E-02

2.906E-02

2.461E-02

2.611E-02

2.457E-02

2.325E-02

1.846E-02

2.051E-02

1.918E-02

1.752E-02

315

1.709E-02

1.792E-02

1.746E-02

1.585E-02

1.499E-02

1.469E-02

1.456E-02

1.254E-02

1.183E-02

1.131E-02

1.097E-02

9.910E-03

365

8.645E-03

9.253E-03

9.267E-03

9.095E-03

7.394E-03

8.199E-03

7.876E-03

7.926E-03

5.937E-03

6.864E-03

6.369E-03

5.951E-03

415

5.447E-03

5.603E-03

5.261E-03

4.411E-03

4.969E-03

4.970E-03

4.477E-03

3.726E-03

4.040E-03

3.806E-03

3.420E-03

2.987E-03

465

3.317E-03

3.340E-03

3.286E-03

2.954E-03

3.438E-03

3.088E-03

2.766E-03

2.448E-03

2.612E-03

2.384E-03

2.266E-03

2.113E-03

515

2.189E-03

2.169E-03

2.139E-03

1.918E-03

2.559E-03

2.441E-03

1.996E-03

1.874E-03

1.862E-03

1.771E-03

1.636E-03

1.412E-03

561

1.860E-03

1.746E-03

1.739E-03

1.433E-03

2.026E-03

1.912E-03

1.734E-03

1.551E-03

1.506E-03

1.377E-03

1.216E-03

1.042E-03



37


Table 5.2: Water tank results for TLD (MTS-7) (IFJ), for four depths and three beam energies. The isocentre dose (d = 100 mm; D = 135 mm; z = 0 mm) is 2 Gy. The lateral distance (D) is measured from the inner face of the water tank at the level of the central axis. In this table, D = 135 – 561 mm

TLD (MTS-7)

6MV

Depth (mm)→

100

150

200

250

100

150

200

250

100

150

200

250

135

2.001E+00

1.500E+00

1.147E+00

8.522E-01

2.017E+00

1.505E+00

1.225E+00

9.476E-01

1.968E+00

1.525E+00

1.253E+00

1.042E+00

150

1.964E+00

1.522E+00

1.148E+00

8.535E-01

2.004E+00

1.565E+00

1.217E+00

9.448E-01

1.978E+00

1.503E+00

1.216E+00

1.018E+00

165

1.931E+00

1.490E+00

1.130E+00

8.363E-01

1.967E+00

1.552E+00

1.202E+00

9.536E-01

1.922E+00

1.473E+00

1.201E+00

9.998E-01

180

1.576E+00

1.372E+00

1.034E+00

7.911E-01

1.820E+00

1.428E+00

1.170E+00

9.133E-01

1.674E+00

1.343E+00

1.126E+00

9.512E-01

185

3.843E-01

7.011E-01

8.658E-01

7.546E-01

7.877E-01

1.170E+00

1.083E+00

8.831E-01

6.837E-01

1.119E+00

1.030E+00

8.995E-01

190

2.186E-01

2.326E-01

2.670E-01

4.422E-01

2.539E-01

3.020E-01

4.935E-01

6.877E-01

2.632E-01

3.119E-01

4.686E-01

6.805E-01

195

1.769E-01

1.777E-01

1.646E-01

1.619E-01

1.736E-01

1.776E-01

1.936E-01

2.091E-01

1.642E-01

1.773E-01

1.940E-01

2.303E-01

205

1.252E-01

1.300E-01

1.222E-01

1.067E-01

1.128E-01

1.151E-01

1.135E-01

1.038E-01

9.805E-02

9.908E-02

1.021E-01

8.697E-02

225

6.920E-02

7.640E-02

7.353E-02

6.565E-02

6.044E-02

6.605E-02

6.591E-02

6.122E-02

5.118E-02

5.380E-02

5.340E-02

5.363E-02

245

4.269E-02

4.700E-02

4.726E-02

4.373E-02

3.774E-02

4.055E-02

4.171E-02

3.920E-02

3.035E-02

3.194E-02

3.269E-02

3.297E-02

275

2.576E-02

2.743E-02

2.715E-02

2.518E-02

2.166E-02

2.306E-02

2.285E-02

2.181E-02

1.738E-02

1.760E-02

1.785E-02

1.773E-02

315

1.459E-02

1.515E-02

1.467E-02

1.378E-02

1.316E-02

1.291E-02

1.269E-02

1.199E-02

1.059E-02

9.836E-03

9.776E-03

9.748E-03

365

6.875E-03

7.614E-03

7.600E-03

7.172E-03

6.426E-03

6.868E-03

6.989E-03

6.649E-03

5.202E-03

5.228E-03

5.333E-03

5.283E-03

415

3.993E-03

4.220E-03

4.003E-03

3.563E-03

4.010E-03

3.861E-03

3.575E-03

3.252E-03

3.259E-03

2.932E-03

2.672E-03

2.481E-03

465

2.485E-03

2.454E-03

2.290E-03

2.056E-03

2.610E-03

2.316E-03

2.141E-03

1.999E-03

2.016E-03

1.744E-03

1.589E-03

1.524E-03

515

1.633E-03

1.557E-03

1.450E-03

1.295E-03

1.803E-03

1.612E-03

1.527E-03

1.373E-03

1.404E-03

1.150E-03

1.065E-03

9.578E-04

561

1.274E-03

1.167E-03

1.055E-03

9.669E-04

1.493E-03

1.335E-03

1.246E-03

1.007E-03

1.073E-03

8.419E-04

8.021E-04

7.038E-04

(d)

12 MV

20 MV


38



Table 5.3: Water tank results for RPL (GD-352M, Asahi Techno Glass Corp) (RBI), for four depths and three beam energies. The isocentre dose (d = 100 mm; D = 135 mm; z = 0 mm) is 2 Gy. The lateral distance (D) is measured from the inner face of the water tank at the level of the central axis. In this table, D = 135 – 561 mm.

RPL Depth (mm)→

6MV (d)

12 MV

20 MV

100

150

200

250

100

150

200

250

100

150

200

250

135

2.267E+00

1.673E+00

1.285E+00

9.309E-01

2.463E+00

1.878E+00

1.500E+00

1.198E+00

2.546E+00

1.954E+00

1.626E+00

1.310E+00

150

2.157E+00

1.649E+00

1.282E+00

9.325E-01

2.426E+00

1.910E+00

1.527E+00

1.203E+00

2.468E+00

1.951E+00

1.611E+00

1.300E+00

165

2.200E+00

1.603E+00

1.231E+00

9.405E-01

2.443E+00

1.889E+00

1.529E+00

1.197E+00

2.434E+00

1.958E+00

1.592E+00

1.303E+00

180

1.980E+00

1.557E+00

1.198E+00

9.045E-01

2.340E+00

1.848E+00

1.488E+00

1.183E+00

2.246E+00

1.814E+00

1.540E+00

1.242E+00

185

4.300E-01

8.593E-01

9.957E-01

8.515E-01

9.853E-01

1.706E+00

1.429E+00

1.178E+00

8.061E-01

1.637E+00

1.490E+00

1.252E+00

190

2.233E-01

2.322E-01

2.532E-01

4.211E-01

2.629E-01

3.340E-01

6.392E-01

9.849E-01

2.658E-01

3.392E-01

6.755E-01

9.999E-01

195

1.845E-01

1.813E-01

1.683E-01

1.541E-01

1.808E-01

1.973E-01

2.020E-01

2.341E-01

1.724E-01

1.883E-01

2.166E-01

2.499E-01

205

1.290E-01

1.338E-01

1.230E-01

1.139E-01

1.223E-01

1.281E-01

1.261E-01

1.191E-01

1.074E-01

1.138E-01

1.184E-01

1.127E-01

225

6.933E-02

7.393E-02

7.278E-02

6.693E-02

6.403E-02

6.952E-02

7.040E-02

6.756E-02

5.534E-02

5.798E-02

5.911E-02

5.779E-02

245

4.310E-02

4.593E-02

4.618E-02

4.365E-02

3.936E-02

4.263E-02

4.366E-02

4.273E-02

3.280E-02

3.511E-02

3.759E-02

3.601E-02

275

2.490E-02

2.700E-02

2.654E-02

2.519E-02

2.306E-02

2.412E-02

2.365E-02

2.336E-02

1.882E-02

1.963E-02

1.969E-02

1.924E-02

315

1.422E-02

1.478E-02

1.458E-02

1.399E-02

1.383E-02

1.353E-02

1.363E-02

1.329E-02

1.158E-02

1.139E-02

1.101E-02

1.033E-02

365

6.704E-03

7.372E-03

7.676E-03

7.366E-03

6.541E-03

7.185E-03

7.298E-03

7.306E-03

5.266E-03

5.912E-03

6.151E-03

5.804E-03

415

3.910E-03

4.114E-03

3.892E-03

3.569E-03

4.211E-03

4.053E-03

3.751E-03

3.428E-03

3.387E-03

3.328E-03

3.079E-03

2.750E-03

465

2.479E-03

2.442E-03

2.348E-03

2.109E-03

2.610E-03

2.472E-03

2.297E-03

2.194E-03

2.057E-03

1.976E-03

1.818E-03

1.672E-03

515

1.771E-03

1.643E-03

1.537E-03

1.366E-03

1.971E-03

1.796E-03

1.632E-03

1.508E-03

1.459E-03

1.373E-03

1.278E-03

1.135E-03

561

1.415E-03

1.359E-03

1.183E-03

1.058E-03

1.691E-03

1.544E-03

1.422E-03

1.280E-03

1.164E-03

1.075E-03

1.000E-03

8.669E-04



39


Table 5.4: Water tank results for TLD-100 and TLD-700 (RBI), for four depths and three beam energies. The isocentre dose (d = 100 mm; D = 135 mm; z = 0 mm) is 2 Gy. The lateral distance (D) is measured from the inner face of the water tank at the level of the central axis. In this table, D = 135 – 561 mm.

TLD

6MV

TLD-100

12 MV

TLD-700

100

150

200

250

100

150

200

250

100

150

200

250

135

1.979E+00

1.473E+00

1.103E+00

8.252E-01

1.988E+00

1.556E+00

1.228E+00

9.517E-01

2.007E+00

1.587E+00

1.292E+00

1.045E+00

150

1.940E+00

1.486E+00

1.113E+00

8.197E-01

2.008E+00

1.563E+00

1.214E+00

9.608E-01

1.996E+00

1.562E+00

1.271E+00

1.023E+00

165 180

1.879E+00

1.427E+00

1.072E+00

8.029E-01

1.991E+00

1.569E+00

1.189E+00

9.402E-01

1.922E+00

1.471E+00

1.233E+00

9.521E-01

1.453E+00

1.293E+00

9.983E-01

7.510E-01

1.753E+00

1.454E+00

1.146E+00

8.984E-01

1.656E+00

1.396E+00

1.070E+00

9.325E-01

185

3.353E-01

5.837E-01

7.762E-01

7.118E-01

5.548E-01

1.040E+00

1.043E+00

8.451E-01

5.805E-01

1.090E+00

1.054E+00

8.748E-01

190

2.089E-01

2.213E-01

2.427E-01

3.110E-01

2.317E-01

2.685E-01

-

5.831E-01

2.530E-01

3.152E-01

4.623E-01

6.022E-01

195

1.677E-01

1.711E-01

1.563E-01

1.414E-01

-

1.678E-01

1.750E-01

1.870E-01

1.588E-01

1.707E-01

1.994E-01

1.870E-01

205

1.194E-01

1.245E-01

1.156E-01

1.017E-01

-

1.120E-01

1.109E-01

1.036E-01

9.516E-02

1.012E-01

1.037E-01

1.005E-01

225

6.686E-02

7.311E-02

7.052E-02

-

5.976E-02

6.469E-02

6.450E-02

6.069E-02

4.973E-02

5.380E-02

5.505E-02

5.167E-02

245

4.246E-02

4.630E-02

4.586E-02

4.186E-02

3.757E-02

4.035E-02

4.150E-02

3.942E-02

3.055E-02

3.286E-02

3.471E-02

3.305E-02

275

2.476E-02

2.658E-02

2.722E-02

2.454E-02

2.218E-02

2.279E-02

2.336E-02

2.179E-02

1.770E-02

1.897E-02

1.797E-02

1.794E-02

315

1.458E-02

1.477E-02

1.492E-02

1.360E-02

1.308E-02

1.312E-02

1.273E-02

1.214E-02

1.096E-02

1.072E-02

1.050E-02

9.800E-03

365

7.053E-03

7.555E-03

8.042E-03

7.348E-03

6.489E-03

7.023E-03

7.191E-03

6.730E-03

5.397E-03

5.721E-03

5.887E-03

5.449E-03

415

4.174E-03

4.224E-03

4.082E-03

3.771E-03

4.083E-03

4.143E-03

3.710E-03

3.297E-03

3.405E-03

3.281E-03

2.946E-03

2.692E-03

465

2.643E-03

2.491E-03

2.460E-03

2.204E-03

2.619E-03

2.511E-03

2.280E-03

2.020E-03

2.202E-03

1.940E-03

1.819E-03

1.606E-03

515

1.735E-03

1.654E-03

1.676E-03

1.431E-03

1.870E-03

1.773E-03

1.574E-03

1.361E-03

1.566E-03

-

1.222E-03

9.896E-04

561

-

1.264E-03

1.402E-03

1.123E-03

1.595E-03

1.459E-03

1.271E-03

1.134E-03

1.191E-03

1.032E-03

-

7.471E-04

Depth (mm)→

(d)

20 MV

TLD-700


40



Table 5.5 Percentage of the fluence and energy fluence for 12 MV in-beam and out-ofbeam up to 400 mm from the beam axis, following Monte Carlo simulation of beam interactions using the PENELOPE code.

% fluence

In-beam

Out-of-beam, < 40 cm from the reference point (isocentre)

~ 15%

< 75%

< 0.7%

< 45%

2.40 MeV

>150 keV

(in the range 0-200 keV) % energy fluence (in the range 0-200 keV) Average energy


41


6. Discussion: water tank results Compared with the ionization chamber measurements, the deviation for the TLDs at the isocentre is less than 1.5%. Readout of OSL dosimeters, after applying the proper correction factor, consistently underestimates the ionization chamber measurement by about 4%. For the RPL dosimeter (type GD-352M) systematic overestimations were encountered for all the radiation qualities used in this work (6, 12 and 20 MV). The overestimations were 13%, 23% and 27% for 6, 12 and 20 MV respectively. For 20 MV, the overestimation was 27.5%, but lower for 6 and 12 MV. This latter effect can be attributed to the influence of the tin cap covering the RPL detector, which introduces a high-atomic number nuclide with which the photons interact to generate secondary electrons, thus depositing additional energy in the RPL material. A correction similar to the one provided for OSL could be calculated if this dosimeter is used for measurements inside the beam without a dedicated calibration. Since the objective is to measure out-of-field doses, corrections for this over-response have not been made as part of this work. Taking into account that the standard uncertainty of the ionization chamber measurement is estimated at ±1.2% and those of passive dosimeters being roughly estimated up to 2.5% for high doses, it can be concluded that TLDs and OSLs allow a reliable measurement of the dose at the isocentre. On the edge of the beam, where a steep variation of dose is seen, the comparison between the dosimeters is meaningless because the sizes of the dosimeters are different so that they integrate different parts of the profile. The absorbed dose out of the beam for a given depth decreases with distance from the beam axis to reach, at a distance of 400 mm, much less than 1% of the absorbed dose at the isocentre (e.g. Figure 5.2). This decrease can be attributed to the absorption of the scattered component of the radiation field in the water as the distance from the beam axis increases. At this stage, it is not possible to distinguish separately the out-of-field doses due to scatter and leakage radiation. Nevertheless, the comparison of absorbed dose as a function of the radiation quality normalised to the same dose at the isocentre shows that out-of-field doses decrease when the incident energy increases (Figure 5.2). Figures 5.3-5.5 highlight the very good agreement between the reference ionization chamber measurements and the TLD measurements in and out of the beam for a given radiation quality and at a given depth. Taking into account the shape and the volume of the dosimeters and the measurement conditions inside the water phantom where the fluence is increasingly isotropic with increasing distance of the dosimeter from the isocentre, it was assumed that the influence of the angular dependence of the dosimeter is negligible. The over-response of OSL dosimeters for energies lower than 200 keV, leads to an overestimation of the absorbed dose in the scatter region (Figure 5.3-5.5) whilst a good agreement is found between TLD, RPL and ionization chamber measurements. This over response of OSL is larger as the distance from the beam axis increases. Table 5.5 (reproduced from Bordy et al. (2013)) provides data, calculated using the Monte Carlo code PENELOPE, on the proportion of the fluence and the energy fluence for energies lower than 200 keV. It is noted that this part of the spectrum is dominant far from the beam axis, explaining why it is necessary to correct the OLS raw results. After correction, OSL results are closer to those obtained with ionization chambers. Nevertheless the spread of the results after applying the correction factor to OSL dosimeters can be larger, in the scatter region, than for the other dosimeters which do not require the use of a correction factor. Indeed, being based on calculations, the correction factor value relies on the precision of the linac 42



head model for calculations. Some discrepancies can exist between the theoretical model and the actual experimental set up which can lead to less accurate correction factor values, especially far from the beam axis. Finally, ionization chamber measurements confirm the appropriateness of the calibration procedure of TLD based dosimeters, for all the measurement positions, and allow the correction of the OSL calibrations in low energy regions far from the beam axis.

Reference Bordy, J.M., Bessiere, I., d'Agostino, E., Domingo, C., d'Errico, F., di Fulvio, A., Knežević, Ž., Miljanić, S., Olko, P., Ostrosky, A., Poumarede, B., Sorel, S., Stolarczyk, L., Vermersse D. 2013. Radiotherapy outof-field dosimetry: Experimental and computational results for photons in a water tank. Radiat. Meas. 57, 29-34.


43


7. Measurements: BOMAB phantom 7.1 BOMAB phantom design The original BOMAB phantom represents an adult human body (170 cm tall) that comprises ten cylinders. Three cylinders with elliptic cross section simulate the trunk, head and pelvis; five circular cross section cylinders represent the human neck, arms and legs. The original BOMAB phantom was first used in the 1950s (Bush, 1949) and is currently the industry standard for calibrating whole body counting systems (HPS, 1995) in North America. Cylinders are filled with a tissue equivalent liquid, in which radionuclides can be internally deposited. Although the main application of the BOMAB phantom is for the testing of whole body counting facilities, it is also used in nuclear medicine and external beam radiotherapy. A modified version of the original BOMAB phantom was adopted for this study and developed at the UNIPI. This modified phantom is henceforth referred to as the BOMAB phantom in this report. This version has fewer cylinders than the original design and the trunk is simulated by a cylinder with elliptical cross section which comprises 15 pipes, through the whole length, to host dosimeters. Its overall dimensions comply with requirements of Reference Man (ICRP, 1975). A comparison between the two versions is shown in Table 7.1. The phantom is made of PMMA, with the exception of trunk flanges, which are polycarbonate. Cylinders are filled with water. These materials are considered sufficiently tissue equivalent over the range of photon energies normally encountered in external photon radiation therapy. A series of PMMA channels (“pipes”) were permanently inserted in the trunk, spaced by 5 cm, in a rectangular lattice geometry. Unlike commercially available versions of the phantom (Nuclear Technology Services, 2014), this BOMAB phantom is transparent (Figure 7.1), allowing visual inspection of dosimeters placed inside the pipes. It can thus be considered, in term of size and materials, as an intermediate design between a water tank and an anthropomorphic phantom. Table 7.1: Dimensional comparison between ANSI standard BOMAB and WG9 BOMAB phantoms. WG9 BOMAB (cm x cm)

ANSI BOMAB (cm x cm)

diameter x length

diameter x length

Head

20 x 30

Diam. 19-14 x 20*

Neck

-

13 x 10

Thorax

30-22 x 60

Diam. 30-20 x 40*

Arms

20 x 40

10.5 x 60

Lumbar

-

Diam. 36-20 x 60*

Thighs

-

16 x 40

Legs

25 x 80

13 x 40

*elliptic cross section: major axis – minor axis x length

44



Figure 7.1: (a) The BOMAB phantom before irradiation. Bi-component glue was used to attach inner PMMA tubes (orange/yellow in the picture) at the bottom flange of the trunk. (b) Cross-section of the phantom trunk. The pipes used for measurements are numbered. From Di Fulvio et al. 2013. The transverse plane section of the BOMAB phantom with denoted pipes filled with dosimeters is shown in Figure 7.1. Pipe 1 corresponds to „prostate“ and pipes 2 and 3 to „bladder“ and „rectum“, respectively. Pipes 4 and 5 are adjacent to the „prostate“ pipe in the coronal plane (Figure 7.2).

Figure 7.2: Simulated prostate, bladder and colon-rectum positions, with BOMAB phantom inner channels. From Di Fulvio et al. 2013. Fifteen pipes were available inside the phantom to host the dosimeters. However, doses were measured in the volume surrounding the planning treatment volume (PTV) and in the most peripheral areas, so only the five central pipes and the four corner ones were loaded with dosimeters. The position of the dosimeters in the pipes was fixed using PMMA spacers. Pipes not loaded with dosimeters were filled with water during irradiations and sealed with bi-component glue on one side and closed by leak proof recessed PMMA caps on the other side. Doses were measured over an axial distance of approximately 50 cm.


45


The prostate has been modeled as a cylinder of 5 cm length and 5 cm diameter, coaxial with pipe 1 (Figure 7.2). In male anatomy, the prostate gland is situated below the bladder, behind the pubic bone and in front of the rectum. For this reason, the cylinder simulating the bladder (5 cm length and 4 cm diameter) is coaxial with pipe 2 (Figure 7.2). The vertical axes of bladder and prostate cylinders are spaced 5 cm apart. The rectum is the terminal part of human intestine and has been simulated as a tube about 12 cm long, surrounding the prostate from pipe 3 (for a length of 5 cm) to pipe 1. The rectum measurement reference point is in pipe 3, in line with the prostate volume (pipe 1). However, dosimeters were placed along the whole length of 1-5 pipes, up to ~40 cm away A CT scan DICOM file (ISO, 2006) of the BOMAB phantom was used to define prostate planning treatment volume (PTV), bladder and rectum as simplified geometries and to plan the radiation therapy treatments accordingly (Figure 7.3).

Figure 7.3 CT scans of the BOMAB phantom trunk. Top left: transverse; top right: sagittal; bottom left: coronal; bottom right: 3-D rendering. The scans show the positions of the simplified organs (prostate, bladder and rectum) and the dose boost region within the treatment volume. From Di Fulvio et al. 2013. The same DICOM file was used for treatment planning in Pisa (SCUH ) and Krakow (COOK), therefore the organs' positions and dimensions were exactly the same in all irradiations. Three metallic fiducial marks were also permanently affixed on the phantom surface allowing the reproducible positioning of the phantom with respect to the beam axis in different facilities.

7.2 Radiotherapy treatment techniques The experiments were carried out at the SCUH (Pisa) and (COOK, Krakow) using two types of linear accelerators: Clinac 2100C (Varian) and Clinac 2300 CD (Varian). Tomotherapy was performed at the POCM, Lucca. Treatment plans were prepared separately in each centre using three different

46



treatment planning systems: CMX XIO Rel. 4.40.05, VARIAN Eclipse External Beam Planning v. 8.6 and TomoHD treatment system – TomoDirect treatment Delivery Mode. CT scans prepared at the University Hospital in Pisa were used for treatment planning with different techniques: single regular field (reference measurements), four and five fields using multileaf-collimator (3DCRT), IMRT, VMAT and Tomotherapy. Details of each technique are given in Table 7.2. Table 7.2. Radiotherapy treatment techniques used in BOMAB experiments. Radiotherapy technique

Single 10 × 10 cm field

Accelerator type

2

4-field 3DCRT 5-field 3DCRT

IMRT

Tomotherapy VMAT (RapidArc)

Varian Clinac 2100C Varian Clinac 2100C Varian Clinac 2300CD Varian Clinac 2300CD Varian Clinac 2300CD Varian Clinac 2300CD

Treatment planning system CMX XiO Rel. 4.40.05 CMX XiO Rel. 4.40.05 Eclipse 8.6 Eclipse 8.6

0

0

Gantry position

MU Number

Emax

251 MU

6 MV

0

218

15 MV

0

240 MU

6 MV

0

199 MU

18 MV

277 MU

6 MV

0 0 0

Eclipse 8.6 Eclipse 8.6

00, 900, 1800, 2700

218 MU

18 MV

Varian Clinac 2100C Varian Clinac 2100C

CMX XiO Rel. 4.40.05 CMX XiO Rel. 4.40.05

266 MU

15 MV

432 MU

6 MV

Varian Clinac 2300CD Varian Clinac 2300CD Tomotherapy HIART Varian Clinac 2100C

Eclipse 8.6

466 MU

6 MV

350 MU

18 MV

TomoDirect

00, 2700, 500, 900 3100 1800, 450, 1030, 2570, 3150 0 0 , 750, 1350, 2550, 2850 0 0 , 750, 1350, 2550, 2850 -

-

6 MV

Eclipse 8.6

-

481 MU

6 MV

Eclipse 8.6

7.3 Photon measurements: dosimetry methods Photon dosimetry methods applied in this work used TL, RPL and OSL materials (Chapter 3.1). The basic principles of the methods, their characteristics, methods of their calibration and use were described in the paper by Knežević, et al. (2013). Some data about these dosimetry methods relevant to the present work are also shown in Table 7.3. The following points are noteworthy: (i) The dimensions and shape of dosimeters can influence their angular dependence. Special attention should be given to positions close to the target edge when a very sharp dose gradient exists. (ii) Dosimeter response is energy dependent. Dosimeters were calibrated in a 60Co field in terms of absorbed dose to water according to the procedure described in IAEA TRS 398 (2000). In the out-of-field region of the water tank exposed to endpoint energies of 6-20 MV there is a large contribution of scattered photons with energies less than 0.2 MeV (Bordy et al., 2013) for which the energy dependence of dosimeters could be different from that of water depending on the effective atomic number of the dosimeter material. It is well known that OSL dosimeter material and RPL glass overestimate dose in that energy range. For OSL, the response was corrected for spectra at different depths in water (Bordy et al., 2013) whereas RPL type GD-352M used in this work has a


47


built-in tin energy compensation filter. For both types of LiF:Mg,Ti TLDs used in this work, it was estimated that there was no need to apply any energy correction. (iii) For endpoint energies greater than about 8 MV, photoneutrons produced by (γ,n) reactions in high-Z materials in the treatment head contribute to the overall dose. The neutron sensitivity of photon dosimeters should be considered. The relative tissue kerma sensitivity, ku, is defined as the ratio of the measured response of the dosimeter material to neutrons (ηn Kn,d) to the measured response to 60Co gamma radiation (ηs Ks,d) relative to the tissue kerma for neutrons (Kn,t) and gamma radiation (Ks,t) respectively (ICRU, 1984, Gibson 1986). Thus:

ku = [(ηn Kn,d)/Kn,t]/[(ηs Ks,d)/Ks,t)]

(7.1)

The neutron sensitivity depends not only on the composition of the detector material itself and on the cross sections of its constituents to different neutron spectra but also on the surrounding media. For irradiations in the BOMAB phantom, recoil protons from PMMA containing 8% hydrogen and 2.2 MeV gamma rays from (n,γ) reactions could be absorbed in dosimeters giving a component of dose originating from neutron interactions in PMMA. According to the manufacturer, the OSL detectors are insensitive to neutrons. For TLD-700 (7LiF:Mg,Ti) it was found that relative neutron sensitivity, ku, varies with neutron energy, ranging from about 4% for Pu-Be neutrons (Krpan et al., 2008) to 7.5% for neutrons of 14.5 MeV (Miljanić et al., 2007). Concerning thermal neutrons, despite the fact that TLD-700 contains only a small amount of 6LiF (~0.007% in mass) in the enriched 7LiF, due to the high cross section for the 6 Li(n,α)3He reaction, response to the thermal neutrons cannot be neglected. Kry et al., 2005 stated that in the energy range of neutrons around medical accelerators, TLD-700 is largely unresponsive to neutrons, detecting only about 1% of the neutron dose. Because the out-of-field neutron dose is typically of the order of 10% of the out-of-field photon dose, the neutron dose contribution produces only a 0.1% error in a measurement of photon dose under the conditions described in this report. They also concluded that the thermal neutrons contribute less than 1% to the dose equivalent to the patient. For RPL, relative neutron sensitivity determined for type SC-1 is somewhat less than for TLD-700 (Miljanić et al., 2008). The sensitivity of MTS-7 (7LiF:Mg,Ti) should be very similar to that of TLD-700 (Knežević, et al., 2013).

48



Table 7.3. Dosimetry methods used for photon dose measurements Institution

Dosimeter Type

Commissariat à l'Énergie Atomique (CEA), Saclay

OSL

Institute of Nuclear Physics (IFJ), Krakow

TLD, type MTS-7 7 ( LiF:Mg,Ti)

Ruđer Bošković Institute (RBI), Zagreb

RPL, type GD-352 M

Dimensions (mm)

Energy dependence

Relative neutron sensitivity, ku (equation (7.3.1))

adapter: 10 x 10 x 2

Corrected for spectra

Insensitive (according to manufacturer)

disc: φ 5 x 1

φ 4.5 x 0.9

holder:φ 4.3 x 14.5 rod:

φ 1.5 x 12

in different depths in water phantom Not corrected

As for TLD-700

Build-in tin energy compensation filter

Thermal: 4.3±0.1a Pu-Be: 0.032±0.005a 14.5 MeV: 0.041±0.005a

Ruđer Bošković Institute (RBI), Zagreb

TLD, type TLD-700 (7LiF:Mg,Ti)

φ 4.5 x 0.9

Not corrected

Thermal: 5.3±3.4b Pu-Be: 0.041±0.007c 14.5 MeV: 0.075±0.003d

a

Miljanić et al., 2008 (RPL type SC-1)

b

Gibson, 1986

c

Krpan et al., 2008

d

Miljanić et al., 2007


49


7.4 Neutron measurements: dosimetry methods Neutron dosimeters used in this work were SDDs and solid state nuclear track detectors (SSNTD). Their working principles, along with calibration and readout procedures can be found in the paper by Di Fulvio et al., 2013 and have been summarized in Chapter 3. In the following, the different types of detectors used, provided by several EURADOS affiliated institutions, are listed, along with a comparison of their sensitivity and the features most relevant for their application in external beam radiotherapy. Two types of superheated emulsion (SE) detectors have been used during WG9 measurement campaigns: SDD by UNIPI, Italy and BDT and BD-PND provided by SCK*EN, Mol, Belgium. These two technologies are in principle very similar; however some differences between the two do exist and have been detailed in Section 3.2. Solid state track detectors for neutron measurement were developed at UAB, Spain, CB-PADC and at the PoliMI, Italy, RDNS. Differences in the application and response of these two detectors rely on the different materials used and etching strategies: electrochemical etching is used for CB-PADC, while standard chemical etching for RDNS (Table 3.6). Selection criteria for dosimetric techniques were identified, taking into account that photoneutron spectra in clinical radiation therapy feature a broad energy range, a high angular spectral dependence and a strong primary pulsed photon component. For this reason, the desired characteristics are summarized as follows: Photon insensitivity is required to avoid false positive counts due to photon background. Both SE and SSNTD are insensitive to photons, since they only respond to high-LET neutron recoils. Small size (compared with the phantom volume) is necessary to allow in-phantom dosimetry and point-like measurements. SDDs and BTI® detector volumes are slightly different (Table 3.5), nevertheless their diameter is less than 2 cm. They are also watertight, so they are suitable for use in water phantoms. Moreover, their cylindrical geometry guarantees a two-axis measurement isotropy. PADC based detectors developed at UAB and PoliMi also comply with this size constraint. Further details have been given in section 3.2.3.1. High neutron sensitivity and wide measurement range are required to achieve a good offaxis accuracy while not saturating in the beam. Two sets of SDDs have been used, with a different amount of C-318, which is the material sensitive to neutrons, resulting in two kerma equivalent response coefficients of 0.2 (±0.03) and 0.5 (±0.08) bubbles µSv-1. Higher sensitivity detectors have been used out of the beam. BTI® detector sensitivity is provided for each vial, and is 3.8 bubbles µSv-1 and 0.6 bubbles µSv-1 on average for BD-PND and BDT, respectively. The precision of sensitivity values was estimated during calibration taking into account counting statistics, source calibration uncertainties, detector temperature fluctuations, discrepancies between detector response and kerma factor, as described in Chapter 3. Characteristics of CB-PADC detectors have been summarised in Table 3.6. Detector response in terms of neutron-induced equivalent doses in the energy spectrum of interest is needed. This is not an easy task and the strong angular and energy dependence of detector response has to be taken into account. Suitable calibration procedures are thus needed in order to estimate in-phantom equivalent doses, particularly for those detectors 50



whose fluence response is not comparable to kerma-equivalent-factor. The calibration strategies adopted have been described in Chapter 3. References Bordy, J.M., d'Agostino, E., Bessiere, I., Domingo, C., d'Errico, F., di Fulvio, A., Knežević, Ž., Miljanić, S., Olko, P., Ostrosky, A., Poumarede, B., Sorel, S., Stolarczyk, L., Vermersse D., Harrison, R., 2013. Radiotherapy out-of-field dosimetry: Experimental and computational results for photons in a water tank. Radiat. Meas. 57, 29-34. Bush, F., 1949. The integral dose received from a uniformly distributed radioactive. The British journal of radiology 22, 96-105. Caresana, M., Ferrarini, M., Fuerstner, M., Mayer, S., 2012. Determination of LET in PADC detectors through the measurement of track parameters. Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 683, 8-15. D'Errico, F., Alberts, W.G., Dietz, E., Gualdrini, G., Kurkdjian, J., Noccioni, P., Siebert, B.R.L., 1996. Neutron ambient dosimetry with superheated drop (bubble) detectors. Radiat Prot Dosim 65, 397400. Di Fulvio, A., Domingo, C., De San Pedro, M., D'Agostino, E., Caresana, M., Tana, L., d'Errico, F., 2013. Superheated emulsions and track etch detectors for photoneutron measurements. Radiat Meas 57, 19-28. Gibson, J.A.B., 1986. The relative tissue kerma sensitivity of thermoluminescent materials to neutrons. Radiat. Prot. Dosim. 15, 253-266. HPS, 1995. Health Physics Society. Specifications for the Bottle Manikin Absorber Phantom. An American National Standard 13 VOL., 35. IAEA, 2000. Absorbed Dose Determination in External Beam Radiotherapy. An International Code of Practice for Dosimetry Based on Standards of Absorbed Dose to Water. Technical Reports Series No. 398. International Atomic Energy Agency, Vienna. ICRP, 1975. Report on the Task Group on Reference Man. ICRP publication 23. ICRU, 1984. Neutron dosimetry for biology and medicine. Report 26, ICRU Publications, Bethesda, MD. Knežević, Ž., Stolarczyk, L., Bessiere, I., Bordy, J. M., Miljanić, S., Olko, P., 2013. Photon dosimetry methods: Optically stimulated luminescence (OSL), thermoluminescence (TL) and radiophotoluminescence (RPL) dosimetry. Radiat. Meas. 57, 9-18. Krpan, K., Miljanić, S., Vekić, B., Deme, S., Szántó, P., Pázmándi T., 2008. TL and PTTL of TLD-100 and TLD-700 after irradiation with Pu-Be neutrons. Radiat. Meas. 43, 1123-1127.


51


Kry, S. F., Salehpour, M., Followill, D. S., Stovall, M., Kuban, D. A., White, R. A., Rosen, I. I., 2005. Out-offield photon and neutron dose equivalents from step-and-shoot intensity-modulated radiation therapy. Int. J. Radiation Oncology Biol. Phys. 62(4), 1204-1216. Miljanić, S., Krpan, K., Blagus, S., 2007. TL and PTTL of TLD-100 and TLD-700 after irradiation with 14.5 MeV neutrons. Nucl. Instrum. Methods Phys. Res. A 574, 510-517. Miljanić, S., Ranogajec-Komor, M., Blagus, S., Pálfalvi, J.K., Pázmándi, Pázmándi T., Deme, S., Szántó, P., 2008. Response of radiophotoluminescent dosimeters to neutrons. Radiat. Meas. 43, 1068-1071. Nuclear Technology Services, I., 2014. Sánchez-Doblado, F., Domingo, C., Gómez, F., Sánchez-Nieto, B., Muñiz, J.L., García-Fusté, M.J., Expósito, M.R., Barquero, R., Hartmann, G., Terrón, J.A., Pena, J., Méndez, R., Gutiérrez, F., Guerre, F.X., Roselló, J., Núñez, L., Brualla-González, L., Manchado, F., Lorente, A., Gallego, E., Capote, R., Planes, D., Lagares, J.I., González-Soto, X., Sansaloni, F., Colmenares, R., Amgarou, K., Morales, E., Bedogni, R., Cano, J.P., Fernández, F., 2012. Estimation of neutron-equivalent dose in organs of patients undergoing radiotherapy by the use of a novel online digital detector. Phys Med Biol 57, 6167-6191.

52



8. Results: BOMAB phantom 8.1 Dosimeter comparisons: photons Prior to irradiations at the reference clinical linac, dosimeters were calibrated in standard calibration conditions using a 60Co source as described in IAEA TRS 398 (2000) (Chapter 3). Then, dosimeters were compared in a reference clinical linac (Saturn 43) beam in a water tank at CEALIST/LNE LNHB, Saclay (chapter 4). Radiation qualities of 6, 12 and 20 MV were used. As discussed in chapter 4, the reference calibration point in the water phantom, i.e. the isocentre, was at a depth of 10 cm in water, on the central axis of the beam, for a field size of 10 x 10 cm2. Doses were measured in the water tank over an axial distance of approximately 50 cm, at positions along a pipe. The beam axis was at 13.5 cm distance from the phantom inner wall. Comparison of TLDs and RPL with a reference ionization chamber (IC) is shown in Figure 8.1 for all three radiation qualities. The mean ratios with standard deviations for doses in the out-of-field region in the range from 1.5 to 150 mGy were within 3%: (i)

6 MV: TLD/IC = 0.982±0.027; RPL/IC = 0.976±0.034;

(ii)

12 MV: TLD/IC = 0.997±0.025; RPL/IC = 1.029±0.025;

(iii)

20 MV: TLD/IC = 0.999±0.046; RPL/IC = 1.027±0.041.

For 6 MV irradiations, TLD-100 (natLiF:Mg,Ti) from RBI was used instead of TLD-700. More details about irradiations and results using the reference clinical linac are presented in the paper by Bordy et al., 2013.

Figure 8.1 Comparison of dosimeters in a reference clinical linac beam in a water tank at CEA-LIST/LNELNHB, Saclay. IONO refers to ionization chamber measurements.


53


For clinical irradiations in the BOMAB phantom, the dosimeters used for dose distribution measurements for different modalities are shown in Table 8.1. Results of the dosimeter comparison are shown for modalities where three types of dosimeters were irradiated: (i) in Figure 8.2 for 1-field 10x10 cm2 in the COOK, Krakow and 5-field CRT in the SCUH, Pisa and (ii) in Figure 8.3 for 6 MV and 18 MV IMRT in the COOK, Krakow. Doses are given for the central „prostate“ pipe (pipe 1). Variation of doses in the out-of-field region (above 1 mGy) was larger than in a water tank in the reference clinical linac beam. Standard deviations of groups of three dosimeters in all 5 pipes per modality were: (i)

7% (between 1 and 150 mGy) for 18MV reference field (COOK);

(ii)

10% (between 3 and 33 mGy) for 15MV 5-field CRT (SCUH);

(iii)

6% (between 1 and 57 mGy) for 6 MV IMRT (COOK) and

(iv)

9% (between 1 and 30 mGy) for 18 MV IMRT (COOK).

Table 8.1. Dosimeters used for different modalities. a)

SCUH, Pisa

Centre

Dosimeter

CEA

OSL

IFJ

TLD (MTS-7)

+

RBI

RPL

+

SCUH

IC

+

6 MV 1field

b)

6 MV 1field

15 MV 1field

15 MV 5field CRT

6 MV IMRT

6 MV VMAT

6 MV tomotherapy

+

+

+

+

+

+

+

+

+

+

+

+

18 MV 1field

6 MV 4field CRT

18 MV 4field CRT

6 MV IMRT

18 MV IMRT

+

+

+

COOK, Kraków

Centre

Dosimeter

CEA

OSL

IFJ

TLD (MTS-7)

RBI

RPL

+

+

+

RBI

TLD (TLD-700) IC

+

+

+

+

+

COOK

+ +

+

+

+

+

+

54

+

+



Figure 8.2 Comparison of dosimeters irradiated in a BOMAB phantom by 18 MV single 10 cm × 10 cm field (COOK,IFJ, Krakow) and a 15 MV 5-field CRT plan (SCUH, Pisa).

Figure 8.3 Comparison of various dosimeters irradiated in a BOMAB phantom using 6 and 18 MV IMRT plans (COOK,IFJ, Krakow).


55


8.2 Dosimeter comparison: neutrons Dosimeters based on superheated emulsions were compared in a reference clinical linac (Saturn 43) beam in a water tank at CEA-LIST/LNE LNHB, Saclay, using radiation qualities of 12 and 20 MV, as described in Chapter 4. Measurements were carried out in a 30 x 30 x 60 cm3 water filled PMMA phantom, with the 10 cm x 10 cm windowed beam axis centred on the (x=150 mm, y=150 mm, z=0 mm) point on the phantom (Figure 8.4). Five frame positions and five pipes were available, on the z and y axis respectively (see Figure 8.4), with a 5o mm spacing.

Figure 8.4: PMMA water tank used for reference measurements, with rails suitable for holding dosimeter pipes. The highlighted plane contains the reference point, at a depth of 100 mm in water, 150 mm from the side of the tank (x=150 mm, y=150 mm, z=100 mm). In this section, the axes z and x correspond to the axes d (depth) and D (lateral distance) respectively in Figure 4.2. The sensitive portion of SDD and BTI® detectors has been centred at 150, 250, 350, 450 and 550 mm along each pipe, on the x axis. As discussed earlier, the reference point in the water phantom is the isocentre. This point is used as a reference to compare different radiation qualities and also neutron measurement systems. At this point, photon doses were kept constant by varying the irradiation time for different qualities of the radiation field. Doses were measured in the water tank over an axial distance (x) of approximately 500 mm. The length and the cross-section of the major axis of the trunk of the BOMAB phantom and the length and height of the PMMA water tank are comparable (Figure 8.5). Spacing between pipes and detector positions inside the pipes in the two phantoms were also designed so that measurements could be compared. The centre of the prostate target volume, used for radiotherapy treatment simulations in the BOMAB phantom, is at a depth of 125 mm, whilst the reference point in the water phantom is at a depth of 150 mm and 135 mm from the inner surface of phantom, where 135 mm is the lateral distance along a pipe (measured as the distance from the inner surface of the phantom, see Chapter 4).

56



Figure 8.5 Trunk of the BOMAB phantom virtually overlapping the PMMA water filled phantom. The reference point is used to report neutron measurement data as ratios between neutron dose equivalent, in μSv, and photon absorbed dose at the isocentre in Gy. The distribution of neutron dose equivalent as a function of penetration depth (z-axis, Figure 8.4) in the water tank phantom along the beam axis is shown in Figure 8.6, for 12 and 20 MV photon irradiation, for BTI® detectors and SDD. An average maximum value of 4.6 (±0.8 1SD) mSv Gy-1 is observed at a depth of 50 mm on the isocentre axis, i.e., 50 mm above the isocentre. The deeper regions at 150 and 200 mm depth appear to be effectively shielded by the neutron attenuation in water, due to the moderating action of hydrogen nuclei. The dose attenuation coefficient (λ) along the z-axis can be calculated as in equation 8.1, for a depth d, where h0 and hx are the entrance neutron dose equivalent and the neutron dose equivalent at a depth d, respectively. ℎ

𝜆 = ln (ℎ0 )

1

(8.1)

𝑥 𝑑

The on-axis dose attenuation coefficient is 0.011 mm-1 and 0.022 mm-1 for 20 and 12 MV respectively. λ at 20 MV compares well with the value estimated by d’Errico et al. (1998) for a Saturne 20 accelerator, 10 cm x 10 cm single field and 20 MV photon radiation quality. In Figure8.6, data acquired with different systems agree within their 1SD uncertainties, which are not shown, but estimated to be in the order of 20%. A detailed evaluation of uncertainty sources in measurements with superheated emulsion detectors can be found in Di Fulvio et al. (2013).


57


Figure 8.6. Neutron dose equivalent along the z-axis for 12 and 20 MV x-ray beam (normalized to 1 Gy of photons at the isocentre) measured with SDD and BD-PND. Irradiations at the CEA-LIST/LNE LNHB facility allowed comparison of responses of different dosimetric systems. Figure 8.7 shows the neutron dose distribution along a phantom pipe (in the x direction) at z=500 mm depth. Measurement points have been centred at 150 (isocentre), 250, 350, 450 and 550 mm along each pipe for SDD and at 150 (isocentre), 350 and 550 mm for BTI® detectors. The average neutron energy and production efficiency increases with maximum energy of the Bremsstrahlung photon spectrum (Tosi et al. 1991). For this reason, 20 MV maximum photon energies result in higher out-of-field neutron doses per unit photon dose at the isocentre, with respect to 12 MV, both on the central axis and along a 50 mm (x=650 mm) off-axis pipe. Figure 8.7 also shows that out-of-field neutron dose is relatively higher for 20 MV than for 12 MV. A similar attenuation trend can be noticed at a depth of z=150 mm inside the phantom (Figure 8.8). In this case, a higher photoneutron dose is delivered on axis where high-energy x-rays generate neutrons through interactions with accelerator structures.

58



Figure 8.7 Neutron dose equivalent over an off-axis (x) distance of 400 mm for 12 and 20 MV x-ray beams (normalized to 1 Gy of photons at the isocentre) measured with SDD and BTI (BD-PND type) at a depth of 50 mm.

Figure 8.8 Neutron dose equivalent over an off-axis (x) distance of 400 mm for 12 and 20 MV x-ray beams (normalized to 1 Gy of photons at the isocentre) measured with SDD and BTI (BD-PND type) at a depth of 150 mm. Measurements carried out in the reference calibration field inside a water tank confirmed that readouts of dosimetric systems based on superheated emulsions compare well, within an estimated measurement uncertainty of ±20% (1 SD), and can be thus used for an extended comparison of different radiation therapy treatment modalities.


59


8.3 Out-of-field doses, photons: comparison with TPS calculations and comparison of modalities

8.3.1 Comparison of doses in different positions (pipes) in the phantom In Figure 8.9 and 8.10, dose distribution measurements in 5 pipes are shown for 6 MV IMRT and 18 MV IMRT, respectively at COOK, Krakow. Each point is a mean value of three dosimeter readings (MTS-7, RPL and TLD-700) from three separate irradiations. The distances between dosimeters were too large to show the realistic shape of the curve in the target volume. Dose distribution measurements with OSL CEA dosimeters for which their spacing in the target volume was smaller, thus showed the realistic shape of the curves as shown in Figures. 8.11 (for 5-field CRT, 15 MV, SCUH,Pisa) and 8.12 (4-field CRT, 18 MV, IFJ & COOK,Krakow). For irradiations by 5-field CRT, 15 MV at SCUH, the dose to „bladder“ is 38.5% and to „rectum“ 24.1% of that measured in the „prostate“. For irradiations by 4-field CRT, 18 MV, IFJ & COOK, doses in „bladder“ and „rectum“ are 52.9% and 53.0% respectively of the „prostate“ dose. Although doses in the treatment volumes for „prostate“, bladder“ and „rectum“ differ significantly, differences in out-of-field doses in different pipes do not show large differences.

Figure 8.9. Dose distribution measurements in 5 pipes for 6 MV IMRT at IFJ & COOK,Krakow. Each point is the mean value of measurements of three dosimeters (MTS-7, RPL and TLD700) from three separate irradiations.

60



Figure 8.10. Dose distribution measurements in 5 pipes for 18 MV IMRT irradiations at IFJ & COOK, Krakow. Each point is the mean value of measurements of three dosimeters (MTS-7, RPL and TLD-700) from three separate irradiations.

Figure 8.11. Dose distribution measurements with OSL CEA dosimeters in 5 pipes for 5-field CRT, 15 MV, in SCUH,Pisa.


61


Figure 8.12 Dose distribution measurements with OSL CEA dosimeters in 5 pipes for 4-field CRT, 18MV, in IFJ & COOK, Krakow.

8.3.2. Comparison of different treatment modalities Comparison of out-of-field doses for different treatment modalities is shown in Figure 8.13 (a) for irradiations at SCUH & POCM and (b) for irradiations at COOK. Results are given for the central (prostate) pipe. For irradiations at SCUH & POCM, mean values of RPL and TLD IFJ were used and for irradiations at IFJ & COOK, results were obtained with TLD IFJ dosimeters. Without 1-field irradiations, the results show that the ratio of maximum (Tomotherapy) to minimum (15 MV 5-field CRT) doses ranges from 2.4 to 3.8 at SCUH & POCM and from 1.6 to 2.4 (the highest for 6 MV IMRT and the lowest for18 MV 4-field CRT). For IMRT modalities, there is a pronounced „hump“ at about 20 cm distance from the isocentre origin from head leakage (Ruben et al., 2011). The same effect is also visible for CRT but less pronounced. Generally, the lowest peripheral doses were obtained for 3DCRT. However, by comparing 18 MV IMRT and 18 MV 4-field CRT at COOK, one can see that DCRT/DIMRT>1 for distances less than 15 cm from the field edge and DCRT/DIMRT < 1 for distances above 15 cm from the field edge. In Table 8.2, out-of-field doses were shown in decreasing order for all modalities used in two centres.

62



Figure 8.13. Comparison of out-field-doses on the PTV (pipe 1) axis for different treatment modalities: (a) irradiations at SCUH (Pisa) and POCM (Lucca) and (b) irradiations at IFJ & COOK(Krakow). For (a), mean values for RPL (RBI) and TLD (IFJ, Krakow) are shown. For (b) data refer to TLD (IFJ, Krakow) only.


63


Table 8.2 Monitor Units for different modalities in decreasing order of out-of-field doses for different modalities. SCUH, Pisa & POCM, Lucca Varian Clinac 2100C 6 MV single 10 cm x 10 cm field (ref)

MU / 2 Gy 251

MU / 2 Gy

COOK, Kraków Varian Clinic 2300CD 6 MV single 10 cm x 10 cm field (ref)

240

18 MV single 10x10 cm field (ref)

199

6 MV IMRT*

466

6 MV 4-field CRT

277

18 MV IMRT

350

18 MV 4-field CRT*

218

6 MV Tomotherapy (POCM) 6 MV VMAT (RapidArc)

481

6 MV IMRT*

432

15 MV 5-field CRT

266

* Pronounced hump at about 22 cm from the isocentre The results of peripheral dose measurements are also shown in Table 8.3. which illustrates mean values of doses in the 5 pipes (mGy per 2 Gy at isocentre ± 1 standard deviation in percent) for different modalities and for 4 distances from the field edge (isocentre). These values represent variations within slices along the longitudinal axis and also differences in out-of-field doses for different modalities. For the calculation, the mean values of 1-3 dosimeters of particular type for a given modality were used (see 8.1). Table 8.3 Mean values of doses in 5 pipes (mGy per fraction of 2Gy at isocentre ± standard deviation in percent) for different modalities and for different distances from the field edge (isocentre). Distance from the field edge (depth of the isocentre in parenthesis) (cm)

Pisa (SCUH & POCM) 5-field CRT 15 MV

IMRT 6 MV

VMAT 6 MV

Tomotherapy 6 MV

9.6 (11.85)

5.68±36.3%

10.88±6.5%

12.57±5.6%

17.41±6.7%

19.6 (21.85)

1.55±13.6%

4.12±19.7%

3.09±0.89%

3.88±4.0%

29.6 (31.85)

0.45±2.6%

1.02±7.7%

1.12±4.0%

1.46±6.7%

39.6 (41.85)

0.27±12.8%

0.61±6.6%

0.66±6.3%

0.94±3.8%

Krakow (COOK) 4-field CRT 6 MV

4-field CRT 18 MV

IMRT 6 MV

IMRT 18 MV

9.6 (11.85)

13.96±5.1%

7.64±6.1%

15.21±5.6%

8.04±3.8%

19.6 (21.85)

4.07±15.1%

3.11±21.3%

6.94±12.6%

5.73±13.6%

29.6 (31.85)

1.14±5.7%

0.90±8.9%

1.76±1.9%

1.36±3.7%

39.6 (41.85)

0.67±2.7%

0.57±5.6%

1.07±2.9%

0.76±17.5%

64



The influence of energy for the same modality for irradiations in COOK (central pipe 1) is shown in Figure 8.14 for: 6 and 18 MV IMRT (left) and 6 and 18 MV 4-fields CRT (right). In Figure 8.15, the comparison of out-of-field doses for different energies is also shown, for 6 and 18 MV, 1-field 10x10 cm2 (Krakow, central pipe 1) (left) and for a water tank, Saclay (Bordy et al., 2013) (right). In all cases as the energy increases, the peripheral dose decreases, giving: (i)

D6MV/D18MV: 1.3-2.2 for IMRT;

(ii)

D6MV/D18MV: 1.1-1.8 for 4-field CRT and

(iii)

D6MV/D18MV: 1.1-1.3 for 1-field 10x10 cm2.

The equivalent results were obtained in water tank in Saclay (Bordy et al., 2013).

(a) 6 MV and 18 MV IMRT (pipe 1, TLD, IFJ, Krakow)

(b) 6 MV and 18 MV 4 field CRT (pipe 1, TLD, IFJ, Krakow) 100

Dose (mGy) for 2Gy to isocentre


100 6 MV IMRT 18 MV IMRT isocentre

10

1

6 MV 4 field CRT 18 MV 4 field CRT isocentre

10

1

0

0 0

5

10

15

20

25

30

35

40

45

50

55

0

60

5

10

15

20

25

30

35

40

45

50

55

60

Distance from inner face (cm)

Distance from inner face (cm)

Figure 8.14 Comparison of out-of-field doses for different energies (irradiations at IFJ & COOK, Krakow, pipe 1, using TLD): (a) 6 and 18 MV IMRT and (b) 6 and 18 MV 4 field CRT.

Figure 8.15 Comparison of out-of-field doses for different energies, left: 6 and 18 MV, 10 cm x 10 cm field (IFJ, Krakow, pipe 1), right: water tank, CEA, Saclay (from Bordy et al., 2013). Comparison of the results in two centres for the same irradiation modalities is interesting. Two examples are shown in Figure 8.16: a) for 6 MV 1-field 10x10 cm2 and b) for 6 MV IMRT. The differences in peripheral doses in two centres, as expected, are small for reference 1-field 10x10


65


cm2 with DKrakow/DPisa in the range 0.9-1.1 (without the furthest point or 0.9-1.2 taking the furthest point into account). For IMRT, differences were larger, DKrakow/DPisa: 1.4-1.8 (taking into account the furthest point: 1.4-2.2). The results are in agreement with the differences in the treatment planning system calculations given in Figure 8.17.

Figure 8.16 Comparison of peripheral doses for the same irradiation conditions at SCUH and COOK, for 6 MV 1-field and 6 MV IMRT.

8.3.3. Comparison of treatment planning system (TPS) calculations and dosimeter measurements Dose profiles from different treatment plans at SCUH & POCM and COOK are given in Figure 8.17. Distance (in cm) is given from the inner face of the posterior transverse surface of the trunk. It is clear that the shape of the TP curves corresponds to the results of peripheral doses measured with dosimeters, but generally dosimeters show larger doses than the TPS as distance from isocentre increases. Dose distributions from TPS are given for distances up to about 15 cm from the isocentre. Beyond this distance, dosimeters show much larger doses than could be predicted by the TPS as shown Figure 8.18, where comparison of TPS and dosimeters for 15 MV 5-fields CRT (SCUH) and 18 MV 4-fields CRT (COOK) is given. Results are given for „prostate“, „bladder“ and „rectum“ axes. The comparisons of TPS and dosimeters are also shown in Figure 8.19 for 6MV at SCUH & COOK and in Figure 8.20 for 6 MV VMAT in Pisa. In this latter case, dosimeters show the closest values to the TPS. The ratios of dosimeters and TPS values in the part of the curve closer to the field edge shown in Table 8.4 are in the range 1.15-2.25. The estimation of the sparing of adjacent sensitive organs from PTV dose data is shown in Figure 8.21. Maximum doses for „bladder“ (pipe 4) and „rectum“ (pipe 5) as the percentage of TP dose for „prostate“ (pipe 1) are given for modalities at SCUH (P) and COOK (K). Generally doses in the rectum pipe are lower than in bladder pipe, with exception of 4-field CRT at COOK as shown in Figure 8.12 and also for IMRT at SCUH. The reasons are in the different gantry rotation angles for treatment delivery, which give the same doses in the rectum and bladder for 4-field CRT and higher doses for the rectum in case of IMRT at SCUH. The best results for sparing bladder and rectum are obtained for IMRT in both hospitals and for 5-field CRT at SCUH. 66




(a) TPS calculations for section through PTV (pipe 1, UHSC, Pisa) 10000 tomotherapy VMAT 6 MV

1000

IMRT 6 MV 5 field CRT

100

10

1 -5

0

5

10

15

20

25

30

Distance (cm)

(b) TPS calculations for section through PTV (pipe 1, IFJ, Krakow) Dose (mGy) for 2Gy to isocentre

10000 IMRT 6 MV IMRT 18 MV 4 field CRT 6 MV

1000

4 field MLC 18 MV 100

10

1 -5

0

5

10

15

20

25

30

Distance (cm)

Figure 8.17. Dose profiles through the PTV axis (pipe 1) for several treatment types (a) SCUH, Pisa (b) COOK, Krakow.


67


15 MV 5-field CRT: UHSC Pisa

18 MV 4-field CRT: IFJ Krakow

(a) PTV: pipe 1 -5

0

(d) PTV: pipe 1

5

10

15

20

25

30

-5

0

5

10

15

20

TPS RPL (RBI)

Dose (mGy)

Dose (mGy)

1000

CEA (OSL)

100

TLD (IFJ) CEA (OSL)

100

10

10

1 -5

0

5

10

15

20

25

1 -5.00

30

0.00

5.00

Distance(cm)

0

5

10.00

15.00

20.00

25.00

30.00

20

25

30

Distance(cm)

(e) Bladder: pipe 4

(b) Bladder: pipe 4 -5

10

15

20

25

30

10000

-5

0

5

10

15

10000

TPS

TPS

TLD (IFJ)

1000

1000

RPL (RBI)

Dose (mGy)

Dose (mGy)

30

TPS

TLD (IFJ)

1000

CEA (OSL)

100

TLD (IFJ) CEA (OSL)

100

10

10

1 -5.00

1 -5

0

5

10

15

20

25

30

20

25

30

0.00

5.00

0

5

15.00

20.00

25.00

30.00

20

25

30

(f) Rectum: pipe 5

(c) Rectum: pipe 5 -5

10.00

Distance(cm)

Distance(cm)

10

15

10000

-5

0

5

10

15

10000

TPS

TPS

TLD (IFJ)

1000

1000

RPL (RBI)

Dose (mGy)

Dose (mGy)

25

10000

10000

CEA (OSL)

100

TLD (IFJ) CEA (OSL)

100

10

10

1 -5

0

5

10

15

20

25

30

1 -5.00

0.00

5.00

10.00

15.00

20.00

25.00

30.00

Distance(cm)

Distance(cm)

Figure 8.18. Comparison of TPS calculations and dose measurements for various dosimeters for 15 MV (5 field treatment) SCUH, Pisa and 18 MV (4 field treatment) at COOK, Krakow. Results are given for the PTV (prostate) axis (a & d),bladder axis (b & e) and rectum axis (c & f).

68



(a) UHSC Pisa -5

0

PTV axis: pipe 1

5

10

15

20

(d) IFZ Krakow 25

-5

30

0

5

PTV axis: pipe 1 10

15

20

25

TPS

TLD (IFJ)

1000

RPL (RBI)

RPL (RBI)

Dose (mGy)

Dose (mGy)

TPS

TLD (IFJ)

1000

100

TLD (RBI)

100

10

10

1 -5

0

5

15

10

20

25

1

30

-5

0

5

10

(b) UHSC Pisa -5

0

5

(e) IFZ Krakow

Bladder axis: pipe 4 10

15

25

20

15

20

25

30

-5

10000

0

5

Bladder axis: pipe 4 10

15

20

25

30

10000 TPS

1000

TPS

RPL (RBI)

100

TLD (IFJ)

1000

TLD (IFJ)

Dose (mGy)

Dose (mGy)

30

Distance(cm)

Distance (cm)

RPL (RBI) TLD (RBI)

100

10

10

1 -5

0

5

10

15

20

25

1

30

-5

Distance (cm)

(c) UHSC Pisa -5

0

5

Rectum axis: pipe 5 10

15

0

5

10

15

20

25

30

Distance (cm)

20

25

(f) IFZ Krakow 30

-5

10000

0

Rectum axis: pipe 5 5

10

15

20

25

30

10000 TPS

TPS

1000

TLD (IFJ)

Dose (mGy)

Dose (mGy)

30

10000

10000

RPL (RBI)

100

10

TLD (IFJ)

1000

RPL (RBI) TLD (RBI)

100

10

1 -5

0

5

10

15

20

25

1

30

-5

Distance (cm)

0

5

10

15

20

25

30

Distance (cm)

Figure 8.19. Comparison of TPS calculations and dose measurements for various dosimeters for 6 MV IMRT at SCUH, Pisa and COOK, Krakow. Results are given for the PTV (prostate) axis (a & d), bladder axis (b & e) and rectum axis (c & f).


69


6 MV VMAT (Varian RapidArc™) SCUH, Pisa

Figure 8.20. Comparison of TPS and dosimeters for 6 MV VMAT at SCUH. Results are given for (a) PTV (prostate), (b) bladder and (c) rectum axes. (d) compares TPS calculations for the same axes.

70



Figure 8.21 Maximum doses for bladder (pipe 4) and rectum (pipe 5) as a percentage of PTV dose for prostate (pipe 1) given for modalities at SCUH, Pisa (P) and COOK, Krakow (K).


71


Table 8.4 Comparison of doses measured by dosimeters and from treatment planning for different modalities per 2 Gy at isocentre. Doses measured at COOK using CRT were measured with TLD IFJ and for all other modalities are mean values of RPL and TLD IFJ. Distance from the field edge Dmeas (depth of the (mGy) isocentre in parenthesis) (cm) 4.60 (6.85)

31.91

9.60(11.85)

7.82

Pisa (SCUH & POCM) 5-field CRT 15 MV

IMRT 6 MV

Dmeas/TPS

TPS (mGy)

14.20

2.25

Dmeas

TPS

(mGy)

(mGy)

42.44

25.10

Dmeas/TPS

1.69

12.13 VMAT 6 MV

Dmeas

TPS

(mGy)

(mGy)

4.60 (6.85)

44.88

39.00

9.60(11.85)

13.79

10.20

Tomotherapy 6 MV Dmeas/TPS

Dmeas

TPS

(mGy)

(mGy)

1.15

76.52

50.53

1.35

19.48

Dmeas/TPS 1.51

Krakow (COOK) 4-field CRT 6 MV Dmeas

TPS

(mGy)

(mGy)

4.60 (6.85)

59.33

47.65

9.60(11.85)

15.10

13.10

4-field CRT 18 MV

Dmeas/TPS

Dmeas

TPS

(mGy)

(mGy)

1.25

35.59

24.80

1.44

1.15

8.42

5.68

1.48

IMRT 6 MV Dmeas

TPS

(mGy)

(mGy)

4.60 (6.85)

56.83

44.60

9.60(11.85)

16.48

Dmeas/TPS

IMRT 18 MV Dmeas/TPS 1.27

Dmeas

TPS

(mGy)

(mGy)

30.24

19.70

Dmeas/TPS 1.54

8.38

72



8.4 Out-of-field doses: neutrons

8.4.1 Comparison of doses in different positions (pipes) in the phantom The planning treatment volume (PTV) and the organs at risk (OAR, i.e. bladder and rectum) simulated for WG9 experimental campaigns have been defined via software (Focal v4.7, Elekta CMS Software, Stockholm, Sweden, 2012) within the trunk of the BOMAB volume, as described in Chapter 7.

Figure 8.22 The BOMAB phantom and Varian Clinac 2300 CD at the SCUH, Pisa. In this position, the angle between phantom sagittal plane and accelerator gantry is 0°. The gantry is allowed a full 360° rotation during the treatment. In the configuration in Figure 8.22 the phantom was placed in a supine position, where the “bladder” pipe is the closest one to the accelerator gantry. The pipes not used in the measurements were filled with water during the irradiations. In external beam radiotherapy, photon dose at treatment volume is maximized while it is kept as low as possible at surrounding organs. Conversely, photoneutrons are generated through interactions of high energy photons with accelerator structures and treatment room, as well as within the patient. This results in a non-negligible whole-body exposure, increasing with irradiation time. The photoneutron dose component is higher, but not concentrated, at the treatment volume with respect to surrounding regions, since neutrons are not shielded as effectively as photons by the accelerator structure. Photoneutron exposure is also proportionally higher with increasing energy of the x-ray primary beam and dependent on the position of the accelerator gantry during the treatment. The full range of irradiation modalities, x-ray beam energies, monitor units per Gy of photon dose at the isocentre and the angular displacement of the accelerator gantry with respect to the BOMAB sagittal axis, used during the simulated clinical treatment campaigns at SCUH, POCM and COOK are reported in Table 7.3 and reported also in Table 8.5 for clarity.


73


Table 8.5. Prostate treatments simulated during WG9 experimental campaigns in at COOK, Krakow and SCUH & POCM, Pisa.

Single field 10x10 cm

2

Energy (MV)

Monitor Unit per Gy at PTV

Gantry Angles

Accelerator / Location

6, 15, 18

SCUH: 126 (6 MV) 109 (15 MV)

0°

Varian Clinac 2300 CD/ SCUH & COOK

COOK: 117 (6 MV), 98 (18 MV) 4-field CRT

6, 18

90 (6 MV), 111 (18 MV)

0°, 90°,180°, 270°

Varian Clinac 2300 CD/ COOK

5-field CRT

15

133

0°, 50°, 90°, 270°, 310°

Varian Clinac 2300 CD/SCUH

IMRT

6, 18

SCUH: 216 (6 MV), COOK:

SCUH: 45°,103°, 180°, 257°, 315°

Varian Clinac 2300 CD/

238 (6 MV), 175 (18 MV)

SCUH & COOK

COOK: 0°, 75°, 135°, 255°, 285°

VMAT (RapidArc)

6

241

-

Varian Clinac 2300 CD/SCUH

Tomotherapy

6

133

-

Tomotherapy/POCM

Photoneutron dose differs significantly in different pipes, depending on the water thickness between the neutron source and measurement point. Water resembles well the human body’s response to neutrons for in terms of the shielding effect, largely due to the moderating action of hydrogen nuclei. A 10x10 cm2 field in fixed position (0° between field axis and phantom sagittal plane) was suitable for studying the water shielding effect in the BOMAB phantom. This kind of field was used for reference and comparison between different facilities during WG9 campaigns. The “bladder” is closer to the phantom surface in front of the gantry and as expected receives a dose higher than “prostate” and “rectum” (Figure 8.23). For a 10x10 cm2 squared field, 15 MV, the “prostate” and “rectum” dose measured in the beam are 72.0% and 52.5% of that measured in the “bladder”, respectively. At the bladder, an average dose of 981 (±5 1SD) µSv/Gy is measured inside the 10x10 cm2 beam field.

74



Figure 8.23. Neutron dose equivalent as a function of distance from the BOMAB phantom bottom face (normalized to 1 Gy of photon dose at the isocentre) measured with SDD and BD-PND in three pipes, in a 10x10 cm2 squared field. The dosimeters placed in the “prostate” channel measure an average out-of-field dose of 5.17 (±0.06 1SD) µSv/Gy. This value was calculated averaging the dose measured at an off-axis distance from the isocentre of at least 30 cm. The out-of-field dose in the “prostate” channel is higher compared to the dose absorbed by detectors placed in the “bladder” and “rectum” pipes at the same location, which measure 3.48 (±0.19 1SD) µSv Gy-1 and 2.52 (±0.15 1SD) µSv Gy-1, respectively. This is probably due to the fact that the “prostate” pipe has a larger field-of-view for primary scattered neutrons, compared to the “bladder” pipe, while being less attenuated by water, compared to the “rectum” pipe. The spatial distribution of photoneutron dose changes when the gantry is rotated during the treatment. IMRT and VMAT techniques use a system of collimator leaves which move into and out of the beam to provide a spatial fluence modulation as the gantry is rotated, to achieve the highest sparing factor, i.e. the ratio between the dose received by tumour and organs at risk (OARs). For irradiations by 5-field CRT, 15 MV at SCUH, the dose to „bladder“ is 42% and to „rectum“ 34% of that measured in „prostate“ (Figure 8.24).


75


Figure 8.24. Neutron dose equivalent as a function of distance from the inner phantom face (normalized to 1 Gy of photon dose at the isocentre) measured with SDD, BD-PND in three pipes, during a 5 field CRT treatment. Compared to the square field, in the clinical treatment based on multi-leaf collimators, the target volume receives the highest photoneutron dose because the gantry is rotated during the treatment and thus there is no single OAR directly irradiated during the whole treatment. The effect of gantry angular displacement on doses in different pipes in the BOMAB phantom is also shown by comparing intensity modulated and multi-field multi-leaf collimator based techniques (Figure 8.25). Advanced treatment planning systems allow the user to select a set of gantry angles for radiation delivery. The gantry was subsequently rotated by 0°, 75°, 135°, 225° and 285° with respect to the phantom sagittal plane for IRMT treatment. Conversely, irradiation angles for the 4 field multi leaf treatment were 0°, 90°, 180° and 270°. Both rectum and bladder are directly irradiated for a fraction of the 4 field CRT treatment; only the bladder is directly irradiated during IMRT. The two treatments were both delivered at COOK using a Varian Clinac 2300 CD operating at 18 MV. For irradiations by 4-field CRT, rectum and bladder receive a comparable photoneutron dose equivalent per unit photon dose delivered at the target volume of 1.25 (+0.05 1SD) and 1.40 (+0.02 1SD) mSv Gy-1, respectively. OARs are better spared in IMRT; bladder and rectum receive 483 (+12 1SD) and 180 (+29 1SD) µSv Gy-1, respectively.

76



Figure 8.25. Dose profiles in "bladder" and "rectum" mimicking pipes – comparison between measurements in 18 MV 4-field CRT and IMRT treatments.


77


8.4.2. Comparison of different treatment modalities and energies Two main parameters are selected by clinicians during the radiation therapy design stage and unequivocally define the treatment protocol for a single therapy fraction: the irradiation energy and the radiation delivery mode, i.e. treatment plan. As detailed in Table 8.5, WG9 undertook several irradiation campaigns, in three European facilities, for four different clinically routinely used treatment plans (IMRT, Tomotherapy, VMAT and 3-D CRT) at accelerator maximum voltage energies of 6, 15 and 18 MV. In this section, we first compare the dose equivalent due to photoneutrons for several treatment plans, operated at the same accelerator voltage and then, for two treatment plans, i.e. IMRT and CRT, we show the photoneutron dose equivalent as a function of the primary photon beam energies. These results and discussion on how they should affect treatment selection criteria are discussed in chapter 9. For the benefit of the reader we report in Table 8.6 the main characteristics of the treatment plans which affect photoneutron dose equivalent to radiosensitive regions, i.e. photon energy, irradiation angles and monitor units per unit photon dose delivered. These parameters directly affect photoneutron production cross section, direct irradiation of radiosensitive organs surrounding the prostate and leakage radiation, respectively. Irradiation angles are not specified for arc therapies, i.e. VMAT and Tomotherapy, as in this case multiple radiation beams sweep around the patient, in complete or partial arc rotations of the gantry.

Table 8.6. Specifications of prostate treatments directly affecting photoneutron dose equivalent to radiosensitive organs. The accelerator gantry angle is defined clockwise with respect to the BOMAB sagittal axis. Treatment modality

Energy (MV)

Monitor Unit per Gy at PTV

Gantry Angles

4-field CRT

6

90

0°, 90°,180°, 270°

4-field CRT

18

111

0°, 90°,180°, 270°

5-field CRT

15

133

0°, 50°, 90°, 270°, 310°

IMRT

6

SCUH: 216, COOK: 238

SCUH: 45°,103°, 180°, 257°, 315° COOK: 0°, 75°, 135°, 255°, 285°

IMRT

18

175

0°, 75°, 135°, 255°, 285°

VMAT (RapidArc)

6

241

n.a.

Tomotherapy

6

133

n.a.

The lowest photon energy used for external radiation therapy is 6 MV, sometimes preferred to higher energies as it is considered free of spurious photoneutrons. However this assumption is not accurate. It is shown in Figure 8.26 that a non-negligible photoneutron dose of tens of microsieverts, per unit photon dose delivered at the treatment volume, is measured at the prostate and at surrounding organs at risk.

78



Figure 8.26. Photoneutron dose equivalent for three treatment modalities carried out at 6 MV, as a function of depth from phantom surface.

Several irradiations at 15 MV were also performed. Figure 8.27 shows the spatial photoneutron dose distribution along the BOMAB pipes for a 5 field multi leaf collimator plan, compared to a reference 10cm x 10 cm single field. The comparison of the conformal treatment to the reference single field shows that photoneutrons are mainly produced by the main photon beam colliding with jaws and the collimator structure and are thus focused at the treatment volume. Figure 8.28 reports numerical data used in Figure 8.29, measured with SDD detectors placed inside the pipes.


79


Figure 8.27. Three dimensional distributions of neutron dose equivalents along BOMAB pipes for two different irradiation geometries at 15 MV: 10 cm × 10 cm single field (right) and 5-field MLC.

Figure 8.28. Neutron dose equivalents along BOMAB pipes at 15 MV for two different irradiation geometries: 10 cm × 10 cm single field (upper figure) and 5-field MLC. 80



Accelerator energies as high as 18 MV are also used and sometimes preferred to lower energies because of steeper dose gradients around the planning target volume, lower out-of-field photon doses and increased skin-sparing. IMRT and 3D CRT treatments based on 4 collimated fields were performed using the BOMAB phantom at 18 MV, and are shown in Figure 8.30. Doses measured with SDD and PADC are shown for the reference irradiation pipes, at both depths of the “prostate” and the “bladder”, i.e. 15 cm and 5 cm along the z-axis. While the doses delivered by the two treatments are comparable at the PTV, neutron peripheral dose absorbed during IMRT treatment is higher compared to the multi-field conformal plan. This will be further discussed in Section 9.

Figure 8.29. Neutron dose equivalents along the BOMAB ‘bladder’ and ‘PTV’ pipes at 18 MV for two different irradiation geometries: IMRT and 4 fields CRT. To summarize the overall difference in treatment modalities, in terms of peripheral dose and keeping the MV constant, Table. 8.7 reports the average peripheral neutron dose (APND, μSv Gy-1 at the isocentre, with a 1 SD uncertainty of about 20%). APND is defined as the average dose measured in each channel at a distance >10 cm from PTV axial plane.


81


Table 8.7. Neutron dose equivalents along BOMAB ‘bladder’, ‘rectum’ and ‘PTV’ pipes at 18 MV for two different irradiation geometries: IMRT and 4 field CRT, both measured at COOK. “PTV” pipe

“ ‘Bladder“ pipe

“Rectum“ pipe

(μSv Gy−1)

(μSv Gy−1)

(μSv Gy−1)

IMRT (18 MV)

26

39

23

4-field MLC (18 MV)

6

7

7

Varying the MV, while keeping the same treatment modality, dramatically affects the amount of photoneutrons produced because the average neutron energy and production efficiency increase with the maximum energy of the Bremsstrahlung photon spectrum. This effect is clearly shown in Figure 8.30, where two irradiations, delivered using the IMRT scheme, are shown for two energies: 6.and 18 MV.

Figure 8.30. Neutron dose equivalents as a function of depth from the phantom surface, using the IMRT protocol at 6 and 18 MV The final comparison shows in Figure 8.31 the difference between the photoneutron dose equivalents measured for the multileaf collimator-based conformal treatment modality at two energies: 15 and 18 MV. Despite the relatively small 3 MV difference in the acceleration voltage, it can be noticed once again that the higher photoneutron production at 18 MV, compared to 15 MV, results once again in a higher overall dose. The spatial dose distribution, however, is similar, so the highest difference may be noticed at the treatment volume.

82



Figure 8.31. Dose profiles in PTV, bladder and rectum in organ-mimicking pipes – comparison between measurements by SDD, BDPND and PADC. TPS photon dose profiles are reported: 15 MV 5-field MLC (left) and 18 MV 4-field MLC (right) treatments are compared.


83


References Bordy, J.M., d'Agostino, E., Bessiere, I., Domingo, C., d'Errico, F., di Fulvio, A., Knežević, Ž., Miljanić, S., Olko, P., Ostrosky, A., Poumarede, B., Sorel, S., Stolarczyk, L., Vermersse D., Harrison, R., 2013. Radiotherapy out-of-field dosimetry: Experimental and computational results for photons in a water tank. Radiat. Meas. 57, 29-34. IAEA, 2000. Absorbed Dose Determination in External Beam Radiotherapy. An International Code of Practice for Dosimetry Based on Standards of Absorbed Dose to Water. Technical Reports Series No. 398. International Atomic Energy Agency, Vienna. Ruben, J. D., Lancaster, C. N., Jones, P., Smith, R. L., 2011. A comparison of out-of-field dose and its constituent components for intensity-modulated radiation therapy versus conformal radiation therapy: D'Errico, F,. Nath, R., Tana, L., Curzio, G., Alberts W.G. 1998 In-phantom dosimetry and spectrometry of photoneutrons from an 18 MV linear accelerator, Med Phys, 25 1717-1724. Di Fulvio, A., Domingo, C., De San Pedro, M., D'Agostino, E., Caresana, M., Tana, L., d'Errico F. Superheated emulsions and track etch detectors for photoneutron measurements 2013 Radiat Meas, 57 19-28. Tosi, G., Torresin, A., Agosteo, S. , Foglio Para, A., Sangiust, V. , Zeni, L., Silari, M., Neutron measurements around medical electron accelerators by active and passive detection techniques 1991Med Phys, 18 54-60.

84



9. Discussion: BOMAB phantom results 9.1 Discussion of measurements and results (photons)

9.1.1. Comparison of dosimetry systems (photons) The comparison of RPL and TL dosimeters under the same irradiation conditions in a water tank showed that dosimeters generally agreed to within 3% in out-of-field regions compared with ionization chamber reference measurements in a reference clinical linac beam. For OSL, dosimeter response was corrected for spectra at different depths in water (Bordy et al., 2013). Variation of dosimeter measurements in BOMAB-like phantoms for different modalities was larger, with standard deviations ranging from 6 to 10%. A possible reason is due to the changing gantry angles during the irradiations that could influence the angular dependence of dosimeters that was not the case in 1-field irradiation under reference conditions. For RPL of the same glass dimensions as in our work, Son et al. (2011) showed that the variation of sensitivity at an angle of 0o was almost 9% lower in comparison to 90o for a 6 MV photon beam. For nanoDot OSL dosimeters (5 mm diameter disk, 0.2 mm thick) when irradiated with the incident photon beam parallel to the plane of the dosimeter, nanoDot response was 4% lower at 6 MV and 3% lower at 18 MV than the response when irradiated with the incident beam normal to the plane of dosimeter (Kerns et al., 2011). For LiF, no data on angular dependence could be found in the literature for dosimeters with dimensions used in these experiments. The neutron contribution to the dose absorbed in dosimeters could be another source of uncertainty, but according to previous published data, sensitivity to neutron absorbed dose probably could be neglected. When comparing different dosimeter types at different irradiation modalities in Figures 8.2 and 8.3 (and in all other irradiations) one can see that there are no systematic deviations between dosimeters, which suggest that the energy dependence of dosimeters is well compensated (in case of RPL) well corrected (in case of OSL) and can be neglected (for LiF based TLDs). In this investigation, the uncertainty in the mean dose decreases when the number of dosimeters of various types (and hence independent measurements) increases.

9.1.2. Components of out-of-field doses and their characteristics for different modalities The new advances in imaging, treatment planning and delivery are providing radiation oncologists with the ability to conform dose closely to the target (tumour) volume while minimizing the dose to organs at risk. However, this transition from 2DRT to 3DCRT and/or IMRT has resulted in clear changes to the dose distribution on which previous clinical experience and second malignancy studies are based. The review by Purdy (2008) shows that in general there is an increase in dose to the patient's target volume that includes the tumour and a limited amount of normal tissue, and an overall reduction in the volume of normal tissue receiving a high dose. However, particularly in the case of IMRT/IGRT, there is a larger volume of normal tissue that is irradiated to low radiation doses. Also compared to 2DRT and 3DCRT, IMRT requires a significantly larger number of monitor units (MUs) to deliver a comparable prescribed dose, which results in an increase in the whole body dose as a result of leakage and scattered radiation. Thus, there is some potential that this era of conformal therapy may actually result in an increased rate of secondary malignancy (Hall and Wuu, 2003). Dose deposited in a patient (or phantom) by x-rays consists of primary and secondary components. Primary dose is delivered by unscattered photons and is confined to the treatment


85


field. Kase et al. (1983) investigated the secondary radiation and their components for 3DCRT. They showed that radiation outside the treatment field arises from scatter generated by the patient (internal scatter), machine (collimator) scatter, and from leakage through the machine's shielding and collimators. They found that internal scatter can be described by a simple exponential function of distance from the central axis for all energies and field sizes. Machine scatter contributes 20 to 40% of total scattered dose depending on machine, field size and distance from the field. Leakage radiation contributes very little dose, but becomes the dominant component at the distances beyond 60 cm from the central axis. Ruben et al. (2011) investigated differences in scatter and leakage between 6 MV IMRT and 3DCRT to describe the relative contribution of internal patient scatter, collimator scatter and head leakage. They found that IMRT results in higher total dose to the patient than does 3DCRT. This increase is small in absolute terms and reductions in internal patient scatter with IMRT are outweighed by increased machine scatter and leakage, at least for small fields. Reductions from IMRT in dose to tissues from internal scatter, which predominates close to the field edge, means that calculations based solely on dose to distant tissues may overestimate carcinogenic risks. The relative contributions of different components of scatter and leakage are likely to vary with field size, beam energy, MU requirements of IMRT, and depth of measurements. Under conditions tested by Ruben et al. (2011), total machine scatter contributed 65% of the secondary dose for IMRT but only 30% for 3DCRT. It is important to point out that collimator scatter and head leakage are also dependent on linear accelerator and collimator design. Chofor et al. (2010) pointed out that any changes in beam head design, possibly capable of reducing the peripheral doses in photon therapy, have become a matter of great interest (Hall 2006). The fraction due to body/phantom scatter is largely unavoidable and slightly depends on photon energy. Two other contributions are technically modifiable and therefore avoidable. Numerous authors have measured out-of-field doses in several phantom designs, including water tanks and similar simple geometrical phantoms, and anthropomorphic phantoms. In an extensive review on the peripheral doses occurring in external photon-beam treatment, Xu et al. (2008) reported 23 studies that considered out-of-field dose from IMRT including Tomotherapy. They summarised some out-of-field measurements for contemporary radiotherapy techniques. Outside the treatment volume at a given distance from the isocentre, out-of-field doses can vary by an order of magnitude or more, depending upon the treatment technique simulated and the linear accelerator employed. However, the methods used to derive out-of-field doses differed from one study to another, making it difficult to assess the variability in out-of-field doses caused by geometry of the specific linac. Joosten et al. (2011) performed measurements on five different linacs and found that out-of-field doses could differ up to a factor 9 for small fields (5 x 5 cm2) and up to a factor of 10 for wedged fields. In our work, the organs’ positions and dimensions were exactly the same in both Pisa and Krakow irradiations and the main goal was to compare different modalities for the same irradiation conditions in these centres for the same type of linac. The results of comparison CRT 6 MV and IMRT 6 MV (Figure 8.13 and Table 8.4, Krakow) show that out-of-field doses for IMRT are lower than that of CRT only at the distances closer to field edge (4.6 cm). They are almost the same up to 14.6 cm and after become higher for IMRT showing a characteristic „bump“ with maximum at about 19.6 cm. For further distances, IMRT gives about 50% higher doses than CRT (for doses higher than about 1 mGy). Ruben et al. (2011) also found a prominent spike in the machine scatter (largely from leakage) over a distance of 5 cm, beginning 86



approximately 15 cm away from the edge of the field. They considered it a product of the treatment head geometry, namely, leakage radiation penetrating through the Y-jaw of the secondary collimator before passing through a gap between the lateral edge of the MLC and the primary collimator. The comparison of the same modalities with different energies (CRT 6 MV and 18 MV; IMRT 6 MV and 18 MV) in the COOK irradiations (Figure 8.14, Tables 8.3 and 8.4) shows that the peripheral photon doses are always significantly lower for higher energies. A similar finding was seen in the reference clinical linac (Saturn 43) beam in a water tank for radiation qualities of 6, 12 and 20 MV (Figure 8.15. right). It was found that when energy increases, water scatter decreases, collimator scatter slightly increases, leakage increases and the combined result is that total peripheral doses decrease as the energy increase (Bordy et al., 2013). The lower photon peripheral doses from 18 MV in comparison with 6 MV photon beams equipped with CRT were obtained by Stern (1999) and Mazonakis et al. (2008). Despite the fact that out-of-field doses decrease when energy increases, there is a further controversy about the optimal energy at which to conduct IMRT (Followill et al., 2007). While high-energy therapy may offer steeper dose gradients around the planning target volume and increased skin-sparing (Followill et al., 2007), it may also lead to an increased risk of secondary malignancy due to the presence of neutrons for energies above 8 MV (Kry et al., 2005b). Kry et al. (2005a) measured out-of-field doses for 6, 10, 15 and 18 MV IMRT. They found large differences in peripheral doses between Siemens and Varian accelerators; peripheral doses were lower, but not significantly, as energy increased. In a recently published paper, Kry et al. (2009) evaluated photon and neutron out-of-field dose equivalents for 6 and 18 MV IMRT using Monte Carlo studies, and showed that there are no significant differences in second cancer risk between them. Taking into account neutron dose components, Howell et al. (2006) calculated effective dose for 6, 15 and 18 MV IMRT and found that 6 MV resulted in the lowest effective dose, while 18 MV resulted in highest effective dose. The comparison of different modalities for irradiations in the Pisa in Krakow irradiations (Figure 8.13, Tables 8.3. and 8.4) shows that the highest peripheral doses were measured for Tomotherapy and somewhat less (but significantly higher than for IMRT) for VMAT. For Tomotherapy, the beam on time needed to deliver a given prescribed dose can be up to 15 times longer than that needed using conventional treatment delivery. Because of that there is concern that this technique has the potential to increase the whole body dose due to increased scatter and leakage as found by Mutic and Low (1998), Wiezorek et. al. (2009) and in our work, but despite that, Ramsay et al., (2006) found the peripheral doses equal to or less than the published peripheral doses for IMRT delivery in most clinical radiotherapy applications, explaining their results by noting that the Tomotherapy delivery system was designed to maximize the shielding for radiation leakage. For VMAT modalities there are currently no published data on peripheral doses.

9.1.3. Characteristics of TPS calculations for different modalities and comparison with dosimeter measurements The objective of delivering a therapeutic dose to a well-defined target while minimizing the dose to the surrounding normal tissue and critical organs requires optimization of conformity of the prescription dose to the planning target volume (PTV), dose homogeneity within the PTV, and dose to the surrounding normal tissue and critical organs. The priority of the treatment plan is to apply


87


the maximum dose to the tumour based on constraints of surrounding organs at risk. These dose constraints are based on clinical experience and aim to minimize side effects (normal tissue complications). Organs at risk identified for prostate cancer are bladder and rectum. Lowering the dose to these organs is an important part of the treatment planning process. Recently there have been a number of published papers dealing with comparison of 3DCRT, IMRT and novel forms of IMRT - VMAT and Tomotherapy - with regard to plan qualities and treatment efficiency for prostate cancer (Aoyama et al.2006; Palma et al., 2008; Wolff et al., 2009; Aznar et al., 2010; Zhang et al., 2010; Tsai et al., 2011) and for other cancer types (Bertelsen et al., 2010; Viellot et al., 2010; Lee et al., 2011; Lu et al., 2012). In all these papers it was claimed that VMAT and Tomotherapy show better treatment efficiency and usually better sparing of OARs. The most pronounced quality of VMAT is much lower MU and delivery time in comparison with conventional IMRT. In this work we compare treatment planning systems for different modalities and from two centres and with the results of dosimeter measurements. Dose profiles for the planning target volumes for prostate in Figure 8.17 show that out-of-field doses are quite different for different modalities and their order from the highest to the lowest values is the same as with dosimetry measurements shown in Figure 8.13. Differences in dosimetry results in Figure 8.16 between 6 MV IMRT in two centres can also be attributed to differences in TPS applied in two centres. In Figure 8.18, comparison of dosimeter readings with the TPS for OARs, bladder and rectum is shown for 3DCRT for SCUH and COOK irradiations. The comparisons of TPS and dosimeters are also shown in Figure 8.19 for 6MV in SCUH and COOK and in Figure 8.20 for 6 MV VMAT at SCUH. Dosimeters show larger values than TPS in all cases (the closest dosimeters values to TPS were for VMAT). Howell et al. (2010a) also determined that Eclipse TPS underestimated doses outside the treatment field by an average of 40% for a clinical treatment delivered on a Varian Clinac 2100. In Table 8.4, comparison of doses measured by dosimeters and from the TPS for different modalities per 2 Gy at isocentre is shown for distances for which comparison was possible by TPS (for further distances, results for TPS are not possible). At a distance of 4.6 cm from the field edge, the ratios of doses measured by dosimeters and that from TPS vary from 1.15 up to 2.25. Accordingly out-offield doses from TPSs should only be used with a clear understanding of the accuracy of dose calculations outside the treatment field. Whereas in-field radiation doses can be accurately and rapidly calculated using commercially available treatment planning systems, these TPSs do not, however, accurately model doses outside the treatment field, nor are they designed for such calculations. Studies that require accurate out-of-field dosimetry should use other dose reconstruction methods, such as measurements or simulated phantom calculations. Measurements in phantoms are accurate over a broad range of doses and closely reproduce the irradiation of a patient. Radiation dose measurements in anthropomorphic phantoms are considered the „gold standard“ in out-of-field dose assessment and have frequently been used to determine out-of-field organ doses in studies of radiation-induced late effects from photon radiotherapy (Howell et al., 2010b). In this study, a BOMAB like phantom was used which enables easier dose measurements with close approximation of clinical situations. The estimation of the sparing of adjacent sensitive organs from PTV data shown in Figure 8.21 is also in agreement with the shapes of the PTV curves in Figure 8.17. These results, contrary to the majority of the above-cited papers show the better results of bladder and rectum sparing for conventional IMRT techniques than for Tomotherapy and VMAT. As a conclusion it is important to point out the importance of investigation and measurement of out-of-field doses especially for the new radiotherapy modalities. 88



9.2 Discussion of measurements and results (neutrons)

9.2.1. Comparison of dosimetry systems (neutrons The neutron detectors used in this work, namely superheated emulsions and track etched detectors, exploit different properties of neutron interactions with matter (Apfel, 1979, Tommasino, 1984), detailed in section 3.2. Thus they show a different response, as a function of neutron energy. The superheated emulsion detectors consist of a suspension of fluoro-/halocarbon droplets, and their response compares well with the fluence-to-kerma equivalent conversion factor for fast neutrons. On the other hand, track etched detector response results from the neutron scattering and capture cross section of materials constituting the detector, i.e. carbon and hydrogen (scattering) or nitrogen (capture), placed on top of the actual detector layer to enhance its response to thermal neutrons. A specific calibration procedure, in terms of neutron equivalent dose, is needed for the assessment of this quantity in an unknown neutron field, like the one generated by high energy photons via photoneutron reactions in medical linacs. A suitable calibration also allows direct comparison of dosimeter readout quantities. The details of the calibration procedure have been discussed in Section 3.3. Superheated emulsion and track etched detectors were at first compared in the 10 x 10 cm2 reference field of the Saturn 43 clinical linac beam, at the calibration facility for radiotherapy dosimeters of the CEA/LIST Laboratoire National Henri Becquerel. Equivalent doses were measured inside a 30 x 30 x 60 cm3 water filled PMMA phantom. In the beam, data acquired with different systems agree within their 1SD uncertainties of about 20%. A combined uncertainty of 20% is generally accepted for neutron dosimetry measurements, as it factors counting statistics, source calibration uncertainties, detector temperature fluctuations, and importantly discrepancies between detector response and kerma factor. It is interesting to notice that at a distance > 20 cm from the isocentre, PADC detectors measure an equivalent dose about 50% higher compared to superheated emulsion based detectors. This can be explained by the fact that a single coefficient for fast neutrons was used in this case to convert from counted tracks to equivalent dose, regardless of the detector position inside the phantom, i.e. the neutron energy spectrum. The neutron spectrum average energy decreases with distance from the isocentre. The energy response of the PADC track detectors deviates significantly from the trend of the kerma equivalent coefficient. PADC detectors are more sensitive to thermal than fast neutrons, because of a nitrogen enriched foil covering the plastic foil. Thus the track density-to-equivalent-dose conversion coefficient has to take into account the impinging neutron energy spectrum. In the BOMAB phantom, irradiated using clinical protocols, the dispersion from the average of neutron equivalent dose measured by each system was about 20%. In this case, different calibration coefficients were determined for PADC and used for the different photoneutron spectra in the phantom. A much better agreement between the three detection instruments was achieved using this approach. Another important difference between superheated emulsions and track etched detector is their angular response. While SE have a cylindrical geometry, which guarantees a two-axis measurement isotropy, PADC response is strongly angle dependent (Domingo, 2013). The thermal component of the out-of-field photoneutron spectrum was approximated as isotropic, because it is produced by scattering reactions with the irradiation room walls and the patient. A PADC detector calibration factor, used to convert readout track density at a distance >20 cm from the isocentre, was derived by irradiating the detectors in a thermal and isotropic reference field at the Physikalisch-Technische


89


Bundesanstalt (Zimbal, 2009). The results presented here show that this assumption is appropriate, as neutron dose equivalent values within 20% (1 SD) are measured by SDD, BTI-BDPND and PADC detectors both at the target volume and outside, for IMRT, VMAT and MLC based conformal treatments. In conclusion, the dosimetry systems and methods used to measure in-phantom neutron equivalent doses in a variety of clinical radiotherapy fields showed consistent and repeatable responses with a ~20% uncertainty (1 SD).

9.2.2. Components of out-of-field doses and their characteristics for different modalities Several parameters characterize an external x-ray radiotherapy treatment: dose delivery rate; total dose at the target volume; number and frequency of treatment fractions; maximum photon energy and radiation delivery modality. These last two in particular affect the energy and spatial distribution of the photoneutron field. In this section, we will relate the main results in terms of out-of-field neutron equivalent dose to different treatment modalities, after clarifying the photoneutron dose dependence from the acceleration energy. Most linacs for external x-ray radiation therapy allow the selection of the electron acceleration voltage, typically between two or three values. In our case, the Varian Clinac 2300 features 6, 15 and 18 MV. High energies, i.e. 15 or 18 MV, are typically preferred to lower ones, i.e. 6 MV, because they show steeper photon dose gradients around the planning target volume and improve skinsparing. This effect is mainly due to reduced water scattering contributions at energies higher than 8 MeV (Bordy, 2013). However, photoneutron production cross section increases with energy, resulting in a neutron equivalent dose at the isocentre of 276 µSv Gy-1 and 576 µSv Gy-1 for an MLC based conformal treatment, at 15 MV and 18 MV respectively. 6 MV radiotherapy fields are typically considered free from a spurious photoneutron component. However in this work a non-negligible photoneutron dose with a 6 MV primary photon beam has been demonstrated, e.g. of 39 µSv Gy-1 for a 10 × 10 cm2 single field. Photoneutrons at energies lower than 8 MV are probably produced by light elements in the accelerator, whose neutron separation energy is low, i.e. 1.66 MeV for beryllium, which typically constitutes the accelerator exit windows. Photoneutron equivalent doses delivered by different treatment modalities, i.e. MLC-based conformal treatments, IMRT, VMAT and helical Tomotherapy, should be compared if delivered at the same acceleration energy. The photoneutron spectrum inside a treatment room results from two main contributions: (i) a direct component which is produced by high energy photon interactions with the x-ray target and the beam flattening filter, typically made by tungsten and steel, respectively; and (ii) an evaporation component, characterized by a Maxwellian spectrum (Tosi, 1991). This radiation, which does not contribute to the treatment, is often referred to as “leakage radiation”, resulting in an unwanted whole body irradiation of the patient. At the patient, a further thermal cloud is generated by fast neutron scattering from soft tissue light elements, such as hydrogen and carbon. New treatment modalities, such as 3D conformal radiotherapy and IMRT, achieve a precise irradiation of the tumour volume, by using multiple beams delivered by a rotating gantry. This approach allows delivering a lower dose to normal tissues around the tumor target, since the beam interaction is overall distributed over a larger volume of normal tissues. However, the time needed to adjust the gantry and couch for each field and the intrinsic segmented delivery pattern leads to an increased number of monitor units and thus irradiation time and whole body exposure by 90



leakage radiation. For example, in Tomotherapy, the beam-on time needed to deliver a certain dose can be up to 15 times longer than that needed in conventional radiotherapy, depending on the slice thickness (i.e. field size) and the number of rotations involved (Ramsey, 2006). We define peripheral secondary neutron dose as the equivalent neutron dose, averaged in all the 5 BOMAB phantom pipes which contained dosimeters, at a distance >10 cm from the field edge. At high energies, i.e. 18 MV, IMRT and 3DCRT were compared, which required 175 MU and 111 MU to deliver a 2 Gy fraction photon dose, respectively. A peripheral secondary neutron dose for IMRT of 30 μSv Gy−1, nearly four times the peripheral dose measured in 3DCRT of 7 μSv Gy−1 was measured. As expected, the number of secondary photoneutrons generated increases with beam-on time, i.e. monitor units. This result confirms the findings of Howell et al. (2006), who compared a conventional four-field plan with gantry angles of 0°, 90°, 180° and 270° with optimized dynamic IMRT with gantry angles of 225°, 285°, 0°, 75°, and 135° at the energies of 6, 15 and 18 MV. They found that IMRT neutron fluences were nearly twice those generated by a 3DCRT protocol, both at the isocentre and at a plane 40 cm above the isocentre. On the other hand, at 6 MV, when compared to VMAT, IMRT shows lower peripheral neutron dose, i.e. a better sparing of organs at risk close to the treatment volume, probably because of the finite number of beam delivering positions for IMRT. Among low energy treatments, Tomotherapy is the less affected by neutrons. In helical Tomotherapy, two aspects mitigate the unwanted photoneutron dose: the reduced number of monitor units per 2Gy photon fraction (Tomotherapy 133 MU, VMAT 241 MU, IMRT 216 MU) and the shape of the collimator. A time effective treatment is achieved by Tomotherapy protocol by delivering the treatment with the gantry and the couch in simultaneous motion. Moreover, similarly to CT, the radiation is delivered slice-by-slice, so that a smaller tungsten multileaf collimator is needed.


91


References Aoyama, H., Westerly, D. C., Rocwell, M. T., Olivera, G., Bentzen, S. M., Patel, R. R., Jaradat, H., Tome, W. A., Ritter, M. A., Mehta, M. P., 2006. Integral radiation dose to normal structures with conformal external beam radiation. Int. J. Radiation Oncology Biol. Phys. 64, 962-967. Apfel, R.E., The superheated drop detector, Nuclear Instruments and Methods, 162 (1979) 603-608. Aznar, M. C., Petersen, P. M., Logadottir, A., Lindberg, H., Korreman, S. S., Kjær-Kristofersen, F., Engelholm, S. A., (2010). Rotational radiotherapy for prostate cancer in clinical practice. Radiotherapy and Oncology 97, 480-484. Bertelsen, A., Hansen, C. R., Johansen, J., Brink, C., (2010). Single Arc Volumetric Arc therapy of head and neck cancer. Radiother. Oncol. 95(2), 142-148. Bordy, J.M., Bessiere, I., d'Agostino, E., Domingo, C., d'Errico, F., di Fulvio, A., Knežević, Ž., Miljanić, S., Olko, P., Ostrosky, A., Poumarede, B., Sorel, S., Stolarczyk, L., Vermersse D., Harrison, R., 2013. Radiotherapy out-of-field dosimetry: Experimental and computational results for photons in a water tank. Radiat. Meas. 57, 29-34. Chofor, N., Harder, D., Rüchmann, A., Wilborn, K. C., Wiezorek, T., Poppe, B. 2010. Experimental study of photon-beam peripheral doses, their components and some possibilitiesfor their reduction. Phys. Med. Biol. 55, 4011-4027. Domingo, C., et al., Estimation of the response function of a PADC based neutron dosimeter in terms of fluence and Hp(10), Radiation Measurements, 50 (2013) 82-86. Followill, DS. Nusslin, P., Orton, C., 2007. IMRT should not be administered at photon energies greater than 10 MV. Med. Phys. 34, 1877-1879. Hall E. J., Wuu, C-S., 2003. Radiation-induced second cancer: the impact of 3D-CRT and IMRT. Int. J. Radiation Oncology Biol. Phys. 56(1), 83-88. Hall, E. J., 2006. Intensity-modulated radiation therapy, protons, and the risk of second cancers. Int. J. Radiation Oncology Biol. Phys. 65(1), 1-7. Howell, R.M. et al., Measurements of secondary neutron dose from 15 Mv and 18 Mv IMRT, Radiat. Prot. Dosim., 115 (1–4) (2005), pp. 508–512 Howell, R. M., Hertel, N. E., Wang, Z., Hutchinson, J., Fullerton, G. D., 2006. Calculation of effective dose from measurements of secondary neutron spectra and scattered photon dose from dynamic MLC IMRT for 6 MV, 15 MV, and 18 MV beam energies. Med. Phys. 33(2) 360-368. Howell, R. M., Scarboro, S. B., Kry, S. F., Yaldo, D. Z., 2010a. Accuracy of out-of-field dose calculations by a commercial treatment planning system. Phys.Med.Biol. 55, 6999-7008. Howell R. M., Scarboro, S. B., Taddei, P. J., Krishnan, S., Kry, S. F., Newhauser, W. D., 2010b. Methodology for determining doses to in-field, out-of-field and partially in-field organs for late effects studies in photon radiotherapy. Phys.Med.Biol. 55, 7009-7023. 92



Joosten, A., Bochud, F., Baechler, S., Levi, F., Mirimanoff, R-O., Moeckly, R. 2011. Variability of a peripheral dose among various linac geometries for second cancer risk assessment. Phys. Med. Biol. 56, 5131-5151. Kase, K. R., Svensson, G. K., Wolbarst, A. B., Marks, M. A., 1983. Measurements of dose from secondary radiation outside a treatment field. International Jornal of Radiation Oncology Biology Physics 9(8), 1177-1183. Kerns, J. R., Kry, S. F., Sahoo, N., Followill, D. S., Ibbott, G. S., 2011. Angular dependence of the nanoDot OSL dosimeter. Med. Phys. 38(7), 3955-3962. Kry, S. F., Salehpour, M., Followill, D. S., Stovall, M., Kuban, D. A., White, R. A., Rosen, I. I., 2005a. Outof-field photon and neutron dose equivalents from step-and-shoot intensity-modulated radiation therapy. Int. J. Radiation Oncology Biol. Phys. 62(4), 1204-1216. Kry, S. F., Salehpour, M., Followill, D. S., Stovall, M., Kuban, D. A., White, R. A., Rosen, I. I., 2005b. The calculated risk of fatal secondary malignancies from intensity -modulated radiation therapy. Int. J. Radiation Oncology Biol. Phys. 62(4), 1195-1203. Kry, S. F. Salehpour, M., Titt, U., White, R. A., Stovall, M., Followill, D. S., 2009. Monte Carlo study show no significant difference in second cancer risk between 6- and 18-MeV intensity modulated radiation therapy. Radiotherapy and Oncology 91, 132-137. Lee, T-F., Chao, P-J., Ting, H-M., Lo, S-H., Wang, Y-W., Tuan, C-C., Fang, F-M., Su, T-J., 2011. Comparative analysis of SmartArc-based dual arc volumetric-modulated arc radiotherapy (VMAT) versus intensity-modulated radiotherapy (IMRT) for nasopharyngeal carcinoma. Journal of Applied Clinical Medical Physics 12(4), 158-174. Lu, S-H., Cheng, J. C-H., Kuo, S-H., Lee, J. J-S., Chen, L-H., Wu, J-K., Chen, W-Y., Chong, F-C., Wu, C-J., Wang, C-W., (2012). Volumetric modulated arc therapy for nasopharyngeal carcinoma: A dosimetric comparison with Tomotherapy and step-and-shoot IMRT. Radiotherapy and Oncology (in press). Mazonakis, M., Zacharopoulou, F., Varveris, H., Damilakis, J., 2008. Peripheral dose measurements for 6 and 18 MV photon beams on a linear accelerator with multileaf collimator. Med. Phys. 35(10), 4396-4403. Mutic, S. Low, D. A., 1998. Whole-body dose from tomotherapy delivery. Int. J. Radiation Oncology Biol. Phys. 42(1), 229-232. Palma, D., Vollans, E., James, K., Nakano, S., Moiseenko, V., Shaffer, R., McKenzie, M., Morris, J., Otto, K., (2008). Volumetric modulated arc therapy for delivery of prostate radiotherapy: comparison with intensity-modulated radiotherapy and three-dimensional radiotherapy. Int. J. Radiation Oncology Biol. Phys. 72(4), 996-1001 Purdy, J., 2008. Dose to normal tissue outside the radiation therapy patient’s treated volume: A review of different radiation therapy techniques. Health Physics 95(5), 666-676. Ramsey, C. R., Seibert, R., Mahan, S. L., Desai, D., Chase, D., 2006. Out-of-field dosimetry measurements for a helical tomotherapy system. J. Appl. Clin. Med. Phys. 7(3), 1-11.


93


Ruben, J. D., Lancaster, C. N., Jones, P., Smith, R. L., 2011. A comparison of out-of-field dose and its constituent components for intensity-modulated radiation therapy versus conformal radiation therapy: implications for carcinogenesis. Int. J. Radiation Oncology Biol. Phys. 81, 1458-1464. Son, K., Jung, H., Shin, S. H., Lee, H-H., Kim, M-S., Ji, Y. H., Kim, K. B., 2011. Evaluation of the dosimetric characteristics of a radiophotoluminescent glass dosimeter for hig-energy photon and electron beams in the field of radiotherapy. Radiation Merasurements 46, 1117-1122. Stern, R. L., 1999. Peripheral dose from a linear accelerator equipped with multileaf collimation. Med. Phys. 26(4), 559-563. Tommasino, L., et al., Different etching processes of damage track detectors for personnel neutron dosimetry, Nuclear Tracks and Radiation Measurements (1982), 8 (1984) 335-339. Tosi, G. et al., Neutron measurements around medical electron accelerators by active and passive detection techniques Med. Phys., 18 (1) (1991), pp. 54–60 Tsai, C-L., Wu, J-K., Chao, H-L., Tsai, Y-C., Cheng, J. C-H., 2011. Treatment and dosimetric advantages between VMAT, IMRT, and helical tomotherapy in prostate cancer. Medical Dosimetry 36, 264-271. Viellot, S., Azria, D., Lemanski, C., Moscardo, C. L., Gourgou, S., Dubois, J-B., Aillères, N., Fenoglietto, P., 2010. Plan comparison of volumetric arc therapy (RapidArc) and conventional intensitymodulated radiation therapy (IMRT) in anal canal cancer. Radiation Oncology, 5:92. Wiezorek, T., Schwahofer, A., Schubert, K., 2009. The influence of different IMRT techniques on the peripheral dose - a comparison between sMLM and helical Tomotherapy. Strahlentherapie und Onkologie, 185, 696-702. Wolff, D. Stieler, F., Welzel, G., Lorenz, F., Abo-Madyan, Y., Mai, s., Polednik, M., Steil, V., Wenz, F., Lohr, F., 2009. Volumetric modulated arc therapy (VMAT) vs. serial tomotherapy, step-and shoot IMRT and 3D-conformal RT for treatment of prostate cancer. Radiotherapy and Oncology 93, 226233. Xu, X. G., Bednarz, B., Paganetti, H., 2008. A review of dosimetry studies on external-beam radiation treatment with respect to second cancer induction. Phys. Med. Biol. 53, R193-R241. Zhang, P., Happersett, L., Hunt, M., Jackson, A., Zelefsky, M., Mageras, G., 2010. Volumetric modulated arc therapy: Planning and evaluation for prostate cancer cases. Int. J. Radiation Oncology Biol. Phys. 76, 1456-1462. Zimbal, A., Personal Communication, (2009)

94



10. Risk factors In this report, the emphasis has been on the dosimetric principles and practice in the measurement of out-of-field doses in radiotherapy. The underlying motivation for this work is to provide a sound foundation for the subsequent estimation of the risk of second cancer induction following radiotherapy. The relationship between dose and risk is however, quite complex and in some areas poorly understood. In this section, we give a brief introduction to some current risk models to illustrate the inherent difficulties and also the advances which have been made in recent years. The text has been adapted in parts from (Harrison, 2013a). A useful schematic graph illustrating dose-risk relationships over a wide range of doses - given by (Hall, 2004) and reproduced subsequently by several other authors - is shown in Figure 10.1.

Figure 10.1: Schematic graph of cancer risk v. equivalent dose (after Hall, 2004) Here, the central portion of the graph derives largely from the atomic bomb survivor data. These data have been extensively analysed and form the basis of the linear no-threshold (LNT) model for radiation protection purposes. If equivalent doses to organs at risk are between approximately 0.05 and 2.5 Sv, then it is reasonable to adopt the LNT model and organ-specific risks may be estimated using appropriate “low dose” risk factors given by committees of several international bodies, for example, ICRP, NCRP, UNSCEAR, BEIR (ICRP, 1991; NCRP, 1993; UNSCEAR, 2000; BEIRVII, 2006; ICRP, 2007). A study carried out by the former UK National Radiological Protection Board (NRPB) on the epidemiology of second cancers also concluded that for most cancer sites studied, the relative risks for the induction of second cancers were comparable to or less than the relative risks derived from the Japanese data, although the differences were sometimes not statistically significant when data were subdivided between cancer types (NRPB, 2000). However, it is important to note that risk factors derived for radiation protection of populations, and especially the working population exposed to low occupational doses, may not necessarily be applicable to radiotherapy patient populations (which may contain a higher proportion of individuals with genetic predisposition to cancer) and are not appropriate for individual patients. These risk data rely on a transfer from


95


Japanese to other populations and some risk estimates are age and sex averaged (although age dependent risks have been given, for example by BEIR VII (BEIRVII, 2006). Much of the motivation for second cancer studies in radiotherapy has derived from the acknowledgement of the “radiation bath” to which all patients are subjected and the realisation that new methods and techniques in radiotherapy (e.g. 3D CRT, IMRT, Tomotherapy and robotic arm techniques) may lead to considerable axial variation in the magnitude of the “radiation bath”. This is reflected in the differing out-of-field doses which have been reported by several groups. For example, (Palm and Johansson, 2007) summarised the dose equivalent per treatment gray as a function of the distance from the field edge, reported by several groups and showing significant differences between treatment methods including “conventional” conformal radiotherapy, IMRT and tomotherapy. Taken over the whole body, this is a highly non-uniform dose distribution (ranging from the prescribed dose within the target volume – tens of Gy – to a few tens of mGy at the patient extremities. Thus, remote from the target volume, organs and tissues, which themselves have varying radiosensitivities, will be subjected to very different doses and therefore second cancer risks. Thus the dose response across almost the whole range of doses shown in Figure 10.1 is required. This necessitates consideration of the breakdown of the LNT hypothesis particularly at higher doses (> 2.5 Sv) and is illustrated in Figure 10.1 as a region where the linear relationship is modulated by cell kill (and other factors such as cell re-population) with the consequent peaking and decrease in risk with increasing dose. This high dose region is particularly important since it has been shown that the majority of second cancers were observed within the Planning Target Volume (PTV) margins (Dörr and Herrmann, 2002). At the other end of the scale, the low dose region (< 0.05 Sv, approximately) has been the subject of study because of various hypotheses which suggest departure from the LNT model (e.g. bystander effects, adaptive responses). This area is of less importance in radiotherapy because the risks in this dose region will be low and likely to be considerably less significant than those arising from the inevitable high dose regions. The organisations which have developed risk factors for radiation protection purposes have usually considered two basic risk models, additive and multiplicative. After a latent period following a single dose exposure, the additive model assumes that the excess probability rate of death from cancer is proportional to the dose. In the multiplicative model, again after a latent period following a single exposure, the excess probability rate of death from cancer is assumed to be proportional both to the dose and also to the background rate of cancer death (ICRP, 1991). These two models together with the choice of a dose and dose rate effectiveness factor (DDREF) to allow for the high dose rates experienced by the Japanese survivors, formed the basis of considerations by ICRP and led to the risk coefficients shown in Table 10.1. These data were also adopted by NCRP and with modification by the US Environmental Protection Agency, have been used to estimate second cancer risks from out-of-field dose measurements (Kry et al., 2007).

96



Table 10.1: ICRP risk factors (ICRP 1990) Organ or tissue

Coefficient of lifetime risk of fatal cancer (MT) (% Sv-1)

Bladder Bone marrow Bone surface Breast Oesophagus Colon Kidney Liver Lung Ovary Skin Stomach Thyroid Remainder Total

0.30 0.50 0.05 0.20 0.30 0.85 0.05 0.15 0.85 0.10 0.02 1.10 0.08 0.50 5.00

The absolute lifetime risk of developing a second cancer, RT , in organ T is given by: (10.1)

𝑅𝑇 = 𝐻𝑇 . 𝑀𝑇

Where HT is the equivalent dose to the organ T and MT is the coefficient of lifetime risk of fatal cancer. The total absolute lifetime risk of developing a second cancer is calculated by summing RT over all irradiated organs. These data have also been used by Miralbell et al. (Miralbell et al., 2002) and Newhauser et al. (Newhauser et al., 2009) for similar purposes, except that these authors used lethality fractions to generate the combined probability of both fatal and non-fatal cancers. More recently, the BEIR VII model for radiocarcinogenisis has been used to calculate second cancer risk estimates following breast radiotherapy and associated imaging (Donovan et al., 2012). In this model, both Excess Relative Risk (ERR) and Excess Absolute Risk (EAR) have been combined to give the Lifetime Attributable Risk (LAR) as follows: 𝑎 𝜂

𝐸𝑅𝑅 𝑎𝑛𝑑 𝐸𝐴𝑅 = 𝛽𝑆 . 𝐷. exp (𝛾 𝑒 ∗ ) . �60�

(10.2)

where:

D = dose (Sv) βS, ϒ and η are organ and sex specific parameters for ERR and EAR given by BEIR VII

a = attained age e = age at exposure EURADOS Report 2017-01

97


e* = (e − 30)/10 for e < 30 and 0 for e > 30 yr

𝐸𝑅𝑅 = �

the rate of disease in an exposed population �−1 the rate of disease in an unexposed population

𝐸𝐴𝑅 = (the rate of disease in an exposed population) − (the rate of disease in an unexposed population) and

𝑆(𝑎)

𝐶 𝐿𝐴𝑅(𝐷,𝑒) = �∑90 𝑎 𝐸𝑅𝑅(𝐷, 𝑒, 𝑎 ) . 𝜆𝐼 . 𝑆(𝑒) . 𝛥𝑎�

𝑘𝐸𝑅𝑅

where:

× �∑90 𝑎 𝐸𝐴𝑅(𝐷, 𝑒, 𝑎 ) .

𝑆(𝑎) 𝑆(𝑒)

. 𝛥𝑎�

𝑘𝐸𝐴𝑅

(10.3)

𝜆𝐶𝐼 = baseline cancer risk

S(a)/S(e) = probability of surviving to the attained age (a) conditional on survival to exposed age (e) Values of kERR and kEAR are given in Table 10.2

Table 10.2: values of k ERR and kEAR Organ

kERR

kEAR

Breast

0

1.0

Thyroid

1.0

0

Lung

0.3

0.7

All other organs given by BEIR VII

0.7

0.3

BEIR VII gives incidence and mortality data for males and females as a function of age at exposure. Several authors have described so-called “biologically based” models in which the population dynamics of normal and radiation-mutated pre-malignant stem cells is modelled following fractionated radiotherapy with realistic absorbed doses (Wheldon et al., 2000; Lindsay et al., 2001; Sachs and Brenner, 2005; Sachs et al., 2007). A useful starting point for these ideas is the concept that ERR is the product of two terms, the first containing radiation parameters such a dose, dose rate and fractionation (the dose terms in brackets in equation 10.4) and the second (term B in equation 10.4) containing dose and fractionation-independent parameters such as time since irradiation, age at irradiation, gender and ethnicity. Thus : 𝐸𝑅𝑅 = {(𝑎𝑑 + 𝑏𝑑2 ) . 𝑒𝑥𝑝 − (𝛼𝑑 + 𝛽𝑑2 )} . B

(10.4) 98



where d is the dose per fraction and a, b, α and β are model parameters. Both deterministic and stochastic models have been developed which indicate the shape of dose response curves following a variety of starting conditions and parameters in three main components of the population dynamics (i) linear quadratic expressions for the initiation of target cells to form pre-malignant cells (ii) linear-quadratic cell survival following irradiation and (iii) cell proliferation. Whether based on deterministic or stochastic arguments, these models have sometimes been referred to collectively as IIP Initiation / Inactivation / Proliferation) models. One of the practical problems with this approach is that the more fundamental the biological starting point, the more parameters (whose values may be inexactly known) are needed to model the subsequent cellular processes following irradiation. However, simplifications may be made which require only the atomic bomb and population data (Sachs et al., 2005). Although not necessarily inconsistent with biologically-based models, some epidemiological studies have indicated dose-risk relationships which suggest a plateau at high doses (Hall and Wuu, 2003). One practical model is described here as an example of how a variety of non-linear dose-response shapes may be accommodated within a model which also takes account of non-uniform organ irradiation. This uses the concept of Organ Equivalent Dose (OED) (Schneider and Kaser-Hotz, 2005a; Schneider et al., 2005b). The starting premise is that various non-uniform dose distributions in an organ are equivalent and correspond to the same OED if they cause the same radiation-induced risk of cancer. Thus for a particular organ and a linear dose response, the OED is given by:  Vi  linear OEDrad  . Di −ther = ∑  i V 

(10.5)

Where:

Vi is the volume of voxel i within the organ V is the total volume of the organ Di is the dose to voxel i. 𝑙𝑖𝑛𝑒𝑎𝑟 For the case of a linear dose response relationship, 𝑂𝐸𝐷𝑟𝑎𝑑−𝑡ℎ𝑒𝑟 is simply the mean organ dose. (The subscript rad-ther has been suggested by Schneider et al. (Schneider et al., 2005a) to avoid confusion with the quantity Equivalent Dose).

Other dose models can be similarly described, i.e. for a linear-exponential dose response:

 Vi  linear −exp OEDrad  . Di . exp − (α org . Di ) −ther = ∑  i V 

(10.6)

and for a plateau response:  Vi  1 − [exp− (δ org . Di )] plateau OEDrad . −ther = ∑  δ org i V 

(10.7)

where αorg and δorg are parameters associated with the organ response. The risk of radiationinduced cancer in the organ (Iorg) is then given by: mod el I org = I 0org . OEDrad −ther


(10.8) 99

R.M. Harrison, A. Di Fulvio, J-M. Bordy, S. Miljanić, L. Stolarczyk and Z. Knežević 𝑜𝑟𝑔

Where 𝐼0

𝑚𝑜𝑑𝑒𝑙 is the low dose risk of radiation-induced cancer per unit dose and 𝑂𝐸𝐷𝑟𝑎𝑑−𝑡ℎ𝑒𝑟 is the 𝑜𝑟𝑔

Organ Equivalent Dose for one of the response models given in equations 10.5-10.7. 𝐼0 is given in several reviews of low dose cancer induction risk, e.g. ICRP, NCRP, UNSCEAR and BEIR VII (ICRP, 1991; NCRP, 1993; UNSCEAR, 2000; BEIRVII, 2006; ICRP, 2007). Schneider has also provided a mechanistic model in which variations in the re-population fraction following irradiation give a family of dose-risk responses which vary from linear-exponential to plateau responses (Schneider, 2009). Other similar models have been proposed (Dasu et al., 2005, Takam et al., 2009) where linear quadratic relationships have been used and allowance made for fractionation. Thus, for a particular organ:

V  Total risk effect = ∑  i  . E ( Di ) i V 

(10.9)

where E(D) is a given non-linear dose-response relationship, such that

  β D2  β D2   E ( D) =  α 1 D + 1  . exp−  α 2 D + 2 n  n   

(10.10)

where α1 and β1 are the linear and quadratic coefficients for the induction of DNA mutations and α2 and β2 are the linear and quadratic coefficients for cell kill. n is the number of fractions over which the total dose D was delivered. All these models are amenable to easy implementation, particularly when dose-volume histogram (DVH) data is available from treatment planning systems. DVH data essentially give the fraction of the organ volume which receives a given dose, i.e values of Di and Vi where the voxels, i, may be of differing volumes. Nevertheless, conventional risk estimation based on the use of dose equivalent or equivalent dose determinations with the applications of appropriate radiation weighting factors may, at least for dose equivalents less than approximately 2.5 Sv, be the most appropriate with the current knowledge base. In contrast, the OED model is open to the criticism that the parameters αorg and δorg have been derived from specific irradiated populations (patients treated for Hodgkin’s disease) but this approach nevertheless provides a promising methodology for risk estimation at the higher doses found closer to the target volume. Second cancer risks may in many cases be acceptably small. Kry et al. showed that the absolute risk of inducing a fatal malignancy following prostate radiotherapy was approximately 2-5%, for several different treatments, accelerator types and endpoint energies, although IMRT gave higher risks than conventional radiotherapy (Kry et al., 2005). Risks should nevertheless be quantified, so that robust risk-benefit judgements may be made. This is particularly relevant for radiotherapy treatments of younger patients who may have long survival prospects and for whom risk factors are higher (BEIRVII, 2006). This is illustrated in Figure 10.2, (taken from (CRUK, 2012) which shows the increasing survival over time for children aged 0-14 years who have received treatment for all forms of lymphoma. For a 10 year old male, the lifetime attributable risk of cancer incidence (for all cancers) is 14.5% Gy-1, compared with 5.9% Gy-1 for a 50 year old male (from BEIR VII data).

100


Dosimetry for second cancer risk estimation in radiotherapy: measurements in water phantoms 5-year survival rates in children: lymphomas age 0-14 y 100 90

survival %

80 70 60 50 40 30 20 10 0

period of diagnosis

Figure 10.2: 5-year survival rates in children treated for lymphomas. From Cancer Research UK (http://info.cancerresearchuk.org/cancerstats.)

Whatever the difficulties and uncertainties in risk estimation, the starting point is the absorbed dose to the irradiated organs. Thus the measurement of out-of-field (sometimes referred to as peripheral) doses, from which specific organ doses may be inferred, is a crucial pre-requisite for risk estimation. The ability to reconstruct organ doses retrospectively will be an essential component of epidemiological studies. A Workshop, held at the Annual Meeting of the European Radiation Dosimetry Group (EURADOS) in Vienna in 2012 reviewed some current progress in out-of-field dosimetry for second cancer risk estimation, largely from the work of EURADOS Working Group 9 and the dosimetric data given in these proceedings (Harrison et al. 2013b) are described in more detail in this report.


101


References Harrison, R.M., 2013a. Introduction to dosimetry and risk estimation of second cancer induction following radiotherapy. Radiation Measurements 57, 1-8 Harrison, R.M. 2013b (Editor) Proceedings of the Workshop: Dosimetry for Second Cancer Risk Estimation. EURADOS Annual Meeting, Vienna 2012. Radiation Measurements. Special Issue 57 ISSN 1350-4487 Hall, E.J., 2004. Henry S. Kaplan Distinguished Scientist Award 2003:The crooked shall be made straight; dose–response relationships for carcinogenesis. Int. J. Radiat. Biol. 80, 327–337. ICRP. 1991. 1990 Recommendations of the International Commission on Radiological Protection Pergammon Press. NCRP. 1993. National Council on Radiation Protection and Measurements. Limitation of Exposure to Ionizing Radiation. NCRP Report 116. Bethesda (MD): National Council on Radiation Protection and Measurements. UNSCEAR. 2000. Report to the general assembly, sources and effects of ionizing radiation, Volume II: Effects, Annex I: Epidemiological evaluation of radiation-induced cancer. BEIRVII. 2006. Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII – Phase 2. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, National Research Council The National Academies Press, Washington DC. ICRP. 2007. ICRP Publication 103: The 2007 Recommendations of the International Commission on Radiological Protection Elsevier. NRPB. 2000. Risks of Second Cancer in Therapeutically Irradiated Populations. National Radiological Protection Board, Chilton, Didcot, Oxon, OX11 0RQ UK. Palm, A., Johansson K.-A., 2007. A review of the impact of photon and proton external beam radiotherapy treatment modalities on the dose distribution in field and out-of-field; implications for the long-term morbidity of cancer survivors. Acta Oncologica 46, 462-473. Dörr, W., Herrmann T., 2002. Second Primary Tumors after Radiotherapy for Malignancies. Treatment-Related Parameters. Strahlentherapie und Onkologie 178, 357–362. Kry, S.F., Followill D., White R.A., Stovall M., Kuban D.A., Salehpour M., 2007. Uncertainty of calculated risk estimates for secondary malignancies after radiotherapy. Int. J. Radiation Oncology Biol. Phys. 68, 1265–1271. Miralbell, R., Lomax A., Cella L., Schneider U., 2002. Potential reduction of the incidence of radiationinduced second cancers by using proton beams in the treatment of pediatric tumors. Int J Radiat Oncol Biol Phys 54, 824-829.

102



Newhauser, W.D., Fontenot J.D., Mahajan A., Kornguth D., Stovall M., Zheng Y., Taddei P.J., Mirkovic D., Mohan R., Cox J.D., Woo S., 2009. The risk of developing a second cancer after receiving craniospinal proton irradiation. Phys. Med. Biol. 54, 2277–2291. Donovan, E., James H., Bonora M., Yarnold J.R., Evans P.M., 2012. Second cancer incidence risk estimates using BEIR VII models for standard and complex external beam radiotherapy for early breast cancer. Medical Physics 39, 5814-5824. Wheldon, E.G., Lindsay K.A., Wheldon T.E., 2000. The dose-response relationship for cancer incidence in a two-stage radiation carcinogenesis model incorporating cellular repopulation. Int. J. Radiat. Biol. 76, 699-710. Lindsay, K.A., Wheldon E.G., Deehan C., Wheldon T.E., 2001. Radiation carcinogenesis modelling for risk of treatment-related second tumours following radiotherapy. Br J Radiol 74, 529-536. Sachs, R.K., Brenner D.J., 2005. Solid tumour risks after high doses of ionizing radiation. Proceedings of the National Academy of Sciences 102, 13040-13045. Sachs, R.K., Shuryak I., Brenner D., Fakir H., Hlatky L., Hahnfeldt P., 2007. Second cancers after fractionated radiotherapy: Stochastic population dynamics effects. Journal of Theoretical Biology 249, 518-531. Hall, J.D., Wuu C.-S., 2003. Radiation-induced second cancers: The impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 56, 83-88. Schneider, U., Kaser-Hotz B., 2005a. Radiation risk estimates after radiotherapy: application of the organ equivalent dose concept to plateau dose-response relationships. Radiation and Environmental Biophysics 44, 235-239. Schneider, U., Zwahlen D., Ross D., Kaser-Hotz B., 2005b. Estimation of radiation-induced cancer from three-dimensional dose distributions: concept of organ equivalent dose. Int. J. Radiation Oncology Biol. Phys. 61, 1510-1515. Schneider, U., 2009. Mechanistic model of radiation-induced cancer after fractionated radiotherapy using the linear-quadratic formula. Med. Phys. 36. Dasu, A., Toma-Dasu I., Olofsson J., Karlsson M., 2005. The use of risk estimation models for the induction of secondary cancers following radiotherapy. Acta Oncologica 44, 339-347. Takam, R., Bezak E., Yeoh E.E., 2009. Risk of second primary cancer following prostate cancer radiotherapy: DVH analysis using the competitive risk model. Phys. Med. Biol. 54, 611–625. Kry, S.F., Salehpour M., Followill D.S., Stovall M., Kuban D.A., White R.A., Rosen I.I., 2005. The calculated risk of fatal secondary malignancies from intensity-modulated radiation therapy. Int.J.Radiation Oncology Biol. Phys. 62, 1195-1203. CRUK. 2012. Cancer Research UK [Online] http://info.cancerresearchuk.org/cancerstats.


103


11. Conclusions The main objective of this work was to measure out-of-field doses under reproducible conditions in a water tank and for several simulated clinical treatments with a BOMAB phantom, using a variety of radiotherapy treatment modalities. In addition, the study provided a valuable exercise in dosimeter comparison and also facilitated the comparison of calibration techniques between participating institutions. For the study of out-of-field doses, the work illustrated several points of interest which will contribute to the next stage of out-of-field dose study using anthropomorphic phantoms. Detailed discussions of the experimental results are given in the appropriate chapters of this report and the following is a brief summary of the main findings. Dose measurements in a water tank (photons) (i)

At the isocentre, TLD measurements agree with ionization chamber measurements to within 1.5%. However, initial readout of OSL dosimeters, after applying a correction factor, consistently underestimates the ionization chamber measurement by about 4%. For the RPL dosimeter (type GD-352M), systematic overestimations are encountered for all the radiation qualities used in this work (6, 12 and 20 MV), due to a high-Z cap. The overestimations are 13%, 23% and 27% for 6, 12 and 20 MV respectively.

(ii)

At the beam edge, dosimeter comparisons cannot be made because of the variation in size in a steep dose gradient.

(iii)

The general characteristics of the spatial variation of out-of-field doses are consistent with previous studies.

(iv)

In the out-of-field region, good agreement was found between TLD, RPL and ionization chamber measurements but OSL overestimates dose and corrections are necessary.

(v)

The TLD types used are consistently reliable dosimeters for all measurement regions.

Dose measurements in a BOMAB phantom (photons) (i)

The observed variation of dosimeter measurements was greater in the BOMAB phantom compared with those in a water tank, probably due to angular dependence (due to irradiation from several gantry angles) compared with a single field in the tank.

(ii)

No systematic differences between dosimeters was observed. This suggests that the energy dependence of the dosimeters is well compensated (in the case of RPL) well corrected (in the case of OSL) or can be neglected (for LiF based TLDs).

(iii)

For comparisons between CRT and IMRT at 6 MV, out-of-field doses for IMRT are lower than those for CRT only at distances close to the field edge (4.6 cm). They are similar up to 14.6 cm and thereafter are higher for IMRT showing a characteristic „peak“ with a maximum at approximately 20 cm from the field edge. This is probably due to leakage through the collimating jaws. For greater distances IMRT doses are approximately 50% higher than for CRT (for doses greater than about 1 mGy). 104



(iv)

For the same modalities with different energies, out-of-field doses are always significantly lower for higher energies.

(v)

The highest out-of-field doses were measured for Tomotherapy and somewhat less (but significantly higher than for IMRT) for VMAT.

(vi)

For all simulations, measured doses were greater than those calculated by the two TPSs used, especially for the SCUH system (CMX Xio Rel. 4.40.05) where order of magnitude differences at distances > approximately 20 cm from the isocentre were observed. However, the best agreement between calculated and measured doses resulted from irradiation using a 6 MV VMAT technique (Varian Eclipse External Beam Planning System (COOK). This is not unexpected, since TPSs are designed primarily to calculate accurate target volume doses, but it does mean that the calculation of out-of-field doses using TPSs should be carefully verified experimentally before clinical use.

Dose measurements in a BOMAB phantom (neutrons) (i)

Dosimetry systems (SDD, BTI-BDPND and PADC detectors ) and methods used to measure in-phantom neutron equivalent doses in a variety of clinical radiotherapy fields showed consistent and repeatable responses with a ~20% uncertainty (1 sd).

(ii)

A non-negligible photoneutron dose was observed for a 6 MV primary photon beam. Photoneutrons at energies lower than 8 MV are probably produced by interaction of photons with light elements (e.g. beryllium) in the accelerator exit window.

(iii)

Peripheral secondary neutron doses for IMRT were nearly four times the peripheral doses measured in 3DCRT.

(iv)

At 6 MV, when compared to VMAT, IMRT showed lower peripheral neutron doses, probably because of the finite number of beam positions for IMRT.

(v)

Of the low energy treatments investigated, Tomotherapy is the least affected by neutron contamination


105